diff --git a/bob/__init__.py b/bob/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..edbb4090fca046b19d22d3982711084621bff3be
--- /dev/null
+++ b/bob/__init__.py
@@ -0,0 +1,4 @@
+# see https://docs.python.org/3/library/pkgutil.html
+from pkgutil import extend_path
+
+__path__ = extend_path(__path__, __name__)
diff --git a/bob/bio/__init__.py b/bob/bio/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..edbb4090fca046b19d22d3982711084621bff3be
--- /dev/null
+++ b/bob/bio/__init__.py
@@ -0,0 +1,4 @@
+# see https://docs.python.org/3/library/pkgutil.html
+from pkgutil import extend_path
+
+__path__ = extend_path(__path__, __name__)
diff --git a/bob/bio/facexzoo/LICENSE.txt b/bob/bio/facexzoo/LICENSE.txt
new file mode 100644
index 0000000000000000000000000000000000000000..dd7083a25926383c625969899052dc2d5bf7bf55
--- /dev/null
+++ b/bob/bio/facexzoo/LICENSE.txt
@@ -0,0 +1,915 @@
+Copyright [2020] JD.com, Inc., JD AI
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+
+----------------------------------------------------------------------------------------------------------
+
+From PyTorch:
+
+Copyright (c) 2016- Facebook, Inc (Adam Paszke)
+Copyright (c) 2014- Facebook, Inc (Soumith Chintala)
+Copyright (c) 2011-2014 Idiap Research Institute (Ronan Collobert)
+Copyright (c) 2012-2014 Deepmind Technologies (Koray Kavukcuoglu)
+Copyright (c) 2011-2012 NEC Laboratories America (Koray Kavukcuoglu)
+Copyright (c) 2011-2013 NYU (Clement Farabet)
+Copyright (c) 2006-2010 NEC Laboratories America (Ronan Collobert, Leon Bottou, Iain Melvin, Jason Weston)
+Copyright (c) 2006 Idiap Research Institute (Samy Bengio)
+Copyright (c) 2001-2004 Idiap Research Institute (Ronan Collobert, Samy Bengio, Johnny Mariethoz)
+
+From Caffe2:
+
+Copyright (c) 2016-present, Facebook Inc. All rights reserved.
+
+All contributions by Facebook:
+Copyright (c) 2016 Facebook Inc.
+
+All contributions by Google:
+Copyright (c) 2015 Google Inc.
+All rights reserved.
+
+All contributions by Yangqing Jia:
+Copyright (c) 2015 Yangqing Jia
+All rights reserved.
+
+All contributions from Caffe:
+Copyright(c) 2013, 2014, 2015, the respective contributors
+All rights reserved.
+
+All other contributions:
+Copyright(c) 2015, 2016 the respective contributors
+All rights reserved.
+
+Caffe2 uses a copyright model similar to Caffe: each contributor holds
+copyright over their contributions to Caffe2. The project versioning records
+all such contribution and copyright details. If a contributor wants to further
+mark their specific copyright on a particular contribution, they should
+indicate their copyright solely in the commit message of the change when it is
+committed.
+
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+1. Redistributions of source code must retain the above copyright
+ notice, this list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright
+ notice, this list of conditions and the following disclaimer in the
+ documentation and/or other materials provided with the distribution.
+
+3. Neither the names of Facebook, Deepmind Technologies, NYU, NEC Laboratories America
+ and IDIAP Research Institute nor the names of its contributors may be
+ used to endorse or promote products derived from this software without
+ specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+POSSIBILITY OF SUCH DAMAGE.
+
+----------------------------------------------------------------------------------------------------------
+
+From insightface:
+
+MIT License
+
+Copyright (c) 2018 Jiankang Deng and Jia Guo
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From ghostnet:
+
+Copyright (c) 2020. Huawei Technologies Co., Ltd.
+ All rights reserved.
+
+ Apache License
+ Version 2.0, January 2004
+ http://www.apache.org/licenses/
+
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+ 1. Definitions.
+
+ "License" shall mean the terms and conditions for use, reproduction,
+ and distribution as defined by Sections 1 through 9 of this document.
+
+ "Licensor" shall mean the copyright owner or entity authorized by
+ the copyright owner that is granting the License.
+
+ "Legal Entity" shall mean the union of the acting entity and all
+ other entities that control, are controlled by, or are under common
+ control with that entity. For the purposes of this definition,
+ "control" means (i) the power, direct or indirect, to cause the
+ direction or management of such entity, whether by contract or
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
+ outstanding shares, or (iii) beneficial ownership of such entity.
+
+ "You" (or "Your") shall mean an individual or Legal Entity
+ exercising permissions granted by this License.
+
+ "Source" form shall mean the preferred form for making modifications,
+ including but not limited to software source code, documentation
+ source, and configuration files.
+
+ "Object" form shall mean any form resulting from mechanical
+ transformation or translation of a Source form, including but
+ not limited to compiled object code, generated documentation,
+ and conversions to other media types.
+
+ "Work" shall mean the work of authorship, whether in Source or
+ Object form, made available under the License, as indicated by a
+ copyright notice that is included in or attached to the work
+ (an example is provided in the Appendix below).
+
+ "Derivative Works" shall mean any work, whether in Source or Object
+ form, that is based on (or derived from) the Work and for which the
+ editorial revisions, annotations, elaborations, or other modifications
+ represent, as a whole, an original work of authorship. For the purposes
+ of this License, Derivative Works shall not include works that remain
+ separable from, or merely link (or bind by name) to the interfaces of,
+ the Work and Derivative Works thereof.
+
+ "Contribution" shall mean any work of authorship, including
+ the original version of the Work and any modifications or additions
+ to that Work or Derivative Works thereof, that is intentionally
+ submitted to Licensor for inclusion in the Work by the copyright owner
+ or by an individual or Legal Entity authorized to submit on behalf of
+ the copyright owner. For the purposes of this definition, "submitted"
+ means any form of electronic, verbal, or written communication sent
+ to the Licensor or its representatives, including but not limited to
+ communication on electronic mailing lists, source code control systems,
+ and issue tracking systems that are managed by, or on behalf of, the
+ Licensor for the purpose of discussing and improving the Work, but
+ excluding communication that is conspicuously marked or otherwise
+ designated in writing by the copyright owner as "Not a Contribution."
+
+ "Contributor" shall mean Licensor and any individual or Legal Entity
+ on behalf of whom a Contribution has been received by Licensor and
+ subsequently incorporated within the Work.
+
+ 2. Grant of Copyright License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ copyright license to reproduce, prepare Derivative Works of,
+ publicly display, publicly perform, sublicense, and distribute the
+ Work and such Derivative Works in Source or Object form.
+
+ 3. Grant of Patent License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ (except as stated in this section) patent license to make, have made,
+ use, offer to sell, sell, import, and otherwise transfer the Work,
+ where such license applies only to those patent claims licensable
+ by such Contributor that are necessarily infringed by their
+ Contribution(s) alone or by combination of their Contribution(s)
+ with the Work to which such Contribution(s) was submitted. If You
+ institute patent litigation against any entity (including a
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
+ or a Contribution incorporated within the Work constitutes direct
+ or contributory patent infringement, then any patent licenses
+ granted to You under this License for that Work shall terminate
+ as of the date such litigation is filed.
+
+ 4. Redistribution. You may reproduce and distribute copies of the
+ Work or Derivative Works thereof in any medium, with or without
+ modifications, and in Source or Object form, provided that You
+ meet the following conditions:
+
+ (a) You must give any other recipients of the Work or
+ Derivative Works a copy of this License; and
+
+ (b) You must cause any modified files to carry prominent notices
+ stating that You changed the files; and
+
+ (c) You must retain, in the Source form of any Derivative Works
+ that You distribute, all copyright, patent, trademark, and
+ attribution notices from the Source form of the Work,
+ excluding those notices that do not pertain to any part of
+ the Derivative Works; and
+
+ (d) If the Work includes a "NOTICE" text file as part of its
+ distribution, then any Derivative Works that You distribute must
+ include a readable copy of the attribution notices contained
+ within such NOTICE file, excluding those notices that do not
+ pertain to any part of the Derivative Works, in at least one
+ of the following places: within a NOTICE text file distributed
+ as part of the Derivative Works; within the Source form or
+ documentation, if provided along with the Derivative Works; or,
+ within a display generated by the Derivative Works, if and
+ wherever such third-party notices normally appear. The contents
+ of the NOTICE file are for informational purposes only and
+ do not modify the License. You may add Your own attribution
+ notices within Derivative Works that You distribute, alongside
+ or as an addendum to the NOTICE text from the Work, provided
+ that such additional attribution notices cannot be construed
+ as modifying the License.
+
+ You may add Your own copyright statement to Your modifications and
+ may provide additional or different license terms and conditions
+ for use, reproduction, or distribution of Your modifications, or
+ for any such Derivative Works as a whole, provided Your use,
+ reproduction, and distribution of the Work otherwise complies with
+ the conditions stated in this License.
+
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
+ any Contribution intentionally submitted for inclusion in the Work
+ by You to the Licensor shall be under the terms and conditions of
+ this License, without any additional terms or conditions.
+ Notwithstanding the above, nothing herein shall supersede or modify
+ the terms of any separate license agreement you may have executed
+ with Licensor regarding such Contributions.
+
+ 6. Trademarks. This License does not grant permission to use the trade
+ names, trademarks, service marks, or product names of the Licensor,
+ except as required for reasonable and customary use in describing the
+ origin of the Work and reproducing the content of the NOTICE file.
+
+ 7. Disclaimer of Warranty. Unless required by applicable law or
+ agreed to in writing, Licensor provides the Work (and each
+ Contributor provides its Contributions) on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+ implied, including, without limitation, any warranties or conditions
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+ PARTICULAR PURPOSE. You are solely responsible for determining the
+ appropriateness of using or redistributing the Work and assume any
+ risks associated with Your exercise of permissions under this License.
+
+ 8. Limitation of Liability. In no event and under no legal theory,
+ whether in tort (including negligence), contract, or otherwise,
+ unless required by applicable law (such as deliberate and grossly
+ negligent acts) or agreed to in writing, shall any Contributor be
+ liable to You for damages, including any direct, indirect, special,
+ incidental, or consequential damages of any character arising as a
+ result of this License or out of the use or inability to use the
+ Work (including but not limited to damages for loss of goodwill,
+ work stoppage, computer failure or malfunction, or any and all
+ other commercial damages or losses), even if such Contributor
+ has been advised of the possibility of such damages.
+
+ 9. Accepting Warranty or Additional Liability. While redistributing
+ the Work or Derivative Works thereof, You may choose to offer,
+ and charge a fee for, acceptance of support, warranty, indemnity,
+ or other liability obligations and/or rights consistent with this
+ License. However, in accepting such obligations, You may act only
+ on Your own behalf and on Your sole responsibility, not on behalf
+ of any other Contributor, and only if You agree to indemnify,
+ defend, and hold each Contributor harmless for any liability
+ incurred by, or claims asserted against, such Contributor by reason
+ of your accepting any such warranty or additional liability.
+
+ END OF TERMS AND CONDITIONS
+
+ APPENDIX: How to apply the Apache License to your work.
+
+ To apply the Apache License to your work, attach the following
+ boilerplate notice, with the fields enclosed by brackets "[]"
+ replaced with your own identifying information. (Don't include
+ the brackets!) The text should be enclosed in the appropriate
+ comment syntax for the file format. We also recommend that a
+ file or class name and description of purpose be included on the
+ same "printed page" as the copyright notice for easier
+ identification within third-party archives.
+
+ Copyright 2020 Huawei Technologies Co., Ltd.
+
+ Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+ You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+ Unless required by applicable law or agreed to in writing, software
+ distributed under the License is distributed on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ See the License for the specific language governing permissions and
+ limitations under the License.
+
+----------------------------------------------------------------------------------------------------------
+
+From HRNet:
+
+MIT License
+
+Copyright (c) 2019 Microsoft Corporation
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From InsightFace_Pytorch:
+
+MIT License
+
+Copyright (c) 2018 TreB1eN
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From pytorch-adacos:
+
+MIT License
+
+Copyright (c) 2019 Takato Kimura
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From CurricularFace:
+
+MIT License
+
+Copyright (c) 2020 HuangYG123
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From Pytorch_Retinaface:
+
+MIT License
+
+Copyright (c) 2019
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From chainercv:
+
+The MIT License
+
+Copyright (c) 2017 Preferred Networks, Inc.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in
+all copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+THE SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From PRNet:
+
+MIT License
+
+Copyright (c) 2018 Yao Feng
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+----------------------------------------------------------------------------------------------------------
+
+From EfficientNet-PyTorch：
+
+ Apache License
+ Version 2.0, January 2004
+ http://www.apache.org/licenses/
+
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+ 1. Definitions.
+
+ "License" shall mean the terms and conditions for use, reproduction,
+ and distribution as defined by Sections 1 through 9 of this document.
+
+ "Licensor" shall mean the copyright owner or entity authorized by
+ the copyright owner that is granting the License.
+
+ "Legal Entity" shall mean the union of the acting entity and all
+ other entities that control, are controlled by, or are under common
+ control with that entity. For the purposes of this definition,
+ "control" means (i) the power, direct or indirect, to cause the
+ direction or management of such entity, whether by contract or
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
+ outstanding shares, or (iii) beneficial ownership of such entity.
+
+ "You" (or "Your") shall mean an individual or Legal Entity
+ exercising permissions granted by this License.
+
+ "Source" form shall mean the preferred form for making modifications,
+ including but not limited to software source code, documentation
+ source, and configuration files.
+
+ "Object" form shall mean any form resulting from mechanical
+ transformation or translation of a Source form, including but
+ not limited to compiled object code, generated documentation,
+ and conversions to other media types.
+
+ "Work" shall mean the work of authorship, whether in Source or
+ Object form, made available under the License, as indicated by a
+ copyright notice that is included in or attached to the work
+ (an example is provided in the Appendix below).
+
+ "Derivative Works" shall mean any work, whether in Source or Object
+ form, that is based on (or derived from) the Work and for which the
+ editorial revisions, annotations, elaborations, or other modifications
+ represent, as a whole, an original work of authorship. For the purposes
+ of this License, Derivative Works shall not include works that remain
+ separable from, or merely link (or bind by name) to the interfaces of,
+ the Work and Derivative Works thereof.
+
+ "Contribution" shall mean any work of authorship, including
+ the original version of the Work and any modifications or additions
+ to that Work or Derivative Works thereof, that is intentionally
+ submitted to Licensor for inclusion in the Work by the copyright owner
+ or by an individual or Legal Entity authorized to submit on behalf of
+ the copyright owner. For the purposes of this definition, "submitted"
+ means any form of electronic, verbal, or written communication sent
+ to the Licensor or its representatives, including but not limited to
+ communication on electronic mailing lists, source code control systems,
+ and issue tracking systems that are managed by, or on behalf of, the
+ Licensor for the purpose of discussing and improving the Work, but
+ excluding communication that is conspicuously marked or otherwise
+ designated in writing by the copyright owner as "Not a Contribution."
+
+ "Contributor" shall mean Licensor and any individual or Legal Entity
+ on behalf of whom a Contribution has been received by Licensor and
+ subsequently incorporated within the Work.
+
+ 2. Grant of Copyright License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ copyright license to reproduce, prepare Derivative Works of,
+ publicly display, publicly perform, sublicense, and distribute the
+ Work and such Derivative Works in Source or Object form.
+
+ 3. Grant of Patent License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ (except as stated in this section) patent license to make, have made,
+ use, offer to sell, sell, import, and otherwise transfer the Work,
+ where such license applies only to those patent claims licensable
+ by such Contributor that are necessarily infringed by their
+ Contribution(s) alone or by combination of their Contribution(s)
+ with the Work to which such Contribution(s) was submitted. If You
+ institute patent litigation against any entity (including a
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
+ or a Contribution incorporated within the Work constitutes direct
+ or contributory patent infringement, then any patent licenses
+ granted to You under this License for that Work shall terminate
+ as of the date such litigation is filed.
+
+ 4. Redistribution. You may reproduce and distribute copies of the
+ Work or Derivative Works thereof in any medium, with or without
+ modifications, and in Source or Object form, provided that You
+ meet the following conditions:
+
+ (a) You must give any other recipients of the Work or
+ Derivative Works a copy of this License; and
+
+ (b) You must cause any modified files to carry prominent notices
+ stating that You changed the files; and
+
+ (c) You must retain, in the Source form of any Derivative Works
+ that You distribute, all copyright, patent, trademark, and
+ attribution notices from the Source form of the Work,
+ excluding those notices that do not pertain to any part of
+ the Derivative Works; and
+
+ (d) If the Work includes a "NOTICE" text file as part of its
+ distribution, then any Derivative Works that You distribute must
+ include a readable copy of the attribution notices contained
+ within such NOTICE file, excluding those notices that do not
+ pertain to any part of the Derivative Works, in at least one
+ of the following places: within a NOTICE text file distributed
+ as part of the Derivative Works; within the Source form or
+ documentation, if provided along with the Derivative Works; or,
+ within a display generated by the Derivative Works, if and
+ wherever such third-party notices normally appear. The contents
+ of the NOTICE file are for informational purposes only and
+ do not modify the License. You may add Your own attribution
+ notices within Derivative Works that You distribute, alongside
+ or as an addendum to the NOTICE text from the Work, provided
+ that such additional attribution notices cannot be construed
+ as modifying the License.
+
+ You may add Your own copyright statement to Your modifications and
+ may provide additional or different license terms and conditions
+ for use, reproduction, or distribution of Your modifications, or
+ for any such Derivative Works as a whole, provided Your use,
+ reproduction, and distribution of the Work otherwise complies with
+ the conditions stated in this License.
+
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
+ any Contribution intentionally submitted for inclusion in the Work
+ by You to the Licensor shall be under the terms and conditions of
+ this License, without any additional terms or conditions.
+ Notwithstanding the above, nothing herein shall supersede or modify
+ the terms of any separate license agreement you may have executed
+ with Licensor regarding such Contributions.
+
+ 6. Trademarks. This License does not grant permission to use the trade
+ names, trademarks, service marks, or product names of the Licensor,
+ except as required for reasonable and customary use in describing the
+ origin of the Work and reproducing the content of the NOTICE file.
+
+ 7. Disclaimer of Warranty. Unless required by applicable law or
+ agreed to in writing, Licensor provides the Work (and each
+ Contributor provides its Contributions) on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+ implied, including, without limitation, any warranties or conditions
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+ PARTICULAR PURPOSE. You are solely responsible for determining the
+ appropriateness of using or redistributing the Work and assume any
+ risks associated with Your exercise of permissions under this License.
+
+ 8. Limitation of Liability. In no event and under no legal theory,
+ whether in tort (including negligence), contract, or otherwise,
+ unless required by applicable law (such as deliberate and grossly
+ negligent acts) or agreed to in writing, shall any Contributor be
+ liable to You for damages, including any direct, indirect, special,
+ incidental, or consequential damages of any character arising as a
+ result of this License or out of the use or inability to use the
+ Work (including but not limited to damages for loss of goodwill,
+ work stoppage, computer failure or malfunction, or any and all
+ other commercial damages or losses), even if such Contributor
+ has been advised of the possibility of such damages.
+
+ 9. Accepting Warranty or Additional Liability. While redistributing
+ the Work or Derivative Works thereof, You may choose to offer,
+ and charge a fee for, acceptance of support, warranty, indemnity,
+ or other liability obligations and/or rights consistent with this
+ License. However, in accepting such obligations, You may act only
+ on Your own behalf and on Your sole responsibility, not on behalf
+ of any other Contributor, and only if You agree to indemnify,
+ defend, and hold each Contributor harmless for any liability
+ incurred by, or claims asserted against, such Contributor by reason
+ of your accepting any such warranty or additional liability.
+
+ END OF TERMS AND CONDITIONS
+
+ APPENDIX: How to apply the Apache License to your work.
+
+ To apply the Apache License to your work, attach the following
+ boilerplate notice, with the fields enclosed by brackets "[]"
+ replaced with your own identifying information. (Don't include
+ the brackets!) The text should be enclosed in the appropriate
+ comment syntax for the file format. We also recommend that a
+ file or class name and description of purpose be included on the
+ same "printed page" as the copyright notice for easier
+ identification within third-party archives.
+
+ Copyright [yyyy] [name of copyright owner]
+
+ Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+ You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+ Unless required by applicable law or agreed to in writing, software
+ distributed under the License is distributed on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ See the License for the specific language governing permissions and
+ limitations under the License.
+
+----------------------------------------------------------------------------------------------------------
+
+From ResNeSt:
+
+ Apache License
+ Version 2.0, January 2004
+ http://www.apache.org/licenses/
+
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+ 1. Definitions.
+
+ "License" shall mean the terms and conditions for use, reproduction,
+ and distribution as defined by Sections 1 through 9 of this document.
+
+ "Licensor" shall mean the copyright owner or entity authorized by
+ the copyright owner that is granting the License.
+
+ "Legal Entity" shall mean the union of the acting entity and all
+ other entities that control, are controlled by, or are under common
+ control with that entity. For the purposes of this definition,
+ "control" means (i) the power, direct or indirect, to cause the
+ direction or management of such entity, whether by contract or
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
+ outstanding shares, or (iii) beneficial ownership of such entity.
+
+ "You" (or "Your") shall mean an individual or Legal Entity
+ exercising permissions granted by this License.
+
+ "Source" form shall mean the preferred form for making modifications,
+ including but not limited to software source code, documentation
+ source, and configuration files.
+
+ "Object" form shall mean any form resulting from mechanical
+ transformation or translation of a Source form, including but
+ not limited to compiled object code, generated documentation,
+ and conversions to other media types.
+
+ "Work" shall mean the work of authorship, whether in Source or
+ Object form, made available under the License, as indicated by a
+ copyright notice that is included in or attached to the work
+ (an example is provided in the Appendix below).
+
+ "Derivative Works" shall mean any work, whether in Source or Object
+ form, that is based on (or derived from) the Work and for which the
+ editorial revisions, annotations, elaborations, or other modifications
+ represent, as a whole, an original work of authorship. For the purposes
+ of this License, Derivative Works shall not include works that remain
+ separable from, or merely link (or bind by name) to the interfaces of,
+ the Work and Derivative Works thereof.
+
+ "Contribution" shall mean any work of authorship, including
+ the original version of the Work and any modifications or additions
+ to that Work or Derivative Works thereof, that is intentionally
+ submitted to Licensor for inclusion in the Work by the copyright owner
+ or by an individual or Legal Entity authorized to submit on behalf of
+ the copyright owner. For the purposes of this definition, "submitted"
+ means any form of electronic, verbal, or written communication sent
+ to the Licensor or its representatives, including but not limited to
+ communication on electronic mailing lists, source code control systems,
+ and issue tracking systems that are managed by, or on behalf of, the
+ Licensor for the purpose of discussing and improving the Work, but
+ excluding communication that is conspicuously marked or otherwise
+ designated in writing by the copyright owner as "Not a Contribution."
+
+ "Contributor" shall mean Licensor and any individual or Legal Entity
+ on behalf of whom a Contribution has been received by Licensor and
+ subsequently incorporated within the Work.
+
+ 2. Grant of Copyright License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ copyright license to reproduce, prepare Derivative Works of,
+ publicly display, publicly perform, sublicense, and distribute the
+ Work and such Derivative Works in Source or Object form.
+
+ 3. Grant of Patent License. Subject to the terms and conditions of
+ this License, each Contributor hereby grants to You a perpetual,
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+ (except as stated in this section) patent license to make, have made,
+ use, offer to sell, sell, import, and otherwise transfer the Work,
+ where such license applies only to those patent claims licensable
+ by such Contributor that are necessarily infringed by their
+ Contribution(s) alone or by combination of their Contribution(s)
+ with the Work to which such Contribution(s) was submitted. If You
+ institute patent litigation against any entity (including a
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
+ or a Contribution incorporated within the Work constitutes direct
+ or contributory patent infringement, then any patent licenses
+ granted to You under this License for that Work shall terminate
+ as of the date such litigation is filed.
+
+ 4. Redistribution. You may reproduce and distribute copies of the
+ Work or Derivative Works thereof in any medium, with or without
+ modifications, and in Source or Object form, provided that You
+ meet the following conditions:
+
+ (a) You must give any other recipients of the Work or
+ Derivative Works a copy of this License; and
+
+ (b) You must cause any modified files to carry prominent notices
+ stating that You changed the files; and
+
+ (c) You must retain, in the Source form of any Derivative Works
+ that You distribute, all copyright, patent, trademark, and
+ attribution notices from the Source form of the Work,
+ excluding those notices that do not pertain to any part of
+ the Derivative Works; and
+
+ (d) If the Work includes a "NOTICE" text file as part of its
+ distribution, then any Derivative Works that You distribute must
+ include a readable copy of the attribution notices contained
+ within such NOTICE file, excluding those notices that do not
+ pertain to any part of the Derivative Works, in at least one
+ of the following places: within a NOTICE text file distributed
+ as part of the Derivative Works; within the Source form or
+ documentation, if provided along with the Derivative Works; or,
+ within a display generated by the Derivative Works, if and
+ wherever such third-party notices normally appear. The contents
+ of the NOTICE file are for informational purposes only and
+ do not modify the License. You may add Your own attribution
+ notices within Derivative Works that You distribute, alongside
+ or as an addendum to the NOTICE text from the Work, provided
+ that such additional attribution notices cannot be construed
+ as modifying the License.
+
+ You may add Your own copyright statement to Your modifications and
+ may provide additional or different license terms and conditions
+ for use, reproduction, or distribution of Your modifications, or
+ for any such Derivative Works as a whole, provided Your use,
+ reproduction, and distribution of the Work otherwise complies with
+ the conditions stated in this License.
+
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
+ any Contribution intentionally submitted for inclusion in the Work
+ by You to the Licensor shall be under the terms and conditions of
+ this License, without any additional terms or conditions.
+ Notwithstanding the above, nothing herein shall supersede or modify
+ the terms of any separate license agreement you may have executed
+ with Licensor regarding such Contributions.
+
+ 6. Trademarks. This License does not grant permission to use the trade
+ names, trademarks, service marks, or product names of the Licensor,
+ except as required for reasonable and customary use in describing the
+ origin of the Work and reproducing the content of the NOTICE file.
+
+ 7. Disclaimer of Warranty. Unless required by applicable law or
+ agreed to in writing, Licensor provides the Work (and each
+ Contributor provides its Contributions) on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+ implied, including, without limitation, any warranties or conditions
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+ PARTICULAR PURPOSE. You are solely responsible for determining the
+ appropriateness of using or redistributing the Work and assume any
+ risks associated with Your exercise of permissions under this License.
+
+ 8. Limitation of Liability. In no event and under no legal theory,
+ whether in tort (including negligence), contract, or otherwise,
+ unless required by applicable law (such as deliberate and grossly
+ negligent acts) or agreed to in writing, shall any Contributor be
+ liable to You for damages, including any direct, indirect, special,
+ incidental, or consequential damages of any character arising as a
+ result of this License or out of the use or inability to use the
+ Work (including but not limited to damages for loss of goodwill,
+ work stoppage, computer failure or malfunction, or any and all
+ other commercial damages or losses), even if such Contributor
+ has been advised of the possibility of such damages.
+
+ 9. Accepting Warranty or Additional Liability. While redistributing
+ the Work or Derivative Works thereof, You may choose to offer,
+ and charge a fee for, acceptance of support, warranty, indemnity,
+ or other liability obligations and/or rights consistent with this
+ License. However, in accepting such obligations, You may act only
+ on Your own behalf and on Your sole responsibility, not on behalf
+ of any other Contributor, and only if You agree to indemnify,
+ defend, and hold each Contributor harmless for any liability
+ incurred by, or claims asserted against, such Contributor by reason
+ of your accepting any such warranty or additional liability.
+
+ END OF TERMS AND CONDITIONS
+
+ APPENDIX: How to apply the Apache License to your work.
+
+ To apply the Apache License to your work, attach the following
+ boilerplate notice, with the fields enclosed by brackets "[]"
+ replaced with your own identifying information. (Don't include
+ the brackets!) The text should be enclosed in the appropriate
+ comment syntax for the file format. We also recommend that a
+ file or class name and description of purpose be included on the
+ same "printed page" as the copyright notice for easier
+ identification within third-party archives.
+
+ Copyright [yyyy] [name of copyright owner]
+
+ Licensed under the Apache License, Version 2.0 (the "License");
+ you may not use this file except in compliance with the License.
+ You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+ Unless required by applicable law or agreed to in writing, software
+ distributed under the License is distributed on an "AS IS" BASIS,
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ See the License for the specific language governing permissions and
+ limitations under the License.
diff --git a/bob/bio/facexzoo/Readme.md b/bob/bio/facexzoo/Readme.md
new file mode 100644
index 0000000000000000000000000000000000000000..6f0901bbf0b5bca38bc016ec92790541211e57e8
--- /dev/null
+++ b/bob/bio/facexzoo/Readme.md
@@ -0,0 +1,2 @@
+Part of the scripts are taken from FaceXZoo reposiroty which is licensed under Apache 2.
+For more details please check the [LICENSE](LICENSE.txt) file.
\ No newline at end of file
diff --git a/bob/bio/facexzoo/__init__.py b/bob/bio/facexzoo/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/backbones/AttentionNets.py b/bob/bio/facexzoo/backbones/AttentionNets.py
new file mode 100644
index 0000000000000000000000000000000000000000..addfd7ee106b034fd191d9477bb4a3c79e2900fd
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/AttentionNets.py
@@ -0,0 +1,240 @@
+"""
+@author: Jun Wang
+@date: 20201019
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/tengshaofeng/ResidualAttentionNetwork-pytorch/tree/master/Residual-Attention-Network/model
+
+import torch
+import torch.nn as nn
+from torch.nn import init
+import functools
+from torch.autograd import Variable
+import numpy as np
+
+class Flatten(nn.Module):
+ def forward(self, x):
+ return x.reshape(x.size(0), -1)
+
+class ResidualBlock(nn.Module):
+ def __init__(self, input_channels, output_channels, stride=1):
+ super(ResidualBlock, self).__init__()
+ self.input_channels = input_channels
+ self.output_channels = output_channels
+ self.stride = stride
+ self.bn1 = nn.BatchNorm2d(input_channels)
+ self.relu = nn.ReLU(inplace=True)
+ self.conv1 = nn.Conv2d(input_channels, output_channels//4, 1, 1, bias = False)
+ self.bn2 = nn.BatchNorm2d(output_channels//4)
+ self.relu = nn.ReLU(inplace=True)
+ self.conv2 = nn.Conv2d(output_channels//4, output_channels//4, 3, stride, padding = 1, bias = False)
+ self.bn3 = nn.BatchNorm2d(output_channels//4)
+ self.relu = nn.ReLU(inplace=True)
+ self.conv3 = nn.Conv2d(output_channels//4, output_channels, 1, 1, bias = False)
+ self.conv4 = nn.Conv2d(input_channels, output_channels , 1, stride, bias = False)
+ def forward(self, x):
+ residual = x
+ out = self.bn1(x)
+ out1 = self.relu(out)
+ out = self.conv1(out1)
+ out = self.bn2(out)
+ out = self.relu(out)
+ out = self.conv2(out)
+ out = self.bn3(out)
+ out = self.relu(out)
+ out = self.conv3(out)
+ if (self.input_channels != self.output_channels) or (self.stride !=1 ):
+ residual = self.conv4(out1)
+ out += residual
+ return out
+
+class AttentionModule_stage1(nn.Module):
+ # input size is 56*56
+ def __init__(self, in_channels, out_channels, size1=(56, 56), size2=(28, 28), size3=(14, 14)):
+ super(AttentionModule_stage1, self).__init__()
+ self.first_residual_blocks = ResidualBlock(in_channels, out_channels)
+ self.trunk_branches = nn.Sequential(
+ ResidualBlock(in_channels, out_channels),
+ ResidualBlock(in_channels, out_channels)
+ )
+ self.mpool1 = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+ self.softmax1_blocks = ResidualBlock(in_channels, out_channels)
+ self.skip1_connection_residual_block = ResidualBlock(in_channels, out_channels)
+ self.mpool2 = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+ self.softmax2_blocks = ResidualBlock(in_channels, out_channels)
+ self.skip2_connection_residual_block = ResidualBlock(in_channels, out_channels)
+ self.mpool3 = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+ self.softmax3_blocks = nn.Sequential(
+ ResidualBlock(in_channels, out_channels),
+ ResidualBlock(in_channels, out_channels)
+ )
+ self.interpolation3 = nn.UpsamplingBilinear2d(size=size3)
+ self.softmax4_blocks = ResidualBlock(in_channels, out_channels)
+ self.interpolation2 = nn.UpsamplingBilinear2d(size=size2)
+ self.softmax5_blocks = ResidualBlock(in_channels, out_channels)
+ self.interpolation1 = nn.UpsamplingBilinear2d(size=size1)
+ self.softmax6_blocks = nn.Sequential(
+ nn.BatchNorm2d(out_channels),
+ nn.ReLU(inplace=True),
+ nn.Conv2d(out_channels, out_channels , kernel_size = 1, stride = 1, bias = False),
+ nn.BatchNorm2d(out_channels),
+ nn.ReLU(inplace=True),
+ nn.Conv2d(out_channels, out_channels , kernel_size = 1, stride = 1, bias = False),
+ nn.Sigmoid()
+ )
+ self.last_blocks = ResidualBlock(in_channels, out_channels)
+ def forward(self, x):
+ x = self.first_residual_blocks(x)
+ out_trunk = self.trunk_branches(x)
+ out_mpool1 = self.mpool1(x)
+ out_softmax1 = self.softmax1_blocks(out_mpool1)
+ out_skip1_connection = self.skip1_connection_residual_block(out_softmax1)
+ out_mpool2 = self.mpool2(out_softmax1)
+ out_softmax2 = self.softmax2_blocks(out_mpool2)
+ out_skip2_connection = self.skip2_connection_residual_block(out_softmax2)
+ out_mpool3 = self.mpool3(out_softmax2)
+ out_softmax3 = self.softmax3_blocks(out_mpool3)
+ #
+ out_interp3 = self.interpolation3(out_softmax3) + out_softmax2
+ # print(out_skip2_connection.data)
+ # print(out_interp3.data)
+ out = out_interp3 + out_skip2_connection
+ out_softmax4 = self.softmax4_blocks(out)
+ out_interp2 = self.interpolation2(out_softmax4) + out_softmax1
+ out = out_interp2 + out_skip1_connection
+ out_softmax5 = self.softmax5_blocks(out)
+ out_interp1 = self.interpolation1(out_softmax5) + out_trunk
+ out_softmax6 = self.softmax6_blocks(out_interp1)
+ out = (1 + out_softmax6) * out_trunk
+ out_last = self.last_blocks(out)
+
+ return out_last
+
+
+class AttentionModule_stage2(nn.Module):
+ # input image size is 28*28
+ def __init__(self, in_channels, out_channels, size1=(28, 28), size2=(14, 14)):
+ super(AttentionModule_stage2, self).__init__()
+ self.first_residual_blocks = ResidualBlock(in_channels, out_channels)
+ self.trunk_branches = nn.Sequential(
+ ResidualBlock(in_channels, out_channels),
+ ResidualBlock(in_channels, out_channels)
+ )
+ self.mpool1 = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+ self.softmax1_blocks = ResidualBlock(in_channels, out_channels)
+ self.skip1_connection_residual_block = ResidualBlock(in_channels, out_channels)
+ self.mpool2 = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+ self.softmax2_blocks = nn.Sequential(
+ ResidualBlock(in_channels, out_channels),
+ ResidualBlock(in_channels, out_channels)
+ )
+ self.interpolation2 = nn.UpsamplingBilinear2d(size=size2)
+ self.softmax3_blocks = ResidualBlock(in_channels, out_channels)
+ self.interpolation1 = nn.UpsamplingBilinear2d(size=size1)
+ self.softmax4_blocks = nn.Sequential(
+ nn.BatchNorm2d(out_channels),
+ nn.ReLU(inplace=True),
+ nn.Conv2d(out_channels, out_channels, kernel_size=1, stride=1, bias=False),
+ nn.BatchNorm2d(out_channels),
+ nn.ReLU(inplace=True),
+ nn.Conv2d(out_channels, out_channels, kernel_size=1, stride=1, bias=False),
+ nn.Sigmoid()
+ )
+ self.last_blocks = ResidualBlock(in_channels, out_channels)
+ def forward(self, x):
+ x = self.first_residual_blocks(x)
+ out_trunk = self.trunk_branches(x)
+ out_mpool1 = self.mpool1(x)
+ out_softmax1 = self.softmax1_blocks(out_mpool1)
+ out_skip1_connection = self.skip1_connection_residual_block(out_softmax1)
+ out_mpool2 = self.mpool2(out_softmax1)
+ out_softmax2 = self.softmax2_blocks(out_mpool2)
+ out_interp2 = self.interpolation2(out_softmax2) + out_softmax1
+ # print(out_skip2_connection.data)
+ # print(out_interp3.data)
+ out = out_interp2 + out_skip1_connection
+ out_softmax3 = self.softmax3_blocks(out)
+ out_interp1 = self.interpolation1(out_softmax3) + out_trunk
+ out_softmax4 = self.softmax4_blocks(out_interp1)
+ out = (1 + out_softmax4) * out_trunk
+ out_last = self.last_blocks(out)
+ return out_last
+
+class AttentionModule_stage3(nn.Module):
+ # input image size is 14*14
+ def __init__(self, in_channels, out_channels, size1=(14, 14)):
+ super(AttentionModule_stage3, self).__init__()
+ self.first_residual_blocks = ResidualBlock(in_channels, out_channels)
+ self.trunk_branches = nn.Sequential(
+ ResidualBlock(in_channels, out_channels),
+ ResidualBlock(in_channels, out_channels)
+ )
+ self.mpool1 = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+ self.softmax1_blocks = nn.Sequential(
+ ResidualBlock(in_channels, out_channels),
+ ResidualBlock(in_channels, out_channels)
+ )
+ self.interpolation1 = nn.UpsamplingBilinear2d(size=size1)
+ self.softmax2_blocks = nn.Sequential(
+ nn.BatchNorm2d(out_channels),
+ nn.ReLU(inplace=True),
+ nn.Conv2d(out_channels, out_channels, kernel_size=1, stride=1, bias=False),
+ nn.BatchNorm2d(out_channels),
+ nn.ReLU(inplace=True),
+ nn.Conv2d(out_channels, out_channels, kernel_size=1, stride=1, bias=False),
+ nn.Sigmoid()
+ )
+ self.last_blocks = ResidualBlock(in_channels, out_channels)
+ def forward(self, x):
+ x = self.first_residual_blocks(x)
+ out_trunk = self.trunk_branches(x)
+ out_mpool1 = self.mpool1(x)
+ out_softmax1 = self.softmax1_blocks(out_mpool1)
+ out_interp1 = self.interpolation1(out_softmax1) + out_trunk
+ out_softmax2 = self.softmax2_blocks(out_interp1)
+ out = (1 + out_softmax2) * out_trunk
+ out_last = self.last_blocks(out)
+ return out_last
+
+class ResidualAttentionNet(nn.Module):
+ def __init__(self, stage1_modules, stage2_modules, stage3_modules, feat_dim, out_h, out_w):
+ super(ResidualAttentionNet, self).__init__()
+ self.conv1 = nn.Sequential(
+ nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3, bias = False),
+ nn.BatchNorm2d(64),
+ nn.ReLU(inplace=True)
+ )
+ attention_modules = []
+
+ attention_modules.append(ResidualBlock(64, 256))
+ # stage 1
+ for i in range(stage1_modules):
+ attention_modules.append(AttentionModule_stage1(256, 256))
+
+ attention_modules.append(ResidualBlock(256, 512, 2))
+ # stage2
+ for i in range(stage2_modules):
+ attention_modules.append(AttentionModule_stage2(512, 512))
+
+ attention_modules.append(ResidualBlock(512, 1024, 2))
+ # stage3
+ for i in range(stage3_modules):
+ attention_modules.append(AttentionModule_stage3(1024, 1024))
+
+ # final residual
+ attention_modules.append(ResidualBlock(1024, 2048, 2))
+ attention_modules.append(ResidualBlock(2048, 2048))
+ attention_modules.append(ResidualBlock(2048, 2048))
+ self.attention_body = nn.Sequential(*attention_modules)
+ # output layer
+ self.output_layer = nn.Sequential(
+ Flatten(),
+ nn.Linear(2048 * out_h * out_w, feat_dim, False),
+ nn.BatchNorm1d(feat_dim))
+ def forward(self, x):
+ out = self.conv1(x)
+ out = self.attention_body(out)
+ out = self.output_layer(out)
+ return out
diff --git a/bob/bio/facexzoo/backbones/EfficientNets.py b/bob/bio/facexzoo/backbones/EfficientNets.py
new file mode 100644
index 0000000000000000000000000000000000000000..19fb8813af1d608ae85ca4de094c1d58ffdd99aa
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/EfficientNets.py
@@ -0,0 +1,1038 @@
+"""
+@author: Jun Wang
+@date: 20201019
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# Github repo: https://github.com/lukemelas/EfficientNet-PyTorch
+
+import re
+import math
+import collections
+from functools import partial
+import torch
+from torch import nn
+from torch.nn import functional as F
+from torch.utils import model_zoo
+from torch.nn import Sequential, BatchNorm1d, BatchNorm2d, Dropout, Module, Linear
+
+
+################################################################################
+### Help functions for model architecture
+################################################################################
+
+# GlobalParams and BlockArgs: Two namedtuples
+# Swish and MemoryEfficientSwish: Two implementations of the method
+# round_filters and round_repeats:
+# Functions to calculate params for scaling model width and depth ! ! !
+# get_width_and_height_from_size and calculate_output_image_size
+# drop_connect: A structural design
+# get_same_padding_conv2d:
+# Conv2dDynamicSamePadding
+# Conv2dStaticSamePadding
+# get_same_padding_maxPool2d:
+# MaxPool2dDynamicSamePadding
+# MaxPool2dStaticSamePadding
+# It's an additional function, not used in EfficientNet,
+# but can be used in other model (such as EfficientDet).
+
+# Parameters for the entire model (stem, all blocks, and head)
+GlobalParams = collections.namedtuple('GlobalParams', [
+ 'width_coefficient', 'depth_coefficient', 'image_size', 'dropout_rate',
+ 'num_classes', 'batch_norm_momentum', 'batch_norm_epsilon',
+ 'drop_connect_rate', 'depth_divisor', 'min_depth', 'include_top'])
+
+# Parameters for an individual model block
+BlockArgs = collections.namedtuple('BlockArgs', [
+ 'num_repeat', 'kernel_size', 'stride', 'expand_ratio',
+ 'input_filters', 'output_filters', 'se_ratio', 'id_skip'])
+
+# Set GlobalParams and BlockArgs's defaults
+GlobalParams.__new__.__defaults__ = (None,) * len(GlobalParams._fields)
+BlockArgs.__new__.__defaults__ = (None,) * len(BlockArgs._fields)
+
+
+# An ordinary implementation of Swish function
+class Swish(nn.Module):
+ def forward(self, x):
+ return x * torch.sigmoid(x)
+
+
+# A memory-efficient implementation of Swish function
+class SwishImplementation(torch.autograd.Function):
+ @staticmethod
+ def forward(ctx, i):
+ result = i * torch.sigmoid(i)
+ ctx.save_for_backward(i)
+ return result
+
+ @staticmethod
+ def backward(ctx, grad_output):
+ i = ctx.saved_tensors[0]
+ sigmoid_i = torch.sigmoid(i)
+ return grad_output * (sigmoid_i * (1 + i * (1 - sigmoid_i)))
+
+class MemoryEfficientSwish(nn.Module):
+ def forward(self, x):
+ return SwishImplementation.apply(x)
+
+
+def round_filters(filters, global_params):
+ """Calculate and round number of filters based on width multiplier.
+ Use width_coefficient, depth_divisor and min_depth of global_params.
+
+ Args:
+ filters (int): Filters number to be calculated.
+ global_params (namedtuple): Global params of the model.
+
+ Returns:
+ new_filters: New filters number after calculating.
+ """
+ multiplier = global_params.width_coefficient
+ if not multiplier:
+ return filters
+ # TODO: modify the params names.
+ # maybe the names (width_divisor,min_width)
+ # are more suitable than (depth_divisor,min_depth).
+ divisor = global_params.depth_divisor
+ min_depth = global_params.min_depth
+ filters *= multiplier
+ min_depth = min_depth or divisor # pay attention to this line when using min_depth
+ # follow the formula transferred from official TensorFlow implementation
+ new_filters = max(min_depth, int(filters + divisor / 2) // divisor * divisor)
+ if new_filters < 0.9 * filters: # prevent rounding by more than 10%
+ new_filters += divisor
+ return int(new_filters)
+
+
+def round_repeats(repeats, global_params):
+ """Calculate module's repeat number of a block based on depth multiplier.
+ Use depth_coefficient of global_params.
+
+ Args:
+ repeats (int): num_repeat to be calculated.
+ global_params (namedtuple): Global params of the model.
+
+ Returns:
+ new repeat: New repeat number after calculating.
+ """
+ multiplier = global_params.depth_coefficient
+ if not multiplier:
+ return repeats
+ # follow the formula transferred from official TensorFlow implementation
+ return int(math.ceil(multiplier * repeats))
+
+
+def drop_connect(inputs, p, training):
+ """Drop connect.
+
+ Args:
+ input (tensor: BCWH): Input of this structure.
+ p (float: 0.0~1.0): Probability of drop connection.
+ training (bool): The running mode.
+
+ Returns:
+ output: Output after drop connection.
+ """
+ assert 0 <= p <= 1, 'p must be in range of [0,1]'
+
+ if not training:
+ return inputs
+
+ batch_size = inputs.shape[0]
+ keep_prob = 1 - p
+
+ # generate binary_tensor mask according to probability (p for 0, 1-p for 1)
+ random_tensor = keep_prob
+ random_tensor += torch.rand([batch_size, 1, 1, 1], dtype=inputs.dtype, device=inputs.device)
+ binary_tensor = torch.floor(random_tensor)
+
+ output = inputs / keep_prob * binary_tensor
+ return output
+
+
+def get_width_and_height_from_size(x):
+ """Obtain height and width from x.
+
+ Args:
+ x (int, tuple or list): Data size.
+
+ Returns:
+ size: A tuple or list (H,W).
+ """
+ if isinstance(x, int):
+ return x, x
+ if isinstance(x, list) or isinstance(x, tuple):
+ return x
+ else:
+ raise TypeError()
+
+
+def calculate_output_image_size(input_image_size, stride):
+ """Calculates the output image size when using Conv2dSamePadding with a stride.
+ Necessary for static padding. Thanks to mannatsingh for pointing this out.
+
+ Args:
+ input_image_size (int, tuple or list): Size of input image.
+ stride (int, tuple or list): Conv2d operation's stride.
+
+ Returns:
+ output_image_size: A list [H,W].
+ """
+ if input_image_size is None:
+ return None
+ image_height, image_width = get_width_and_height_from_size(input_image_size)
+ stride = stride if isinstance(stride, int) else stride[0]
+ image_height = int(math.ceil(image_height / stride))
+ image_width = int(math.ceil(image_width / stride))
+ return [image_height, image_width]
+
+
+# Note:
+# The following 'SamePadding' functions make output size equal ceil(input size/stride).
+# Only when stride equals 1, can the output size be the same as input size.
+# Don't be confused by their function names ! ! !
+
+def get_same_padding_conv2d(image_size=None):
+ """Chooses static padding if you have specified an image size, and dynamic padding otherwise.
+ Static padding is necessary for ONNX exporting of models.
+
+ Args:
+ image_size (int or tuple): Size of the image.
+
+ Returns:
+ Conv2dDynamicSamePadding or Conv2dStaticSamePadding.
+ """
+ if image_size is None:
+ return Conv2dDynamicSamePadding
+ else:
+ return partial(Conv2dStaticSamePadding, image_size=image_size)
+
+
+class Conv2dDynamicSamePadding(nn.Conv2d):
+ """2D Convolutions like TensorFlow, for a dynamic image size.
+ The padding is operated in forward function by calculating dynamically.
+ """
+
+ # Tips for 'SAME' mode padding.
+ # Given the following:
+ # i: width or height
+ # s: stride
+ # k: kernel size
+ # d: dilation
+ # p: padding
+ # Output after Conv2d:
+ # o = floor((i+p-((k-1)*d+1))/s+1)
+ # If o equals i, i = floor((i+p-((k-1)*d+1))/s+1),
+ # => p = (i-1)*s+((k-1)*d+1)-i
+
+ def __init__(self, in_channels, out_channels, kernel_size, stride=1, dilation=1, groups=1, bias=True):
+ super().__init__(in_channels, out_channels, kernel_size, stride, 0, dilation, groups, bias)
+ self.stride = self.stride if len(self.stride) == 2 else [self.stride[0]] * 2
+
+ def forward(self, x):
+ ih, iw = x.size()[-2:]
+ kh, kw = self.weight.size()[-2:]
+ sh, sw = self.stride
+ oh, ow = math.ceil(ih / sh), math.ceil(iw / sw) # change the output size according to stride ! ! !
+ pad_h = max((oh - 1) * self.stride[0] + (kh - 1) * self.dilation[0] + 1 - ih, 0)
+ pad_w = max((ow - 1) * self.stride[1] + (kw - 1) * self.dilation[1] + 1 - iw, 0)
+ if pad_h > 0 or pad_w > 0:
+ x = F.pad(x, [pad_w // 2, pad_w - pad_w // 2, pad_h // 2, pad_h - pad_h // 2])
+ return F.conv2d(x, self.weight, self.bias, self.stride, self.padding, self.dilation, self.groups)
+
+
+class Conv2dStaticSamePadding(nn.Conv2d):
+ """2D Convolutions like TensorFlow's 'SAME' mode, with the given input image size.
+ The padding mudule is calculated in construction function, then used in forward.
+ """
+
+ # With the same calculation as Conv2dDynamicSamePadding
+
+ def __init__(self, in_channels, out_channels, kernel_size, stride=1, image_size=None, **kwargs):
+ super().__init__(in_channels, out_channels, kernel_size, stride, **kwargs)
+ self.stride = self.stride if len(self.stride) == 2 else [self.stride[0]] * 2
+
+ # Calculate padding based on image size and save it
+ assert image_size is not None
+ ih, iw = (image_size, image_size) if isinstance(image_size, int) else image_size
+ kh, kw = self.weight.size()[-2:]
+ sh, sw = self.stride
+ oh, ow = math.ceil(ih / sh), math.ceil(iw / sw)
+ pad_h = max((oh - 1) * self.stride[0] + (kh - 1) * self.dilation[0] + 1 - ih, 0)
+ pad_w = max((ow - 1) * self.stride[1] + (kw - 1) * self.dilation[1] + 1 - iw, 0)
+ if pad_h > 0 or pad_w > 0:
+ self.static_padding = nn.ZeroPad2d((pad_w // 2, pad_w - pad_w // 2,
+ pad_h // 2, pad_h - pad_h // 2))
+ else:
+ self.static_padding = nn.Identity()
+
+ def forward(self, x):
+ x = self.static_padding(x)
+ x = F.conv2d(x, self.weight, self.bias, self.stride, self.padding, self.dilation, self.groups)
+ return x
+
+
+def get_same_padding_maxPool2d(image_size=None):
+ """Chooses static padding if you have specified an image size, and dynamic padding otherwise.
+ Static padding is necessary for ONNX exporting of models.
+
+ Args:
+ image_size (int or tuple): Size of the image.
+
+ Returns:
+ MaxPool2dDynamicSamePadding or MaxPool2dStaticSamePadding.
+ """
+ if image_size is None:
+ return MaxPool2dDynamicSamePadding
+ else:
+ return partial(MaxPool2dStaticSamePadding, image_size=image_size)
+
+
+class MaxPool2dDynamicSamePadding(nn.MaxPool2d):
+ """2D MaxPooling like TensorFlow's 'SAME' mode, with a dynamic image size.
+ The padding is operated in forward function by calculating dynamically.
+ """
+
+ def __init__(self, kernel_size, stride, padding=0, dilation=1, return_indices=False, ceil_mode=False):
+ super().__init__(kernel_size, stride, padding, dilation, return_indices, ceil_mode)
+ self.stride = [self.stride] * 2 if isinstance(self.stride, int) else self.stride
+ self.kernel_size = [self.kernel_size] * 2 if isinstance(self.kernel_size, int) else self.kernel_size
+ self.dilation = [self.dilation] * 2 if isinstance(self.dilation, int) else self.dilation
+
+ def forward(self, x):
+ ih, iw = x.size()[-2:]
+ kh, kw = self.kernel_size
+ sh, sw = self.stride
+ oh, ow = math.ceil(ih / sh), math.ceil(iw / sw)
+ pad_h = max((oh - 1) * self.stride[0] + (kh - 1) * self.dilation[0] + 1 - ih, 0)
+ pad_w = max((ow - 1) * self.stride[1] + (kw - 1) * self.dilation[1] + 1 - iw, 0)
+ if pad_h > 0 or pad_w > 0:
+ x = F.pad(x, [pad_w // 2, pad_w - pad_w // 2, pad_h // 2, pad_h - pad_h // 2])
+ return F.max_pool2d(x, self.kernel_size, self.stride, self.padding,
+ self.dilation, self.ceil_mode, self.return_indices)
+
+class MaxPool2dStaticSamePadding(nn.MaxPool2d):
+ """2D MaxPooling like TensorFlow's 'SAME' mode, with the given input image size.
+ The padding mudule is calculated in construction function, then used in forward.
+ """
+
+ def __init__(self, kernel_size, stride, image_size=None, **kwargs):
+ super().__init__(kernel_size, stride, **kwargs)
+ self.stride = [self.stride] * 2 if isinstance(self.stride, int) else self.stride
+ self.kernel_size = [self.kernel_size] * 2 if isinstance(self.kernel_size, int) else self.kernel_size
+ self.dilation = [self.dilation] * 2 if isinstance(self.dilation, int) else self.dilation
+
+ # Calculate padding based on image size and save it
+ assert image_size is not None
+ ih, iw = (image_size, image_size) if isinstance(image_size, int) else image_size
+ kh, kw = self.kernel_size
+ sh, sw = self.stride
+ oh, ow = math.ceil(ih / sh), math.ceil(iw / sw)
+ pad_h = max((oh - 1) * self.stride[0] + (kh - 1) * self.dilation[0] + 1 - ih, 0)
+ pad_w = max((ow - 1) * self.stride[1] + (kw - 1) * self.dilation[1] + 1 - iw, 0)
+ if pad_h > 0 or pad_w > 0:
+ self.static_padding = nn.ZeroPad2d((pad_w // 2, pad_w - pad_w // 2, pad_h // 2, pad_h - pad_h // 2))
+ else:
+ self.static_padding = nn.Identity()
+
+ def forward(self, x):
+ x = self.static_padding(x)
+ x = F.max_pool2d(x, self.kernel_size, self.stride, self.padding,
+ self.dilation, self.ceil_mode, self.return_indices)
+ return x
+
+
+################################################################################
+### Helper functions for loading model params
+################################################################################
+
+# BlockDecoder: A Class for encoding and decoding BlockArgs
+# efficientnet_params: A function to query compound coefficient
+# get_model_params and efficientnet:
+# Functions to get BlockArgs and GlobalParams for efficientnet
+# url_map and url_map_advprop: Dicts of url_map for pretrained weights
+# load_pretrained_weights: A function to load pretrained weights
+
+class BlockDecoder(object):
+ """Block Decoder for readability,
+ straight from the official TensorFlow repository.
+ """
+
+ @staticmethod
+ def _decode_block_string(block_string):
+ """Get a block through a string notation of arguments.
+
+ Args:
+ block_string (str): A string notation of arguments.
+ Examples: 'r1_k3_s11_e1_i32_o16_se0.25_noskip'.
+
+ Returns:
+ BlockArgs: The namedtuple defined at the top of this file.
+ """
+ assert isinstance(block_string, str)
+
+ ops = block_string.split('_')
+ options = {}
+ for op in ops:
+ splits = re.split(r'(\d.*)', op)
+ if len(splits) >= 2:
+ key, value = splits[:2]
+ options[key] = value
+
+ # Check stride
+ assert (('s' in options and len(options['s']) == 1) or
+ (len(options['s']) == 2 and options['s'][0] == options['s'][1]))
+
+ return BlockArgs(
+ num_repeat=int(options['r']),
+ kernel_size=int(options['k']),
+ stride=[int(options['s'][0])],
+ expand_ratio=int(options['e']),
+ input_filters=int(options['i']),
+ output_filters=int(options['o']),
+ se_ratio=float(options['se']) if 'se' in options else None,
+ id_skip=('noskip' not in block_string))
+
+ @staticmethod
+ def _encode_block_string(block):
+ """Encode a block to a string.
+
+ Args:
+ block (namedtuple): A BlockArgs type argument.
+
+ Returns:
+ block_string: A String form of BlockArgs.
+ """
+ args = [
+ 'r%d' % block.num_repeat,
+ 'k%d' % block.kernel_size,
+ 's%d%d' % (block.strides[0], block.strides[1]),
+ 'e%s' % block.expand_ratio,
+ 'i%d' % block.input_filters,
+ 'o%d' % block.output_filters
+ ]
+ if 0 < block.se_ratio <= 1:
+ args.append('se%s' % block.se_ratio)
+ if block.id_skip is False:
+ args.append('noskip')
+ return '_'.join(args)
+
+ @staticmethod
+ def decode(string_list):
+ """Decode a list of string notations to specify blocks inside the network.
+
+ Args:
+ string_list (list[str]): A list of strings, each string is a notation of block.
+
+ Returns:
+ blocks_args: A list of BlockArgs namedtuples of block args.
+ """
+ assert isinstance(string_list, list)
+ blocks_args = []
+ for block_string in string_list:
+ blocks_args.append(BlockDecoder._decode_block_string(block_string))
+ return blocks_args
+
+ @staticmethod
+ def encode(blocks_args):
+ """Encode a list of BlockArgs to a list of strings.
+
+ Args:
+ blocks_args (list[namedtuples]): A list of BlockArgs namedtuples of block args.
+
+ Returns:
+ block_strings: A list of strings, each string is a notation of block.
+ """
+ block_strings = []
+ for block in blocks_args:
+ block_strings.append(BlockDecoder._encode_block_string(block))
+ return block_strings
+
+
+def efficientnet_params(model_name):
+ """Map EfficientNet model name to parameter coefficients.
+
+ Args:
+ model_name (str): Model name to be queried.
+
+ Returns:
+ params_dict[model_name]: A (width,depth,res,dropout) tuple.
+ """
+ """
+ params_dict = {
+ # Coefficients: width,depth,res,dropout
+ 'efficientnet-b0': (1.0, 1.0, 224, 0.2),
+ 'efficientnet-b1': (1.0, 1.1, 240, 0.2),
+ 'efficientnet-b2': (1.1, 1.2, 260, 0.3),
+ 'efficientnet-b3': (1.2, 1.4, 300, 0.3),
+ 'efficientnet-b4': (1.4, 1.8, 380, 0.4),
+ 'efficientnet-b5': (1.6, 2.2, 456, 0.4),
+ 'efficientnet-b6': (1.8, 2.6, 528, 0.5),
+ 'efficientnet-b7': (2.0, 3.1, 600, 0.5),
+ 'efficientnet-b8': (2.2, 3.6, 672, 0.5),
+ 'efficientnet-l2': (4.3, 5.3, 800, 0.5),
+ }
+ """
+ params_dict = {
+ # Coefficients: width,depth,res,dropout
+ 'efficientnet-b0': (1.0, 1.0, 112, 0.2),
+ 'efficientnet-b1': (1.0, 1.1, 112, 0.2),
+ 'efficientnet-b2': (1.1, 1.2, 112, 0.3),
+ 'efficientnet-b3': (1.2, 1.4, 112, 0.3),
+ 'efficientnet-b4': (1.4, 1.8, 112, 0.4),
+ 'efficientnet-b5': (1.6, 2.2, 112, 0.4),
+ 'efficientnet-b6': (1.8, 2.6, 112, 0.5),
+ 'efficientnet-b7': (2.0, 3.1, 112, 0.5),
+ 'efficientnet-b8': (2.2, 3.6, 112, 0.5),
+ 'efficientnet-l2': (4.3, 5.3, 112, 0.5),
+ }
+ return params_dict[model_name]
+
+
+def efficientnet(width_coefficient=None, depth_coefficient=None, image_size=None,
+ dropout_rate=0.2, drop_connect_rate=0.2, num_classes=1000, include_top=True):
+ """Create BlockArgs and GlobalParams for efficientnet model.
+
+ Args:
+ width_coefficient (float)
+ depth_coefficient (float)
+ image_size (int)
+ dropout_rate (float)
+ drop_connect_rate (float)
+ num_classes (int)
+
+ Meaning as the name suggests.
+
+ Returns:
+ blocks_args, global_params.
+ """
+
+ # Blocks args for the whole model(efficientnet-b0 by default)
+ # It will be modified in the construction of EfficientNet Class according to model
+ blocks_args = [
+ 'r1_k3_s11_e1_i32_o16_se0.25',
+ 'r2_k3_s22_e6_i16_o24_se0.25',
+ 'r2_k5_s22_e6_i24_o40_se0.25',
+ 'r3_k3_s22_e6_i40_o80_se0.25',
+ 'r3_k5_s11_e6_i80_o112_se0.25',
+ 'r4_k5_s22_e6_i112_o192_se0.25',
+ 'r1_k3_s11_e6_i192_o320_se0.25',
+ ]
+ blocks_args = BlockDecoder.decode(blocks_args)
+
+ global_params = GlobalParams(
+ width_coefficient=width_coefficient,
+ depth_coefficient=depth_coefficient,
+ image_size=image_size,
+ dropout_rate=dropout_rate,
+
+ num_classes=num_classes,
+ batch_norm_momentum=0.99,
+ batch_norm_epsilon=1e-3,
+ drop_connect_rate=drop_connect_rate,
+ depth_divisor=8,
+ min_depth=None,
+ include_top=include_top,
+ )
+
+ return blocks_args, global_params
+
+
+def get_model_params(model_name, override_params):
+ """Get the block args and global params for a given model name.
+
+ Args:
+ model_name (str): Model's name.
+ override_params (dict): A dict to modify global_params.
+
+ Returns:
+ blocks_args, global_params
+ """
+ if model_name.startswith('efficientnet'):
+ w, d, s, p = efficientnet_params(model_name)
+ # note: all models have drop connect rate = 0.2
+ blocks_args, global_params = efficientnet(
+ width_coefficient=w, depth_coefficient=d, dropout_rate=p, image_size=s)
+ else:
+ raise NotImplementedError('model name is not pre-defined: {}'.format(model_name))
+ if override_params:
+ # ValueError will be raised here if override_params has fields not included in global_params.
+ global_params = global_params._replace(**override_params)
+ return blocks_args, global_params
+
+
+# train with Standard methods
+# check more details in paper(EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks)
+url_map = {
+ 'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b0-355c32eb.pth',
+ 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b1-f1951068.pth',
+ 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b2-8bb594d6.pth',
+ 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b3-5fb5a3c3.pth',
+ 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b4-6ed6700e.pth',
+ 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b5-b6417697.pth',
+ 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b6-c76e70fd.pth',
+ 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/efficientnet-b7-dcc49843.pth',
+}
+
+# train with Adversarial Examples(AdvProp)
+# check more details in paper(Adversarial Examples Improve Image Recognition)
+url_map_advprop = {
+ 'efficientnet-b0': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b0-b64d5a18.pth',
+ 'efficientnet-b1': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b1-0f3ce85a.pth',
+ 'efficientnet-b2': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b2-6e9d97e5.pth',
+ 'efficientnet-b3': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b3-cdd7c0f4.pth',
+ 'efficientnet-b4': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b4-44fb3a87.pth',
+ 'efficientnet-b5': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b5-86493f6b.pth',
+ 'efficientnet-b6': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b6-ac80338e.pth',
+ 'efficientnet-b7': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b7-4652b6dd.pth',
+ 'efficientnet-b8': 'https://github.com/lukemelas/EfficientNet-PyTorch/releases/download/1.0/adv-efficientnet-b8-22a8fe65.pth',
+}
+
+# TODO: add the petrained weights url map of 'efficientnet-l2'
+
+
+def load_pretrained_weights(model, model_name, weights_path=None, load_fc=True, advprop=False):
+ """Loads pretrained weights from weights path or download using url.
+
+ Args:
+ model (Module): The whole model of efficientnet.
+ model_name (str): Model name of efficientnet.
+ weights_path (None or str):
+ str: path to pretrained weights file on the local disk.
+ None: use pretrained weights downloaded from the Internet.
+ load_fc (bool): Whether to load pretrained weights for fc layer at the end of the model.
+ advprop (bool): Whether to load pretrained weights
+ trained with advprop (valid when weights_path is None).
+ """
+ if isinstance(weights_path, str):
+ state_dict = torch.load(weights_path)
+ else:
+ # AutoAugment or Advprop (different preprocessing)
+ url_map_ = url_map_advprop if advprop else url_map
+ state_dict = model_zoo.load_url(url_map_[model_name])
+
+ if load_fc:
+ ret = model.load_state_dict(state_dict, strict=False)
+ assert not ret.missing_keys, 'Missing keys when loading pretrained weights: {}'.format(ret.missing_keys)
+ else:
+ state_dict.pop('_fc.weight')
+ state_dict.pop('_fc.bias')
+ ret = model.load_state_dict(state_dict, strict=False)
+ assert set(ret.missing_keys) == set(
+ ['_fc.weight', '_fc.bias']), 'Missing keys when loading pretrained weights: {}'.format(ret.missing_keys)
+ assert not ret.unexpected_keys, 'Missing keys when loading pretrained weights: {}'.format(ret.unexpected_keys)
+
+ print('Loaded pretrained weights for {}'.format(model_name))
+
+class Flatten(Module):
+ def forward(self, input):
+ return input.reshape(input.size(0), -1)
+
+######################################### backbone ##################################################################
+
+VALID_MODELS = (
+ 'efficientnet-b0', 'efficientnet-b1', 'efficientnet-b2', 'efficientnet-b3',
+ 'efficientnet-b4', 'efficientnet-b5', 'efficientnet-b6', 'efficientnet-b7',
+ 'efficientnet-b8',
+
+ # Support the construction of 'efficientnet-l2' without pretrained weights
+ 'efficientnet-l2'
+)
+
+
+class MBConvBlock(nn.Module):
+ """Mobile Inverted Residual Bottleneck Block.
+
+ Args:
+ block_args (namedtuple): BlockArgs, defined in utils.py.
+ global_params (namedtuple): GlobalParam, defined in utils.py.
+ image_size (tuple or list): [image_height, image_width].
+
+ References:
+ [1] https://arxiv.org/abs/1704.04861 (MobileNet v1)
+ [2] https://arxiv.org/abs/1801.04381 (MobileNet v2)
+ [3] https://arxiv.org/abs/1905.02244 (MobileNet v3)
+ """
+
+ def __init__(self, block_args, global_params, image_size=None):
+ super().__init__()
+ self._block_args = block_args
+ self._bn_mom = 1 - global_params.batch_norm_momentum # pytorch's difference from tensorflow
+ self._bn_eps = global_params.batch_norm_epsilon
+ self.has_se = (self._block_args.se_ratio is not None) and (0 < self._block_args.se_ratio <= 1)
+ self.id_skip = block_args.id_skip # whether to use skip connection and drop connect
+
+ # Expansion phase (Inverted Bottleneck)
+ inp = self._block_args.input_filters # number of input channels
+ oup = self._block_args.input_filters * self._block_args.expand_ratio # number of output channels
+ if self._block_args.expand_ratio != 1:
+ Conv2d = get_same_padding_conv2d(image_size=image_size)
+ self._expand_conv = Conv2d(in_channels=inp, out_channels=oup, kernel_size=1, bias=False)
+ self._bn0 = nn.BatchNorm2d(num_features=oup, momentum=self._bn_mom, eps=self._bn_eps)
+ # image_size = calculate_output_image_size(image_size, 1) <-- this wouldn't modify image_size
+
+ # Depthwise convolution phase
+ k = self._block_args.kernel_size
+ s = self._block_args.stride
+ Conv2d = get_same_padding_conv2d(image_size=image_size)
+ self._depthwise_conv = Conv2d(
+ in_channels=oup, out_channels=oup, groups=oup, # groups makes it depthwise
+ kernel_size=k, stride=s, bias=False)
+ self._bn1 = nn.BatchNorm2d(num_features=oup, momentum=self._bn_mom, eps=self._bn_eps)
+ image_size = calculate_output_image_size(image_size, s)
+
+ # Squeeze and Excitation layer, if desired
+ if self.has_se:
+ Conv2d = get_same_padding_conv2d(image_size=(1, 1))
+ num_squeezed_channels = max(1, int(self._block_args.input_filters * self._block_args.se_ratio))
+ self._se_reduce = Conv2d(in_channels=oup, out_channels=num_squeezed_channels, kernel_size=1)
+ self._se_expand = Conv2d(in_channels=num_squeezed_channels, out_channels=oup, kernel_size=1)
+
+ # Pointwise convolution phase
+ final_oup = self._block_args.output_filters
+ Conv2d = get_same_padding_conv2d(image_size=image_size)
+ self._project_conv = Conv2d(in_channels=oup, out_channels=final_oup, kernel_size=1, bias=False)
+ self._bn2 = nn.BatchNorm2d(num_features=final_oup, momentum=self._bn_mom, eps=self._bn_eps)
+ self._swish = MemoryEfficientSwish()
+
+ def forward(self, inputs, drop_connect_rate=None):
+ """MBConvBlock's forward function.
+
+ Args:
+ inputs (tensor): Input tensor.
+ drop_connect_rate (bool): Drop connect rate (float, between 0 and 1).
+
+ Returns:
+ Output of this block after processing.
+ """
+
+ # Expansion and Depthwise Convolution
+ x = inputs
+ if self._block_args.expand_ratio != 1:
+ x = self._expand_conv(inputs)
+ x = self._bn0(x)
+ x = self._swish(x)
+
+ x = self._depthwise_conv(x)
+ x = self._bn1(x)
+ x = self._swish(x)
+
+ # Squeeze and Excitation
+ if self.has_se:
+ x_squeezed = F.adaptive_avg_pool2d(x, 1)
+ x_squeezed = self._se_reduce(x_squeezed)
+ x_squeezed = self._swish(x_squeezed)
+ x_squeezed = self._se_expand(x_squeezed)
+ x = torch.sigmoid(x_squeezed) * x
+
+ # Pointwise Convolution
+ x = self._project_conv(x)
+ x = self._bn2(x)
+
+ # Skip connection and drop connect
+ input_filters, output_filters = self._block_args.input_filters, self._block_args.output_filters
+ if self.id_skip and self._block_args.stride == 1 and input_filters == output_filters:
+ # The combination of skip connection and drop connect brings about stochastic depth.
+ if drop_connect_rate:
+ x = drop_connect(x, p=drop_connect_rate, training=self.training)
+ x = x + inputs # skip connection
+ return x
+
+ def set_swish(self, memory_efficient=True):
+ """Sets swish function as memory efficient (for training) or standard (for export).
+
+ Args:
+ memory_efficient (bool): Whether to use memory-efficient version of swish.
+ """
+ self._swish = MemoryEfficientSwish() if memory_efficient else Swish()
+
+
+class EfficientNet(nn.Module):
+ """EfficientNet model.
+ Most easily loaded with the .from_name or .from_pretrained methods.
+
+ Args:
+ blocks_args (list[namedtuple]): A list of BlockArgs to construct blocks.
+ global_params (namedtuple): A set of GlobalParams shared between blocks.
+
+ References:
+ [1] https://arxiv.org/abs/1905.11946 (EfficientNet)
+
+ Example:
+
+
+ import torch
+ >>> from efficientnet.model import EfficientNet
+ >>> inputs = torch.rand(1, 3, 224, 224)
+ >>> model = EfficientNet.from_pretrained('efficientnet-b0')
+ >>> model.eval()
+ >>> outputs = model(inputs)
+ """
+
+ def __init__(self, out_h, out_w, feat_dim, blocks_args=None, global_params=None):
+ super().__init__()
+ assert isinstance(blocks_args, list), 'blocks_args should be a list'
+ assert len(blocks_args) > 0, 'block args must be greater than 0'
+ self._global_params = global_params
+ self._blocks_args = blocks_args
+
+ # Batch norm parameters
+ bn_mom = 1 - self._global_params.batch_norm_momentum
+ bn_eps = self._global_params.batch_norm_epsilon
+
+ # Get stem static or dynamic convolution depending on image size
+ image_size = global_params.image_size
+ Conv2d = get_same_padding_conv2d(image_size=image_size)
+
+ # Stem
+ in_channels = 3 # rgb
+ out_channels = round_filters(32, self._global_params) # number of output channels
+ #self._conv_stem = Conv2d(in_channels, out_channels, kernel_size=3, stride=2, bias=False)
+ self._conv_stem = Conv2d(in_channels, out_channels, kernel_size=3, stride=1, bias=False)
+ self._bn0 = nn.BatchNorm2d(num_features=out_channels, momentum=bn_mom, eps=bn_eps)
+ image_size = calculate_output_image_size(image_size, 2)
+
+ # Build blocks
+ self._blocks = nn.ModuleList([])
+ for block_args in self._blocks_args:
+
+ # Update block input and output filters based on depth multiplier.
+ block_args = block_args._replace(
+ input_filters=round_filters(block_args.input_filters, self._global_params),
+ output_filters=round_filters(block_args.output_filters, self._global_params),
+ num_repeat=round_repeats(block_args.num_repeat, self._global_params)
+ )
+
+ # The first block needs to take care of stride and filter size increase.
+ self._blocks.append(MBConvBlock(block_args, self._global_params, image_size=image_size))
+ image_size = calculate_output_image_size(image_size, block_args.stride)
+ if block_args.num_repeat > 1: # modify block_args to keep same output size
+ block_args = block_args._replace(input_filters=block_args.output_filters, stride=1)
+ for _ in range(block_args.num_repeat - 1):
+ self._blocks.append(MBConvBlock(block_args, self._global_params, image_size=image_size))
+ # image_size = calculate_output_image_size(image_size, block_args.stride) # stride = 1
+
+ # Head
+ in_channels = block_args.output_filters # output of final block
+ out_channels = round_filters(1280, self._global_params)
+ #out_channels = round_filters(512, self._global_params)
+ Conv2d = get_same_padding_conv2d(image_size=image_size)
+ self._conv_head = Conv2d(in_channels, out_channels, kernel_size=1, bias=False)
+ self._bn1 = nn.BatchNorm2d(num_features=out_channels, momentum=bn_mom, eps=bn_eps)
+
+ # Final linear layer
+ self._avg_pooling = nn.AdaptiveAvgPool2d(1)
+ self._dropout = nn.Dropout(self._global_params.dropout_rate)
+ self._fc = nn.Linear(out_channels, self._global_params.num_classes)
+ self._swish = MemoryEfficientSwish()
+ self.output_layer = Sequential(
+ BatchNorm2d(1280),
+ #BatchNorm2d(512),
+ Dropout(self._global_params.dropout_rate),
+ Flatten(),
+ Linear(1280 * out_h * out_w, feat_dim),
+ #Linear(512 * out_h * out_w, feat_dim),
+ BatchNorm1d(feat_dim))
+
+
+ def set_swish(self, memory_efficient=True):
+ """Sets swish function as memory efficient (for training) or standard (for export).
+
+ Args:
+ memory_efficient (bool): Whether to use memory-efficient version of swish.
+
+ """
+ self._swish = MemoryEfficientSwish() if memory_efficient else Swish()
+ for block in self._blocks:
+ block.set_swish(memory_efficient)
+
+ def extract_endpoints(self, inputs):
+ """Use convolution layer to extract features
+ from reduction levels i in [1, 2, 3, 4, 5].
+
+ Args:
+ inputs (tensor): Input tensor.
+
+ Returns:
+ Dictionary of last intermediate features
+ with reduction levels i in [1, 2, 3, 4, 5].
+ Example:
+ >>> import torch
+ >>> from efficientnet.model import EfficientNet
+ >>> inputs = torch.rand(1, 3, 224, 224)
+ >>> model = EfficientNet.from_pretrained('efficientnet-b0')
+ >>> endpoints = model.extract_endpoints(inputs)
+ >>> print(endpoints['reduction_1'].shape) # torch.Size([1, 16, 112, 112])
+ >>> print(endpoints['reduction_2'].shape) # torch.Size([1, 24, 56, 56])
+ >>> print(endpoints['reduction_3'].shape) # torch.Size([1, 40, 28, 28])
+ >>> print(endpoints['reduction_4'].shape) # torch.Size([1, 112, 14, 14])
+ >>> print(endpoints['reduction_5'].shape) # torch.Size([1, 1280, 7, 7])
+ """
+ endpoints = dict()
+
+ # Stem
+ x = self._swish(self._bn0(self._conv_stem(inputs)))
+ prev_x = x
+
+ # Blocks
+ for idx, block in enumerate(self._blocks):
+ drop_connect_rate = self._global_params.drop_connect_rate
+ if drop_connect_rate:
+ drop_connect_rate *= float(idx) / len(self._blocks) # scale drop connect_rate
+ x = block(x, drop_connect_rate=drop_connect_rate)
+ if prev_x.size(2) > x.size(2):
+ endpoints['reduction_{}'.format(len(endpoints)+1)] = prev_x
+ prev_x = x
+
+ # Head
+ x = self._swish(self._bn1(self._conv_head(x)))
+ endpoints['reduction_{}'.format(len(endpoints)+1)] = x
+
+ return endpoints
+
+ def extract_features(self, inputs):
+ """use convolution layer to extract feature .
+
+ Args:
+ inputs (tensor): Input tensor.
+
+ Returns:
+ Output of the final convolution
+ layer in the efficientnet model.
+ """
+ # Stem
+ x = self._swish(self._bn0(self._conv_stem(inputs)))
+ # Blocks
+ for idx, block in enumerate(self._blocks):
+ drop_connect_rate = self._global_params.drop_connect_rate
+ if drop_connect_rate:
+ drop_connect_rate *= float(idx) / len(self._blocks) # scale drop connect_rate
+ x = block(x, drop_connect_rate=drop_connect_rate)
+
+ # Head
+ x = self._swish(self._bn1(self._conv_head(x)))
+
+ return x
+
+ def forward(self, inputs):
+ """EfficientNet's forward function.
+ Calls extract_features to extract features, applies final linear layer, and returns logits.
+
+ Args:
+ inputs (tensor): Input tensor.
+
+ Returns:
+ Output of this model after processing.
+ """
+ # Convolution layers
+ x = self.extract_features(inputs)
+ '''
+ # Pooling and final linear layer
+ x = self._avg_pooling(x)
+ if self._global_params.include_top:
+ x = x.flatten(start_dim=1)
+ x = self._dropout(x)
+ #x = self._fc(x)
+ '''
+ x = self.output_layer(x)
+ return x
+
+ @classmethod
+ def from_name(cls, model_name, in_channels=3, **override_params):
+ """create an efficientnet model according to name.
+
+ Args:
+ model_name (str): Name for efficientnet.
+ in_channels (int): Input data's channel number.
+ override_params (other key word params):
+ Params to override model's global_params.
+ Optional key:
+ 'width_coefficient', 'depth_coefficient',
+ 'image_size', 'dropout_rate',
+ 'num_classes', 'batch_norm_momentum',
+ 'batch_norm_epsilon', 'drop_connect_rate',
+ 'depth_divisor', 'min_depth'
+
+ Returns:
+ An efficientnet model.
+ """
+ cls._check_model_name_is_valid(model_name)
+ blocks_args, global_params = get_model_params(model_name, override_params)
+ model = cls(blocks_args, global_params)
+ model._change_in_channels(in_channels)
+ return model
+
+ @classmethod
+ def from_pretrained(cls, model_name, weights_path=None, advprop=False,
+ in_channels=3, num_classes=1000, **override_params):
+ """create an efficientnet model according to name.
+
+ Args:
+ model_name (str): Name for efficientnet.
+ weights_path (None or str):
+ str: path to pretrained weights file on the local disk.
+ None: use pretrained weights downloaded from the Internet.
+ advprop (bool):
+ Whether to load pretrained weights
+ trained with advprop (valid when weights_path is None).
+ in_channels (int): Input data's channel number.
+ num_classes (int):
+ Number of categories for classification.
+ It controls the output size for final linear layer.
+ override_params (other key word params):
+ Params to override model's global_params.
+ Optional key:
+ 'width_coefficient', 'depth_coefficient',
+ 'image_size', 'dropout_rate',
+ 'batch_norm_momentum',
+ 'batch_norm_epsilon', 'drop_connect_rate',
+ 'depth_divisor', 'min_depth'
+
+ Returns:
+ A pretrained efficientnet model.
+ """
+ model = cls.from_name(model_name, num_classes=num_classes, **override_params)
+ load_pretrained_weights(model, model_name, weights_path=weights_path, load_fc=(num_classes == 1000), advprop=advprop)
+ model._change_in_channels(in_channels)
+ return model
+
+ @classmethod
+ def get_image_size(cls, model_name):
+ """Get the input image size for a given efficientnet model.
+
+ Args:
+ model_name (str): Name for efficientnet.
+
+ Returns:
+ Input image size (resolution).
+ """
+ cls._check_model_name_is_valid(model_name)
+ _, _, res, _ = efficientnet_params(model_name)
+ return res
+
+ @classmethod
+ def _check_model_name_is_valid(cls, model_name):
+ """Validates model name.
+
+ Args:
+ model_name (str): Name for efficientnet.
+
+ Returns:
+ bool: Is a valid name or not.
+ """
+ if model_name not in VALID_MODELS:
+ raise ValueError('model_name should be one of: ' + ', '.join(VALID_MODELS))
+
+ def _change_in_channels(self, in_channels):
+ """Adjust model's first convolution layer to in_channels, if in_channels not equals 3.
+
+ Args:
+ in_channels (int): Input data's channel number.
+ """
+ if in_channels != 3:
+ Conv2d = get_same_padding_conv2d(image_size=self._global_params.image_size)
+ out_channels = round_filters(32, self._global_params)
+ self._conv_stem = Conv2d(in_channels, out_channels, kernel_size=3, stride=2, bias=False)
+
diff --git a/bob/bio/facexzoo/backbones/GhostNet.py b/bob/bio/facexzoo/backbones/GhostNet.py
new file mode 100644
index 0000000000000000000000000000000000000000..d96462084720b0e9bd96f2c70f85db4db07a6610
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/GhostNet.py
@@ -0,0 +1,248 @@
+"""
+@author: Jun Wang
+@date: 20210121
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/huawei-noah/ghostnet/blob/master/ghostnet_pytorch/ghostnet.py
+
+# 2020.06.09-Changed for building GhostNet
+# Huawei Technologies Co., Ltd. <foss@huawei.com>
+"""
+Creates a GhostNet Model as defined in:
+GhostNet: More Features from Cheap Operations By Kai Han, Yunhe Wang, Qi Tian, Jianyuan Guo, Chunjing Xu, Chang Xu.
+https://arxiv.org/abs/1911.11907
+Modified from https://github.com/d-li14/mobilenetv3.pytorch and https://github.com/rwightman/pytorch-image-models
+"""
+import math
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from torch.nn import Sequential, BatchNorm2d, Dropout, Module, Linear, BatchNorm1d
+
+__all__ = ['ghost_net']
+
+
+class Flatten(Module):
+ def forward(self, input):
+ return input.reshape(input.size(0), -1)
+
+def _make_divisible(v, divisor, min_value=None):
+ """
+ This function is taken from the original tf repo.
+ It ensures that all layers have a channel number that is divisible by 8
+ It can be seen here:
+ https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet.py
+ """
+ if min_value is None:
+ min_value = divisor
+ new_v = max(min_value, int(v + divisor / 2) // divisor * divisor)
+ # Make sure that round down does not go down by more than 10%.
+ if new_v < 0.9 * v:
+ new_v += divisor
+ return new_v
+
+
+def hard_sigmoid(x, inplace: bool = False):
+ if inplace:
+ return x.add_(3.).clamp_(0., 6.).div_(6.)
+ else:
+ return F.relu6(x + 3.) / 6.
+
+
+class SqueezeExcite(nn.Module):
+ def __init__(self, in_chs, se_ratio=0.25, reduced_base_chs=None,
+ act_layer=nn.ReLU, gate_fn=hard_sigmoid, divisor=4, **_):
+ super(SqueezeExcite, self).__init__()
+ self.gate_fn = gate_fn
+ reduced_chs = _make_divisible((reduced_base_chs or in_chs) * se_ratio, divisor)
+ self.avg_pool = nn.AdaptiveAvgPool2d(1)
+ self.conv_reduce = nn.Conv2d(in_chs, reduced_chs, 1, bias=True)
+ self.act1 = act_layer(inplace=True)
+ self.conv_expand = nn.Conv2d(reduced_chs, in_chs, 1, bias=True)
+
+ def forward(self, x):
+ x_se = self.avg_pool(x)
+ x_se = self.conv_reduce(x_se)
+ x_se = self.act1(x_se)
+ x_se = self.conv_expand(x_se)
+ x = x * self.gate_fn(x_se)
+ return x
+
+
+class ConvBnAct(nn.Module):
+ def __init__(self, in_chs, out_chs, kernel_size,
+ stride=1, act_layer=nn.ReLU):
+ super(ConvBnAct, self).__init__()
+ self.conv = nn.Conv2d(in_chs, out_chs, kernel_size, stride, kernel_size//2, bias=False)
+ self.bn1 = nn.BatchNorm2d(out_chs)
+ self.act1 = act_layer(inplace=True)
+
+ def forward(self, x):
+ x = self.conv(x)
+ x = self.bn1(x)
+ x = self.act1(x)
+ return x
+
+
+class GhostModule(nn.Module):
+ def __init__(self, inp, oup, kernel_size=1, ratio=2, dw_size=3, stride=1, relu=True):
+ super(GhostModule, self).__init__()
+ self.oup = oup
+ init_channels = math.ceil(oup / ratio)
+ new_channels = init_channels*(ratio-1)
+
+ self.primary_conv = nn.Sequential(
+ nn.Conv2d(inp, init_channels, kernel_size, stride, kernel_size//2, bias=False),
+ nn.BatchNorm2d(init_channels),
+ nn.ReLU(inplace=True) if relu else nn.Sequential(),
+ )
+
+ self.cheap_operation = nn.Sequential(
+ nn.Conv2d(init_channels, new_channels, dw_size, 1, dw_size//2, groups=init_channels, bias=False),
+ nn.BatchNorm2d(new_channels),
+ nn.ReLU(inplace=True) if relu else nn.Sequential(),
+ )
+
+ def forward(self, x):
+ x1 = self.primary_conv(x)
+ x2 = self.cheap_operation(x1)
+ out = torch.cat([x1,x2], dim=1)
+ return out[:,:self.oup,:,:]
+
+
+class GhostBottleneck(nn.Module):
+ """ Ghost bottleneck w/ optional SE"""
+
+ def __init__(self, in_chs, mid_chs, out_chs, dw_kernel_size=3,
+ stride=1, act_layer=nn.ReLU, se_ratio=0.):
+ super(GhostBottleneck, self).__init__()
+ has_se = se_ratio is not None and se_ratio > 0.
+ self.stride = stride
+
+ # Point-wise expansion
+ self.ghost1 = GhostModule(in_chs, mid_chs, relu=True)
+
+ # Depth-wise convolution
+ if self.stride > 1:
+ self.conv_dw = nn.Conv2d(mid_chs, mid_chs, dw_kernel_size, stride=stride,
+ padding=(dw_kernel_size-1)//2,
+ groups=mid_chs, bias=False)
+ self.bn_dw = nn.BatchNorm2d(mid_chs)
+
+ # Squeeze-and-excitation
+ if has_se:
+ self.se = SqueezeExcite(mid_chs, se_ratio=se_ratio)
+ else:
+ self.se = None
+
+ # Point-wise linear projection
+ self.ghost2 = GhostModule(mid_chs, out_chs, relu=False)
+
+ # shortcut
+ if (in_chs == out_chs and self.stride == 1):
+ self.shortcut = nn.Sequential()
+ else:
+ self.shortcut = nn.Sequential(
+ nn.Conv2d(in_chs, in_chs, dw_kernel_size, stride=stride,
+ padding=(dw_kernel_size-1)//2, groups=in_chs, bias=False),
+ nn.BatchNorm2d(in_chs),
+ nn.Conv2d(in_chs, out_chs, 1, stride=1, padding=0, bias=False),
+ nn.BatchNorm2d(out_chs),
+ )
+
+
+ def forward(self, x):
+ residual = x
+
+ # 1st ghost bottleneck
+ x = self.ghost1(x)
+
+ # Depth-wise convolution
+ if self.stride > 1:
+ x = self.conv_dw(x)
+ x = self.bn_dw(x)
+
+ # Squeeze-and-excitation
+ if self.se is not None:
+ x = self.se(x)
+
+ # 2nd ghost bottleneck
+ x = self.ghost2(x)
+
+ x += self.shortcut(residual)
+ return x
+
+
+class GhostNet(nn.Module):
+ def __init__(self, width=1.0, drop_ratio=0.2, feat_dim=512, out_h=7, out_w=7):
+ super(GhostNet, self).__init__()
+ # setting of inverted residual blocks
+ self.cfgs = [
+ # k, t, c, SE, s
+ # stage1
+ [[3, 16, 16, 0, 1]],
+ # stage2
+ [[3, 48, 24, 0, 2]],
+ [[3, 72, 24, 0, 1]],
+ # stage3
+ [[5, 72, 40, 0.25, 2]],
+ [[5, 120, 40, 0.25, 1]],
+ # stage4
+ [[3, 240, 80, 0, 2]],
+ [[3, 200, 80, 0, 1],
+ [3, 184, 80, 0, 1],
+ [3, 184, 80, 0, 1],
+ [3, 480, 112, 0.25, 1],
+ [3, 672, 112, 0.25, 1]
+ ],
+ # stage5
+ [[5, 672, 160, 0.25, 2]],
+ [[5, 960, 160, 0, 1],
+ [5, 960, 160, 0.25, 1],
+ [5, 960, 160, 0, 1],
+ [5, 960, 160, 0.25, 1]
+ ]
+ ]
+
+ # building first layer
+ output_channel = _make_divisible(16 * width, 4)
+ #self.conv_stem = nn.Conv2d(3, output_channel, 3, 2, 1, bias=False)
+ self.conv_stem = nn.Conv2d(3, output_channel, 3, 1, 1, bias=False)
+ self.bn1 = nn.BatchNorm2d(output_channel)
+ self.act1 = nn.ReLU(inplace=True)
+ input_channel = output_channel
+
+ # building inverted residual blocks
+ stages = []
+ block = GhostBottleneck
+ for cfg in self.cfgs:
+ layers = []
+ for k, exp_size, c, se_ratio, s in cfg:
+ output_channel = _make_divisible(c * width, 4)
+ hidden_channel = _make_divisible(exp_size * width, 4)
+ layers.append(block(input_channel, hidden_channel, output_channel, k, s,
+ se_ratio=se_ratio))
+ input_channel = output_channel
+ stages.append(nn.Sequential(*layers))
+
+ output_channel = _make_divisible(exp_size * width, 4)
+ stages.append(nn.Sequential(ConvBnAct(input_channel, output_channel, 1)))
+ input_channel = output_channel
+
+ self.blocks = nn.Sequential(*stages)
+
+ self.output_layer = Sequential(BatchNorm2d(960),
+ Dropout(drop_ratio),
+ Flatten(),
+ Linear(960 * out_h * out_w, feat_dim), # for eye
+ BatchNorm1d(feat_dim))
+
+ def forward(self, x):
+ x = self.conv_stem(x)
+ x = self.bn1(x)
+ x = self.act1(x)
+ x = self.blocks(x)
+ x = self.output_layer(x)
+ return x
diff --git a/bob/bio/facexzoo/backbones/HRNet.py b/bob/bio/facexzoo/backbones/HRNet.py
new file mode 100644
index 0000000000000000000000000000000000000000..46b2d3afd6de37aede837e9cc56158658845c225
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/HRNet.py
@@ -0,0 +1,527 @@
+"""
+@author: Hanbin Dai, Jun Wang
+@date: 20201020
+@contact: daihanbin.ac@gmail.com, jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/HRNet/HRNet-Image-Classification/blob/master/lib/models/cls_hrnet.py
+
+import os
+import logging
+import functools
+
+import numpy as np
+
+import torch
+import torch.nn as nn
+import torch._utils
+import torch.nn.functional as F
+from torch.nn import Sequential, Module, Linear, BatchNorm1d
+
+BN_MOMENTUM = 0.1
+logger = logging.getLogger(__name__)
+
+class Flatten(Module):
+ def forward(self, x):
+ return x.reshape(x.size(0), -1)
+
+def conv3x3(in_planes, out_planes, stride=1):
+ """3x3 convolution with padding"""
+ return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
+ padding=1, bias=False)
+
+
+class BasicBlock(nn.Module):
+ expansion = 1
+
+ def __init__(self, inplanes, planes, stride=1, downsample=None):
+ super(BasicBlock, self).__init__()
+ self.conv1 = conv3x3(inplanes, planes, stride)
+ self.bn1 = nn.BatchNorm2d(planes, momentum=BN_MOMENTUM)
+ self.relu = nn.ReLU(inplace=True)
+ self.conv2 = conv3x3(planes, planes)
+ self.bn2 = nn.BatchNorm2d(planes, momentum=BN_MOMENTUM)
+ self.downsample = downsample
+ self.stride = stride
+
+ def forward(self, x):
+ residual = x
+
+ out = self.conv1(x)
+ out = self.bn1(out)
+ out = self.relu(out)
+
+ out = self.conv2(out)
+ out = self.bn2(out)
+
+ if self.downsample is not None:
+ residual = self.downsample(x)
+
+ out += residual
+ out = self.relu(out)
+
+ return out
+
+
+class Bottleneck(nn.Module):
+ expansion = 4
+
+ def __init__(self, inplanes, planes, stride=1, downsample=None):
+ super(Bottleneck, self).__init__()
+ self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
+ self.bn1 = nn.BatchNorm2d(planes, momentum=BN_MOMENTUM)
+ self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
+ padding=1, bias=False)
+ self.bn2 = nn.BatchNorm2d(planes, momentum=BN_MOMENTUM)
+ self.conv3 = nn.Conv2d(planes, planes * self.expansion, kernel_size=1,
+ bias=False)
+ self.bn3 = nn.BatchNorm2d(planes * self.expansion,
+ momentum=BN_MOMENTUM)
+ self.relu = nn.ReLU(inplace=True)
+ self.downsample = downsample
+ self.stride = stride
+
+ def forward(self, x):
+ residual = x
+
+ out = self.conv1(x)
+ out = self.bn1(out)
+ out = self.relu(out)
+
+ out = self.conv2(out)
+ out = self.bn2(out)
+ out = self.relu(out)
+
+ out = self.conv3(out)
+ out = self.bn3(out)
+
+ if self.downsample is not None:
+ residual = self.downsample(x)
+
+ out += residual
+ out = self.relu(out)
+
+ return out
+
+
+class HighResolutionModule(nn.Module):
+ def __init__(self, num_branches, blocks, num_blocks, num_inchannels,
+ num_channels, fuse_method, multi_scale_output=True):
+ super(HighResolutionModule, self).__init__()
+ self._check_branches(
+ num_branches, blocks, num_blocks, num_inchannels, num_channels)
+
+ self.num_inchannels = num_inchannels
+ self.fuse_method = fuse_method
+ self.num_branches = num_branches
+
+ self.multi_scale_output = multi_scale_output
+
+ self.branches = self._make_branches(
+ num_branches, blocks, num_blocks, num_channels)
+ self.fuse_layers = self._make_fuse_layers()
+ self.relu = nn.ReLU(False)
+
+ def _check_branches(self, num_branches, blocks, num_blocks,
+ num_inchannels, num_channels):
+ if num_branches != len(num_blocks):
+ error_msg = 'NUM_BRANCHES({}) <> NUM_BLOCKS({})'.format(
+ num_branches, len(num_blocks))
+ logger.error(error_msg)
+ raise ValueError(error_msg)
+
+ if num_branches != len(num_channels):
+ error_msg = 'NUM_BRANCHES({}) <> NUM_CHANNELS({})'.format(
+ num_branches, len(num_channels))
+ logger.error(error_msg)
+ raise ValueError(error_msg)
+
+ if num_branches != len(num_inchannels):
+ error_msg = 'NUM_BRANCHES({}) <> NUM_INCHANNELS({})'.format(
+ num_branches, len(num_inchannels))
+ logger.error(error_msg)
+ raise ValueError(error_msg)
+
+ def _make_one_branch(self, branch_index, block, num_blocks, num_channels,
+ stride=1):
+ downsample = None
+ if stride != 1 or \
+ self.num_inchannels[branch_index] != num_channels[branch_index] * block.expansion:
+ downsample = nn.Sequential(
+ nn.Conv2d(self.num_inchannels[branch_index],
+ num_channels[branch_index] * block.expansion,
+ kernel_size=1, stride=stride, bias=False),
+ nn.BatchNorm2d(num_channels[branch_index] * block.expansion,
+ momentum=BN_MOMENTUM),
+ )
+
+ layers = []
+ layers.append(block(self.num_inchannels[branch_index],
+ num_channels[branch_index], stride, downsample))
+ self.num_inchannels[branch_index] = \
+ num_channels[branch_index] * block.expansion
+ for i in range(1, num_blocks[branch_index]):
+ layers.append(block(self.num_inchannels[branch_index],
+ num_channels[branch_index]))
+
+ return nn.Sequential(*layers)
+
+ def _make_branches(self, num_branches, block, num_blocks, num_channels):
+ branches = []
+
+ for i in range(num_branches):
+ branches.append(
+ self._make_one_branch(i, block, num_blocks, num_channels))
+
+ return nn.ModuleList(branches)
+
+ def _make_fuse_layers(self):
+ if self.num_branches == 1:
+ return None
+
+ num_branches = self.num_branches
+ num_inchannels = self.num_inchannels
+ fuse_layers = []
+ for i in range(num_branches if self.multi_scale_output else 1):
+ fuse_layer = []
+ for j in range(num_branches):
+ if j > i:
+ fuse_layer.append(nn.Sequential(
+ nn.Conv2d(num_inchannels[j],
+ num_inchannels[i],
+ 1,
+ 1,
+ 0,
+ bias=False),
+ nn.BatchNorm2d(num_inchannels[i],
+ momentum=BN_MOMENTUM),
+ nn.Upsample(scale_factor=2**(j-i), mode='nearest')))
+ elif j == i:
+ fuse_layer.append(None)
+ else:
+ conv3x3s = []
+ for k in range(i-j):
+ if k == i - j - 1:
+ num_outchannels_conv3x3 = num_inchannels[i]
+ conv3x3s.append(nn.Sequential(
+ nn.Conv2d(num_inchannels[j],
+ num_outchannels_conv3x3,
+ 3, 2, 1, bias=False),
+ nn.BatchNorm2d(num_outchannels_conv3x3,
+ momentum=BN_MOMENTUM)))
+ else:
+ num_outchannels_conv3x3 = num_inchannels[j]
+ conv3x3s.append(nn.Sequential(
+ nn.Conv2d(num_inchannels[j],
+ num_outchannels_conv3x3,
+ 3, 2, 1, bias=False),
+ nn.BatchNorm2d(num_outchannels_conv3x3,
+ momentum=BN_MOMENTUM),
+ nn.ReLU(False)))
+ fuse_layer.append(nn.Sequential(*conv3x3s))
+ fuse_layers.append(nn.ModuleList(fuse_layer))
+
+ return nn.ModuleList(fuse_layers)
+
+ def get_num_inchannels(self):
+ return self.num_inchannels
+
+ def forward(self, x):
+ if self.num_branches == 1:
+ return [self.branches[0](x[0])]
+
+ for i in range(self.num_branches):
+ x[i] = self.branches[i](x[i])
+
+ x_fuse = []
+ for i in range(len(self.fuse_layers)):
+ y = x[0] if i == 0 else self.fuse_layers[i][0](x[0])
+ for j in range(1, self.num_branches):
+ if i == j:
+ y = y + x[j]
+ else:
+ y = y + self.fuse_layers[i][j](x[j])
+ x_fuse.append(self.relu(y))
+
+ return x_fuse
+
+
+blocks_dict = {
+ 'BASIC': BasicBlock,
+ 'BOTTLENECK': Bottleneck
+}
+
+
+class HighResolutionNet(nn.Module):
+
+ def __init__(self, cfg, **kwargs):
+ super(HighResolutionNet, self).__init__()
+
+ self.conv1 = nn.Conv2d(3, 64, kernel_size=3, stride=2, padding=1,
+ bias=False)
+ self.bn1 = nn.BatchNorm2d(64, momentum=BN_MOMENTUM)
+ #self.conv2 = nn.Conv2d(64, 64, kernel_size=3, stride=2, padding=1, bias=False)
+ self.conv2 = nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1, bias=False)
+
+ self.bn2 = nn.BatchNorm2d(64, momentum=BN_MOMENTUM)
+ self.relu = nn.ReLU(inplace=True)
+
+ self.stage1_cfg = cfg['MODEL']['EXTRA']['STAGE1']
+ num_channels = self.stage1_cfg['NUM_CHANNELS'][0]
+ block = blocks_dict[self.stage1_cfg['BLOCK']]
+ num_blocks = self.stage1_cfg['NUM_BLOCKS'][0]
+ self.layer1 = self._make_layer(block, 64, num_channels, num_blocks)
+ stage1_out_channel = block.expansion*num_channels
+
+ self.stage2_cfg = cfg['MODEL']['EXTRA']['STAGE2']
+ num_channels = self.stage2_cfg['NUM_CHANNELS']
+ block = blocks_dict[self.stage2_cfg['BLOCK']]
+ num_channels = [
+ num_channels[i] * block.expansion for i in range(len(num_channels))]
+ self.transition1 = self._make_transition_layer(
+ [stage1_out_channel], num_channels)
+ self.stage2, pre_stage_channels = self._make_stage(
+ self.stage2_cfg, num_channels)
+
+ self.stage3_cfg = cfg['MODEL']['EXTRA']['STAGE3']
+ num_channels = self.stage3_cfg['NUM_CHANNELS']
+ block = blocks_dict[self.stage3_cfg['BLOCK']]
+ num_channels = [
+ num_channels[i] * block.expansion for i in range(len(num_channels))]
+ self.transition2 = self._make_transition_layer(
+ pre_stage_channels, num_channels)
+ self.stage3, pre_stage_channels = self._make_stage(
+ self.stage3_cfg, num_channels)
+
+ self.stage4_cfg = cfg['MODEL']['EXTRA']['STAGE4']
+ num_channels = self.stage4_cfg['NUM_CHANNELS']
+ block = blocks_dict[self.stage4_cfg['BLOCK']]
+ num_channels = [
+ num_channels[i] * block.expansion for i in range(len(num_channels))]
+ self.transition3 = self._make_transition_layer(
+ pre_stage_channels, num_channels)
+ self.stage4, pre_stage_channels = self._make_stage(
+ self.stage4_cfg, num_channels, multi_scale_output=True)
+
+ # Classification Head
+ self.incre_modules, self.downsamp_modules, \
+ self.final_layer = self._make_head(pre_stage_channels)
+
+ #self.classifier = nn.Linear(2048, 1000)
+ self.output_layer = Sequential(Flatten(),
+ Linear(2048 * cfg['MODEL']['out_h'] * cfg['MODEL']['out_w'],
+ cfg['MODEL']['feat_dim'], False),
+ BatchNorm1d(512))
+
+ def _make_head(self, pre_stage_channels):
+ head_block = Bottleneck
+ head_channels = [32, 64, 128, 256]
+
+ # Increasing the #channels on each resolution
+ # from C, 2C, 4C, 8C to 128, 256, 512, 1024
+ incre_modules = []
+ for i, channels in enumerate(pre_stage_channels):
+ incre_module = self._make_layer(head_block,
+ channels,
+ head_channels[i],
+ 1,
+ stride=1)
+ incre_modules.append(incre_module)
+ incre_modules = nn.ModuleList(incre_modules)
+
+ # downsampling modules
+ downsamp_modules = []
+ for i in range(len(pre_stage_channels)-1):
+ in_channels = head_channels[i] * head_block.expansion
+ out_channels = head_channels[i+1] * head_block.expansion
+
+ downsamp_module = nn.Sequential(
+ nn.Conv2d(in_channels=in_channels,
+ out_channels=out_channels,
+ kernel_size=3,
+ stride=2,
+ padding=1),
+ nn.BatchNorm2d(out_channels, momentum=BN_MOMENTUM),
+ nn.ReLU(inplace=True)
+ )
+
+ downsamp_modules.append(downsamp_module)
+ downsamp_modules = nn.ModuleList(downsamp_modules)
+
+ final_layer = nn.Sequential(
+ nn.Conv2d(
+ in_channels=head_channels[3] * head_block.expansion,
+ out_channels=2048,
+ kernel_size=1,
+ stride=1,
+ padding=0
+ ),
+ nn.BatchNorm2d(2048, momentum=BN_MOMENTUM),
+ nn.ReLU(inplace=True)
+ )
+
+ return incre_modules, downsamp_modules, final_layer
+
+ def _make_transition_layer(
+ self, num_channels_pre_layer, num_channels_cur_layer):
+ num_branches_cur = len(num_channels_cur_layer)
+ num_branches_pre = len(num_channels_pre_layer)
+
+ transition_layers = []
+ for i in range(num_branches_cur):
+ if i < num_branches_pre:
+ if num_channels_cur_layer[i] != num_channels_pre_layer[i]:
+ transition_layers.append(nn.Sequential(
+ nn.Conv2d(num_channels_pre_layer[i],
+ num_channels_cur_layer[i],
+ 3,
+ 1,
+ 1,
+ bias=False),
+ nn.BatchNorm2d(
+ num_channels_cur_layer[i], momentum=BN_MOMENTUM),
+ nn.ReLU(inplace=True)))
+ else:
+ transition_layers.append(None)
+ else:
+ conv3x3s = []
+ for j in range(i+1-num_branches_pre):
+ inchannels = num_channels_pre_layer[-1]
+ outchannels = num_channels_cur_layer[i] \
+ if j == i-num_branches_pre else inchannels
+ conv3x3s.append(nn.Sequential(
+ nn.Conv2d(
+ inchannels, outchannels, 3, 2, 1, bias=False),
+ nn.BatchNorm2d(outchannels, momentum=BN_MOMENTUM),
+ nn.ReLU(inplace=True)))
+ transition_layers.append(nn.Sequential(*conv3x3s))
+
+ return nn.ModuleList(transition_layers)
+
+ def _make_layer(self, block, inplanes, planes, blocks, stride=1):
+ downsample = None
+ if stride != 1 or inplanes != planes * block.expansion:
+ downsample = nn.Sequential(
+ nn.Conv2d(inplanes, planes * block.expansion,
+ kernel_size=1, stride=stride, bias=False),
+ nn.BatchNorm2d(planes * block.expansion, momentum=BN_MOMENTUM),
+ )
+
+ layers = []
+ layers.append(block(inplanes, planes, stride, downsample))
+ inplanes = planes * block.expansion
+ for i in range(1, blocks):
+ layers.append(block(inplanes, planes))
+
+ return nn.Sequential(*layers)
+
+ def _make_stage(self, layer_config, num_inchannels,
+ multi_scale_output=True):
+ num_modules = layer_config['NUM_MODULES']
+ num_branches = layer_config['NUM_BRANCHES']
+ num_blocks = layer_config['NUM_BLOCKS']
+ num_channels = layer_config['NUM_CHANNELS']
+ block = blocks_dict[layer_config['BLOCK']]
+ fuse_method = layer_config['FUSE_METHOD']
+
+ modules = []
+ for i in range(num_modules):
+ # multi_scale_output is only used last module
+ if not multi_scale_output and i == num_modules - 1:
+ reset_multi_scale_output = False
+ else:
+ reset_multi_scale_output = True
+
+ modules.append(
+ HighResolutionModule(num_branches,
+ block,
+ num_blocks,
+ num_inchannels,
+ num_channels,
+ fuse_method,
+ reset_multi_scale_output)
+ )
+ num_inchannels = modules[-1].get_num_inchannels()
+
+ return nn.Sequential(*modules), num_inchannels
+
+ def forward(self, x):
+ x = self.conv1(x)
+ x = self.bn1(x)
+ x = self.relu(x)
+ x = self.conv2(x)
+ x = self.bn2(x)
+ x = self.relu(x)
+ x = self.layer1(x)
+
+ x_list = []
+ for i in range(self.stage2_cfg['NUM_BRANCHES']):
+ if self.transition1[i] is not None:
+ x_list.append(self.transition1[i](x))
+ else:
+ x_list.append(x)
+ y_list = self.stage2(x_list)
+
+ x_list = []
+ for i in range(self.stage3_cfg['NUM_BRANCHES']):
+ if self.transition2[i] is not None:
+ x_list.append(self.transition2[i](y_list[-1]))
+ else:
+ x_list.append(y_list[i])
+ y_list = self.stage3(x_list)
+
+ x_list = []
+ for i in range(self.stage4_cfg['NUM_BRANCHES']):
+ if self.transition3[i] is not None:
+ x_list.append(self.transition3[i](y_list[-1]))
+ else:
+ x_list.append(y_list[i])
+ y_list = self.stage4(x_list)
+
+ # Classification Head
+ y = self.incre_modules[0](y_list[0])
+ for i in range(len(self.downsamp_modules)):
+ y = self.incre_modules[i+1](y_list[i+1]) + \
+ self.downsamp_modules[i](y)
+
+ y = self.final_layer(y)
+ '''
+ if torch._C._get_tracing_state():
+ y = y.flatten(start_dim=2).mean(dim=2)
+ else:
+ y = F.avg_pool2d(y, kernel_size=y.size()
+ [2:]).view(y.size(0), -1)
+
+ y = self.classifier(y)
+ '''
+ y = self.output_layer(y)
+ return y
+
+ def init_weights(self, pretrained='',):
+ logger.info('=> init weights from normal distribution')
+ for m in self.modules():
+ if isinstance(m, nn.Conv2d):
+ nn.init.kaiming_normal_(
+ m.weight, mode='fan_out', nonlinearity='relu')
+ elif isinstance(m, nn.BatchNorm2d):
+ nn.init.constant_(m.weight, 1)
+ nn.init.constant_(m.bias, 0)
+ if os.path.isfile(pretrained):
+ pretrained_dict = torch.load(pretrained)
+ logger.info('=> loading pretrained model {}'.format(pretrained))
+ model_dict = self.state_dict()
+ pretrained_dict = {k: v for k, v in pretrained_dict.items()
+ if k in model_dict.keys()}
+ for k, _ in pretrained_dict.items():
+ logger.info(
+ '=> loading {} pretrained model {}'.format(k, pretrained))
+ model_dict.update(pretrained_dict)
+ self.load_state_dict(model_dict)
+
+
+def get_cls_net(config, **kwargs):
+ model = HighResolutionNet(config, **kwargs)
+ model.init_weights()
+ return model
+
diff --git a/bob/bio/facexzoo/backbones/LightCNN.py b/bob/bio/facexzoo/backbones/LightCNN.py
new file mode 100644
index 0000000000000000000000000000000000000000..851882a6b59f169547ce6ac1b7faa19983cefa01
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/LightCNN.py
@@ -0,0 +1,200 @@
+'''
+ implement Light CNN
+ @author: Alfred Xiang Wu
+ @date: 2017.07.04
+'''
+
+import math
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Flatten(nn.Module):
+ def forward(self, input):
+ return input.reshape(input.size(0), -1)
+
+class mfm(nn.Module):
+ def __init__(self, in_channels, out_channels, kernel_size=3, stride=1, padding=1, type=1):
+ super(mfm, self).__init__()
+ self.out_channels = out_channels
+ if type == 1:
+ self.filter = nn.Conv2d(in_channels, 2*out_channels, kernel_size=kernel_size, stride=stride, padding=padding)
+ else:
+ self.filter = nn.Linear(in_channels, 2*out_channels)
+
+ def forward(self, x):
+ x = self.filter(x)
+ out = torch.split(x, self.out_channels, 1)
+ return torch.max(out[0], out[1])
+
+class group(nn.Module):
+ def __init__(self, in_channels, out_channels, kernel_size, stride, padding):
+ super(group, self).__init__()
+ self.conv_a = mfm(in_channels, in_channels, 1, 1, 0)
+ self.conv = mfm(in_channels, out_channels, kernel_size, stride, padding)
+
+ def forward(self, x):
+ x = self.conv_a(x)
+ x = self.conv(x)
+ return x
+
+class resblock(nn.Module):
+ def __init__(self, in_channels, out_channels):
+ super(resblock, self).__init__()
+ self.conv1 = mfm(in_channels, out_channels, kernel_size=3, stride=1, padding=1)
+ self.conv2 = mfm(in_channels, out_channels, kernel_size=3, stride=1, padding=1)
+
+ def forward(self, x):
+ res = x
+ out = self.conv1(x)
+ out = self.conv2(out)
+ out = out + res
+ return out
+
+class network_9layers(nn.Module):
+ def __init__(self, num_classes=79077):
+ super(network_9layers, self).__init__()
+ self.features = nn.Sequential(
+ mfm(1, 48, 5, 1, 2),
+ nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True),
+ group(48, 96, 3, 1, 1),
+ nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True),
+ group(96, 192, 3, 1, 1),
+ nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True),
+ group(192, 128, 3, 1, 1),
+ group(128, 128, 3, 1, 1),
+ nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True),
+ )
+ self.fc1 = mfm(8*8*128, 256, type=0)
+ self.fc2 = nn.Linear(256, num_classes)
+
+ def forward(self, x):
+ x = self.features(x)
+ x = x.view(x.size(0), -1)
+ x = self.fc1(x)
+ x = F.dropout(x, training=self.training)
+ out = self.fc2(x)
+ return out, x
+
+class network_29layers(nn.Module):
+ def __init__(self, block, layers, num_classes=79077):
+ super(network_29layers, self).__init__()
+ #self.conv1 = mfm(1, 48, 5, 1, 2)
+ self.conv1 = mfm(3, 48, 5, 1, 2)
+ self.pool1 = nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True)
+ self.block1 = self._make_layer(block, layers[0], 48, 48)
+ self.group1 = group(48, 96, 3, 1, 1)
+ self.pool2 = nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True)
+ self.block2 = self._make_layer(block, layers[1], 96, 96)
+ self.group2 = group(96, 192, 3, 1, 1)
+ self.pool3 = nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True)
+ self.block3 = self._make_layer(block, layers[2], 192, 192)
+ self.group3 = group(192, 128, 3, 1, 1)
+ self.block4 = self._make_layer(block, layers[3], 128, 128)
+ self.group4 = group(128, 128, 3, 1, 1)
+ self.pool4 = nn.MaxPool2d(kernel_size=2, stride=2, ceil_mode=True)
+ self.fc = mfm(8*8*128, 256, type=0)
+ self.fc2 = nn.Linear(256, num_classes)
+
+ def _make_layer(self, block, num_blocks, in_channels, out_channels):
+ layers = []
+ for i in range(0, num_blocks):
+ layers.append(block(in_channels, out_channels))
+ return nn.Sequential(*layers)
+
+ def forward(self, x):
+ x = self.conv1(x)
+ x = self.pool1(x)
+
+ x = self.block1(x)
+ x = self.group1(x)
+ x = self.pool2(x)
+
+ x = self.block2(x)
+ x = self.group2(x)
+ x = self.pool3(x)
+
+ x = self.block3(x)
+ x = self.group3(x)
+ x = self.block4(x)
+ x = self.group4(x)
+ x = self.pool4(x)
+
+ x = x.view(x.size(0), -1)
+ fc = self.fc(x)
+ fc = F.dropout(fc, training=self.training)
+ out = self.fc2(fc)
+ return out, fc
+
+
+class network_29layers_v2(nn.Module):
+ def __init__(self, block, layers, drop_ratio, out_h, out_w, feat_dim):
+ super(network_29layers_v2, self).__init__()
+ #self.conv1 = mfm(1, 48, 5, 1, 2)
+ self.conv1 = mfm(3, 48, 5, 1, 2)
+ self.block1 = self._make_layer(block, layers[0], 48, 48)
+ self.group1 = group(48, 96, 3, 1, 1)
+ self.block2 = self._make_layer(block, layers[1], 96, 96)
+ self.group2 = group(96, 192, 3, 1, 1)
+ self.block3 = self._make_layer(block, layers[2], 192, 192)
+ self.group3 = group(192, 128, 3, 1, 1)
+ self.block4 = self._make_layer(block, layers[3], 128, 128)
+ self.group4 = group(128, 128, 3, 1, 1)
+ #self.fc = nn.Linear(8*8*128, 256)
+ #self.fc2 = nn.Linear(256, num_classes, bias=False)
+ self.output_layer = nn.Sequential(nn.BatchNorm2d(128),
+ nn.Dropout(drop_ratio),
+ Flatten(),
+ nn.Linear(128 * out_h * out_w, feat_dim),
+ nn.BatchNorm1d(feat_dim))
+
+ def _make_layer(self, block, num_blocks, in_channels, out_channels):
+ layers = []
+ for i in range(0, num_blocks):
+ layers.append(block(in_channels, out_channels))
+ return nn.Sequential(*layers)
+
+ def forward(self, x):
+ x = self.conv1(x)
+ x = F.max_pool2d(x, 2) + F.avg_pool2d(x, 2)
+
+ x = self.block1(x)
+ x = self.group1(x)
+ x = F.max_pool2d(x, 2) + F.avg_pool2d(x, 2)
+
+ x = self.block2(x)
+ x = self.group2(x)
+ x = F.max_pool2d(x, 2) + F.avg_pool2d(x, 2)
+
+ x = self.block3(x)
+ x = self.group3(x)
+ x = self.block4(x)
+ x = self.group4(x)
+ x = F.max_pool2d(x, 2) + F.avg_pool2d(x, 2) # 7*7
+
+ #x = x.view(x.size(0), -1)
+ #fc = self.fc(x)
+ #x = F.dropout(fc, training=self.training)
+ #out = self.fc2(x)
+ #return out, fc
+ x = self.output_layer(x)
+ return x
+
+
+def LightCNN_9Layers(drop_ratio, out_h, out_w, feat_dim):
+ model = network_9layers(drop_ratio, out_h, out_w, feat_dim)
+ return model
+
+def LightCNN_29Layers(drop_ratio, out_h, out_w, feat_dim):
+ model = network_29layers(resblock, [1, 2, 3, 4], drop_ratio, out_h, out_w, feat_dim)
+ return model
+
+def LightCNN_29Layers_v2(drop_ratio, out_h, out_w, feat_dim):
+ model = network_29layers_v2(resblock, [1, 2, 3, 4], drop_ratio, out_h, out_w, feat_dim)
+ return model
+
+def LightCNN(depth, drop_ratio, out_h, out_w, feat_dim):
+ if depth == 9:
+ return LightCNN_9Layers(drop_ratio, out_h, out_w, feat_dim)
+ elif depth == 29:
+ return LightCNN_29Layers_v2(drop_ratio, out_h, out_w, feat_dim)
diff --git a/bob/bio/facexzoo/backbones/MobileFaceNets.py b/bob/bio/facexzoo/backbones/MobileFaceNets.py
new file mode 100644
index 0000000000000000000000000000000000000000..2785fc4d817a3e50b861b28f59c20841c3d0d24c
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/MobileFaceNets.py
@@ -0,0 +1,101 @@
+"""
+@author: Jun Wang
+@date: 20201019
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/TreB1eN/InsightFace_Pytorch/blob/master/model.py
+
+from torch.nn import Linear, Conv2d, BatchNorm1d, BatchNorm2d, PReLU, Sequential, Module
+import torch
+
+class Flatten(Module):
+ def forward(self, input):
+ return input.view(input.size(0), -1)
+
+class Conv_block(Module):
+ def __init__(self, in_c, out_c, kernel=(1, 1), stride=(1, 1), padding=(0, 0), groups=1):
+ super(Conv_block, self).__init__()
+ self.conv = Conv2d(in_c, out_channels=out_c, kernel_size=kernel, groups=groups, stride=stride, padding=padding, bias=False)
+ self.bn = BatchNorm2d(out_c)
+ self.prelu = PReLU(out_c)
+ def forward(self, x):
+ x = self.conv(x)
+ x = self.bn(x)
+ x = self.prelu(x)
+ return x
+
+class Linear_block(Module):
+ def __init__(self, in_c, out_c, kernel=(1, 1), stride=(1, 1), padding=(0, 0), groups=1):
+ super(Linear_block, self).__init__()
+ self.conv = Conv2d(in_c, out_channels=out_c, kernel_size=kernel, groups=groups, stride=stride, padding=padding, bias=False)
+ self.bn = BatchNorm2d(out_c)
+ def forward(self, x):
+ x = self.conv(x)
+ x = self.bn(x)
+ return x
+
+class Depth_Wise(Module):
+ def __init__(self, in_c, out_c, residual = False, kernel=(3, 3), stride=(2, 2), padding=(1, 1), groups=1):
+ super(Depth_Wise, self).__init__()
+ self.conv = Conv_block(in_c, out_c=groups, kernel=(1, 1), padding=(0, 0), stride=(1, 1))
+ self.conv_dw = Conv_block(groups, groups, groups=groups, kernel=kernel, padding=padding, stride=stride)
+ self.project = Linear_block(groups, out_c, kernel=(1, 1), padding=(0, 0), stride=(1, 1))
+ self.residual = residual
+ def forward(self, x):
+ if self.residual:
+ short_cut = x
+ x = self.conv(x)
+ x = self.conv_dw(x)
+ x = self.project(x)
+ if self.residual:
+ output = short_cut + x
+ else:
+ output = x
+ return output
+
+class Residual(Module):
+ def __init__(self, c, num_block, groups, kernel=(3, 3), stride=(1, 1), padding=(1, 1)):
+ super(Residual, self).__init__()
+ modules = []
+ for _ in range(num_block):
+ modules.append(Depth_Wise(c, c, residual=True, kernel=kernel, padding=padding, stride=stride, groups=groups))
+ self.model = Sequential(*modules)
+ def forward(self, x):
+ return self.model(x)
+
+class MobileFaceNet(Module):
+ def __init__(self, embedding_size, out_h, out_w):
+ super(MobileFaceNet, self).__init__()
+ self.conv1 = Conv_block(3, 64, kernel=(3, 3), stride=(2, 2), padding=(1, 1))
+ self.conv2_dw = Conv_block(64, 64, kernel=(3, 3), stride=(1, 1), padding=(1, 1), groups=64)
+ self.conv_23 = Depth_Wise(64, 64, kernel=(3, 3), stride=(2, 2), padding=(1, 1), groups=128)
+ self.conv_3 = Residual(64, num_block=4, groups=128, kernel=(3, 3), stride=(1, 1), padding=(1, 1))
+ self.conv_34 = Depth_Wise(64, 128, kernel=(3, 3), stride=(2, 2), padding=(1, 1), groups=256)
+ self.conv_4 = Residual(128, num_block=6, groups=256, kernel=(3, 3), stride=(1, 1), padding=(1, 1))
+ self.conv_45 = Depth_Wise(128, 128, kernel=(3, 3), stride=(2, 2), padding=(1, 1), groups=512)
+ self.conv_5 = Residual(128, num_block=2, groups=256, kernel=(3, 3), stride=(1, 1), padding=(1, 1))
+ self.conv_6_sep = Conv_block(128, 512, kernel=(1, 1), stride=(1, 1), padding=(0, 0))
+ #self.conv_6_dw = Linear_block(512, 512, groups=512, kernel=(7,7), stride=(1, 1), padding=(0, 0))
+ #self.conv_6_dw = Linear_block(512, 512, groups=512, kernel=(4,7), stride=(1, 1), padding=(0, 0))
+ self.conv_6_dw = Linear_block(512, 512, groups=512, kernel=(out_h, out_w), stride=(1, 1), padding=(0, 0))
+ self.conv_6_flatten = Flatten()
+ self.linear = Linear(512, embedding_size, bias=False)
+ self.bn = BatchNorm1d(embedding_size)
+
+ def forward(self, x):
+ out = self.conv1(x)
+ out = self.conv2_dw(out)
+ out = self.conv_23(out)
+ out = self.conv_3(out)
+ out = self.conv_34(out)
+ out = self.conv_4(out)
+ out = self.conv_45(out)
+ out = self.conv_5(out)
+ out = self.conv_6_sep(out)
+ out = self.conv_6_dw(out)
+ out = self.conv_6_flatten(out)
+ out = self.linear(out)
+ out = self.bn(out)
+ return out
diff --git a/bob/bio/facexzoo/backbones/ReXNets.py b/bob/bio/facexzoo/backbones/ReXNets.py
new file mode 100644
index 0000000000000000000000000000000000000000..5bc87b41e7767835a3e0256074d9f20d7e5101c8
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/ReXNets.py
@@ -0,0 +1,195 @@
+"""
+@author: Jun Wang
+@date: 20210322
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/clovaai/rexnet/blob/master/rexnetv1.py
+"""
+ReXNet
+Copyright (c) 2020-present NAVER Corp.
+MIT license
+"""
+
+import torch
+import torch.nn as nn
+from math import ceil
+
+class Flatten(nn.Module):
+ def forward(self, input):
+ return input.reshape(input.size(0), -1)
+
+# Memory-efficient Siwsh using torch.jit.script borrowed from the code in (https://twitter.com/jeremyphoward/status/1188251041835315200)
+# Currently use memory-efficient Swish as default:
+USE_MEMORY_EFFICIENT_SWISH = True
+
+if USE_MEMORY_EFFICIENT_SWISH:
+ @torch.jit.script
+ def swish_fwd(x):
+ return x.mul(torch.sigmoid(x))
+
+
+ @torch.jit.script
+ def swish_bwd(x, grad_output):
+ x_sigmoid = torch.sigmoid(x)
+ return grad_output * (x_sigmoid * (1. + x * (1. - x_sigmoid)))
+
+
+ class SwishJitImplementation(torch.autograd.Function):
+ @staticmethod
+ def forward(ctx, x):
+ ctx.save_for_backward(x)
+ return swish_fwd(x)
+
+ @staticmethod
+ def backward(ctx, grad_output):
+ x = ctx.saved_tensors[0]
+ return swish_bwd(x, grad_output)
+
+
+ def swish(x, inplace=False):
+ return SwishJitImplementation.apply(x)
+
+else:
+ def swish(x, inplace=False):
+ return x.mul_(x.sigmoid()) if inplace else x.mul(x.sigmoid())
+
+
+class Swish(nn.Module):
+ def __init__(self, inplace=True):
+ super(Swish, self).__init__()
+ self.inplace = inplace
+
+ def forward(self, x):
+ return swish(x, self.inplace)
+
+
+def ConvBNAct(out, in_channels, channels, kernel=1, stride=1, pad=0,
+ num_group=1, active=True, relu6=False):
+ out.append(nn.Conv2d(in_channels, channels, kernel,
+ stride, pad, groups=num_group, bias=False))
+ out.append(nn.BatchNorm2d(channels))
+ if active:
+ out.append(nn.ReLU6(inplace=True) if relu6 else nn.ReLU(inplace=True))
+
+
+def ConvBNSwish(out, in_channels, channels, kernel=1, stride=1, pad=0, num_group=1):
+ out.append(nn.Conv2d(in_channels, channels, kernel,
+ stride, pad, groups=num_group, bias=False))
+ out.append(nn.BatchNorm2d(channels))
+ out.append(Swish())
+
+
+class SE(nn.Module):
+ def __init__(self, in_channels, channels, se_ratio=12):
+ super(SE, self).__init__()
+ self.avg_pool = nn.AdaptiveAvgPool2d(1)
+ self.fc = nn.Sequential(
+ nn.Conv2d(in_channels, channels // se_ratio, kernel_size=1, padding=0),
+ nn.BatchNorm2d(channels // se_ratio),
+ nn.ReLU(inplace=True),
+ nn.Conv2d(channels // se_ratio, channels, kernel_size=1, padding=0),
+ nn.Sigmoid()
+ )
+
+ def forward(self, x):
+ y = self.avg_pool(x)
+ y = self.fc(y)
+ return x * y
+
+
+class LinearBottleneck(nn.Module):
+ def __init__(self, in_channels, channels, t, stride, use_se=True, se_ratio=12,
+ **kwargs):
+ super(LinearBottleneck, self).__init__(**kwargs)
+ self.use_shortcut = stride == 1 and in_channels <= channels
+ self.in_channels = in_channels
+ self.out_channels = channels
+
+ out = []
+ if t != 1:
+ dw_channels = in_channels * t
+ ConvBNSwish(out, in_channels=in_channels, channels=dw_channels)
+ else:
+ dw_channels = in_channels
+
+ ConvBNAct(out, in_channels=dw_channels, channels=dw_channels, kernel=3, stride=stride, pad=1,
+ num_group=dw_channels, active=False)
+
+ if use_se:
+ out.append(SE(dw_channels, dw_channels, se_ratio))
+
+ out.append(nn.ReLU6())
+ ConvBNAct(out, in_channels=dw_channels, channels=channels, active=False, relu6=True)
+ self.out = nn.Sequential(*out)
+
+ def forward(self, x):
+ out = self.out(x)
+ if self.use_shortcut:
+ out[:, 0:self.in_channels] += x
+
+ return out
+
+class ReXNetV1(nn.Module):
+ def __init__(self, input_ch=16, final_ch=180, width_mult=1.0, depth_mult=1.0,
+ use_se=True, se_ratio=12, out_h=7, out_w=7, feat_dim=512,
+ dropout_ratio=0.2, bn_momentum=0.9):
+ super(ReXNetV1, self).__init__()
+
+ layers = [1, 2, 2, 3, 3, 5]
+ strides = [1, 2, 2, 2, 1, 2]
+ use_ses = [False, False, True, True, True, True]
+
+ layers = [ceil(element * depth_mult) for element in layers]
+ strides = sum([[element] + [1] * (layers[idx] - 1)
+ for idx, element in enumerate(strides)], [])
+ if use_se:
+ use_ses = sum([[element] * layers[idx] for idx, element in enumerate(use_ses)], [])
+ else:
+ use_ses = [False] * sum(layers[:])
+ ts = [1] * layers[0] + [6] * sum(layers[1:])
+
+ self.depth = sum(layers[:]) * 3
+ stem_channel = 32 / width_mult if width_mult < 1.0 else 32
+ inplanes = input_ch / width_mult if width_mult < 1.0 else input_ch
+
+ features = []
+ in_channels_group = []
+ channels_group = []
+
+ # The following channel configuration is a simple instance to make each layer become an expand layer.
+ for i in range(self.depth // 3):
+ if i == 0:
+ in_channels_group.append(int(round(stem_channel * width_mult)))
+ channels_group.append(int(round(inplanes * width_mult)))
+ else:
+ in_channels_group.append(int(round(inplanes * width_mult)))
+ inplanes += final_ch / (self.depth // 3 * 1.0)
+ channels_group.append(int(round(inplanes * width_mult)))
+
+ #ConvBNSwish(features, 3, int(round(stem_channel * width_mult)), kernel=3, stride=2, pad=1)
+ ConvBNSwish(features, 3, int(round(stem_channel * width_mult)), kernel=3, stride=1, pad=1)
+
+ for block_idx, (in_c, c, t, s, se) in enumerate(zip(in_channels_group, channels_group, ts, strides, use_ses)):
+ features.append(LinearBottleneck(in_channels=in_c,
+ channels=c,
+ t=t,
+ stride=s,
+ use_se=se, se_ratio=se_ratio))
+
+ #pen_channels = int(1280 * width_mult)
+ pen_channels = int(512 * width_mult)
+ ConvBNSwish(features, c, pen_channels)
+
+ #features.append(nn.AdaptiveAvgPool2d(1))
+ self.features = nn.Sequential(*features)
+ self.output_layer = nn.Sequential(nn.BatchNorm2d(512),
+ nn.Dropout(dropout_ratio),
+ Flatten(),
+ nn.Linear(512 * out_h * out_w, feat_dim),
+ nn.BatchNorm1d(feat_dim))
+ def forward(self, x):
+ x = self.features(x)
+ x = self.output_layer(x)
+ return x
diff --git a/bob/bio/facexzoo/backbones/Readme.md b/bob/bio/facexzoo/backbones/Readme.md
new file mode 100644
index 0000000000000000000000000000000000000000..e1f78b98db4ea17caa1651f92ab898563edd7dea
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/Readme.md
@@ -0,0 +1,3 @@
+# Source:
+- https://github.com/JDAI-CV/FaceX-Zoo/tree/main/backbone
+- https://gitlab.idiap.ch/bob/bob.learn.pytorch/-/tree/master/bob/learn/pytorch/architectures/facexzoo
\ No newline at end of file
diff --git a/bob/bio/facexzoo/backbones/RepVGG.py b/bob/bio/facexzoo/backbones/RepVGG.py
new file mode 100644
index 0000000000000000000000000000000000000000..afa46ccf57c1959da72e75f9bd173bf4fb55bcea
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/RepVGG.py
@@ -0,0 +1,312 @@
+"""
+@author: Jun Wang
+@date: 20210910
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/DingXiaoH/RepVGG/edit/main/repvgg.py
+
+import torch.nn as nn
+import numpy as np
+import torch
+import copy
+
+class Flatten(nn.Module):
+ def forward(self, input):
+ return input.reshape(input.size(0), -1)
+
+def conv_bn(in_channels, out_channels, kernel_size, stride, padding, groups=1):
+ result = nn.Sequential()
+ result.add_module('conv', nn.Conv2d(in_channels=in_channels, out_channels=out_channels,
+ kernel_size=kernel_size, stride=stride, padding=padding, groups=groups, bias=False))
+ result.add_module('bn', nn.BatchNorm2d(num_features=out_channels))
+ return result
+
+class RepVGGBlock(nn.Module):
+
+ def __init__(self, in_channels, out_channels, kernel_size,
+ stride=1, padding=0, dilation=1, groups=1, padding_mode='zeros', deploy=False, use_se=False):
+ super(RepVGGBlock, self).__init__()
+ self.deploy = deploy
+ self.groups = groups
+ self.in_channels = in_channels
+
+ assert kernel_size == 3
+ assert padding == 1
+
+ padding_11 = padding - kernel_size // 2
+
+ self.nonlinearity = nn.ReLU()
+ '''
+ if use_se:
+ self.se = SEBlock(out_channels, internal_neurons=out_channels // 16)
+ else:
+ self.se = nn.Identity()
+ '''
+ self.se = nn.Identity()
+
+ if deploy:
+ self.rbr_reparam = nn.Conv2d(in_channels=in_channels, out_channels=out_channels, kernel_size=kernel_size, stride=stride,
+ padding=padding, dilation=dilation, groups=groups, bias=True, padding_mode=padding_mode)
+
+ else:
+ self.rbr_identity = nn.BatchNorm2d(num_features=in_channels) if out_channels == in_channels and stride == 1 else None
+ self.rbr_dense = conv_bn(in_channels=in_channels, out_channels=out_channels, kernel_size=kernel_size, stride=stride, padding=padding, groups=groups)
+ self.rbr_1x1 = conv_bn(in_channels=in_channels, out_channels=out_channels, kernel_size=1, stride=stride, padding=padding_11, groups=groups)
+ #print('RepVGG Block, identity = ', self.rbr_identity)
+
+
+ def forward(self, inputs):
+ if hasattr(self, 'rbr_reparam'):
+ return self.nonlinearity(self.se(self.rbr_reparam(inputs)))
+
+ if self.rbr_identity is None:
+ id_out = 0
+ else:
+ id_out = self.rbr_identity(inputs)
+
+ return self.nonlinearity(self.se(self.rbr_dense(inputs) + self.rbr_1x1(inputs) + id_out))
+
+
+ # Optional. This improves the accuracy and facilitates quantization.
+ # 1. Cancel the original weight decay on rbr_dense.conv.weight and rbr_1x1.conv.weight.
+ # 2. Use like this.
+ # loss = criterion(....)
+ # for every RepVGGBlock blk:
+ # loss += weight_decay_coefficient * 0.5 * blk.get_cust_L2()
+ # optimizer.zero_grad()
+ # loss.backward()
+ def get_custom_L2(self):
+ K3 = self.rbr_dense.conv.weight
+ K1 = self.rbr_1x1.conv.weight
+ t3 = (self.rbr_dense.bn.weight / ((self.rbr_dense.bn.running_var + self.rbr_dense.bn.eps).sqrt())).reshape(-1, 1, 1, 1).detach()
+ t1 = (self.rbr_1x1.bn.weight / ((self.rbr_1x1.bn.running_var + self.rbr_1x1.bn.eps).sqrt())).reshape(-1, 1, 1, 1).detach()
+
+ l2_loss_circle = (K3 ** 2).sum() - (K3[:, :, 1:2, 1:2] ** 2).sum() # The L2 loss of the "circle" of weights in 3x3 kernel. Use regular L2 on them.
+ eq_kernel = K3[:, :, 1:2, 1:2] * t3 + K1 * t1 # The equivalent resultant central point of 3x3 kernel.
+ l2_loss_eq_kernel = (eq_kernel ** 2 / (t3 ** 2 + t1 ** 2)).sum() # Normalize for an L2 coefficient comparable to regular L2.
+ return l2_loss_eq_kernel + l2_loss_circle
+
+
+
+# This func derives the equivalent kernel and bias in a DIFFERENTIABLE way.
+# You can get the equivalent kernel and bias at any time and do whatever you want,
+ # for example, apply some penalties or constraints during training, just like you do to the other models.
+# May be useful for quantization or pruning.
+ def get_equivalent_kernel_bias(self):
+ kernel3x3, bias3x3 = self._fuse_bn_tensor(self.rbr_dense)
+ kernel1x1, bias1x1 = self._fuse_bn_tensor(self.rbr_1x1)
+ kernelid, biasid = self._fuse_bn_tensor(self.rbr_identity)
+ return kernel3x3 + self._pad_1x1_to_3x3_tensor(kernel1x1) + kernelid, bias3x3 + bias1x1 + biasid
+
+ def _pad_1x1_to_3x3_tensor(self, kernel1x1):
+ if kernel1x1 is None:
+ return 0
+ else:
+ return torch.nn.functional.pad(kernel1x1, [1,1,1,1])
+
+ def _fuse_bn_tensor(self, branch):
+ if branch is None:
+ return 0, 0
+ if isinstance(branch, nn.Sequential):
+ kernel = branch.conv.weight
+ running_mean = branch.bn.running_mean
+ running_var = branch.bn.running_var
+ gamma = branch.bn.weight
+ beta = branch.bn.bias
+ eps = branch.bn.eps
+ else:
+ assert isinstance(branch, nn.BatchNorm2d)
+ if not hasattr(self, 'id_tensor'):
+ input_dim = self.in_channels // self.groups
+ kernel_value = np.zeros((self.in_channels, input_dim, 3, 3), dtype=np.float32)
+ for i in range(self.in_channels):
+ kernel_value[i, i % input_dim, 1, 1] = 1
+ self.id_tensor = torch.from_numpy(kernel_value).to(branch.weight.device)
+ kernel = self.id_tensor
+ running_mean = branch.running_mean
+ running_var = branch.running_var
+ gamma = branch.weight
+ beta = branch.bias
+ eps = branch.eps
+ std = (running_var + eps).sqrt()
+ t = (gamma / std).reshape(-1, 1, 1, 1)
+ return kernel * t, beta - running_mean * gamma / std
+
+ def switch_to_deploy(self):
+ if hasattr(self, 'rbr_reparam'):
+ return
+ kernel, bias = self.get_equivalent_kernel_bias()
+ self.rbr_reparam = nn.Conv2d(in_channels=self.rbr_dense.conv.in_channels, out_channels=self.rbr_dense.conv.out_channels,
+ kernel_size=self.rbr_dense.conv.kernel_size, stride=self.rbr_dense.conv.stride,
+ padding=self.rbr_dense.conv.padding, dilation=self.rbr_dense.conv.dilation, groups=self.rbr_dense.conv.groups, bias=True)
+ self.rbr_reparam.weight.data = kernel
+ self.rbr_reparam.bias.data = bias
+ for para in self.parameters():
+ para.detach_()
+ self.__delattr__('rbr_dense')
+ self.__delattr__('rbr_1x1')
+ if hasattr(self, 'rbr_identity'):
+ self.__delattr__('rbr_identity')
+ if hasattr(self, 'id_tensor'):
+ self.__delattr__('id_tensor')
+ self.deploy = True
+
+
+
+class RepVGG(nn.Module):
+
+ def __init__(self, num_blocks, width_multiplier, feat_dim=512, out_h=7, out_w=7, override_groups_map=None, deploy=False, use_se=False):
+ super(RepVGG, self).__init__()
+
+ assert len(width_multiplier) == 4
+
+ self.deploy = deploy
+ self.override_groups_map = override_groups_map or dict()
+ self.use_se = use_se
+
+ assert 0 not in self.override_groups_map
+
+ self.in_planes = min(64, int(64 * width_multiplier[0]))
+
+ #self.stage0 = RepVGGBlock(in_channels=3, out_channels=self.in_planes, kernel_size=3, stride=2, padding=1, deploy=self.deploy, use_se=self.use_se)
+ self.stage0 = RepVGGBlock(in_channels=3, out_channels=self.in_planes, kernel_size=3, stride=1, padding=1, deploy=self.deploy, use_se=self.use_se)
+ self.cur_layer_idx = 1
+ self.stage1 = self._make_stage(int(64 * width_multiplier[0]), num_blocks[0], stride=2)
+ self.stage2 = self._make_stage(int(128 * width_multiplier[1]), num_blocks[1], stride=2)
+ self.stage3 = self._make_stage(int(256 * width_multiplier[2]), num_blocks[2], stride=2)
+ self.stage4 = self._make_stage(int(512 * width_multiplier[3]), num_blocks[3], stride=2)
+ self.output_layer = nn.Sequential(nn.BatchNorm2d(int(512*width_multiplier[3])),
+ Flatten(),
+ nn.Linear(int(512 * width_multiplier[3]) * out_h * out_w, feat_dim), # for eye
+ nn.BatchNorm1d(feat_dim))
+ def _make_stage(self, planes, num_blocks, stride):
+ strides = [stride] + [1]*(num_blocks-1)
+ blocks = []
+ for stride in strides:
+ cur_groups = self.override_groups_map.get(self.cur_layer_idx, 1)
+ blocks.append(RepVGGBlock(in_channels=self.in_planes, out_channels=planes, kernel_size=3,
+ stride=stride, padding=1, groups=cur_groups, deploy=self.deploy, use_se=self.use_se))
+ self.in_planes = planes
+ self.cur_layer_idx += 1
+ return nn.Sequential(*blocks)
+
+ def forward(self, x):
+ out = self.stage0(x)
+ out = self.stage1(out)
+ out = self.stage2(out)
+ out = self.stage3(out)
+ out = self.stage4(out)
+ out = self.output_layer(out)
+ return out
+
+optional_groupwise_layers = [2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26]
+g2_map = {l: 2 for l in optional_groupwise_layers}
+g4_map = {l: 4 for l in optional_groupwise_layers}
+
+def create_RepVGG_A0(deploy=False):
+ return RepVGG(num_blocks=[2, 4, 14, 1],
+ width_multiplier=[0.75, 0.75, 0.75, 2.5], override_groups_map=None, deploy=deploy)
+
+def create_RepVGG_A1(deploy=False):
+ return RepVGG(num_blocks=[2, 4, 14, 1],
+ width_multiplier=[1, 1, 1, 2.5], override_groups_map=None, deploy=deploy)
+
+def create_RepVGG_A2(deploy=False):
+ return RepVGG(num_blocks=[2, 4, 14, 1],
+ width_multiplier=[1.5, 1.5, 1.5, 2.75], override_groups_map=None, deploy=deploy)
+
+def create_RepVGG_B0(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[1, 1, 1, 2.5], override_groups_map=None, deploy=deploy)
+
+def create_RepVGG_B1(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[2, 2, 2, 4], override_groups_map=None, deploy=deploy)
+
+def create_RepVGG_B1g2(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[2, 2, 2, 4], override_groups_map=g2_map, deploy=deploy)
+
+def create_RepVGG_B1g4(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[2, 2, 2, 4], override_groups_map=g4_map, deploy=deploy)
+
+
+def create_RepVGG_B2(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[2.5, 2.5, 2.5, 5], override_groups_map=None, deploy=deploy)
+
+def create_RepVGG_B2g2(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[2.5, 2.5, 2.5, 5], override_groups_map=g2_map, deploy=deploy)
+
+def create_RepVGG_B2g4(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[2.5, 2.5, 2.5, 5], override_groups_map=g4_map, deploy=deploy)
+
+
+def create_RepVGG_B3(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[3, 3, 3, 5], override_groups_map=None, deploy=deploy)
+
+def create_RepVGG_B3g2(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[3, 3, 3, 5], override_groups_map=g2_map, deploy=deploy)
+
+def create_RepVGG_B3g4(deploy=False):
+ return RepVGG(num_blocks=[4, 6, 16, 1],
+ width_multiplier=[3, 3, 3, 5], override_groups_map=g4_map, deploy=deploy)
+
+def create_RepVGG_D2se(deploy=False):
+ return RepVGG(num_blocks=[8, 14, 24, 1],
+ width_multiplier=[2.5, 2.5, 2.5, 5], override_groups_map=None, deploy=deploy, use_se=True)
+
+
+func_dict = {
+'RepVGG-A0': create_RepVGG_A0,
+'RepVGG-A1': create_RepVGG_A1,
+'RepVGG-A2': create_RepVGG_A2,
+'RepVGG-B0': create_RepVGG_B0,
+'RepVGG-B1': create_RepVGG_B1,
+'RepVGG-B1g2': create_RepVGG_B1g2,
+'RepVGG-B1g4': create_RepVGG_B1g4,
+'RepVGG-B2': create_RepVGG_B2,
+'RepVGG-B2g2': create_RepVGG_B2g2,
+'RepVGG-B2g4': create_RepVGG_B2g4,
+'RepVGG-B3': create_RepVGG_B3,
+'RepVGG-B3g2': create_RepVGG_B3g2,
+'RepVGG-B3g4': create_RepVGG_B3g4,
+'RepVGG-D2se': create_RepVGG_D2se, # Updated at April 25, 2021. This is not reported in the CVPR paper.
+}
+def get_RepVGG_func_by_name(name):
+ return func_dict[name]
+
+
+
+# Use this for converting a RepVGG model or a bigger model with RepVGG as its component
+# Use like this
+# model = create_RepVGG_A0(deploy=False)
+# train model or load weights
+# repvgg_model_convert(model, save_path='repvgg_deploy.pth')
+# If you want to preserve the original model, call with do_copy=True
+
+# ====================== for using RepVGG as the backbone of a bigger model, e.g., PSPNet, the pseudo code will be like
+# train_backbone = create_RepVGG_B2(deploy=False)
+# train_backbone.load_state_dict(torch.load('RepVGG-B2-train.pth'))
+# train_pspnet = build_pspnet(backbone=train_backbone)
+# segmentation_train(train_pspnet)
+# deploy_pspnet = repvgg_model_convert(train_pspnet)
+# segmentation_test(deploy_pspnet)
+# ===================== example_pspnet.py shows an example
+
+def repvgg_model_convert(model:torch.nn.Module, save_path=None, do_copy=True):
+ if do_copy:
+ model = copy.deepcopy(model)
+ for module in model.modules():
+ if hasattr(module, 'switch_to_deploy'):
+ module.switch_to_deploy()
+ if save_path is not None:
+ torch.save(model.state_dict(), save_path)
+ return model
diff --git a/bob/bio/facexzoo/backbones/ResNets.py b/bob/bio/facexzoo/backbones/ResNets.py
new file mode 100644
index 0000000000000000000000000000000000000000..5a76e7e78eed77ed509f44f7120c6edcc1f235b1
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/ResNets.py
@@ -0,0 +1,142 @@
+"""
+@author: Jun Wang
+@date: 20201019
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/TreB1eN/InsightFace_Pytorch/blob/master/model.py
+
+from torch.nn import Linear, Conv2d, BatchNorm1d, BatchNorm2d, PReLU, ReLU, Sigmoid, Dropout2d, Dropout, AvgPool2d, MaxPool2d, AdaptiveAvgPool2d, Sequential, Module, Parameter
+import torch.nn.functional as F
+import torch
+from collections import namedtuple
+
+
+class Flatten(Module):
+ def forward(self, input):
+ return input.view(input.size(0), -1)
+
+class SEModule(Module):
+ def __init__(self, channels, reduction):
+ super(SEModule, self).__init__()
+ self.avg_pool = AdaptiveAvgPool2d(1)
+ self.fc1 = Conv2d(
+ channels, channels // reduction, kernel_size=1, padding=0 ,bias=False)
+ self.relu = ReLU(inplace=True)
+ self.fc2 = Conv2d(
+ channels // reduction, channels, kernel_size=1, padding=0 ,bias=False)
+ self.sigmoid = Sigmoid()
+
+ def forward(self, x):
+ module_input = x
+ x = self.avg_pool(x)
+ x = self.fc1(x)
+ x = self.relu(x)
+ x = self.fc2(x)
+ x = self.sigmoid(x)
+ return module_input * x
+
+class bottleneck_IR(Module):
+ def __init__(self, in_channel, depth, stride):
+ super(bottleneck_IR, self).__init__()
+ if in_channel == depth:
+ self.shortcut_layer = MaxPool2d(1, stride)
+ else:
+ self.shortcut_layer = Sequential(
+ Conv2d(in_channel, depth, (1, 1), stride ,bias=False), BatchNorm2d(depth))
+ self.res_layer = Sequential(
+ BatchNorm2d(in_channel),
+ Conv2d(in_channel, depth, (3, 3), (1, 1), 1 ,bias=False), PReLU(depth),
+ Conv2d(depth, depth, (3, 3), stride, 1 ,bias=False), BatchNorm2d(depth))
+
+ def forward(self, x):
+ shortcut = self.shortcut_layer(x)
+ res = self.res_layer(x)
+ return res + shortcut
+
+class bottleneck_IR_SE(Module):
+ def __init__(self, in_channel, depth, stride):
+ super(bottleneck_IR_SE, self).__init__()
+ if in_channel == depth:
+ self.shortcut_layer = MaxPool2d(1, stride)
+ else:
+ self.shortcut_layer = Sequential(
+ Conv2d(in_channel, depth, (1, 1), stride ,bias=False),
+ BatchNorm2d(depth))
+ self.res_layer = Sequential(
+ BatchNorm2d(in_channel),
+ Conv2d(in_channel, depth, (3,3), (1,1),1 ,bias=False),
+ PReLU(depth),
+ Conv2d(depth, depth, (3,3), stride, 1 ,bias=False),
+ BatchNorm2d(depth),
+ SEModule(depth,16)
+ )
+ def forward(self,x):
+ shortcut = self.shortcut_layer(x)
+ res = self.res_layer(x)
+ return res + shortcut
+
+class Bottleneck(namedtuple('Block', ['in_channel', 'depth', 'stride'])):
+ '''A named tuple describing a ResNet block.'''
+
+def get_block(in_channel, depth, num_units, stride = 2):
+ return [Bottleneck(in_channel, depth, stride)] + [Bottleneck(depth, depth, 1) for i in range(num_units-1)]
+
+def get_blocks(num_layers):
+ if num_layers == 50:
+ blocks = [
+ get_block(in_channel=64, depth=64, num_units = 3),
+ get_block(in_channel=64, depth=128, num_units=4),
+ get_block(in_channel=128, depth=256, num_units=14),
+ get_block(in_channel=256, depth=512, num_units=3)
+ ]
+ elif num_layers == 100:
+ blocks = [
+ get_block(in_channel=64, depth=64, num_units=3),
+ get_block(in_channel=64, depth=128, num_units=13),
+ get_block(in_channel=128, depth=256, num_units=30),
+ get_block(in_channel=256, depth=512, num_units=3)
+ ]
+ elif num_layers == 152:
+ blocks = [
+ get_block(in_channel=64, depth=64, num_units=3),
+ get_block(in_channel=64, depth=128, num_units=8),
+ get_block(in_channel=128, depth=256, num_units=36),
+ get_block(in_channel=256, depth=512, num_units=3)
+ ]
+ return blocks
+
+#class Backbone(Module):
+class Resnet(Module):
+ def __init__(self, num_layers, drop_ratio, mode='ir', feat_dim=512, out_h=7, out_w=7):
+ super(Resnet, self).__init__()
+ assert num_layers in [50, 100, 152], 'num_layers should be 50,100, or 152'
+ assert mode in ['ir', 'ir_se'], 'mode should be ir or ir_se'
+ blocks = get_blocks(num_layers)
+ if mode == 'ir':
+ unit_module = bottleneck_IR
+ elif mode == 'ir_se':
+ unit_module = bottleneck_IR_SE
+ self.input_layer = Sequential(Conv2d(3, 64, (3, 3), 1, 1 ,bias=False),
+ BatchNorm2d(64),
+ PReLU(64))
+ self.output_layer = Sequential(BatchNorm2d(512),
+ Dropout(drop_ratio),
+ Flatten(),
+ Linear(512 * out_h * out_w, feat_dim), # for eye
+ BatchNorm1d(feat_dim))
+ modules = []
+ for block in blocks:
+ for bottleneck in block:
+ modules.append(
+ unit_module(bottleneck.in_channel,
+ bottleneck.depth,
+ bottleneck.stride))
+ self.body = Sequential(*modules)
+
+ def forward(self,x):
+ x = self.input_layer(x)
+ x = self.body(x)
+ x = self.output_layer(x)
+ return x
diff --git a/bob/bio/facexzoo/backbones/Swin_Transformer.py b/bob/bio/facexzoo/backbones/Swin_Transformer.py
new file mode 100644
index 0000000000000000000000000000000000000000..15cf70b58d40655de03dac34fd2d5ca556805a0e
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/Swin_Transformer.py
@@ -0,0 +1,592 @@
+# --------------------------------------------------------
+# Swin Transformer
+# Copyright (c) 2021 Microsoft
+# Licensed under The MIT License [see LICENSE for details]
+# Written by Ze Liu
+# --------------------------------------------------------
+
+import torch
+import torch.nn as nn
+import torch.utils.checkpoint as checkpoint
+from timm.models.layers import DropPath, to_2tuple, trunc_normal_
+
+class Flatten(nn.Module):
+ def forward(self, input):
+ return input.view(input.size(0), -1)
+
+
+class Mlp(nn.Module):
+ def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
+ super().__init__()
+ out_features = out_features or in_features
+ hidden_features = hidden_features or in_features
+ self.fc1 = nn.Linear(in_features, hidden_features)
+ self.act = act_layer()
+ self.fc2 = nn.Linear(hidden_features, out_features)
+ self.drop = nn.Dropout(drop)
+
+ def forward(self, x):
+ x = self.fc1(x)
+ x = self.act(x)
+ x = self.drop(x)
+ x = self.fc2(x)
+ x = self.drop(x)
+ return x
+
+
+def window_partition(x, window_size):
+ """
+ Args:
+ x: (B, H, W, C)
+ window_size (int): window size
+
+ Returns:
+ windows: (num_windows*B, window_size, window_size, C)
+ """
+ B, H, W, C = x.shape
+ x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
+ windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
+ return windows
+
+
+def window_reverse(windows, window_size, H, W):
+ """
+ Args:
+ windows: (num_windows*B, window_size, window_size, C)
+ window_size (int): Window size
+ H (int): Height of image
+ W (int): Width of image
+
+ Returns:
+ x: (B, H, W, C)
+ """
+ B = int(windows.shape[0] / (H * W / window_size / window_size))
+ x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
+ x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
+ return x
+
+
+class WindowAttention(nn.Module):
+ r""" Window based multi-head self attention (W-MSA) module with relative position bias.
+ It supports both of shifted and non-shifted window.
+
+ Args:
+ dim (int): Number of input channels.
+ window_size (tuple[int]): The height and width of the window.
+ num_heads (int): Number of attention heads.
+ qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
+ attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
+ proj_drop (float, optional): Dropout ratio of output. Default: 0.0
+ """
+
+ def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0.):
+
+ super().__init__()
+ self.dim = dim
+ self.window_size = window_size # Wh, Ww
+ self.num_heads = num_heads
+ head_dim = dim // num_heads
+ self.scale = qk_scale or head_dim ** -0.5
+
+ # define a parameter table of relative position bias
+ self.relative_position_bias_table = nn.Parameter(
+ torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads)) # 2*Wh-1 * 2*Ww-1, nH
+
+ # get pair-wise relative position index for each token inside the window
+ coords_h = torch.arange(self.window_size[0])
+ coords_w = torch.arange(self.window_size[1])
+ coords = torch.stack(torch.meshgrid([coords_h, coords_w])) # 2, Wh, Ww
+ coords_flatten = torch.flatten(coords, 1) # 2, Wh*Ww
+ relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :] # 2, Wh*Ww, Wh*Ww
+ relative_coords = relative_coords.permute(1, 2, 0).contiguous() # Wh*Ww, Wh*Ww, 2
+ relative_coords[:, :, 0] += self.window_size[0] - 1 # shift to start from 0
+ relative_coords[:, :, 1] += self.window_size[1] - 1
+ relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
+ relative_position_index = relative_coords.sum(-1) # Wh*Ww, Wh*Ww
+ self.register_buffer("relative_position_index", relative_position_index)
+
+ self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
+ self.attn_drop = nn.Dropout(attn_drop)
+ self.proj = nn.Linear(dim, dim)
+ self.proj_drop = nn.Dropout(proj_drop)
+
+ trunc_normal_(self.relative_position_bias_table, std=.02)
+ self.softmax = nn.Softmax(dim=-1)
+
+ def forward(self, x, mask=None):
+ """
+ Args:
+ x: input features with shape of (num_windows*B, N, C)
+ mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
+ """
+ B_, N, C = x.shape
+ qkv = self.qkv(x).reshape(B_, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
+ q, k, v = qkv[0], qkv[1], qkv[2] # make torchscript happy (cannot use tensor as tuple)
+
+ q = q * self.scale
+ attn = (q @ k.transpose(-2, -1))
+
+ relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
+ self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1) # Wh*Ww,Wh*Ww,nH
+ relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous() # nH, Wh*Ww, Wh*Ww
+ attn = attn + relative_position_bias.unsqueeze(0)
+
+ if mask is not None:
+ nW = mask.shape[0]
+ attn = attn.view(B_ // nW, nW, self.num_heads, N, N) + mask.unsqueeze(1).unsqueeze(0)
+ attn = attn.view(-1, self.num_heads, N, N)
+ attn = self.softmax(attn)
+ else:
+ attn = self.softmax(attn)
+
+ attn = self.attn_drop(attn)
+
+ x = (attn @ v).transpose(1, 2).reshape(B_, N, C)
+ x = self.proj(x)
+ x = self.proj_drop(x)
+ return x
+
+ def extra_repr(self) -> str:
+ return f'dim={self.dim}, window_size={self.window_size}, num_heads={self.num_heads}'
+
+ def flops(self, N):
+ # calculate flops for 1 window with token length of N
+ flops = 0
+ # qkv = self.qkv(x)
+ flops += N * self.dim * 3 * self.dim
+ # attn = (q @ k.transpose(-2, -1))
+ flops += self.num_heads * N * (self.dim // self.num_heads) * N
+ # x = (attn @ v)
+ flops += self.num_heads * N * N * (self.dim // self.num_heads)
+ # x = self.proj(x)
+ flops += N * self.dim * self.dim
+ return flops
+
+
+class SwinTransformerBlock(nn.Module):
+ r""" Swin Transformer Block.
+
+ Args:
+ dim (int): Number of input channels.
+ input_resolution (tuple[int]): Input resulotion.
+ num_heads (int): Number of attention heads.
+ window_size (int): Window size.
+ shift_size (int): Shift size for SW-MSA.
+ mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+ qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
+ drop (float, optional): Dropout rate. Default: 0.0
+ attn_drop (float, optional): Attention dropout rate. Default: 0.0
+ drop_path (float, optional): Stochastic depth rate. Default: 0.0
+ act_layer (nn.Module, optional): Activation layer. Default: nn.GELU
+ norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+ """
+
+ def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
+ mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
+ act_layer=nn.GELU, norm_layer=nn.LayerNorm):
+ super().__init__()
+ self.dim = dim
+ self.input_resolution = input_resolution
+ self.num_heads = num_heads
+ self.window_size = window_size
+ self.shift_size = shift_size
+ self.mlp_ratio = mlp_ratio
+ if min(self.input_resolution) <= self.window_size:
+ # if window size is larger than input resolution, we don't partition windows
+ self.shift_size = 0
+ self.window_size = min(self.input_resolution)
+ assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
+
+ self.norm1 = norm_layer(dim)
+ self.attn = WindowAttention(
+ dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
+ qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)
+
+ self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
+ self.norm2 = norm_layer(dim)
+ mlp_hidden_dim = int(dim * mlp_ratio)
+ self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)
+
+ if self.shift_size > 0:
+ # calculate attention mask for SW-MSA
+ H, W = self.input_resolution
+ img_mask = torch.zeros((1, H, W, 1)) # 1 H W 1
+ h_slices = (slice(0, -self.window_size),
+ slice(-self.window_size, -self.shift_size),
+ slice(-self.shift_size, None))
+ w_slices = (slice(0, -self.window_size),
+ slice(-self.window_size, -self.shift_size),
+ slice(-self.shift_size, None))
+ cnt = 0
+ for h in h_slices:
+ for w in w_slices:
+ img_mask[:, h, w, :] = cnt
+ cnt += 1
+
+ mask_windows = window_partition(img_mask, self.window_size) # nW, window_size, window_size, 1
+ mask_windows = mask_windows.view(-1, self.window_size * self.window_size)
+ attn_mask = mask_windows.unsqueeze(1) - mask_windows.unsqueeze(2)
+ attn_mask = attn_mask.masked_fill(attn_mask != 0, float(-100.0)).masked_fill(attn_mask == 0, float(0.0))
+ else:
+ attn_mask = None
+
+ self.register_buffer("attn_mask", attn_mask)
+
+ def forward(self, x):
+ H, W = self.input_resolution
+ B, L, C = x.shape
+ assert L == H * W, "input feature has wrong size"
+
+ shortcut = x
+ x = self.norm1(x)
+ x = x.view(B, H, W, C)
+
+ # cyclic shift
+ if self.shift_size > 0:
+ shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
+ else:
+ shifted_x = x
+
+ # partition windows
+ x_windows = window_partition(shifted_x, self.window_size) # nW*B, window_size, window_size, C
+ x_windows = x_windows.view(-1, self.window_size * self.window_size, C) # nW*B, window_size*window_size, C
+
+ # W-MSA/SW-MSA
+ attn_windows = self.attn(x_windows, mask=self.attn_mask) # nW*B, window_size*window_size, C
+
+ # merge windows
+ attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
+ shifted_x = window_reverse(attn_windows, self.window_size, H, W) # B H' W' C
+
+ # reverse cyclic shift
+ if self.shift_size > 0:
+ x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))
+ else:
+ x = shifted_x
+ x = x.view(B, H * W, C)
+
+ # FFN
+ x = shortcut + self.drop_path(x)
+ x = x + self.drop_path(self.mlp(self.norm2(x)))
+
+ return x
+
+ def extra_repr(self) -> str:
+ return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
+ f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"
+
+ def flops(self):
+ flops = 0
+ H, W = self.input_resolution
+ # norm1
+ flops += self.dim * H * W
+ # W-MSA/SW-MSA
+ nW = H * W / self.window_size / self.window_size
+ flops += nW * self.attn.flops(self.window_size * self.window_size)
+ # mlp
+ flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
+ # norm2
+ flops += self.dim * H * W
+ return flops
+
+
+class PatchMerging(nn.Module):
+ r""" Patch Merging Layer.
+
+ Args:
+ input_resolution (tuple[int]): Resolution of input feature.
+ dim (int): Number of input channels.
+ norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+ """
+
+ def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
+ super().__init__()
+ self.input_resolution = input_resolution
+ self.dim = dim
+ self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
+ self.norm = norm_layer(4 * dim)
+
+ def forward(self, x):
+ """
+ x: B, H*W, C
+ """
+ H, W = self.input_resolution
+ B, L, C = x.shape
+ assert L == H * W, "input feature has wrong size"
+ assert H % 2 == 0 and W % 2 == 0, f"x size ({H}*{W}) are not even."
+
+ x = x.view(B, H, W, C)
+
+ x0 = x[:, 0::2, 0::2, :] # B H/2 W/2 C
+ x1 = x[:, 1::2, 0::2, :] # B H/2 W/2 C
+ x2 = x[:, 0::2, 1::2, :] # B H/2 W/2 C
+ x3 = x[:, 1::2, 1::2, :] # B H/2 W/2 C
+ x = torch.cat([x0, x1, x2, x3], -1) # B H/2 W/2 4*C
+ x = x.view(B, -1, 4 * C) # B H/2*W/2 4*C
+
+ x = self.norm(x)
+ x = self.reduction(x)
+
+ return x
+
+ def extra_repr(self) -> str:
+ return f"input_resolution={self.input_resolution}, dim={self.dim}"
+
+ def flops(self):
+ H, W = self.input_resolution
+ flops = H * W * self.dim
+ flops += (H // 2) * (W // 2) * 4 * self.dim * 2 * self.dim
+ return flops
+
+
+class BasicLayer(nn.Module):
+ """ A basic Swin Transformer layer for one stage.
+
+ Args:
+ dim (int): Number of input channels.
+ input_resolution (tuple[int]): Input resolution.
+ depth (int): Number of blocks.
+ num_heads (int): Number of attention heads.
+ window_size (int): Local window size.
+ mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+ qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
+ drop (float, optional): Dropout rate. Default: 0.0
+ attn_drop (float, optional): Attention dropout rate. Default: 0.0
+ drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
+ norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+ downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
+ use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
+ """
+
+ def __init__(self, dim, input_resolution, depth, num_heads, window_size,
+ mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
+ drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
+
+ super().__init__()
+ self.dim = dim
+ self.input_resolution = input_resolution
+ self.depth = depth
+ self.use_checkpoint = use_checkpoint
+
+ # build blocks
+ self.blocks = nn.ModuleList([
+ SwinTransformerBlock(dim=dim, input_resolution=input_resolution,
+ num_heads=num_heads, window_size=window_size,
+ shift_size=0 if (i % 2 == 0) else window_size // 2,
+ mlp_ratio=mlp_ratio,
+ qkv_bias=qkv_bias, qk_scale=qk_scale,
+ drop=drop, attn_drop=attn_drop,
+ drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path,
+ norm_layer=norm_layer)
+ for i in range(depth)])
+
+ # patch merging layer
+ if downsample is not None:
+ self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
+ else:
+ self.downsample = None
+
+ def forward(self, x):
+ for blk in self.blocks:
+ if self.use_checkpoint:
+ x = checkpoint.checkpoint(blk, x)
+ else:
+ x = blk(x)
+ if self.downsample is not None:
+ x = self.downsample(x)
+ return x
+
+ def extra_repr(self) -> str:
+ return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
+
+ def flops(self):
+ flops = 0
+ for blk in self.blocks:
+ flops += blk.flops()
+ if self.downsample is not None:
+ flops += self.downsample.flops()
+ return flops
+
+
+class PatchEmbed(nn.Module):
+ r""" Image to Patch Embedding
+
+ Args:
+ img_size (int): Image size. Default: 224.
+ patch_size (int): Patch token size. Default: 4.
+ in_chans (int): Number of input image channels. Default: 3.
+ embed_dim (int): Number of linear projection output channels. Default: 96.
+ norm_layer (nn.Module, optional): Normalization layer. Default: None
+ """
+
+ def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
+ super().__init__()
+ img_size = to_2tuple(img_size)
+ patch_size = to_2tuple(patch_size)
+ patches_resolution = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
+ self.img_size = img_size
+ self.patch_size = patch_size
+ self.patches_resolution = patches_resolution
+ self.num_patches = patches_resolution[0] * patches_resolution[1]
+
+ self.in_chans = in_chans
+ self.embed_dim = embed_dim
+
+ self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
+ if norm_layer is not None:
+ self.norm = norm_layer(embed_dim)
+ else:
+ self.norm = None
+
+ def forward(self, x):
+ B, C, H, W = x.shape
+ # FIXME look at relaxing size constraints
+ assert H == self.img_size[0] and W == self.img_size[1], \
+ f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
+ x = self.proj(x).flatten(2).transpose(1, 2) # B Ph*Pw C
+ if self.norm is not None:
+ x = self.norm(x)
+ return x
+
+ def flops(self):
+ Ho, Wo = self.patches_resolution
+ flops = Ho * Wo * self.embed_dim * self.in_chans * (self.patch_size[0] * self.patch_size[1])
+ if self.norm is not None:
+ flops += Ho * Wo * self.embed_dim
+ return flops
+
+
+class SwinTransformer(nn.Module):
+ r""" Swin Transformer
+ A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` -
+ https://arxiv.org/pdf/2103.14030
+
+ Args:
+ img_size (int | tuple(int)): Input image size. Default 224
+ patch_size (int | tuple(int)): Patch size. Default: 4
+ in_chans (int): Number of input image channels. Default: 3
+ num_classes (int): Number of classes for classification head. Default: 1000
+ embed_dim (int): Patch embedding dimension. Default: 96
+ depths (tuple(int)): Depth of each Swin Transformer layer.
+ num_heads (tuple(int)): Number of attention heads in different layers.
+ window_size (int): Window size. Default: 7
+ mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4
+ qkv_bias (bool): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float): Override default qk scale of head_dim ** -0.5 if set. Default: None
+ drop_rate (float): Dropout rate. Default: 0
+ attn_drop_rate (float): Attention dropout rate. Default: 0
+ drop_path_rate (float): Stochastic depth rate. Default: 0.1
+ norm_layer (nn.Module): Normalization layer. Default: nn.LayerNorm.
+ ape (bool): If True, add absolute position embedding to the patch embedding. Default: False
+ patch_norm (bool): If True, add normalization after patch embedding. Default: True
+ use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False
+ """
+
+ def __init__(self, img_size=224, patch_size=4, in_chans=3,
+ embed_dim=96, depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24],
+ window_size=7, mlp_ratio=4., qkv_bias=True, qk_scale=None,
+ drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
+ norm_layer=nn.LayerNorm, ape=False, patch_norm=True,
+ use_checkpoint=False, **kwargs):
+ super().__init__()
+
+ self.num_layers = len(depths)
+ self.embed_dim = embed_dim
+ self.ape = ape
+ self.patch_norm = patch_norm
+ self.num_features = int(embed_dim * 2 ** (self.num_layers - 1))
+ self.mlp_ratio = mlp_ratio
+
+ # split image into non-overlapping patches
+ self.patch_embed = PatchEmbed(
+ img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim,
+ norm_layer=norm_layer if self.patch_norm else None)
+ num_patches = self.patch_embed.num_patches
+ patches_resolution = self.patch_embed.patches_resolution
+ self.patches_resolution = patches_resolution
+
+ # absolute position embedding
+ if self.ape:
+ self.absolute_pos_embed = nn.Parameter(torch.zeros(1, num_patches, embed_dim))
+ trunc_normal_(self.absolute_pos_embed, std=.02)
+
+ self.pos_drop = nn.Dropout(p=drop_rate)
+
+ # stochastic depth
+ dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))] # stochastic depth decay rule
+
+ # build layers
+ self.layers = nn.ModuleList()
+ for i_layer in range(self.num_layers):
+ layer = BasicLayer(dim=int(embed_dim * 2 ** i_layer),
+ input_resolution=(patches_resolution[0] // (2 ** i_layer),
+ patches_resolution[1] // (2 ** i_layer)),
+ depth=depths[i_layer],
+ num_heads=num_heads[i_layer],
+ window_size=window_size,
+ mlp_ratio=self.mlp_ratio,
+ qkv_bias=qkv_bias, qk_scale=qk_scale,
+ drop=drop_rate, attn_drop=attn_drop_rate,
+ drop_path=dpr[sum(depths[:i_layer]):sum(depths[:i_layer + 1])],
+ norm_layer=norm_layer,
+ downsample=PatchMerging if (i_layer < self.num_layers - 1) else None,
+ use_checkpoint=use_checkpoint)
+ self.layers.append(layer)
+
+ #self.norm = norm_layer(self.num_features)
+ #self.avgpool = nn.AdaptiveAvgPool1d(1)
+ self.output_layer = nn.Sequential(norm_layer(self.num_features),
+ Flatten(),
+ nn.Linear(49*768, 512),
+ nn.BatchNorm1d(512))
+
+ self.apply(self._init_weights)
+
+ def _init_weights(self, m):
+ if isinstance(m, nn.Linear):
+ trunc_normal_(m.weight, std=.02)
+ if isinstance(m, nn.Linear) and m.bias is not None:
+ nn.init.constant_(m.bias, 0)
+ elif isinstance(m, nn.LayerNorm):
+ nn.init.constant_(m.bias, 0)
+ nn.init.constant_(m.weight, 1.0)
+
+ @torch.jit.ignore
+ def no_weight_decay(self):
+ return {'absolute_pos_embed'}
+
+ @torch.jit.ignore
+ def no_weight_decay_keywords(self):
+ return {'relative_position_bias_table'}
+
+ def forward_features(self, x):
+ x = self.patch_embed(x)
+ if self.ape:
+ x = x + self.absolute_pos_embed
+ x = self.pos_drop(x)
+
+ for layer in self.layers:
+ x = layer(x)
+
+ #x = self.norm(x) # B L C --> [128, 49, 768]
+ #x = self.avgpool(x.transpose(1, 2)) # B C 1 --> [128, 768, 1]
+ #x = torch.flatten(x, 1)
+ x = self.output_layer(x)
+ return x
+
+ def forward(self, x):
+ x = self.forward_features(x) #[128,768]
+ return x
+ '''
+ def flops(self):
+ flops = 0
+ flops += self.patch_embed.flops()
+ for i, layer in enumerate(self.layers):
+ flops += layer.flops()
+ flops += self.num_features * self.patches_resolution[0] * self.patches_resolution[1] // (2 ** self.num_layers)
+ flops += self.num_features * self.num_classes
+ return flops
+ '''
diff --git a/bob/bio/facexzoo/backbones/TF_NAS.py b/bob/bio/facexzoo/backbones/TF_NAS.py
new file mode 100644
index 0000000000000000000000000000000000000000..e8a6738f9de4fb5ee49dea10dbfe0a95b6c2b04e
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/TF_NAS.py
@@ -0,0 +1,492 @@
+"""
+@author: Yibo Hu, Jun Wang
+@date: 20201019
+@contact: jun21wangustc@gmail.com
+"""
+
+import sys
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from collections import OrderedDict
+
+
+def channel_shuffle(x, groups):
+ assert groups > 1
+ batchsize, num_channels, height, width = x.size()
+ assert (num_channels % groups == 0)
+ channels_per_group = num_channels // groups
+ # reshape
+ x = x.view(batchsize, groups, channels_per_group, height, width)
+ # transpose
+ x = torch.transpose(x, 1, 2).contiguous()
+ # flatten
+ x = x.view(batchsize, -1, height, width)
+ return x
+
+
+def get_same_padding(kernel_size):
+ if isinstance(kernel_size, tuple):
+ assert len(kernel_size) == 2, 'invalid kernel size: {}'.format(kernel_size)
+ p1 = get_same_padding(kernel_size[0])
+ p2 = get_same_padding(kernel_size[1])
+ return p1, p2
+ assert isinstance(kernel_size, int), 'kernel size should be either `int` or `tuple`'
+ assert kernel_size % 2 > 0, 'kernel size should be odd number'
+ return kernel_size // 2
+
+
+class Swish(nn.Module):
+ def __init__(self, inplace=False):
+ super(Swish, self).__init__()
+ self.inplace = inplace
+
+ def forward(self, x):
+ if self.inplace:
+ return x.mul_(x.sigmoid())
+ else:
+ return x * x.sigmoid()
+
+
+class HardSwish(nn.Module):
+ def __init__(self, inplace=False):
+ super(HardSwish, self).__init__()
+ self.inplace = inplace
+
+ def forward(self, x):
+ if self.inplace:
+ return x.mul_(F.relu6(x + 3., inplace=True) / 6.)
+ else:
+ return x * F.relu6(x + 3.) /6.
+
+
+class BasicLayer(nn.Module):
+
+ def __init__(
+ self,
+ in_channels,
+ out_channels,
+ use_bn=True,
+ affine = True,
+ act_func='relu6',
+ ops_order='weight_bn_act'):
+ super(BasicLayer, self).__init__()
+
+ self.in_channels = in_channels
+ self.out_channels = out_channels
+ self.use_bn = use_bn
+ self.affine = affine
+ self.act_func = act_func
+ self.ops_order = ops_order
+
+ """ add modules """
+ # batch norm
+ if self.use_bn:
+ if self.bn_before_weight:
+ self.bn = nn.BatchNorm2d(in_channels, affine=affine, track_running_stats=affine)
+ else:
+ self.bn = nn.BatchNorm2d(out_channels, affine=affine, track_running_stats=affine)
+ else:
+ self.bn = None
+ # activation
+ if act_func == 'relu':
+ if self.ops_list[0] == 'act':
+ self.act = nn.ReLU(inplace=False)
+ else:
+ self.act = nn.ReLU(inplace=True)
+ elif act_func == 'relu6':
+ if self.ops_list[0] == 'act':
+ self.act = nn.ReLU6(inplace=False)
+ else:
+ self.act = nn.ReLU6(inplace=True)
+ elif act_func == 'swish':
+ if self.ops_list[0] == 'act':
+ self.act = Swish(inplace=False)
+ else:
+ self.act = Swish(inplace=True)
+ elif act_func == 'h-swish':
+ if self.ops_list[0] == 'act':
+ self.act = HardSwish(inplace=False)
+ else:
+ self.act = HardSwish(inplace=True)
+ else:
+ self.act = None
+
+ @property
+ def ops_list(self):
+ return self.ops_order.split('_')
+
+ @property
+ def bn_before_weight(self):
+ for op in self.ops_list:
+ if op == 'bn':
+ return True
+ elif op == 'weight':
+ return False
+ raise ValueError('Invalid ops_order: %s' % self.ops_order)
+
+ def weight_call(self, x):
+ raise NotImplementedError
+
+ def forward(self, x):
+ for op in self.ops_list:
+ if op == 'weight':
+ x = self.weight_call(x)
+ elif op == 'bn':
+ if self.bn is not None:
+ x = self.bn(x)
+ elif op == 'act':
+ if self.act is not None:
+ x = self.act(x)
+ else:
+ raise ValueError('Unrecognized op: %s' % op)
+ return x
+
+
+class ConvLayer(BasicLayer):
+
+ def __init__(
+ self,
+ in_channels,
+ out_channels,
+ kernel_size=3,
+ stride=1,
+ groups=1,
+ has_shuffle=False,
+ bias=False,
+ use_bn=True,
+ affine=True,
+ act_func='relu6',
+ ops_order='weight_bn_act'):
+ super(ConvLayer, self).__init__(
+ in_channels,
+ out_channels,
+ use_bn,
+ affine,
+ act_func,
+ ops_order)
+
+ self.kernel_size = kernel_size
+ self.stride = stride
+ self.groups = groups
+ self.has_shuffle = has_shuffle
+ self.bias = bias
+
+ padding = get_same_padding(self.kernel_size)
+ self.conv = nn.Conv2d(
+ in_channels,
+ out_channels,
+ kernel_size=self.kernel_size,
+ stride=self.stride,
+ padding=padding,
+ groups=self.groups,
+ bias=self.bias)
+
+ def weight_call(self, x):
+ x = self.conv(x)
+ if self.has_shuffle and self.groups > 1:
+ x = channel_shuffle(x, self.groups)
+ return x
+
+
+class LinearLayer(nn.Module):
+
+ def __init__(
+ self,
+ in_features,
+ out_features,
+ bias=True,
+ use_bn=False,
+ affine=False,
+ act_func=None,
+ ops_order='weight_bn_act'):
+ super(LinearLayer, self).__init__()
+
+ self.in_features = in_features
+ self.out_features = out_features
+ self.bias = bias
+ self.use_bn = use_bn
+ self.affine = affine
+ self.act_func = act_func
+ self.ops_order = ops_order
+
+ """ add modules """
+ # batch norm
+ if self.use_bn:
+ if self.bn_before_weight:
+ self.bn = nn.BatchNorm1d(in_features, affine=affine, track_running_stats=affine)
+ else:
+ self.bn = nn.BatchNorm1d(out_features, affine=affine, track_running_stats=affine)
+ else:
+ self.bn = None
+ # activation
+ if act_func == 'relu':
+ if self.ops_list[0] == 'act':
+ self.act = nn.ReLU(inplace=False)
+ else:
+ self.act = nn.ReLU(inplace=True)
+ elif act_func == 'relu6':
+ if self.ops_list[0] == 'act':
+ self.act = nn.ReLU6(inplace=False)
+ else:
+ self.act = nn.ReLU6(inplace=True)
+ elif act_func == 'tanh':
+ self.act = nn.Tanh()
+ elif act_func == 'sigmoid':
+ self.act = nn.Sigmoid()
+ else:
+ self.act = None
+ # linear
+ self.linear = nn.Linear(self.in_features, self.out_features, self.bias)
+
+ @property
+ def ops_list(self):
+ return self.ops_order.split('_')
+
+ @property
+ def bn_before_weight(self):
+ for op in self.ops_list:
+ if op == 'bn':
+ return True
+ elif op == 'weight':
+ return False
+ raise ValueError('Invalid ops_order: %s' % self.ops_order)
+
+ def forward(self, x):
+ for op in self.ops_list:
+ if op == 'weight':
+ x = self.linear(x)
+ elif op == 'bn':
+ if self.bn is not None:
+ x = self.bn(x)
+ elif op == 'act':
+ if self.act is not None:
+ x = self.act(x)
+ else:
+ raise ValueError('Unrecognized op: %s' % op)
+ return x
+
+
+class MBInvertedResBlock(nn.Module):
+
+ def __init__(
+ self,
+ in_channels,
+ mid_channels,
+ se_channels,
+ out_channels,
+ kernel_size=3,
+ stride=1,
+ groups=1,
+ has_shuffle=False,
+ bias=False,
+ use_bn=True,
+ affine=True,
+ act_func='relu6'):
+ super(MBInvertedResBlock, self).__init__()
+
+ self.in_channels = in_channels
+ self.mid_channels = mid_channels
+ self.se_channels = se_channels
+ self.out_channels = out_channels
+ self.kernel_size = kernel_size
+ self.stride = stride
+ self.groups = groups
+ self.has_shuffle = has_shuffle
+ self.bias = bias
+ self.use_bn = use_bn
+ self.affine = affine
+ self.act_func = act_func
+
+ # inverted bottleneck
+ if mid_channels > in_channels:
+ inverted_bottleneck = OrderedDict([
+ ('conv', nn.Conv2d(in_channels, mid_channels, 1, 1, 0, groups=groups, bias=bias)),
+ ])
+ if use_bn:
+ inverted_bottleneck['bn'] = nn.BatchNorm2d(mid_channels, affine=affine, track_running_stats=affine)
+ if act_func == 'relu':
+ inverted_bottleneck['act'] = nn.ReLU(inplace=True)
+ elif act_func == 'relu6':
+ inverted_bottleneck['act'] = nn.ReLU6(inplace=True)
+ elif act_func == 'swish':
+ inverted_bottleneck['act'] = Swish(inplace=True)
+ elif act_func == 'h-swish':
+ inverted_bottleneck['act'] = HardSwish(inplace=True)
+ self.inverted_bottleneck = nn.Sequential(inverted_bottleneck)
+ else:
+ self.inverted_bottleneck = None
+ self.mid_channels = in_channels
+ mid_channels = in_channels
+
+ # depthwise convolution
+ padding = get_same_padding(self.kernel_size)
+ depth_conv = OrderedDict([
+ ('conv',
+ nn.Conv2d(
+ mid_channels,
+ mid_channels,
+ kernel_size,
+ stride,
+ padding,
+ groups=mid_channels,
+ bias=bias)),
+ ])
+ if use_bn:
+ depth_conv['bn'] = nn.BatchNorm2d(mid_channels, affine=affine, track_running_stats=affine)
+ if act_func == 'relu':
+ depth_conv['act'] = nn.ReLU(inplace=True)
+ elif act_func == 'relu6':
+ depth_conv['act'] = nn.ReLU6(inplace=True)
+ elif act_func == 'swish':
+ depth_conv['act'] = Swish(inplace=True)
+ elif act_func == 'h-swish':
+ depth_conv['act'] = HardSwish(inplace=True)
+ self.depth_conv = nn.Sequential(depth_conv)
+
+ # se model
+ if se_channels > 0:
+ squeeze_excite = OrderedDict([
+ ('conv_reduce', nn.Conv2d(mid_channels, se_channels, 1, 1, 0, groups=groups, bias=True)),
+ ])
+ if act_func == 'relu':
+ squeeze_excite['act'] = nn.ReLU(inplace=True)
+ elif act_func == 'relu6':
+ squeeze_excite['act'] = nn.ReLU6(inplace=True)
+ elif act_func == 'swish':
+ squeeze_excite['act'] = Swish(inplace=True)
+ elif act_func == 'h-swish':
+ squeeze_excite['act'] = HardSwish(inplace=True)
+ squeeze_excite['conv_expand'] = nn.Conv2d(se_channels, mid_channels, 1, 1, 0, groups=groups, bias=True)
+ self.squeeze_excite = nn.Sequential(squeeze_excite)
+ else:
+ self.squeeze_excite = None
+ self.se_channels = 0
+
+ # pointwise linear
+ point_linear = OrderedDict([
+ ('conv', nn.Conv2d(mid_channels, out_channels, 1, 1, 0, groups=groups, bias=bias)),
+ ])
+ if use_bn:
+ point_linear['bn'] = nn.BatchNorm2d(out_channels, affine=affine, track_running_stats=affine)
+ self.point_linear = nn.Sequential(point_linear)
+
+ # residual flag
+ self.has_residual = (in_channels == out_channels) and (stride == 1)
+
+ def forward(self, x):
+ res = x
+
+ if self.inverted_bottleneck is not None:
+ x = self.inverted_bottleneck(x)
+ if self.has_shuffle and self.groups > 1:
+ x = channel_shuffle(x, self.groups)
+
+ x = self.depth_conv(x)
+ if self.squeeze_excite is not None:
+ x_se = F.adaptive_avg_pool2d(x, 1)
+ x = x * torch.sigmoid(self.squeeze_excite(x_se))
+
+ x = self.point_linear(x)
+ if self.has_shuffle and self.groups > 1:
+ x = channel_shuffle(x, self.groups)
+
+ if self.has_residual:
+ x += res
+
+ return x
+
+class Flatten(nn.Module):
+ def forward(self, x):
+ return x.reshape(x.size(0), -1)
+
+
+class TF_NAS_A(nn.Module):
+ def __init__(self, out_h, out_w, feat_dim, drop_ratio=0.0):
+ super(TF_NAS_A, self).__init__()
+ self.drop_ratio = drop_ratio
+
+ self.first_stem = ConvLayer(3, 32, kernel_size=3, stride=1, act_func='relu')
+ self.second_stem = MBInvertedResBlock(32, 32, 8, 16, kernel_size=3, stride=1, act_func='relu')
+ self.stage1 = nn.Sequential(
+ MBInvertedResBlock(16, 83, 32, 24, kernel_size=3, stride=2, act_func='relu'),
+ MBInvertedResBlock(24, 128, 0, 24, kernel_size=5, stride=1, act_func='relu'),
+ )
+ self.stage2 = nn.Sequential(
+ MBInvertedResBlock(24, 138, 48, 40, kernel_size=3, stride=2, act_func='swish'),
+ MBInvertedResBlock(40, 297, 0, 40, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(40, 170, 80, 40, kernel_size=5, stride=1, act_func='swish'),
+ )
+ self.stage3 = nn.Sequential(
+ MBInvertedResBlock(40, 248, 80, 80, kernel_size=5, stride=2, act_func='swish'),
+ MBInvertedResBlock(80, 500, 0, 80, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(80, 424, 0, 80, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(80, 477, 0, 80, kernel_size=3, stride=1, act_func='swish'),
+ )
+ self.stage4 = nn.Sequential(
+ MBInvertedResBlock(80, 504, 160, 112, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(112, 796, 0, 112, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(112, 723, 224, 112, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(112, 555, 224, 112, kernel_size=3, stride=1, act_func='swish'),
+ )
+ self.stage5 = nn.Sequential(
+ MBInvertedResBlock(112, 813, 0, 192, kernel_size=3, stride=2, act_func='swish'),
+ MBInvertedResBlock(192, 1370, 0, 192, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(192, 1138, 384, 192, kernel_size=3, stride=1, act_func='swish'),
+ MBInvertedResBlock(192, 1359, 384, 192, kernel_size=3, stride=1, act_func='swish'),
+ )
+ self.stage6 = nn.Sequential(
+ MBInvertedResBlock(192, 1203, 384, 320, kernel_size=5, stride=1, act_func='swish'),
+ )
+ self.feature_mix_layer = ConvLayer(320, 1280, kernel_size=1, stride=1, act_func='none')
+ self.output_layer = nn.Sequential(
+ nn.Dropout(self.drop_ratio),
+ Flatten(),
+ nn.Linear(1280 * out_h * out_w, feat_dim),
+ nn.BatchNorm1d(feat_dim))
+
+ self._initialization()
+
+ def forward(self, x):
+ x = self.first_stem(x)
+ x = self.second_stem(x)
+ for block in self.stage1:
+ x = block(x)
+ for block in self.stage2:
+ x = block(x)
+ for block in self.stage3:
+ x = block(x)
+ for block in self.stage4:
+ x = block(x)
+ for block in self.stage5:
+ x = block(x)
+ for block in self.stage6:
+ x = block(x)
+ x = self.feature_mix_layer(x)
+ x = self.output_layer(x)
+ return x
+
+ def _initialization(self):
+ for m in self.modules():
+ if isinstance(m, nn.Conv2d):
+ if m.bias is not None:
+ nn.init.constant_(m.bias, 0)
+ elif isinstance(m, nn.Linear):
+ if m.bias is not None:
+ nn.init.constant_(m.bias, 0)
+ elif isinstance(m, nn.BatchNorm2d):
+ if m.weight is not None:
+ nn.init.constant_(m.weight, 1)
+ if m.bias is not None:
+ nn.init.constant_(m.bias, 0)
+
+
+if __name__ == '__main__':
+ x = torch.rand((2,3,112,112))
+ net = TF_NAS_A(7, 7, 512, drop_ratio=0.0)
+
+ x = x.cuda()
+ net = net.cuda()
+
+ out = net(x)
+ print(out.size())
diff --git a/bob/bio/facexzoo/backbones/__init__.py b/bob/bio/facexzoo/backbones/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..0487c5de8615241b4a6b8dafd66bc3246661a955
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/__init__.py
@@ -0,0 +1 @@
+from .models import FaceXZooModelFactory
diff --git a/bob/bio/facexzoo/backbones/backbone_conf.yaml b/bob/bio/facexzoo/backbones/backbone_conf.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..b3023dc570569cd95f1823c26646f351a9911d2a
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/backbone_conf.yaml
@@ -0,0 +1,285 @@
+MobileFaceNet:
+ feat_dim: 512
+ #out_h: 4
+ out_h: 7
+ out_w: 7
+
+# ResNet:
+# depth: 152
+# drop_ratio: 0.4
+# net_mode: ir_se
+# feat_dim: 512
+# out_h: 7
+# out_w: 7
+
+# according to the log of model:
+ResNet50_ir:
+ depth: 50
+ drop_ratio: 0.4
+ net_mode: ir
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+# according to the log of model:
+ResNet152_irse:
+ depth: 152
+ drop_ratio: 0.4
+ net_mode: ir_se
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+
+EfficientNet_B0:
+ width: 1.0
+ depth: 1.0
+ image_size: 110
+ drop_ratio: 0.2
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+
+HRNet:
+ NAME: cls_hrnet
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+ IMAGE_SIZE:
+ - 112
+ - 112
+ EXTRA:
+ STAGE1:
+ NUM_MODULES: 1
+ NUM_RANCHES: 1
+ BLOCK: BOTTLENECK
+ NUM_BLOCKS:
+ - 4
+ NUM_CHANNELS:
+ - 64
+ FUSE_METHOD: SUM
+ STAGE2:
+ NUM_MODULES: 1
+ NUM_BRANCHES: 2
+ BLOCK: BASIC
+ NUM_BLOCKS:
+ - 4
+ - 4
+ NUM_CHANNELS:
+ - 18
+ - 36
+ FUSE_METHOD: SUM
+ STAGE3:
+ NUM_MODULES: 4
+ NUM_BRANCHES: 3
+ BLOCK: BASIC
+ NUM_BLOCKS:
+ - 4
+ - 4
+ - 4
+ NUM_CHANNELS:
+ - 18
+ - 36
+ - 72
+ FUSE_METHOD: SUM
+ STAGE4:
+ NUM_MODULES: 3
+ NUM_BRANCHES: 4
+ BLOCK: BASIC
+ NUM_BLOCKS:
+ - 4
+ - 4
+ - 4
+ - 4
+ NUM_CHANNELS:
+ - 18
+ - 36
+ - 72
+ - 144
+ FUSE_METHOD: SUM
+
+GhostNet:
+ width: 1.0
+ drop_ratio: 0.2
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+
+# AttentionNet:
+# stage1_modules: 1
+# stage2_modules: 2
+# stage3_modules: 3
+# feat_dim: 512
+# out_h: 7
+# out_w: 7
+
+# https://github.com/JDAI-CV/FaceX-Zoo/issues/96#issuecomment-929808352
+AttentionNet56:
+ #AttentionNet:
+ stage1_modules: 1
+ stage2_modules: 1
+ stage3_modules: 1
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+# https://github.com/JDAI-CV/FaceX-Zoo/issues/96#issuecomment-929808352
+AttentionNet92:
+ #AttentionNet:
+ stage1_modules: 1
+ stage2_modules: 2
+ stage3_modules: 3
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+TF_NAS_A:
+ feat_dim: 512
+ drop_ratio: 0.2
+ out_h: 7
+ out_w: 7
+
+ResNeSt50:
+ depth: 50
+ drop_ratio: 0.4
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+ReXNet_1:
+ input_ch: 16
+ final_ch: 180
+ width_mult: 1.0
+ depth_mult: 1.0
+ use_se: 0
+ se_ratio: 12
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+ dropout_ratio: 0.2
+
+
+LightCNN29:
+ depth: 29
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+ dropout_ratio: 0.2
+
+# RepVGG:
+# blocks1: 4
+# blocks2: 6
+# blocks3: 16
+# blocks4: 1
+# width1: 2
+# width2: 2
+# width3: 2
+# width4: 4
+# out_h: 7
+# out_w: 7
+# feat_dim: 512
+
+# according to the log of model:
+RepVGG_A0:
+ blocks1: 2
+ blocks2: 4
+ blocks3: 14
+ blocks4: 1
+ width1: 0.75
+ width2: 0.75
+ width3: 0.75
+ width4: 2.5
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+
+# according to the log of model:
+RepVGG_B0:
+ blocks1: 4
+ blocks2: 6
+ blocks3: 16
+ blocks4: 1
+ width1: 1
+ width2: 1
+ width3: 1
+ width4: 2.5
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+
+# according to the log of model:
+RepVGG_B1:
+ blocks1: 4
+ blocks2: 6
+ blocks3: 16
+ blocks4: 1
+ width1: 2
+ width2: 2
+ width3: 2
+ width4: 4
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+
+
+# SwinTransformer:
+# img_size: 224
+# patch_size: 4
+# in_chans: 3
+# embed_dim: 96
+# depths:
+# - 2
+# - 2
+# - 18
+# - 2
+# num_heads:
+# - 3
+# - 6
+# - 12
+# - 24
+# window_size: 7
+# mlp_ratio: 4.0
+# drop_rate: 0.0
+# drop_path_rate: 0.3
+
+
+# according to the log of model:
+SwinTransformer_S:
+ img_size: 224
+ patch_size: 4
+ in_chans: 3
+ embed_dim: 96
+ depths:
+ - 2
+ - 2
+ - 18
+ - 2
+ num_heads:
+ - 3
+ - 6
+ - 12
+ - 24
+ window_size: 7
+ mlp_ratio: 4.0
+ drop_rate: 0.0
+ drop_path_rate: 0.3
+
+SwinTransformer_T:
+ img_size: 224
+ patch_size: 4
+ in_chans: 3
+ embed_dim: 96
+ depths:
+ - 2
+ - 2
+ - 6
+ - 2
+ num_heads:
+ - 3
+ - 6
+ - 12
+ - 24
+ window_size: 7
+ mlp_ratio: 4.0
+ drop_rate: 0.0
+ drop_path_rate: 0.2
\ No newline at end of file
diff --git a/bob/bio/facexzoo/backbones/backbone_def.py b/bob/bio/facexzoo/backbones/backbone_def.py
new file mode 100644
index 0000000000000000000000000000000000000000..1ccd916bb04b73772830e44da472dcb17dbf9036
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/backbone_def.py
@@ -0,0 +1,243 @@
+# https://github.com/JDAI-CV/FaceX-Zoo/blob/main/backbone/backbone_def.py
+
+"""
+@author: Jun Wang
+@date: 20201019
+@contact: jun21wangustc@gmail.com
+"""
+
+import sys
+import yaml
+
+from .ResNets import Resnet
+from .MobileFaceNets import MobileFaceNet
+from .EfficientNets import EfficientNet
+from .EfficientNets import efficientnet
+from .HRNet import HighResolutionNet
+from .GhostNet import GhostNet
+from .AttentionNets import ResidualAttentionNet
+from .TF_NAS import TF_NAS_A
+from .resnest.resnest import ResNeSt
+from .ReXNets import ReXNetV1
+from .LightCNN import LightCNN
+from .RepVGG import RepVGG
+from .Swin_Transformer import SwinTransformer
+
+
+
+class BackboneFactory:
+ """Factory to produce backbone according the backbone_conf.yaml.
+
+ Attributes:
+ backbone_type(str): which backbone will produce.
+ backbone_param(dict): parsed params and it's value.
+ """
+ def __init__(self, backbone_type, backbone_conf_file):
+ self.backbone_type = backbone_type
+ with open(backbone_conf_file) as f:
+ backbone_conf = yaml.safe_load(f)
+ self.backbone_param = backbone_conf[backbone_type]
+ print('backbone param:')
+ print(self.backbone_param)
+
+ def get_backbone(self):
+
+ if self.backbone_type == 'MobileFaceNet':
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ backbone = MobileFaceNet(feat_dim, out_h, out_w)
+ elif self.backbone_type == 'ResNet50_ir':
+ depth = self.backbone_param['depth'] # depth of the ResNet, e.g. 50, 100, 152.
+ drop_ratio = self.backbone_param['drop_ratio'] # drop out ratio.
+ net_mode = self.backbone_param['net_mode'] # 'ir' for improved by resnt, 'ir_se' for SE-ResNet.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ backbone = Resnet(depth, drop_ratio, net_mode, feat_dim, out_h, out_w)
+ elif self.backbone_type == 'ResNet152_irse':
+ depth = self.backbone_param['depth'] # depth of the ResNet, e.g. 50, 100, 152.
+ drop_ratio = self.backbone_param['drop_ratio'] # drop out ratio.
+ net_mode = self.backbone_param['net_mode'] # 'ir' for improved by resnt, 'ir_se' for SE-ResNet.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ backbone = Resnet(depth, drop_ratio, net_mode, feat_dim, out_h, out_w)
+ elif self.backbone_type == 'HRNet':
+ config = {}
+ config['MODEL'] = self.backbone_param
+ backbone = HighResolutionNet(config)
+ elif self.backbone_type == 'EfficientNet_B0':
+ width = self.backbone_param['width'] # width for EfficientNet, e.g. 1.0, 1.2, 1.4, ...
+ depth = self.backbone_param['depth'] # depth for EfficientNet, e.g. 1.0, 1.2, 1.4, ...
+ image_size = self.backbone_param['image_size'] # input image size, e.g. 112.
+ drop_ratio = self.backbone_param['drop_ratio'] # drop out ratio.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ blocks_args, global_params = efficientnet(
+ width_coefficient=width, depth_coefficient=depth,
+ dropout_rate=drop_ratio, image_size=image_size)
+ backbone = EfficientNet(out_h, out_w, feat_dim, blocks_args, global_params)
+ elif self.backbone_type == 'TF_NAS_A':
+ drop_ratio = self.backbone_param['drop_ratio'] # drop out ratio.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ backbone = TF_NAS_A(out_h, out_w, feat_dim, drop_ratio)
+ elif self.backbone_type == 'GhostNet':
+ width = self.backbone_param['width']
+ drop_ratio = self.backbone_param['drop_ratio'] # drop out ratio.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final feature
+ backbone = GhostNet(width, drop_ratio, feat_dim, out_h, out_w)
+ elif self.backbone_type == 'AttentionNet56':
+ stage1_modules = self.backbone_param['stage1_modules'] # the number of attention modules in stage1.
+ stage2_modules = self.backbone_param['stage2_modules'] # the number of attention modules in stage2.
+ stage3_modules = self.backbone_param['stage3_modules'] # the number of attention modules in stage3.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ backbone = ResidualAttentionNet(
+ stage1_modules, stage2_modules, stage3_modules,
+ feat_dim, out_h, out_w)
+ elif self.backbone_type == 'AttentionNet92':
+ stage1_modules = self.backbone_param['stage1_modules'] # the number of attention modules in stage1.
+ stage2_modules = self.backbone_param['stage2_modules'] # the number of attention modules in stage2.
+ stage3_modules = self.backbone_param['stage3_modules'] # the number of attention modules in stage3.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ backbone = ResidualAttentionNet(
+ stage1_modules, stage2_modules, stage3_modules,
+ feat_dim, out_h, out_w)
+ elif self.backbone_type == 'ResNeSt50':
+ depth = self.backbone_param['depth'] # depth of the ResNet, e.g. 50, 100, 152.
+ drop_ratio = self.backbone_param['drop_ratio'] # drop out ratio.
+ feat_dim = self.backbone_param['feat_dim'] # dimension of the output features, e.g. 512.
+ out_h = self.backbone_param['out_h'] # height of the feature map before the final features.
+ out_w = self.backbone_param['out_w'] # width of the feature map before the final features.
+ backbone = ResNeSt(depth, drop_ratio, feat_dim, out_h, out_w)
+ elif self.backbone_type == 'ReXNet_1':
+ input_ch = self.backbone_param['input_ch']
+ final_ch = self.backbone_param['final_ch']
+ width_mult = self.backbone_param['width_mult']
+ depth_mult = self.backbone_param['depth_mult']
+ use_se = True if self.backbone_param['use_se']==1 else False
+ se_ratio = self.backbone_param['se_ratio']
+ out_h = self.backbone_param['out_h']
+ out_w = self.backbone_param['out_w']
+ feat_dim = self.backbone_param['feat_dim']
+ dropout_ratio = self.backbone_param['dropout_ratio']
+ backbone = ReXNetV1(input_ch, final_ch, width_mult, depth_mult, use_se, se_ratio,
+ out_h, out_w, feat_dim, dropout_ratio)
+ elif self.backbone_type == 'LightCNN29':
+ depth = self.backbone_param['depth']
+ out_h = self.backbone_param['out_h']
+ out_w = self.backbone_param['out_w']
+ feat_dim = self.backbone_param['feat_dim']
+ drop_ratio = self.backbone_param['dropout_ratio']
+ backbone = LightCNN(depth, drop_ratio, out_h, out_w, feat_dim)
+ elif self.backbone_type == 'RepVGG_A0':
+ blocks1 = self.backbone_param['blocks1']
+ blocks2 = self.backbone_param['blocks2']
+ blocks3 = self.backbone_param['blocks3']
+ blocks4 = self.backbone_param['blocks4']
+ width1 = self.backbone_param['width1']
+ width2 = self.backbone_param['width2']
+ width3 = self.backbone_param['width3']
+ width4 = self.backbone_param['width4']
+ out_h = self.backbone_param['out_h']
+ out_w = self.backbone_param['out_w']
+ feat_dim = self.backbone_param['feat_dim']
+ backbone = RepVGG([blocks1, blocks2, blocks3, blocks4],
+ [width1, width2, width3, width4],
+ feat_dim, out_h, out_w)
+ elif self.backbone_type == 'RepVGG_B0':
+ blocks1 = self.backbone_param['blocks1']
+ blocks2 = self.backbone_param['blocks2']
+ blocks3 = self.backbone_param['blocks3']
+ blocks4 = self.backbone_param['blocks4']
+ width1 = self.backbone_param['width1']
+ width2 = self.backbone_param['width2']
+ width3 = self.backbone_param['width3']
+ width4 = self.backbone_param['width4']
+ out_h = self.backbone_param['out_h']
+ out_w = self.backbone_param['out_w']
+ feat_dim = self.backbone_param['feat_dim']
+ backbone = RepVGG([blocks1, blocks2, blocks3, blocks4],
+ [width1, width2, width3, width4],
+ feat_dim, out_h, out_w)
+ elif self.backbone_type == 'RepVGG_B1':
+ blocks1 = self.backbone_param['blocks1']
+ blocks2 = self.backbone_param['blocks2']
+ blocks3 = self.backbone_param['blocks3']
+ blocks4 = self.backbone_param['blocks4']
+ width1 = self.backbone_param['width1']
+ width2 = self.backbone_param['width2']
+ width3 = self.backbone_param['width3']
+ width4 = self.backbone_param['width4']
+ out_h = self.backbone_param['out_h']
+ out_w = self.backbone_param['out_w']
+ feat_dim = self.backbone_param['feat_dim']
+ backbone = RepVGG([blocks1, blocks2, blocks3, blocks4],
+ [width1, width2, width3, width4],
+ feat_dim, out_h, out_w)
+
+ elif self.backbone_type == 'SwinTransformer_S':
+ img_size = self.backbone_param['img_size']
+ patch_size= self.backbone_param['patch_size']
+ in_chans = self.backbone_param['in_chans']
+ embed_dim = self.backbone_param['embed_dim']
+ depths = self.backbone_param['depths']
+ num_heads = self.backbone_param['num_heads']
+ window_size = self.backbone_param['window_size']
+ mlp_ratio = self.backbone_param['mlp_ratio']
+ drop_rate = self.backbone_param['drop_rate']
+ drop_path_rate = self.backbone_param['drop_path_rate']
+ backbone = SwinTransformer(img_size=img_size,
+ patch_size=patch_size,
+ in_chans=in_chans,
+ embed_dim=embed_dim,
+ depths=depths,
+ num_heads=num_heads,
+ window_size=window_size,
+ mlp_ratio=mlp_ratio,
+ qkv_bias=True,
+ qk_scale=None,
+ drop_rate=drop_rate,
+ drop_path_rate=drop_path_rate,
+ ape=False,
+ patch_norm=True,
+ use_checkpoint=False)
+
+ elif self.backbone_type == 'SwinTransformer_T':
+ img_size = self.backbone_param['img_size']
+ patch_size= self.backbone_param['patch_size']
+ in_chans = self.backbone_param['in_chans']
+ embed_dim = self.backbone_param['embed_dim']
+ depths = self.backbone_param['depths']
+ num_heads = self.backbone_param['num_heads']
+ window_size = self.backbone_param['window_size']
+ mlp_ratio = self.backbone_param['mlp_ratio']
+ drop_rate = self.backbone_param['drop_rate']
+ drop_path_rate = self.backbone_param['drop_path_rate']
+ backbone = SwinTransformer(img_size=img_size,
+ patch_size=patch_size,
+ in_chans=in_chans,
+ embed_dim=embed_dim,
+ depths=depths,
+ num_heads=num_heads,
+ window_size=window_size,
+ mlp_ratio=mlp_ratio,
+ qkv_bias=True,
+ qk_scale=None,
+ drop_rate=drop_rate,
+ drop_path_rate=drop_path_rate,
+ ape=False,
+ patch_norm=True,
+ use_checkpoint=False)
+ else:
+ pass
+ return backbone
diff --git a/bob/bio/facexzoo/backbones/models.py b/bob/bio/facexzoo/backbones/models.py
new file mode 100644
index 0000000000000000000000000000000000000000..1f1b1910bfe9bf2f21138993430a9703b8f58b4e
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/models.py
@@ -0,0 +1,70 @@
+# https://gitlab.idiap.ch/bob/bob.learn.pytorch/-/blob/master/bob/learn/pytorch/architectures/facexzoo/models.py
+
+import pkg_resources
+from bob.extension.download import get_file
+from bob.extension import rc
+# from bob.learn.pytorch.architectures.facexzoo.backbone_def import BackboneFactory
+from bob.bio.facexzoo.backbones.backbone_def import BackboneFactory
+
+# def_backbone_conf = pkg_resources.resource_filename( 'bob.learn.pytorch', 'architectures/facexzoo/backbone_conf.yaml')
+def_backbone_conf = pkg_resources.resource_filename( 'bob.bio.facexzoo', 'backbones/backbone_conf.yaml')
+
+info = {
+ 'AttentionNet': ['AttentionNet-f4c6f908.pt.tar.gz','49e435d8d9c075a4f613336090eac242'],
+ 'ResNeSt':['ResNeSt-e8b132d4.pt.tar.gz','51eef17ef7c17d1b22bbc13022393f31'],
+ 'MobileFaceNet': ['MobileFaceNet-ca475a8d.pt.tar.gz', 'e5fc0ae59d1a290b58a297b37f015e11'],
+ 'ResNet': ['ResNet-e07e7fa1.pt.tar.gz','13596dfeeb7f40c4b746ad2f0b271c36'],
+ 'EfficientNet':['EfficientNet-5aed534e.pt.tar.gz','31c827017fe2029c1ab57371c8e5abf4'],
+ 'TF-NAS':['TF-NAS-709d8562.pt.tar.gz','f96fe2683970140568a17c09fff24fab'],
+ 'HRNet':['HRNet-edc4da11.pt.tar.gz','5ed9920e004af440b623339a7008a758'],
+ 'ReXNet':['ReXNet-7c45620c.pt.tar.gz','b24cf257a25486c52fde5626007b324b'],
+ 'GhostNet':['GhostNet-5f026295.pt.tar.gz','9edb8327c62b62197ad023f21bd865bc']
+ }
+
+class FaceXZooModelFactory():
+ def __init__(self, arch='MobileFaceNet', head='MV-Softmax', backbone_conf= def_backbone_conf, info=info):
+ self.arch = arch
+ self.head = head
+ self.backbone_conf = backbone_conf
+ self.info=info
+
+ #assert(self.arch in self.info.keys())
+
+ def get_model(self):
+ return BackboneFactory(self.arch, self.backbone_conf).get_backbone()
+
+ def get_checkpoint_name(self):
+ #return self.info[self.arch][0]
+
+ if self.head == 'MV-Softmax':
+ checkpoint_name = self.arch
+ else:
+ checkpoint_name = self.head
+
+ return checkpoint_name
+
+ def get_facexzoo_file(self):
+ '''
+ #
+ urls = [
+ "https://www.idiap.ch/software/bob/data/bob/bob.learn.pytorch/facexzoomodels/{}".format(self.info[self.arch][0]),
+ "http://www.idiap.ch/software/bob/data/bob/bob.learn.pytorch/facexzoomodels/{}".format(self.info[self.arch][0]),
+ ]
+
+ return get_file(
+ self.info[self.arch][0],
+ urls,
+ cache_subdir="data/pytorch/{}/".format(self.info[self.arch][0]),
+ file_hash=self.info[self.arch][1],
+ extract=True,
+ )
+ '''
+ import os
+ # root_dir = os.path.dirname(os.getcwd())
+ root_dir = rc.get("facexzoo.checkpoints.directory")
+ if self.head == 'MV-Softmax':
+ model_path = f'{root_dir}/models/backbones/{self.arch}/Epoch_17.pt'
+ else:
+ model_path = f'{root_dir}/models/heads/{self.head}/Epoch_17.pt'
+
+ return model_path
\ No newline at end of file
diff --git a/bob/bio/facexzoo/backbones/resnest/__init__.py b/bob/bio/facexzoo/backbones/resnest/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..2acf216b90720c266e9582ab16b4204a8a072a25
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/resnest/__init__.py
@@ -0,0 +1,2 @@
+from .resnest import *
+from .ablation import *
diff --git a/bob/bio/facexzoo/backbones/resnest/ablation.py b/bob/bio/facexzoo/backbones/resnest/ablation.py
new file mode 100644
index 0000000000000000000000000000000000000000..00743ccdcf8c909b262c37476488c92ba947fde5
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/resnest/ablation.py
@@ -0,0 +1,106 @@
+##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+## Created by: Hang Zhang
+## Email: zhanghang0704@gmail.com
+## Copyright (c) 2020
+##
+## LICENSE file in the root directory of this source tree
+##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+"""ResNeSt ablation study models"""
+
+import torch
+from .resnet import ResNet, Bottleneck
+
+__all__ = ['resnest50_fast_1s1x64d', 'resnest50_fast_2s1x64d', 'resnest50_fast_4s1x64d',
+ 'resnest50_fast_1s2x40d', 'resnest50_fast_2s2x40d', 'resnest50_fast_4s2x40d',
+ 'resnest50_fast_1s4x24d']
+
+_url_format = 'https://s3.us-west-1.wasabisys.com/resnest/torch/{}-{}.pth'
+
+_model_sha256 = {name: checksum for checksum, name in [
+ ('d8fbf808', 'resnest50_fast_1s1x64d'),
+ ('44938639', 'resnest50_fast_2s1x64d'),
+ ('f74f3fc3', 'resnest50_fast_4s1x64d'),
+ ('32830b84', 'resnest50_fast_1s2x40d'),
+ ('9d126481', 'resnest50_fast_2s2x40d'),
+ ('41d14ed0', 'resnest50_fast_4s2x40d'),
+ ('d4a4f76f', 'resnest50_fast_1s4x24d'),
+ ]}
+
+def short_hash(name):
+ if name not in _model_sha256:
+ raise ValueError('Pretrained model for {name} is not available.'.format(name=name))
+ return _model_sha256[name][:8]
+
+resnest_model_urls = {name: _url_format.format(name, short_hash(name)) for
+ name in _model_sha256.keys()
+}
+
+def resnest50_fast_1s1x64d(pretrained=False, root='~/.encoding/models', **kwargs):
+ model = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=1, groups=1, bottleneck_width=64,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=True, **kwargs)
+ if pretrained:
+ model.load_state_dict(torch.hub.load_state_dict_from_url(
+ resnest_model_urls['resnest50_fast_1s1x64d'], progress=True, check_hash=True))
+ return model
+
+def resnest50_fast_2s1x64d(pretrained=False, root='~/.encoding/models', **kwargs):
+ model = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=2, groups=1, bottleneck_width=64,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=True, **kwargs)
+ if pretrained:
+ model.load_state_dict(torch.hub.load_state_dict_from_url(
+ resnest_model_urls['resnest50_fast_2s1x64d'], progress=True, check_hash=True))
+ return model
+
+def resnest50_fast_4s1x64d(pretrained=False, root='~/.encoding/models', **kwargs):
+ model = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=4, groups=1, bottleneck_width=64,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=True, **kwargs)
+ if pretrained:
+ model.load_state_dict(torch.hub.load_state_dict_from_url(
+ resnest_model_urls['resnest50_fast_4s1x64d'], progress=True, check_hash=True))
+ return model
+
+def resnest50_fast_1s2x40d(pretrained=False, root='~/.encoding/models', **kwargs):
+ model = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=1, groups=2, bottleneck_width=40,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=True, **kwargs)
+ if pretrained:
+ model.load_state_dict(torch.hub.load_state_dict_from_url(
+ resnest_model_urls['resnest50_fast_1s2x40d'], progress=True, check_hash=True))
+ return model
+
+def resnest50_fast_2s2x40d(pretrained=False, root='~/.encoding/models', **kwargs):
+ model = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=2, groups=2, bottleneck_width=40,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=True, **kwargs)
+ if pretrained:
+ model.load_state_dict(torch.hub.load_state_dict_from_url(
+ resnest_model_urls['resnest50_fast_2s2x40d'], progress=True, check_hash=True))
+ return model
+
+def resnest50_fast_4s2x40d(pretrained=False, root='~/.encoding/models', **kwargs):
+ model = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=4, groups=2, bottleneck_width=40,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=True, **kwargs)
+ if pretrained:
+ model.load_state_dict(torch.hub.load_state_dict_from_url(
+ resnest_model_urls['resnest50_fast_4s2x40d'], progress=True, check_hash=True))
+ return model
+
+def resnest50_fast_1s4x24d(pretrained=False, root='~/.encoding/models', **kwargs):
+ model = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=1, groups=4, bottleneck_width=24,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=True, **kwargs)
+ if pretrained:
+ model.load_state_dict(torch.hub.load_state_dict_from_url(
+ resnest_model_urls['resnest50_fast_1s4x24d'], progress=True, check_hash=True))
+ return model
diff --git a/bob/bio/facexzoo/backbones/resnest/resnest.py b/bob/bio/facexzoo/backbones/resnest/resnest.py
new file mode 100644
index 0000000000000000000000000000000000000000..dcc673957a1e816f7b73931cab18f63ceefdcd11
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/resnest/resnest.py
@@ -0,0 +1,60 @@
+"""
+@author: Jun Wang
+@date: 20210301
+@contact: jun21wangustc@gmail.com
+"""
+
+# based on:
+# https://github.com/zhanghang1989/ResNeSt/blob/master/resnest/torch/resnest.py
+
+import torch
+import torch.nn as nn
+from .resnet import ResNet, Bottleneck
+
+class Flatten(nn.Module):
+ def forward(self, input):
+ return input.view(input.size(0), -1)
+
+def l2_norm(input,axis=1):
+ norm = torch.norm(input,2,axis,True)
+ output = torch.div(input, norm)
+ return output
+
+class ResNeSt(nn.Module):
+ def __init__(self, num_layers, drop_ratio, feat_dim, out_h=7, out_w=7):
+ super(ResNeSt, self).__init__()
+ self.input_layer = nn.Sequential(nn.Conv2d(3, 64, (3, 3), 1, 1 ,bias=False),
+ nn.BatchNorm2d(64),
+ nn.PReLU(64))
+ self.output_layer = nn.Sequential(nn.BatchNorm2d(2048),
+ nn.Dropout(drop_ratio),
+ Flatten(),
+ nn.Linear(2048 * out_h * out_w, feat_dim),
+ nn.BatchNorm1d(feat_dim))
+ if num_layers == 50:
+ self.body = ResNet(Bottleneck, [3, 4, 6, 3],
+ radix=2, groups=1, bottleneck_width=64,
+ deep_stem=True, stem_width=32, avg_down=True,
+ avd=True, avd_first=False)
+ elif num_layers == 101:
+ self.body = ResNet(Bottleneck, [3, 4, 23, 3],
+ radix=2, groups=1, bottleneck_width=64,
+ deep_stem=True, stem_width=64, avg_down=True,
+ avd=True, avd_first=False)
+ elif num_layers == 200:
+ self.body = ResNet(Bottleneck, [3, 24, 36, 3],
+ radix=2, groups=1, bottleneck_width=64,
+ deep_stem=True, stem_width=64, avg_down=True,
+ avd=True, avd_first=False)
+ elif num_layers == 269:
+ self.body = ResNet(Bottleneck, [3, 30, 48, 8],
+ radix=2, groups=1, bottleneck_width=64,
+ deep_stem=True, stem_width=64, avg_down=True,
+ avd=True, avd_first=False)
+ else:
+ pass
+ def forward(self, x):
+ x = self.input_layer(x)
+ x = self.body(x)
+ x = self.output_layer(x)
+ return l2_norm(x)
diff --git a/bob/bio/facexzoo/backbones/resnest/resnet.py b/bob/bio/facexzoo/backbones/resnest/resnet.py
new file mode 100644
index 0000000000000000000000000000000000000000..1ae6083a388cf3eb7b8a73197e13fb783fdce8fe
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/resnest/resnet.py
@@ -0,0 +1,310 @@
+##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+## Created by: Hang Zhang
+## Email: zhanghang0704@gmail.com
+## Copyright (c) 2020
+##
+## LICENSE file in the root directory of this source tree
+##+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+"""ResNet variants"""
+import math
+import torch
+import torch.nn as nn
+
+from .splat import SplAtConv2d
+
+__all__ = ['ResNet', 'Bottleneck']
+
+class DropBlock2D(object):
+ def __init__(self, *args, **kwargs):
+ raise NotImplementedError
+
+class GlobalAvgPool2d(nn.Module):
+ def __init__(self):
+ """Global average pooling over the input's spatial dimensions"""
+ super(GlobalAvgPool2d, self).__init__()
+
+ def forward(self, inputs):
+ return nn.functional.adaptive_avg_pool2d(inputs, 1).view(inputs.size(0), -1)
+
+class Bottleneck(nn.Module):
+ """ResNet Bottleneck
+ """
+ # pylint: disable=unused-argument
+ expansion = 4
+ def __init__(self, inplanes, planes, stride=1, downsample=None,
+ radix=1, cardinality=1, bottleneck_width=64,
+ avd=False, avd_first=False, dilation=1, is_first=False,
+ rectified_conv=False, rectify_avg=False,
+ norm_layer=None, dropblock_prob=0.0, last_gamma=False):
+ super(Bottleneck, self).__init__()
+ group_width = int(planes * (bottleneck_width / 64.)) * cardinality
+ self.conv1 = nn.Conv2d(inplanes, group_width, kernel_size=1, bias=False)
+ self.bn1 = norm_layer(group_width)
+ self.dropblock_prob = dropblock_prob
+ self.radix = radix
+ self.avd = avd and (stride > 1 or is_first)
+ self.avd_first = avd_first
+
+ if self.avd:
+ self.avd_layer = nn.AvgPool2d(3, stride, padding=1)
+ stride = 1
+
+ if dropblock_prob > 0.0:
+ self.dropblock1 = DropBlock2D(dropblock_prob, 3)
+ if radix == 1:
+ self.dropblock2 = DropBlock2D(dropblock_prob, 3)
+ self.dropblock3 = DropBlock2D(dropblock_prob, 3)
+
+ if radix >= 1:
+ self.conv2 = SplAtConv2d(
+ group_width, group_width, kernel_size=3,
+ stride=stride, padding=dilation,
+ dilation=dilation, groups=cardinality, bias=False,
+ radix=radix, rectify=rectified_conv,
+ rectify_avg=rectify_avg,
+ norm_layer=norm_layer,
+ dropblock_prob=dropblock_prob)
+ elif rectified_conv:
+ from rfconv import RFConv2d
+ self.conv2 = RFConv2d(
+ group_width, group_width, kernel_size=3, stride=stride,
+ padding=dilation, dilation=dilation,
+ groups=cardinality, bias=False,
+ average_mode=rectify_avg)
+ self.bn2 = norm_layer(group_width)
+ else:
+ self.conv2 = nn.Conv2d(
+ group_width, group_width, kernel_size=3, stride=stride,
+ padding=dilation, dilation=dilation,
+ groups=cardinality, bias=False)
+ self.bn2 = norm_layer(group_width)
+
+ self.conv3 = nn.Conv2d(
+ group_width, planes * 4, kernel_size=1, bias=False)
+ self.bn3 = norm_layer(planes*4)
+
+ if last_gamma:
+ from torch.nn.init import zeros_
+ zeros_(self.bn3.weight)
+ self.relu = nn.ReLU(inplace=True)
+ self.downsample = downsample
+ self.dilation = dilation
+ self.stride = stride
+
+ def forward(self, x):
+ residual = x
+
+ out = self.conv1(x)
+ out = self.bn1(out)
+ if self.dropblock_prob > 0.0:
+ out = self.dropblock1(out)
+ out = self.relu(out)
+
+ if self.avd and self.avd_first:
+ out = self.avd_layer(out)
+
+ out = self.conv2(out)
+ if self.radix == 0:
+ out = self.bn2(out)
+ if self.dropblock_prob > 0.0:
+ out = self.dropblock2(out)
+ out = self.relu(out)
+
+ if self.avd and not self.avd_first:
+ out = self.avd_layer(out)
+
+ out = self.conv3(out)
+ out = self.bn3(out)
+ if self.dropblock_prob > 0.0:
+ out = self.dropblock3(out)
+
+ if self.downsample is not None:
+ residual = self.downsample(x)
+
+ out += residual
+ out = self.relu(out)
+
+ return out
+
+class ResNet(nn.Module):
+ """ResNet Variants
+
+ Parameters
+ ----------
+ block : Block
+ Class for the residual block. Options are BasicBlockV1, BottleneckV1.
+ layers : list of int
+ Numbers of layers in each block
+ classes : int, default 1000
+ Number of classification classes.
+ dilated : bool, default False
+ Applying dilation strategy to pretrained ResNet yielding a stride-8 model,
+ typically used in Semantic Segmentation.
+ norm_layer : object
+ Normalization layer used in backbone network (default: :class:`mxnet.gluon.nn.BatchNorm`;
+ for Synchronized Cross-GPU BachNormalization).
+
+ Reference:
+
+ - He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
+
+ - Yu, Fisher, and Vladlen Koltun. "Multi-scale context aggregation by dilated convolutions."
+ """
+ # pylint: disable=unused-variable
+ def __init__(self, block, layers, radix=1, groups=1, bottleneck_width=64,
+ num_classes=1000, dilated=False, dilation=1,
+ deep_stem=False, stem_width=64, avg_down=False,
+ rectified_conv=False, rectify_avg=False,
+ avd=False, avd_first=False,
+ final_drop=0.0, dropblock_prob=0,
+ last_gamma=False, norm_layer=nn.BatchNorm2d):
+ self.cardinality = groups
+ self.bottleneck_width = bottleneck_width
+ # ResNet-D params
+ self.inplanes = stem_width*2 if deep_stem else 64
+ self.avg_down = avg_down
+ self.last_gamma = last_gamma
+ # ResNeSt params
+ self.radix = radix
+ self.avd = avd
+ self.avd_first = avd_first
+
+ super(ResNet, self).__init__()
+ self.rectified_conv = rectified_conv
+ self.rectify_avg = rectify_avg
+ if rectified_conv:
+ from rfconv import RFConv2d
+ conv_layer = RFConv2d
+ else:
+ conv_layer = nn.Conv2d
+ conv_kwargs = {'average_mode': rectify_avg} if rectified_conv else {}
+ '''
+ if deep_stem:
+ self.conv1 = nn.Sequential(
+ conv_layer(3, stem_width, kernel_size=3, stride=2, padding=1, bias=False, **conv_kwargs),
+ norm_layer(stem_width),
+ nn.ReLU(inplace=True),
+ conv_layer(stem_width, stem_width, kernel_size=3, stride=1, padding=1, bias=False, **conv_kwargs),
+ norm_layer(stem_width),
+ nn.ReLU(inplace=True),
+ conv_layer(stem_width, stem_width*2, kernel_size=3, stride=1, padding=1, bias=False, **conv_kwargs),
+ )
+ else:
+ self.conv1 = conv_layer(3, 64, kernel_size=7, stride=2, padding=3,
+ bias=False, **conv_kwargs)
+ self.bn1 = norm_layer(self.inplanes)
+ self.relu = nn.ReLU(inplace=True)
+ self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+ '''
+ #self.layer1 = self._make_layer(block, 64, layers[0], norm_layer=norm_layer, is_first=False)
+ self.layer1 = self._make_layer(block, 64, layers[0], stride=2, norm_layer=norm_layer, is_first=False)
+ self.layer2 = self._make_layer(block, 128, layers[1], stride=2, norm_layer=norm_layer)
+ if dilated or dilation == 4:
+ self.layer3 = self._make_layer(block, 256, layers[2], stride=1,
+ dilation=2, norm_layer=norm_layer,
+ dropblock_prob=dropblock_prob)
+ self.layer4 = self._make_layer(block, 512, layers[3], stride=1,
+ dilation=4, norm_layer=norm_layer,
+ dropblock_prob=dropblock_prob)
+ elif dilation==2:
+ self.layer3 = self._make_layer(block, 256, layers[2], stride=2,
+ dilation=1, norm_layer=norm_layer,
+ dropblock_prob=dropblock_prob)
+ self.layer4 = self._make_layer(block, 512, layers[3], stride=1,
+ dilation=2, norm_layer=norm_layer,
+ dropblock_prob=dropblock_prob)
+ else:
+ self.layer3 = self._make_layer(block, 256, layers[2], stride=2,
+ norm_layer=norm_layer,
+ dropblock_prob=dropblock_prob)
+ self.layer4 = self._make_layer(block, 512, layers[3], stride=2,
+ norm_layer=norm_layer,
+ dropblock_prob=dropblock_prob)
+ '''
+ self.avgpool = GlobalAvgPool2d()
+ self.drop = nn.Dropout(final_drop) if final_drop > 0.0 else None
+ self.fc = nn.Linear(512 * block.expansion, num_classes)
+
+ for m in self.modules():
+ if isinstance(m, nn.Conv2d):
+ n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
+ m.weight.data.normal_(0, math.sqrt(2. / n))
+ elif isinstance(m, norm_layer):
+ m.weight.data.fill_(1)
+ m.bias.data.zero_()
+ '''
+ def _make_layer(self, block, planes, blocks, stride=1, dilation=1, norm_layer=None,
+ dropblock_prob=0.0, is_first=True):
+ downsample = None
+ if stride != 1 or self.inplanes != planes * block.expansion:
+ down_layers = []
+ if self.avg_down:
+ if dilation == 1:
+ down_layers.append(nn.AvgPool2d(kernel_size=stride, stride=stride,
+ ceil_mode=True, count_include_pad=False))
+ else:
+ down_layers.append(nn.AvgPool2d(kernel_size=1, stride=1,
+ ceil_mode=True, count_include_pad=False))
+ down_layers.append(nn.Conv2d(self.inplanes, planes * block.expansion,
+ kernel_size=1, stride=1, bias=False))
+ else:
+ down_layers.append(nn.Conv2d(self.inplanes, planes * block.expansion,
+ kernel_size=1, stride=stride, bias=False))
+ down_layers.append(norm_layer(planes * block.expansion))
+ downsample = nn.Sequential(*down_layers)
+
+ layers = []
+ if dilation == 1 or dilation == 2:
+ layers.append(block(self.inplanes, planes, stride, downsample=downsample,
+ radix=self.radix, cardinality=self.cardinality,
+ bottleneck_width=self.bottleneck_width,
+ avd=self.avd, avd_first=self.avd_first,
+ dilation=1, is_first=is_first, rectified_conv=self.rectified_conv,
+ rectify_avg=self.rectify_avg,
+ norm_layer=norm_layer, dropblock_prob=dropblock_prob,
+ last_gamma=self.last_gamma))
+ elif dilation == 4:
+ layers.append(block(self.inplanes, planes, stride, downsample=downsample,
+ radix=self.radix, cardinality=self.cardinality,
+ bottleneck_width=self.bottleneck_width,
+ avd=self.avd, avd_first=self.avd_first,
+ dilation=2, is_first=is_first, rectified_conv=self.rectified_conv,
+ rectify_avg=self.rectify_avg,
+ norm_layer=norm_layer, dropblock_prob=dropblock_prob,
+ last_gamma=self.last_gamma))
+ else:
+ raise RuntimeError("=> unknown dilation size: {}".format(dilation))
+
+ self.inplanes = planes * block.expansion
+ for i in range(1, blocks):
+ layers.append(block(self.inplanes, planes,
+ radix=self.radix, cardinality=self.cardinality,
+ bottleneck_width=self.bottleneck_width,
+ avd=self.avd, avd_first=self.avd_first,
+ dilation=dilation, rectified_conv=self.rectified_conv,
+ rectify_avg=self.rectify_avg,
+ norm_layer=norm_layer, dropblock_prob=dropblock_prob,
+ last_gamma=self.last_gamma))
+
+ return nn.Sequential(*layers)
+
+ def forward(self, x):
+ '''
+ x = self.conv1(x)
+ x = self.bn1(x)
+ x = self.relu(x)
+ x = self.maxpool(x)
+ '''
+ x = self.layer1(x)
+ x = self.layer2(x)
+ x = self.layer3(x)
+ x = self.layer4(x)
+ '''
+ x = self.avgpool(x)
+ #x = x.view(x.size(0), -1)
+ x = torch.flatten(x, 1)
+ if self.drop:
+ x = self.drop(x)
+ x = self.fc(x)
+ '''
+ return x
diff --git a/bob/bio/facexzoo/backbones/resnest/splat.py b/bob/bio/facexzoo/backbones/resnest/splat.py
new file mode 100644
index 0000000000000000000000000000000000000000..c3f21b19ac75534521b9a0eae957e8ee454f1cd4
--- /dev/null
+++ b/bob/bio/facexzoo/backbones/resnest/splat.py
@@ -0,0 +1,99 @@
+"""Split-Attention"""
+
+import torch
+from torch import nn
+import torch.nn.functional as F
+from torch.nn import Conv2d, Module, Linear, BatchNorm2d, ReLU
+from torch.nn.modules.utils import _pair
+
+__all__ = ['SplAtConv2d']
+
+class SplAtConv2d(Module):
+ """Split-Attention Conv2d
+ """
+ def __init__(self, in_channels, channels, kernel_size, stride=(1, 1), padding=(0, 0),
+ dilation=(1, 1), groups=1, bias=True,
+ radix=2, reduction_factor=4,
+ rectify=False, rectify_avg=False, norm_layer=None,
+ dropblock_prob=0.0, **kwargs):
+ super(SplAtConv2d, self).__init__()
+ padding = _pair(padding)
+ self.rectify = rectify and (padding[0] > 0 or padding[1] > 0)
+ self.rectify_avg = rectify_avg
+ inter_channels = max(in_channels*radix//reduction_factor, 32)
+ self.radix = radix
+ self.cardinality = groups
+ self.channels = channels
+ self.dropblock_prob = dropblock_prob
+ if self.rectify:
+ from rfconv import RFConv2d
+ self.conv = RFConv2d(in_channels, channels*radix, kernel_size, stride, padding, dilation,
+ groups=groups*radix, bias=bias, average_mode=rectify_avg, **kwargs)
+ else:
+ self.conv = Conv2d(in_channels, channels*radix, kernel_size, stride, padding, dilation,
+ groups=groups*radix, bias=bias, **kwargs)
+ self.use_bn = norm_layer is not None
+ if self.use_bn:
+ self.bn0 = norm_layer(channels*radix)
+ self.relu = ReLU(inplace=True)
+ self.fc1 = Conv2d(channels, inter_channels, 1, groups=self.cardinality)
+ if self.use_bn:
+ self.bn1 = norm_layer(inter_channels)
+ self.fc2 = Conv2d(inter_channels, channels*radix, 1, groups=self.cardinality)
+ if dropblock_prob > 0.0:
+ self.dropblock = DropBlock2D(dropblock_prob, 3)
+ self.rsoftmax = rSoftMax(radix, groups)
+
+ def forward(self, x):
+ x = self.conv(x)
+ if self.use_bn:
+ x = self.bn0(x)
+ if self.dropblock_prob > 0.0:
+ x = self.dropblock(x)
+ x = self.relu(x)
+
+ batch, rchannel = x.shape[:2]
+ if self.radix > 1:
+ if torch.__version__ < '1.5':
+ splited = torch.split(x, int(rchannel//self.radix), dim=1)
+ else:
+ splited = torch.split(x, rchannel//self.radix, dim=1)
+ gap = sum(splited)
+ else:
+ gap = x
+ gap = F.adaptive_avg_pool2d(gap, 1)
+ gap = self.fc1(gap)
+
+ if self.use_bn:
+ gap = self.bn1(gap)
+ gap = self.relu(gap)
+
+ atten = self.fc2(gap)
+ atten = self.rsoftmax(atten).view(batch, -1, 1, 1)
+
+ if self.radix > 1:
+ if torch.__version__ < '1.5':
+ attens = torch.split(atten, int(rchannel//self.radix), dim=1)
+ else:
+ attens = torch.split(atten, rchannel//self.radix, dim=1)
+ out = sum([att*split for (att, split) in zip(attens, splited)])
+ else:
+ out = atten * x
+ return out.contiguous()
+
+class rSoftMax(nn.Module):
+ def __init__(self, radix, cardinality):
+ super().__init__()
+ self.radix = radix
+ self.cardinality = cardinality
+
+ def forward(self, x):
+ batch = x.size(0)
+ if self.radix > 1:
+ x = x.view(batch, self.cardinality, self.radix, -1).transpose(1, 2)
+ x = F.softmax(x, dim=1)
+ x = x.reshape(batch, -1)
+ else:
+ x = torch.sigmoid(x)
+ return x
+
diff --git a/bob/bio/facexzoo/config/attention_net.py b/bob/bio/facexzoo/config/attention_net.py
new file mode 100644
index 0000000000000000000000000000000000000000..fe7d9cb088f872ec3130cde40933c8367e373fa6
--- /dev/null
+++ b/bob/bio/facexzoo/config/attention_net.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import AttentionNet
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return AttentionNet(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/config/efficient_net.py b/bob/bio/facexzoo/config/efficient_net.py
new file mode 100644
index 0000000000000000000000000000000000000000..d521ecc049e7f617fd966438a9644ea04d7bdc5b
--- /dev/null
+++ b/bob/bio/facexzoo/config/efficient_net.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import EfficientNet
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return EfficientNet(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/config/ghost_net.py b/bob/bio/facexzoo/config/ghost_net.py
new file mode 100644
index 0000000000000000000000000000000000000000..8ed58ea4ad305d81af3652b0fe4a7d0f61b6aea5
--- /dev/null
+++ b/bob/bio/facexzoo/config/ghost_net.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import GhostNet
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return GhostNet(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/config/hr_net.py b/bob/bio/facexzoo/config/hr_net.py
new file mode 100644
index 0000000000000000000000000000000000000000..eb330256030d270f5ee8998fb8ac86e2f24433e7
--- /dev/null
+++ b/bob/bio/facexzoo/config/hr_net.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import HRNet
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return HRNet(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/config/mobile_facenet.py b/bob/bio/facexzoo/config/mobile_facenet.py
new file mode 100644
index 0000000000000000000000000000000000000000..cfc3f98bd074ddf599e9b96dd977975d858f0ae5
--- /dev/null
+++ b/bob/bio/facexzoo/config/mobile_facenet.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import MobileFaceNet
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return MobileFaceNet(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/config/resne_st.py b/bob/bio/facexzoo/config/resne_st.py
new file mode 100644
index 0000000000000000000000000000000000000000..116d9a6394ebcc4f8f1435cd363367e8d9a0189f
--- /dev/null
+++ b/bob/bio/facexzoo/config/resne_st.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import ResNeSt
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return ResNeSt(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/config/rex_net.py b/bob/bio/facexzoo/config/rex_net.py
new file mode 100644
index 0000000000000000000000000000000000000000..e8cc81c6dc8cf323eb9d7ea5a6db945049b5bf98
--- /dev/null
+++ b/bob/bio/facexzoo/config/rex_net.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import ReXNet
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return ReXNet(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/config/tf_nas.py b/bob/bio/facexzoo/config/tf_nas.py
new file mode 100644
index 0000000000000000000000000000000000000000..000caa7d8fed7e89757e30ffaf2213fc91bd44bc
--- /dev/null
+++ b/bob/bio/facexzoo/config/tf_nas.py
@@ -0,0 +1,15 @@
+from bob.bio.face.embeddings.pytorch import TF_NAS
+from bob.bio.face.utils import lookup_config_from_database
+
+
+annotation_type, fixed_positions, memory_demanding = lookup_config_from_database(
+ locals().get("database")
+)
+
+
+def load(annotation_type, fixed_positions=None, memory_demanding=False):
+ return TF_NAS(annotation_type, fixed_positions, memory_demanding)
+
+
+pipeline = load(annotation_type, fixed_positions, memory_demanding)
+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/.gitkeep b/bob/bio/facexzoo/models/backbones/AttentionNet56/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..eb3651de373f9a80020fe29bb41f5f1dfefd813a
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.9811666666666667 | 0.0024222477062879537 |
+| Epoch_17_batch_5999.pt | 0.9811666666666667 | 0.002318071782780554 |
+| Epoch_15_batch_5999.pt | 0.9808333333333333 | 0.0023339946152888865 |
+| Epoch_15.pt | 0.9806666666666668 | 0.002238826853289986 |
+| Epoch_13_batch_2999.pt | 0.9806666666666667 | 0.002411392712690075 |
+| Epoch_14_batch_2999.pt | 0.9805000000000001 | 0.002460176647276053 |
+| Epoch_15_batch_2999.pt | 0.9803333333333333 | 0.002380476142847622 |
+| Epoch_12_batch_2999.pt | 0.9801666666666667 | 0.0022966696231771513 |
+| Epoch_16_batch_2999.pt | 0.9801666666666667 | 0.0023498095536174015 |
+| Epoch_16_batch_5999.pt | 0.9801666666666667 | 0.0024400212506664135 |
+| Epoch_17_batch_2999.pt | 0.9800000000000001 | 0.002208289657150194 |
+| Epoch_14.pt | 0.9798333333333333 | 0.0019945914523351368 |
+| Epoch_11.pt | 0.9796666666666667 | 0.00216310248154797 |
+| Epoch_14_batch_5999.pt | 0.9796666666666667 | 0.0024444444444444466 |
+| Epoch_17.pt | 0.9796666666666665 | 0.002301368353023107 |
+| Epoch_13_batch_5999.pt | 0.9795 | 0.002208988372452335 |
+| Epoch_13.pt | 0.9791666666666666 | 0.0022532555322970224 |
+| Epoch_11_batch_2999.pt | 0.9785 | 0.002010005835013944 |
+| Epoch_10_batch_2999.pt | 0.9783333333333335 | 0.002033667246413678 |
+| Epoch_10.pt | 0.9781666666666669 | 0.0019633996718969203 |
+| Epoch_11_batch_5999.pt | 0.9778333333333332 | 0.0019412672455173456 |
+| Epoch_10_batch_5999.pt | 0.9768333333333334 | 0.0020253029037286853 |
+| Epoch_12_batch_5999.pt | 0.9768333333333332 | 0.00218651873935049 |
+| Epoch_12.pt | 0.9766666666666666 | 0.0018592445034090477 |
+| Epoch_7_batch_2999.pt | 0.9731666666666667 | 0.0016564501684306187 |
+| Epoch_8.pt | 0.9731666666666667 | 0.002965334698907227 |
+| Epoch_8_batch_5999.pt | 0.9728333333333335 | 0.003268348017796163 |
+| Epoch_9_batch_5999.pt | 0.9724999999999999 | 0.0025489043882120995 |
+| Epoch_6_batch_5999.pt | 0.9716666666666665 | 0.0027102913150334395 |
+| Epoch_5_batch_5999.pt | 0.9708333333333334 | 0.0027470522922391394 |
+| Epoch_6.pt | 0.9701666666666668 | 0.002923404914871749 |
+| Epoch_8_batch_2999.pt | 0.9696666666666667 | 0.0020608041101101574 |
+| Epoch_7_batch_5999.pt | 0.9693333333333334 | 0.003124969135650052 |
+| Epoch_9_batch_2999.pt | 0.9693333333333334 | 0.002888888888888889 |
+| Epoch_5.pt | 0.9691666666666666 | 0.0028246052930858143 |
+| Epoch_6_batch_2999.pt | 0.9686666666666668 | 0.0022194427061598054 |
+| Epoch_5_batch_2999.pt | 0.9683333333333334 | 0.002908056072956086 |
+| Epoch_3_batch_2999.pt | 0.9678333333333334 | 0.002855036701673244 |
+| Epoch_9.pt | 0.9665000000000001 | 0.0033742854753467154 |
+| Epoch_4.pt | 0.9658333333333333 | 0.0033724556023947082 |
+| Epoch_4_batch_5999.pt | 0.9654999999999999 | 0.0026533230969465705 |
+| Epoch_4_batch_2999.pt | 0.9650000000000001 | 0.003191423692521133 |
+| Epoch_3_batch_5999.pt | 0.9636666666666667 | 0.0034318767136623336 |
+| Epoch_2_batch_5999.pt | 0.961 | 0.003173968190463492 |
+| Epoch_3.pt | 0.9603333333333334 | 0.0036834632120557377 |
+| Epoch_2.pt | 0.9586666666666666 | 0.004380089632094897 |
+| Epoch_2_batch_2999.pt | 0.9555 | 0.002778333277788891 |
+| Epoch_1_batch_5999.pt | 0.9451666666666666 | 0.003253203549323857 |
+| Epoch_1.pt | 0.9403333333333332 | 0.003863408589047492 |
+| Epoch_1_batch_2999.pt | 0.9273333333333333 | 0.0047544650878521425 |
+| Epoch_7.pt | 0.9211666666666666 | 0.003881341883268571 |
+| Epoch_0.pt | 0.8788333333333332 | 0.006206577554358891 |
+| Epoch_0_batch_5999.pt | 0.8743333333333332 | 0.007540835333758729 |
+| Epoch_0_batch_2999.pt | 0.6866666666666666 | 0.006075654717552141 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..9306a9d00bb6a0533a7b55591e69260aa26799af
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9565000000000001 | 0.003820429165989819 |
+| Epoch_11.pt | 0.9561666666666667 | 0.003352261075899024 |
+| Epoch_13_batch_5999.pt | 0.9561666666666667 | 0.0037354656608920606 |
+| Epoch_11_batch_5999.pt | 0.9560000000000001 | 0.0036951753662924263 |
+| Epoch_13.pt | 0.9560000000000001 | 0.003661612678507579 |
+| Epoch_15_batch_5999.pt | 0.9560000000000001 | 0.003567687635111632 |
+| Epoch_17_batch_5999.pt | 0.9560000000000001 | 0.0038184089505421334 |
+| Epoch_16_batch_5999.pt | 0.9559999999999998 | 0.0036531738272830294 |
+| Epoch_13_batch_2999.pt | 0.9558333333333333 | 0.003670452590924213 |
+| Epoch_15_batch_2999.pt | 0.9558333333333333 | 0.003670452590924213 |
+| Epoch_15.pt | 0.9558333333333333 | 0.0037863464914197212 |
+| Epoch_17.pt | 0.9558333333333333 | 0.003695592971013915 |
+| Epoch_14_batch_5999.pt | 0.9556666666666667 | 0.003777777777777784 |
+| Epoch_17_batch_2999.pt | 0.9556666666666667 | 0.0036021255727660788 |
+| Epoch_10_batch_5999.pt | 0.9555 | 0.003583656316193365 |
+| Epoch_11_batch_2999.pt | 0.9553333333333333 | 0.003741657386773944 |
+| Epoch_12_batch_2999.pt | 0.9553333333333333 | 0.003807075933113742 |
+| Epoch_12_batch_5999.pt | 0.9551666666666666 | 0.0038606114784648245 |
+| Epoch_16.pt | 0.9550000000000001 | 0.003865006029094689 |
+| Epoch_14_batch_2999.pt | 0.9548333333333334 | 0.003994208770670823 |
+| Epoch_14.pt | 0.9548333333333334 | 0.003663719354177233 |
+| Epoch_10_batch_2999.pt | 0.9548333333333332 | 0.003860611478464824 |
+| Epoch_10.pt | 0.9548333333333332 | 0.003900379848548468 |
+| Epoch_12.pt | 0.9545 | 0.00418735604138028 |
+| Epoch_7_batch_2999.pt | 0.9501666666666667 | 0.004703885969055426 |
+| Epoch_9_batch_5999.pt | 0.9501666666666667 | 0.003195772670726596 |
+| Epoch_7_batch_5999.pt | 0.95 | 0.0040292143032152825 |
+| Epoch_9_batch_2999.pt | 0.95 | 0.0037433067839828904 |
+| Epoch_8.pt | 0.9495000000000001 | 0.003913020367320065 |
+| Epoch_8_batch_2999.pt | 0.9491666666666667 | 0.0038429830234491657 |
+| Epoch_6_batch_2999.pt | 0.9486666666666667 | 0.004192880503136264 |
+| Epoch_4_batch_2999.pt | 0.9483333333333335 | 0.004044505494044734 |
+| Epoch_6_batch_5999.pt | 0.9481666666666667 | 0.003986474044646721 |
+| Epoch_6.pt | 0.9480000000000001 | 0.0036413265795942097 |
+| Epoch_5_batch_5999.pt | 0.9478333333333333 | 0.004267664813801918 |
+| Epoch_5_batch_2999.pt | 0.9466666666666667 | 0.004346134936801763 |
+| Epoch_8_batch_5999.pt | 0.9465 | 0.0037552432480269307 |
+| Epoch_4_batch_5999.pt | 0.9461666666666668 | 0.0031822229981377063 |
+| Epoch_4.pt | 0.9458333333333334 | 0.003289057803470151 |
+| Epoch_3_batch_5999.pt | 0.944 | 0.0031544599036840808 |
+| Epoch_3_batch_2999.pt | 0.9433333333333334 | 0.003333333333333324 |
+| Epoch_5.pt | 0.9433333333333334 | 0.004097575314352386 |
+| Epoch_9.pt | 0.9428333333333333 | 0.004513696577099908 |
+| Epoch_2_batch_5999.pt | 0.9406666666666668 | 0.003984538017120241 |
+| Epoch_3.pt | 0.9406666666666667 | 0.003929942040850531 |
+| Epoch_2.pt | 0.9395000000000001 | 0.004157768276015032 |
+| Epoch_2_batch_2999.pt | 0.9381666666666668 | 0.0038606114784648167 |
+| Epoch_1.pt | 0.9321666666666666 | 0.003913020367320074 |
+| Epoch_1_batch_5999.pt | 0.93 | 0.0030932024237944602 |
+| Epoch_1_batch_2999.pt | 0.9145 | 0.0033797691498344776 |
+| Epoch_7.pt | 0.9099999999999999 | 0.005061351988413502 |
+| Epoch_0_batch_5999.pt | 0.8826666666666668 | 0.005001234415523066 |
+| Epoch_0.pt | 0.8808333333333334 | 0.004369153942451501 |
+| Epoch_0_batch_2999.pt | 0.7048333333333333 | 0.009439377725880656 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3fdf437aefa4e5ce82b6f54aed4c492e2b1b4205
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.8918333333333333 | 0.005094477765552005 |
+| Epoch_17.pt | 0.8915000000000001 | 0.006067775641416538 |
+| Epoch_14_batch_2999.pt | 0.8911666666666667 | 0.005415170733603339 |
+| Epoch_16_batch_5999.pt | 0.8911666666666667 | 0.005426557920468327 |
+| Epoch_13_batch_2999.pt | 0.8906666666666668 | 0.006005141829720733 |
+| Epoch_13_batch_5999.pt | 0.8906666666666666 | 0.005943146275696148 |
+| Epoch_15_batch_2999.pt | 0.8906666666666666 | 0.005989703098641574 |
+| Epoch_15.pt | 0.8901666666666668 | 0.005990991179149072 |
+| Epoch_16_batch_2999.pt | 0.89 | 0.005643744488533468 |
+| Epoch_16.pt | 0.8898333333333334 | 0.005267240943842783 |
+| Epoch_14_batch_5999.pt | 0.8896666666666668 | 0.005836242660741736 |
+| Epoch_17_batch_2999.pt | 0.8896666666666666 | 0.005481731726019566 |
+| Epoch_13.pt | 0.8895 | 0.0059992283454408345 |
+| Epoch_17_batch_5999.pt | 0.8893333333333334 | 0.00603590082468492 |
+| Epoch_15_batch_5999.pt | 0.8889999999999999 | 0.005922336876105083 |
+| Epoch_12_batch_2999.pt | 0.8873333333333335 | 0.00563389214848321 |
+| Epoch_12.pt | 0.8871666666666667 | 0.005627588567621101 |
+| Epoch_12_batch_5999.pt | 0.8861666666666667 | 0.005842849380986033 |
+| Epoch_11.pt | 0.8853333333333332 | 0.004998765279645333 |
+| Epoch_10_batch_2999.pt | 0.8851666666666667 | 0.005580223014261071 |
+| Epoch_10.pt | 0.8846666666666666 | 0.005470459388929409 |
+| Epoch_11_batch_2999.pt | 0.8843333333333334 | 0.005785251576419916 |
+| Epoch_11_batch_5999.pt | 0.8833333333333334 | 0.0056382730917178955 |
+| Epoch_10_batch_5999.pt | 0.8826666666666666 | 0.005715476066494082 |
+| Epoch_9_batch_5999.pt | 0.8656666666666666 | 0.005283330412430552 |
+| Epoch_7_batch_5999.pt | 0.8620000000000001 | 0.005453507196359127 |
+| Epoch_6_batch_5999.pt | 0.8608333333333332 | 0.00489551318890227 |
+| Epoch_8_batch_5999.pt | 0.8591666666666666 | 0.00490810616967884 |
+| Epoch_9_batch_2999.pt | 0.8581666666666667 | 0.007480220832549239 |
+| Epoch_4_batch_5999.pt | 0.8575000000000002 | 0.005399188324915262 |
+| Epoch_8_batch_2999.pt | 0.8548333333333333 | 0.005202385492715321 |
+| Epoch_6_batch_2999.pt | 0.8536666666666667 | 0.006688851974796629 |
+| Epoch_8.pt | 0.8536666666666667 | 0.006439653391485378 |
+| Epoch_4_batch_2999.pt | 0.8535 | 0.006188649506641129 |
+| Epoch_7_batch_2999.pt | 0.8505 | 0.006858580279969766 |
+| Epoch_5_batch_2999.pt | 0.8500000000000002 | 0.0056710222911888165 |
+| Epoch_4.pt | 0.8493333333333333 | 0.004382907316292446 |
+| Epoch_3_batch_2999.pt | 0.8465000000000001 | 0.0052906276320244605 |
+| Epoch_5.pt | 0.8465 | 0.0051306991798461595 |
+| Epoch_6.pt | 0.8459999999999999 | 0.006750400079592905 |
+| Epoch_2_batch_2999.pt | 0.8458333333333334 | 0.005812130779458671 |
+| Epoch_3.pt | 0.845 | 0.006034878050666784 |
+| Epoch_9.pt | 0.8448333333333334 | 0.006365658248623751 |
+| Epoch_3_batch_5999.pt | 0.8445 | 0.0062856391289398145 |
+| Epoch_5_batch_5999.pt | 0.8418333333333333 | 0.007033649282003063 |
+| Epoch_2_batch_5999.pt | 0.8416666666666666 | 0.007079792143143992 |
+| Epoch_2.pt | 0.8356666666666666 | 0.004863570806275398 |
+| Epoch_1_batch_5999.pt | 0.8258333333333333 | 0.0068729654677575445 |
+| Epoch_1.pt | 0.8211666666666666 | 0.007445480856545478 |
+| Epoch_1_batch_2999.pt | 0.8041666666666666 | 0.006555790956225525 |
+| Epoch_7.pt | 0.7866666666666665 | 0.009142223302544095 |
+| Epoch_0.pt | 0.7416666666666666 | 0.008138462256546656 |
+| Epoch_0_batch_5999.pt | 0.7373333333333333 | 0.008700078047045971 |
+| Epoch_0_batch_2999.pt | 0.6186666666666667 | 0.006848447602771095 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..afa654d8ed47577d9d139e1ba238ba153e97eaaf
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+------------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+------------------------+
+| Epoch_17.pt | 0.9988333333333334 | 0.0005000000000000021 |
+| Epoch_11.pt | 0.9986666666666666 | 0.0005443310539518171 |
+| Epoch_14_batch_2999.pt | 0.9986666666666666 | 0.0005443310539518171 |
+| Epoch_16.pt | 0.9986666666666666 | 0.00048432210483785436 |
+| Epoch_12_batch_5999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_13.pt | 0.9984999999999999 | 0.000524110062892033 |
+| Epoch_15_batch_5999.pt | 0.9984999999999999 | 0.000524110062892033 |
+| Epoch_15.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_11_batch_5999.pt | 0.9983333333333334 | 0.0006573421981221791 |
+| Epoch_13_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_13_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_14_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_14.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_15_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_16_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_16_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_17_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_17_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_10_batch_5999.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_10.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_12_batch_2999.pt | 0.9981666666666665 | 0.0005800170282728054 |
+| Epoch_10_batch_2999.pt | 0.998 | 0.0007370277311900908 |
+| Epoch_11_batch_2999.pt | 0.998 | 0.0006478835438716985 |
+| Epoch_12.pt | 0.998 | 0.0007370277311900908 |
+| Epoch_9_batch_2999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_7_batch_5999.pt | 0.9973333333333333 | 0.0009026709338484372 |
+| Epoch_5_batch_5999.pt | 0.9971666666666665 | 0.0008624541497922201 |
+| Epoch_4_batch_5999.pt | 0.9970000000000001 | 0.0010482201257840675 |
+| Epoch_4.pt | 0.9970000000000001 | 0.0011600340565456147 |
+| Epoch_8_batch_2999.pt | 0.9970000000000001 | 0.00116003405654562 |
+| Epoch_6_batch_2999.pt | 0.9968333333333333 | 0.001228519132638666 |
+| Epoch_3.pt | 0.9966666666666667 | 0.0007856742013183885 |
+| Epoch_6_batch_5999.pt | 0.9966666666666667 | 0.0010540925533894607 |
+| Epoch_6.pt | 0.9966666666666667 | 0.0009938079899999065 |
+| Epoch_7_batch_2999.pt | 0.9966666666666667 | 0.0010829771494232203 |
+| Epoch_4_batch_2999.pt | 0.9964999999999999 | 0.0009111788592698183 |
+| Epoch_3_batch_5999.pt | 0.9963333333333333 | 0.0011600340565456205 |
+| Epoch_9_batch_5999.pt | 0.9963333333333333 | 0.001077262190536964 |
+| Epoch_3_batch_2999.pt | 0.9961666666666666 | 0.0012184284555256323 |
+| Epoch_5_batch_2999.pt | 0.9961666666666666 | 0.000897527467855748 |
+| Epoch_5.pt | 0.9959999999999999 | 0.0009686442096757059 |
+| Epoch_8.pt | 0.9958333333333333 | 0.0012729376930432875 |
+| Epoch_9.pt | 0.9958333333333332 | 0.0008695819912499205 |
+| Epoch_2_batch_2999.pt | 0.9953333333333333 | 0.0009875771574795078 |
+| Epoch_2_batch_5999.pt | 0.9951666666666666 | 0.0009444444444444449 |
+| Epoch_2.pt | 0.9950000000000001 | 0.0007027283689263051 |
+| Epoch_8_batch_5999.pt | 0.9948333333333332 | 0.0013933262448871625 |
+| Epoch_1_batch_5999.pt | 0.9946666666666666 | 0.0011055415967851294 |
+| Epoch_1.pt | 0.9945 | 0.001112499133027818 |
+| Epoch_1_batch_2999.pt | 0.9936666666666667 | 0.0011863420280034847 |
+| Epoch_7.pt | 0.9914999999999999 | 0.001458267194267416 |
+| Epoch_0_batch_5999.pt | 0.9844999999999999 | 0.0020645449084513334 |
+| Epoch_0.pt | 0.9831666666666665 | 0.001580162517036429 |
+| Epoch_0_batch_2999.pt | 0.9235000000000001 | 0.0038606114784648154 |
++------------------------+--------------------+------------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..5c352cd1d388ad8980316dc82ed8f3561a81dc26
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.9651666666666667 | 0.002490103870112779 |
+| Epoch_15.pt | 0.9641666666666667 | 0.0026323017201071394 |
+| Epoch_16_batch_5999.pt | 0.9640000000000001 | 0.0023856567281759903 |
+| Epoch_17_batch_5999.pt | 0.9640000000000001 | 0.0026550673656330105 |
+| Epoch_14_batch_2999.pt | 0.9628333333333334 | 0.002222916558193615 |
+| Epoch_16_batch_2999.pt | 0.9626666666666667 | 0.0020964402515681346 |
+| Epoch_17.pt | 0.9626666666666667 | 0.002608154354290107 |
+| Epoch_17_batch_2999.pt | 0.9620000000000001 | 0.0024190601174530276 |
+| Epoch_15_batch_2999.pt | 0.9618333333333334 | 0.0026229048075806383 |
+| Epoch_13.pt | 0.9616666666666667 | 0.0027777777777777805 |
+| Epoch_12_batch_5999.pt | 0.9613333333333335 | 0.0022194427061597963 |
+| Epoch_16.pt | 0.9613333333333334 | 0.0028631330503833662 |
+| Epoch_13_batch_2999.pt | 0.9611666666666668 | 0.0029297326385411587 |
+| Epoch_12.pt | 0.9603333333333334 | 0.002419060117453026 |
+| Epoch_14_batch_5999.pt | 0.9601666666666666 | 0.0023100692095879816 |
+| Epoch_14.pt | 0.9601666666666666 | 0.0027827732858596 |
+| Epoch_11_batch_5999.pt | 0.959 | 0.002372684056006953 |
+| Epoch_13_batch_5999.pt | 0.9586666666666666 | 0.002662033011269099 |
+| Epoch_10.pt | 0.9576666666666667 | 0.0025117010123238514 |
+| Epoch_11.pt | 0.9570000000000001 | 0.002830608711746006 |
+| Epoch_12_batch_2999.pt | 0.9561666666666667 | 0.002485141027371671 |
+| Epoch_11_batch_2999.pt | 0.9558333333333332 | 0.0027805541680538254 |
+| Epoch_10_batch_5999.pt | 0.9549999999999998 | 0.0026874192494328528 |
+| Epoch_10_batch_2999.pt | 0.9526666666666668 | 0.0031348302177035235 |
+| Epoch_7_batch_5999.pt | 0.9345000000000001 | 0.0027783332777888805 |
+| Epoch_7_batch_2999.pt | 0.9336666666666666 | 0.003395712619985799 |
+| Epoch_6_batch_5999.pt | 0.9326666666666666 | 0.0034623192113073043 |
+| Epoch_9_batch_2999.pt | 0.9321666666666667 | 0.0032683480177961625 |
+| Epoch_8.pt | 0.9313333333333335 | 0.0030812054719693382 |
+| Epoch_8_batch_5999.pt | 0.9306666666666665 | 0.00455555555555555 |
+| Epoch_5_batch_5999.pt | 0.9298333333333332 | 0.003688905622452004 |
+| Epoch_9_batch_5999.pt | 0.9291666666666666 | 0.0026902888917340532 |
+| Epoch_4_batch_2999.pt | 0.9278333333333334 | 0.003002570914860375 |
+| Epoch_6_batch_2999.pt | 0.9260000000000002 | 0.004480410034145773 |
+| Epoch_5_batch_2999.pt | 0.9229999999999998 | 0.0034228715112776353 |
+| Epoch_8_batch_2999.pt | 0.9223333333333332 | 0.002197079992587245 |
+| Epoch_3_batch_5999.pt | 0.922 | 0.0042946995755750415 |
+| Epoch_4_batch_5999.pt | 0.9215 | 0.004070747800382753 |
+| Epoch_3_batch_2999.pt | 0.9166666666666666 | 0.004120110270608708 |
+| Epoch_5.pt | 0.9133333333333334 | 0.0021081851067789163 |
+| Epoch_6.pt | 0.9131666666666666 | 0.0027492984514797003 |
+| Epoch_4.pt | 0.9126666666666667 | 0.0033351846710674704 |
+| Epoch_2_batch_5999.pt | 0.9115 | 0.003578485092263985 |
+| Epoch_2_batch_2999.pt | 0.9045 | 0.005346338310781812 |
+| Epoch_3.pt | 0.9039999999999999 | 0.004575835618216452 |
+| Epoch_2.pt | 0.901 | 0.0046361435655611245 |
+| Epoch_9.pt | 0.8958333333333334 | 0.0034805455794837915 |
+| Epoch_1_batch_5999.pt | 0.8796666666666667 | 0.005604231205741548 |
+| Epoch_1.pt | 0.8779999999999999 | 0.004586614986788952 |
+| Epoch_7.pt | 0.8673333333333334 | 0.004995059287330891 |
+| Epoch_1_batch_2999.pt | 0.85 | 0.005246986201385598 |
+| Epoch_0.pt | 0.769 | 0.004869912666846916 |
+| Epoch_0_batch_5999.pt | 0.7619999999999999 | 0.0047777777777777775 |
+| Epoch_0_batch_2999.pt | 0.6128333333333333 | 0.008864724097981464 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0b7d70172c6c8e200221d102680103c4f31e9125
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.9571666666666665 | 0.0022777777777777852 |
+| Epoch_16_batch_5999.pt | 0.9570000000000001 | 0.0020757268546965973 |
+| Epoch_15_batch_5999.pt | 0.9564999999999999 | 0.0024273391416615017 |
+| Epoch_13_batch_2999.pt | 0.9563333333333333 | 0.0021343747458109565 |
+| Epoch_16.pt | 0.9561666666666667 | 0.0025585730977043257 |
+| Epoch_17_batch_2999.pt | 0.9560000000000001 | 0.0022525705480792597 |
+| Epoch_13.pt | 0.9550000000000001 | 0.002408831487630976 |
+| Epoch_14_batch_2999.pt | 0.9546666666666667 | 0.0024570382652773356 |
+| Epoch_15.pt | 0.9546666666666667 | 0.001822357718539635 |
+| Epoch_17_batch_5999.pt | 0.9543333333333335 | 0.002252570548079252 |
+| Epoch_16_batch_2999.pt | 0.9541666666666666 | 0.0017612074982385963 |
+| Epoch_17.pt | 0.954 | 0.0018954135676924428 |
+| Epoch_14.pt | 0.9533333333333334 | 0.002208289657150204 |
+| Epoch_12_batch_2999.pt | 0.9530000000000001 | 0.0026270200927859793 |
+| Epoch_14_batch_5999.pt | 0.9526666666666668 | 0.0021401511426953628 |
+| Epoch_15_batch_2999.pt | 0.9518333333333334 | 0.002388888888888892 |
+| Epoch_12_batch_5999.pt | 0.9518333333333333 | 0.002025302903728689 |
+| Epoch_12.pt | 0.951 | 0.002560381915956204 |
+| Epoch_11_batch_2999.pt | 0.9508333333333333 | 0.001960253196041929 |
+| Epoch_11.pt | 0.9505000000000001 | 0.0025706078447242857 |
+| Epoch_11_batch_5999.pt | 0.9501666666666667 | 0.001915659961062967 |
+| Epoch_10.pt | 0.9491666666666667 | 0.0030555555555555557 |
+| Epoch_10_batch_5999.pt | 0.9488333333333333 | 0.00271086064416797 |
+| Epoch_10_batch_2999.pt | 0.9478333333333333 | 0.003249406403531052 |
+| Epoch_7_batch_2999.pt | 0.9253333333333332 | 0.003150543750835077 |
+| Epoch_9_batch_5999.pt | 0.925 | 0.003478327964999669 |
+| Epoch_8_batch_5999.pt | 0.9245000000000001 | 0.00305353468525807 |
+| Epoch_6_batch_5999.pt | 0.9235 | 0.0012533904636309414 |
+| Epoch_7_batch_5999.pt | 0.9233333333333332 | 0.0028974232912011757 |
+| Epoch_6_batch_2999.pt | 0.9226666666666666 | 0.00223882685328999 |
+| Epoch_8.pt | 0.9221666666666666 | 0.0022641870969238712 |
+| Epoch_9_batch_2999.pt | 0.9211666666666666 | 0.003960062974932611 |
+| Epoch_4_batch_2999.pt | 0.9163333333333334 | 0.003395712619985814 |
+| Epoch_5_batch_2999.pt | 0.9158333333333333 | 0.003260784575739825 |
+| Epoch_8_batch_2999.pt | 0.915 | 0.0017568209223157636 |
+| Epoch_6.pt | 0.9136666666666666 | 0.003030707043774638 |
+| Epoch_5_batch_5999.pt | 0.9121666666666666 | 0.0033059056770692965 |
+| Epoch_5.pt | 0.9115 | 0.002490103870112779 |
+| Epoch_9.pt | 0.9081666666666666 | 0.00340161544774259 |
+| Epoch_4_batch_5999.pt | 0.9051666666666666 | 0.002242270674512282 |
+| Epoch_3_batch_2999.pt | 0.9046666666666667 | 0.0035986966090448117 |
+| Epoch_4.pt | 0.9046666666666667 | 0.0047648403649759 |
+| Epoch_3_batch_5999.pt | 0.9040000000000001 | 0.0028240588949197407 |
+| Epoch_3.pt | 0.9 | 0.004479032082388084 |
+| Epoch_2_batch_2999.pt | 0.8976666666666666 | 0.003250830852961734 |
+| Epoch_2.pt | 0.8973333333333334 | 0.004000000000000001 |
+| Epoch_2_batch_5999.pt | 0.8945000000000001 | 0.0031725092300598306 |
+| Epoch_1_batch_5999.pt | 0.8803333333333333 | 0.0030912061651652326 |
+| Epoch_1.pt | 0.8728333333333333 | 0.0026764865489879806 |
+| Epoch_1_batch_2999.pt | 0.8511666666666666 | 0.004135438531988092 |
+| Epoch_7.pt | 0.8048333333333334 | 0.004055555555555552 |
+| Epoch_0_batch_5999.pt | 0.7933333333333334 | 0.006136311676215136 |
+| Epoch_0.pt | 0.791 | 0.005265775827143462 |
+| Epoch_0_batch_2999.pt | 0.6616666666666666 | 0.005533288710217216 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f1495cc8ce6a155d75bd3819bac12d99fcfe3e98
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9913333333333334 | 0.0016629588385661935 |
+| Epoch_16_batch_5999.pt | 0.991 | 0.0017427096823731277 |
+| Epoch_14.pt | 0.9908333333333333 | 0.0018798804795209572 |
+| Epoch_15_batch_2999.pt | 0.9906666666666666 | 0.0017950549357115021 |
+| Epoch_14_batch_5999.pt | 0.9901666666666668 | 0.0017471316881684884 |
+| Epoch_13_batch_2999.pt | 0.9898333333333333 | 0.0015406027359846739 |
+| Epoch_14_batch_2999.pt | 0.9898333333333333 | 0.0017293758240303765 |
+| Epoch_15_batch_5999.pt | 0.9898333333333333 | 0.00183333333333334 |
+| Epoch_16_batch_2999.pt | 0.9898333333333333 | 0.0017471316881684936 |
+| Epoch_13_batch_5999.pt | 0.9896666666666667 | 0.0015869840952317477 |
+| Epoch_17_batch_2999.pt | 0.9896666666666667 | 0.001644294287438752 |
+| Epoch_12_batch_5999.pt | 0.9890000000000001 | 0.001409841948938834 |
+| Epoch_15.pt | 0.9889999999999999 | 0.0017777777777777757 |
+| Epoch_17_batch_5999.pt | 0.9889999999999999 | 0.0017603310575283184 |
+| Epoch_10.pt | 0.9886666666666665 | 0.001527525231651948 |
+| Epoch_13.pt | 0.9885000000000002 | 0.0017471316881684839 |
+| Epoch_16.pt | 0.9884999999999999 | 0.001931704294523746 |
+| Epoch_11.pt | 0.9880000000000001 | 0.0019531550923607703 |
+| Epoch_11_batch_5999.pt | 0.9879999999999999 | 0.001921290718421177 |
+| Epoch_10_batch_2999.pt | 0.9876666666666667 | 0.0015555555555555548 |
+| Epoch_12.pt | 0.9875 | 0.0019444444444444455 |
+| Epoch_12_batch_2999.pt | 0.9874999999999998 | 0.0014958791130929123 |
+| Epoch_11_batch_2999.pt | 0.9868333333333332 | 0.0015406027359846691 |
+| Epoch_10_batch_5999.pt | 0.9863333333333333 | 0.001527525231651942 |
+| Epoch_8_batch_2999.pt | 0.9768333333333334 | 0.002085369375458162 |
+| Epoch_6_batch_5999.pt | 0.9768333333333332 | 0.0020100058350139404 |
+| Epoch_7_batch_5999.pt | 0.976 | 0.002051798368068817 |
+| Epoch_9_batch_2999.pt | 0.975 | 0.0022498285257018425 |
+| Epoch_9_batch_5999.pt | 0.9748333333333333 | 0.0022831913986704985 |
+| Epoch_7_batch_2999.pt | 0.9743333333333333 | 0.002197079992587244 |
+| Epoch_5_batch_5999.pt | 0.9740000000000002 | 0.0019277057303219388 |
+| Epoch_6_batch_2999.pt | 0.9731666666666665 | 0.0022005891467042566 |
+| Epoch_8_batch_5999.pt | 0.9731666666666665 | 0.001899480110808749 |
+| Epoch_4_batch_2999.pt | 0.9718333333333333 | 0.0024145904235667428 |
+| Epoch_8.pt | 0.9715 | 0.0025633937766798513 |
+| Epoch_6.pt | 0.9708333333333332 | 0.0018130187635645265 |
+| Epoch_4_batch_5999.pt | 0.9706666666666667 | 0.0021401511426953623 |
+| Epoch_5_batch_2999.pt | 0.97 | 0.0030429030972509183 |
+| Epoch_9.pt | 0.9661666666666667 | 0.0020794408080236555 |
+| Epoch_3_batch_2999.pt | 0.9651666666666667 | 0.002040485296911888 |
+| Epoch_3_batch_5999.pt | 0.9648333333333333 | 0.002255993389360772 |
+| Epoch_5.pt | 0.9646666666666667 | 0.002060804110110164 |
+| Epoch_4.pt | 0.9644999999999999 | 0.001192828364087998 |
+| Epoch_3.pt | 0.9608333333333332 | 0.002584975583140411 |
+| Epoch_2_batch_5999.pt | 0.9591666666666667 | 0.0032702361450581024 |
+| Epoch_2_batch_2999.pt | 0.9566666666666667 | 0.0017391639824998343 |
+| Epoch_2.pt | 0.9555 | 0.0024094720491334956 |
+| Epoch_1_batch_5999.pt | 0.945 | 0.003181738014061409 |
+| Epoch_1.pt | 0.9426666666666665 | 0.0015355861067872473 |
+| Epoch_1_batch_2999.pt | 0.9241666666666669 | 0.0035246048723470936 |
+| Epoch_7.pt | 0.899 | 0.004061259307221564 |
+| Epoch_0.pt | 0.8720000000000001 | 0.004640136226182466 |
+| Epoch_0_batch_5999.pt | 0.8705 | 0.004720914851703747 |
+| Epoch_0_batch_2999.pt | 0.7423333333333334 | 0.0052774853720168515 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..baf24e175edf2c31fa375bc0a466119a5ec7ac6c
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.9683333333333334 | 0.00226350542082926 |
+| Epoch_14_batch_2999.pt | 0.9678333333333333 | 0.0021235016504018195 |
+| Epoch_16_batch_2999.pt | 0.9678333333333333 | 0.0023180717827805623 |
+| Epoch_17_batch_2999.pt | 0.9678333333333333 | 0.0022089883724523353 |
+| Epoch_12_batch_5999.pt | 0.967 | 0.0016996731711976015 |
+| Epoch_16_batch_5999.pt | 0.9668333333333334 | 0.0022831913986705006 |
+| Epoch_14.pt | 0.9668333333333333 | 0.002040485296911887 |
+| Epoch_12_batch_2999.pt | 0.9666666666666668 | 0.003113094605804871 |
+| Epoch_13_batch_2999.pt | 0.9666666666666668 | 0.002472066162365224 |
+| Epoch_17.pt | 0.9661666666666668 | 0.0025343321617682358 |
+| Epoch_13.pt | 0.9661666666666667 | 0.0022777777777777818 |
+| Epoch_13_batch_5999.pt | 0.9656666666666667 | 0.0023726840560069646 |
+| Epoch_15_batch_5999.pt | 0.9656666666666667 | 0.002051798368068822 |
+| Epoch_15_batch_2999.pt | 0.9646666666666668 | 0.0022054925823643606 |
+| Epoch_11_batch_5999.pt | 0.9646666666666667 | 0.0023280363155285554 |
+| Epoch_15.pt | 0.9644999999999999 | 0.0023966282900136715 |
+| Epoch_10.pt | 0.9643333333333333 | 0.002211083193570265 |
+| Epoch_12.pt | 0.9641666666666666 | 0.0023207331749117125 |
+| Epoch_14_batch_5999.pt | 0.9638333333333333 | 0.0023180717827805614 |
+| Epoch_11_batch_2999.pt | 0.9631666666666666 | 0.0024017740356908133 |
+| Epoch_11.pt | 0.9630000000000001 | 0.002719386277893432 |
+| Epoch_16.pt | 0.9630000000000001 | 0.0021343747458109504 |
+| Epoch_10_batch_5999.pt | 0.9618333333333332 | 0.0023759338644825712 |
+| Epoch_10_batch_2999.pt | 0.9615 | 0.0019790570145063195 |
+| Epoch_7_batch_2999.pt | 0.9470000000000001 | 0.003449816599168892 |
+| Epoch_7_batch_5999.pt | 0.9450000000000001 | 0.0027666443551086117 |
+| Epoch_8_batch_5999.pt | 0.9448333333333332 | 0.0038042374035044376 |
+| Epoch_6_batch_2999.pt | 0.9443333333333334 | 0.00315445990368409 |
+| Epoch_9_batch_5999.pt | 0.9436666666666668 | 0.003265986323710902 |
+| Epoch_6_batch_5999.pt | 0.9434999999999999 | 0.0025147711772791695 |
+| Epoch_8_batch_2999.pt | 0.943 | 0.002105255035721824 |
+| Epoch_9_batch_2999.pt | 0.9423333333333332 | 0.002952295612352533 |
+| Epoch_5_batch_5999.pt | 0.9388333333333334 | 0.003702268240343582 |
+| Epoch_4_batch_2999.pt | 0.9378333333333334 | 0.0034251250315107374 |
+| Epoch_8.pt | 0.9358333333333333 | 0.0029840108893323973 |
+| Epoch_4_batch_5999.pt | 0.9355 | 0.0031822229981377055 |
+| Epoch_6.pt | 0.9338333333333335 | 0.003434124324561236 |
+| Epoch_5_batch_2999.pt | 0.933 | 0.003666666666666666 |
+| Epoch_3_batch_5999.pt | 0.9316666666666666 | 0.0036767538017276214 |
+| Epoch_5.pt | 0.9309999999999998 | 0.0031150768375280344 |
+| Epoch_3_batch_2999.pt | 0.9285 | 0.003500000000000002 |
+| Epoch_4.pt | 0.9259999999999999 | 0.004158881616047318 |
+| Epoch_2.pt | 0.9248333333333333 | 0.0037552432480269316 |
+| Epoch_2_batch_5999.pt | 0.9243333333333332 | 0.0032603112780269375 |
+| Epoch_9.pt | 0.9238333333333333 | 0.0028442057191489407 |
+| Epoch_3.pt | 0.9221666666666668 | 0.0025221243250702604 |
+| Epoch_2_batch_2999.pt | 0.9146666666666666 | 0.0018392161508052128 |
+| Epoch_1.pt | 0.9058333333333335 | 0.0023207331749117133 |
+| Epoch_1_batch_5999.pt | 0.905 | 0.002249828525701841 |
+| Epoch_1_batch_2999.pt | 0.8805000000000002 | 0.004216736202032426 |
+| Epoch_7.pt | 0.869 | 0.003593547028621393 |
+| Epoch_0.pt | 0.8281666666666666 | 0.005662580225905397 |
+| Epoch_0_batch_5999.pt | 0.8206666666666667 | 0.00517114501254227 |
+| Epoch_0_batch_2999.pt | 0.6966666666666667 | 0.004274529791482515 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet56/log.log b/bob/bio/facexzoo/models/backbones/AttentionNet56/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..3acbcdcda98995568eb04fd775b8536c990c72f8
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet56/log.log
@@ -0,0 +1,655 @@
+INFO 2020-12-06 22:21:17 train.py: 177] Start optimization.
+INFO 2020-12-06 22:21:17 train.py: 178] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='AttentionNet', batch_size=512, data_root='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-att56', train_file='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7fb66fac16d8>)
+backbone param:
+{'stage1_modules': 1, 'stage2_modules': 1, 'stage3_modules': 1, 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-12-06 22:21:40 train.py: 79] Epoch 0, iter 0/6416, lr 0.100000, loss 16.339117
+INFO 2020-12-06 22:29:02 train.py: 79] Epoch 0, iter 200/6416, lr 0.100000, loss 15.653760
+INFO 2020-12-06 22:36:24 train.py: 79] Epoch 0, iter 400/6416, lr 0.100000, loss 15.389072
+INFO 2020-12-06 22:43:45 train.py: 79] Epoch 0, iter 600/6416, lr 0.100000, loss 15.345332
+INFO 2020-12-06 22:51:06 train.py: 79] Epoch 0, iter 800/6416, lr 0.100000, loss 15.305819
+INFO 2020-12-06 22:58:28 train.py: 79] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.241742
+INFO 2020-12-06 23:05:50 train.py: 79] Epoch 0, iter 1200/6416, lr 0.100000, loss 15.116363
+INFO 2020-12-06 23:13:12 train.py: 79] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.905524
+INFO 2020-12-06 23:20:33 train.py: 79] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.609928
+INFO 2020-12-06 23:27:55 train.py: 79] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.315692
+INFO 2020-12-06 23:35:17 train.py: 79] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.953385
+INFO 2020-12-06 23:42:39 train.py: 79] Epoch 0, iter 2200/6416, lr 0.100000, loss 13.549322
+INFO 2020-12-06 23:50:01 train.py: 79] Epoch 0, iter 2400/6416, lr 0.100000, loss 13.170653
+INFO 2020-12-06 23:57:23 train.py: 79] Epoch 0, iter 2600/6416, lr 0.100000, loss 12.746714
+INFO 2020-12-07 00:04:45 train.py: 79] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.351781
+INFO 2020-12-07 00:12:05 train.py: 92] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-12-07 00:12:08 train.py: 79] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.031275
+INFO 2020-12-07 00:19:29 train.py: 79] Epoch 0, iter 3200/6416, lr 0.100000, loss 11.797055
+INFO 2020-12-07 00:26:50 train.py: 79] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.682382
+INFO 2020-12-07 00:34:10 train.py: 79] Epoch 0, iter 3600/6416, lr 0.100000, loss 11.715891
+INFO 2020-12-07 00:41:30 train.py: 79] Epoch 0, iter 3800/6416, lr 0.100000, loss 11.886867
+INFO 2020-12-07 00:48:50 train.py: 79] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.140014
+INFO 2020-12-07 00:56:10 train.py: 79] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.420920
+INFO 2020-12-07 01:03:29 train.py: 79] Epoch 0, iter 4400/6416, lr 0.100000, loss 12.726578
+INFO 2020-12-07 01:10:48 train.py: 79] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.040385
+INFO 2020-12-07 01:18:07 train.py: 79] Epoch 0, iter 4800/6416, lr 0.100000, loss 13.254581
+INFO 2020-12-07 01:25:25 train.py: 79] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.402957
+INFO 2020-12-07 01:32:43 train.py: 79] Epoch 0, iter 5200/6416, lr 0.100000, loss 13.488013
+INFO 2020-12-07 01:40:01 train.py: 79] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.496982
+INFO 2020-12-07 01:47:18 train.py: 79] Epoch 0, iter 5600/6416, lr 0.100000, loss 13.414288
+INFO 2020-12-07 01:54:36 train.py: 79] Epoch 0, iter 5800/6416, lr 0.100000, loss 13.259584
+INFO 2020-12-07 02:01:51 train.py: 92] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-12-07 02:01:53 train.py: 79] Epoch 0, iter 6000/6416, lr 0.100000, loss 13.092949
+INFO 2020-12-07 02:09:10 train.py: 79] Epoch 0, iter 6200/6416, lr 0.100000, loss 12.860123
+INFO 2020-12-07 02:16:26 train.py: 79] Epoch 0, iter 6400/6416, lr 0.100000, loss 12.605660
+INFO 2020-12-07 02:16:58 train.py: 97] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-12-07 02:17:01 train.py: 79] Epoch 1, iter 0/6416, lr 0.100000, loss 12.393160
+INFO 2020-12-07 02:24:17 train.py: 79] Epoch 1, iter 200/6416, lr 0.100000, loss 12.085337
+INFO 2020-12-07 02:31:33 train.py: 79] Epoch 1, iter 400/6416, lr 0.100000, loss 11.759532
+INFO 2020-12-07 02:38:49 train.py: 79] Epoch 1, iter 600/6416, lr 0.100000, loss 11.514332
+INFO 2020-12-07 02:46:05 train.py: 79] Epoch 1, iter 800/6416, lr 0.100000, loss 11.240436
+INFO 2020-12-07 02:53:20 train.py: 79] Epoch 1, iter 1000/6416, lr 0.100000, loss 11.003754
+INFO 2020-12-07 03:00:36 train.py: 79] Epoch 1, iter 1200/6416, lr 0.100000, loss 10.700621
+INFO 2020-12-07 03:07:52 train.py: 79] Epoch 1, iter 1400/6416, lr 0.100000, loss 10.439210
+INFO 2020-12-07 03:15:08 train.py: 79] Epoch 1, iter 1600/6416, lr 0.100000, loss 10.217995
+INFO 2020-12-07 03:22:24 train.py: 79] Epoch 1, iter 1800/6416, lr 0.100000, loss 10.008241
+INFO 2020-12-07 03:29:39 train.py: 79] Epoch 1, iter 2000/6416, lr 0.100000, loss 9.750344
+INFO 2020-12-07 03:36:55 train.py: 79] Epoch 1, iter 2200/6416, lr 0.100000, loss 9.543679
+INFO 2020-12-07 03:44:11 train.py: 79] Epoch 1, iter 2400/6416, lr 0.100000, loss 9.361696
+INFO 2020-12-07 03:51:26 train.py: 79] Epoch 1, iter 2600/6416, lr 0.100000, loss 9.178392
+INFO 2020-12-07 03:58:42 train.py: 79] Epoch 1, iter 2800/6416, lr 0.100000, loss 8.976297
+INFO 2020-12-07 04:05:56 train.py: 92] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-12-07 04:05:58 train.py: 79] Epoch 1, iter 3000/6416, lr 0.100000, loss 8.794548
+INFO 2020-12-07 04:13:14 train.py: 79] Epoch 1, iter 3200/6416, lr 0.100000, loss 8.647182
+INFO 2020-12-07 04:20:30 train.py: 79] Epoch 1, iter 3400/6416, lr 0.100000, loss 8.476677
+INFO 2020-12-07 04:27:45 train.py: 79] Epoch 1, iter 3600/6416, lr 0.100000, loss 8.388841
+INFO 2020-12-07 04:35:01 train.py: 79] Epoch 1, iter 3800/6416, lr 0.100000, loss 8.202265
+INFO 2020-12-07 04:42:16 train.py: 79] Epoch 1, iter 4000/6416, lr 0.100000, loss 8.063694
+INFO 2020-12-07 04:49:31 train.py: 79] Epoch 1, iter 4200/6416, lr 0.100000, loss 7.975410
+INFO 2020-12-07 04:56:48 train.py: 79] Epoch 1, iter 4400/6416, lr 0.100000, loss 7.846756
+INFO 2020-12-07 05:04:03 train.py: 79] Epoch 1, iter 4600/6416, lr 0.100000, loss 7.762870
+INFO 2020-12-07 05:11:20 train.py: 79] Epoch 1, iter 4800/6416, lr 0.100000, loss 7.624160
+INFO 2020-12-07 05:18:36 train.py: 79] Epoch 1, iter 5000/6416, lr 0.100000, loss 7.570746
+INFO 2020-12-07 05:25:52 train.py: 79] Epoch 1, iter 5200/6416, lr 0.100000, loss 7.451804
+INFO 2020-12-07 05:33:08 train.py: 79] Epoch 1, iter 5400/6416, lr 0.100000, loss 7.377990
+INFO 2020-12-07 05:40:24 train.py: 79] Epoch 1, iter 5600/6416, lr 0.100000, loss 7.248031
+INFO 2020-12-07 05:47:40 train.py: 79] Epoch 1, iter 5800/6416, lr 0.100000, loss 7.187469
+INFO 2020-12-07 05:54:54 train.py: 92] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-12-07 05:54:56 train.py: 79] Epoch 1, iter 6000/6416, lr 0.100000, loss 7.126930
+INFO 2020-12-07 06:02:12 train.py: 79] Epoch 1, iter 6200/6416, lr 0.100000, loss 7.025377
+INFO 2020-12-07 06:09:27 train.py: 79] Epoch 1, iter 6400/6416, lr 0.100000, loss 6.978082
+INFO 2020-12-07 06:09:59 train.py: 97] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-12-07 06:10:02 train.py: 79] Epoch 2, iter 0/6416, lr 0.100000, loss 6.911716
+INFO 2020-12-07 06:17:18 train.py: 79] Epoch 2, iter 200/6416, lr 0.100000, loss 6.299496
+INFO 2020-12-07 06:24:34 train.py: 79] Epoch 2, iter 400/6416, lr 0.100000, loss 6.312017
+INFO 2020-12-07 06:31:50 train.py: 79] Epoch 2, iter 600/6416, lr 0.100000, loss 6.325150
+INFO 2020-12-07 06:39:06 train.py: 79] Epoch 2, iter 800/6416, lr 0.100000, loss 6.360390
+INFO 2020-12-07 06:46:22 train.py: 79] Epoch 2, iter 1000/6416, lr 0.100000, loss 6.392142
+INFO 2020-12-07 06:53:38 train.py: 79] Epoch 2, iter 1200/6416, lr 0.100000, loss 6.364653
+INFO 2020-12-07 07:00:54 train.py: 79] Epoch 2, iter 1400/6416, lr 0.100000, loss 6.360372
+INFO 2020-12-07 07:08:10 train.py: 79] Epoch 2, iter 1600/6416, lr 0.100000, loss 6.294883
+INFO 2020-12-07 07:15:26 train.py: 79] Epoch 2, iter 1800/6416, lr 0.100000, loss 6.293037
+INFO 2020-12-07 07:22:42 train.py: 79] Epoch 2, iter 2000/6416, lr 0.100000, loss 6.283850
+INFO 2020-12-07 07:29:58 train.py: 79] Epoch 2, iter 2200/6416, lr 0.100000, loss 6.241851
+INFO 2020-12-07 07:37:14 train.py: 79] Epoch 2, iter 2400/6416, lr 0.100000, loss 6.213096
+INFO 2020-12-07 07:44:29 train.py: 79] Epoch 2, iter 2600/6416, lr 0.100000, loss 6.207854
+INFO 2020-12-07 07:51:45 train.py: 79] Epoch 2, iter 2800/6416, lr 0.100000, loss 6.195715
+INFO 2020-12-07 07:59:00 train.py: 92] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-12-07 07:59:02 train.py: 79] Epoch 2, iter 3000/6416, lr 0.100000, loss 6.097220
+INFO 2020-12-07 08:06:17 train.py: 79] Epoch 2, iter 3200/6416, lr 0.100000, loss 6.087660
+INFO 2020-12-07 08:13:33 train.py: 79] Epoch 2, iter 3400/6416, lr 0.100000, loss 6.057884
+INFO 2020-12-07 08:20:49 train.py: 79] Epoch 2, iter 3600/6416, lr 0.100000, loss 5.996236
+INFO 2020-12-07 08:28:04 train.py: 79] Epoch 2, iter 3800/6416, lr 0.100000, loss 5.990073
+INFO 2020-12-07 08:35:20 train.py: 79] Epoch 2, iter 4000/6416, lr 0.100000, loss 5.945493
+INFO 2020-12-07 08:42:36 train.py: 79] Epoch 2, iter 4200/6416, lr 0.100000, loss 5.928440
+INFO 2020-12-07 08:49:52 train.py: 79] Epoch 2, iter 4400/6416, lr 0.100000, loss 5.873710
+INFO 2020-12-07 08:57:08 train.py: 79] Epoch 2, iter 4600/6416, lr 0.100000, loss 5.872127
+INFO 2020-12-07 09:04:24 train.py: 79] Epoch 2, iter 4800/6416, lr 0.100000, loss 5.819641
+INFO 2020-12-07 09:11:40 train.py: 79] Epoch 2, iter 5000/6416, lr 0.100000, loss 5.801713
+INFO 2020-12-07 09:18:56 train.py: 79] Epoch 2, iter 5200/6416, lr 0.100000, loss 5.765310
+INFO 2020-12-07 09:26:12 train.py: 79] Epoch 2, iter 5400/6416, lr 0.100000, loss 5.751568
+INFO 2020-12-07 09:33:28 train.py: 79] Epoch 2, iter 5600/6416, lr 0.100000, loss 5.715537
+INFO 2020-12-07 09:40:44 train.py: 79] Epoch 2, iter 5800/6416, lr 0.100000, loss 5.674472
+INFO 2020-12-07 09:47:58 train.py: 92] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-12-07 09:48:01 train.py: 79] Epoch 2, iter 6000/6416, lr 0.100000, loss 5.651606
+INFO 2020-12-07 09:55:16 train.py: 79] Epoch 2, iter 6200/6416, lr 0.100000, loss 5.597177
+INFO 2020-12-07 10:02:31 train.py: 79] Epoch 2, iter 6400/6416, lr 0.100000, loss 5.579819
+INFO 2020-12-07 10:03:03 train.py: 97] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-12-07 10:03:06 train.py: 79] Epoch 3, iter 0/6416, lr 0.100000, loss 5.477883
+INFO 2020-12-07 10:10:22 train.py: 79] Epoch 3, iter 200/6416, lr 0.100000, loss 5.022878
+INFO 2020-12-07 10:17:38 train.py: 79] Epoch 3, iter 400/6416, lr 0.100000, loss 5.046607
+INFO 2020-12-07 10:24:53 train.py: 79] Epoch 3, iter 600/6416, lr 0.100000, loss 5.062072
+INFO 2020-12-07 10:32:09 train.py: 79] Epoch 3, iter 800/6416, lr 0.100000, loss 5.149184
+INFO 2020-12-07 10:39:25 train.py: 79] Epoch 3, iter 1000/6416, lr 0.100000, loss 5.172674
+INFO 2020-12-07 10:46:41 train.py: 79] Epoch 3, iter 1200/6416, lr 0.100000, loss 5.196896
+INFO 2020-12-07 10:53:57 train.py: 79] Epoch 3, iter 1400/6416, lr 0.100000, loss 5.239215
+INFO 2020-12-07 11:01:13 train.py: 79] Epoch 3, iter 1600/6416, lr 0.100000, loss 5.242085
+INFO 2020-12-07 11:08:29 train.py: 79] Epoch 3, iter 1800/6416, lr 0.100000, loss 5.241448
+INFO 2020-12-07 11:15:44 train.py: 79] Epoch 3, iter 2000/6416, lr 0.100000, loss 5.248835
+INFO 2020-12-07 11:23:00 train.py: 79] Epoch 3, iter 2200/6416, lr 0.100000, loss 5.259196
+INFO 2020-12-07 11:30:16 train.py: 79] Epoch 3, iter 2400/6416, lr 0.100000, loss 5.236755
+INFO 2020-12-07 11:37:32 train.py: 79] Epoch 3, iter 2600/6416, lr 0.100000, loss 5.191518
+INFO 2020-12-07 11:44:47 train.py: 79] Epoch 3, iter 2800/6416, lr 0.100000, loss 5.186061
+INFO 2020-12-07 11:52:02 train.py: 92] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-12-07 11:52:04 train.py: 79] Epoch 3, iter 3000/6416, lr 0.100000, loss 5.180946
+INFO 2020-12-07 11:59:19 train.py: 79] Epoch 3, iter 3200/6416, lr 0.100000, loss 5.216998
+INFO 2020-12-07 12:06:34 train.py: 79] Epoch 3, iter 3400/6416, lr 0.100000, loss 5.146158
+INFO 2020-12-07 12:13:49 train.py: 79] Epoch 3, iter 3600/6416, lr 0.100000, loss 5.140103
+INFO 2020-12-07 12:21:04 train.py: 79] Epoch 3, iter 3800/6416, lr 0.100000, loss 5.166954
+INFO 2020-12-07 12:28:19 train.py: 79] Epoch 3, iter 4000/6416, lr 0.100000, loss 5.095518
+INFO 2020-12-07 12:35:34 train.py: 79] Epoch 3, iter 4200/6416, lr 0.100000, loss 5.100386
+INFO 2020-12-07 12:42:48 train.py: 79] Epoch 3, iter 4400/6416, lr 0.100000, loss 5.069518
+INFO 2020-12-07 12:50:04 train.py: 79] Epoch 3, iter 4600/6416, lr 0.100000, loss 5.064836
+INFO 2020-12-07 12:57:19 train.py: 79] Epoch 3, iter 4800/6416, lr 0.100000, loss 5.075561
+INFO 2020-12-07 13:04:34 train.py: 79] Epoch 3, iter 5000/6416, lr 0.100000, loss 5.044867
+INFO 2020-12-07 13:11:49 train.py: 79] Epoch 3, iter 5200/6416, lr 0.100000, loss 5.020993
+INFO 2020-12-07 13:19:05 train.py: 79] Epoch 3, iter 5400/6416, lr 0.100000, loss 5.007873
+INFO 2020-12-07 13:26:21 train.py: 79] Epoch 3, iter 5600/6416, lr 0.100000, loss 5.001257
+INFO 2020-12-07 13:33:36 train.py: 79] Epoch 3, iter 5800/6416, lr 0.100000, loss 4.998086
+INFO 2020-12-07 13:40:50 train.py: 92] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-12-07 13:40:53 train.py: 79] Epoch 3, iter 6000/6416, lr 0.100000, loss 4.974132
+INFO 2020-12-07 13:48:08 train.py: 79] Epoch 3, iter 6200/6416, lr 0.100000, loss 4.970709
+INFO 2020-12-07 13:55:23 train.py: 79] Epoch 3, iter 6400/6416, lr 0.100000, loss 4.920613
+INFO 2020-12-07 13:55:55 train.py: 97] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-12-07 13:55:58 train.py: 79] Epoch 4, iter 0/6416, lr 0.100000, loss 4.905231
+INFO 2020-12-07 14:03:14 train.py: 79] Epoch 4, iter 200/6416, lr 0.100000, loss 4.430804
+INFO 2020-12-07 14:10:30 train.py: 79] Epoch 4, iter 400/6416, lr 0.100000, loss 4.386273
+INFO 2020-12-07 14:17:45 train.py: 79] Epoch 4, iter 600/6416, lr 0.100000, loss 4.456437
+INFO 2020-12-07 14:25:01 train.py: 79] Epoch 4, iter 800/6416, lr 0.100000, loss 4.541948
+INFO 2020-12-07 14:32:17 train.py: 79] Epoch 4, iter 1000/6416, lr 0.100000, loss 4.576958
+INFO 2020-12-07 14:39:33 train.py: 79] Epoch 4, iter 1200/6416, lr 0.100000, loss 4.655973
+INFO 2020-12-07 14:46:49 train.py: 79] Epoch 4, iter 1400/6416, lr 0.100000, loss 4.660767
+INFO 2020-12-07 14:54:04 train.py: 79] Epoch 4, iter 1600/6416, lr 0.100000, loss 4.659213
+INFO 2020-12-07 15:01:20 train.py: 79] Epoch 4, iter 1800/6416, lr 0.100000, loss 4.659288
+INFO 2020-12-07 15:08:35 train.py: 79] Epoch 4, iter 2000/6416, lr 0.100000, loss 4.679884
+INFO 2020-12-07 15:15:51 train.py: 79] Epoch 4, iter 2200/6416, lr 0.100000, loss 4.691229
+INFO 2020-12-07 15:23:06 train.py: 79] Epoch 4, iter 2400/6416, lr 0.100000, loss 4.709726
+INFO 2020-12-07 15:30:22 train.py: 79] Epoch 4, iter 2600/6416, lr 0.100000, loss 4.684563
+INFO 2020-12-07 15:37:38 train.py: 79] Epoch 4, iter 2800/6416, lr 0.100000, loss 4.664523
+INFO 2020-12-07 15:44:51 train.py: 92] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-12-07 15:44:53 train.py: 79] Epoch 4, iter 3000/6416, lr 0.100000, loss 4.673299
+INFO 2020-12-07 15:52:08 train.py: 79] Epoch 4, iter 3200/6416, lr 0.100000, loss 4.679122
+INFO 2020-12-07 15:59:23 train.py: 79] Epoch 4, iter 3400/6416, lr 0.100000, loss 4.673668
+INFO 2020-12-07 16:06:38 train.py: 79] Epoch 4, iter 3600/6416, lr 0.100000, loss 4.690486
+INFO 2020-12-07 16:13:53 train.py: 79] Epoch 4, iter 3800/6416, lr 0.100000, loss 4.711592
+INFO 2020-12-07 16:21:07 train.py: 79] Epoch 4, iter 4000/6416, lr 0.100000, loss 4.679128
+INFO 2020-12-07 16:28:22 train.py: 79] Epoch 4, iter 4200/6416, lr 0.100000, loss 4.642389
+INFO 2020-12-07 16:35:37 train.py: 79] Epoch 4, iter 4400/6416, lr 0.100000, loss 4.640824
+INFO 2020-12-07 16:42:53 train.py: 79] Epoch 4, iter 4600/6416, lr 0.100000, loss 4.956016
+INFO 2020-12-07 16:50:08 train.py: 79] Epoch 4, iter 4800/6416, lr 0.100000, loss 4.703107
+INFO 2020-12-07 16:57:24 train.py: 79] Epoch 4, iter 5000/6416, lr 0.100000, loss 4.654803
+INFO 2020-12-07 17:04:39 train.py: 79] Epoch 4, iter 5200/6416, lr 0.100000, loss 4.646274
+INFO 2020-12-07 17:11:55 train.py: 79] Epoch 4, iter 5400/6416, lr 0.100000, loss 4.609154
+INFO 2020-12-07 17:19:11 train.py: 79] Epoch 4, iter 5600/6416, lr 0.100000, loss 4.582357
+INFO 2020-12-07 17:26:26 train.py: 79] Epoch 4, iter 5800/6416, lr 0.100000, loss 4.605782
+INFO 2020-12-07 17:33:41 train.py: 92] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-12-07 17:33:43 train.py: 79] Epoch 4, iter 6000/6416, lr 0.100000, loss 4.570646
+INFO 2020-12-07 17:40:58 train.py: 79] Epoch 4, iter 6200/6416, lr 0.100000, loss 4.556395
+INFO 2020-12-07 17:48:13 train.py: 79] Epoch 4, iter 6400/6416, lr 0.100000, loss 4.555425
+INFO 2020-12-07 17:48:45 train.py: 97] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-12-07 17:48:48 train.py: 79] Epoch 5, iter 0/6416, lr 0.100000, loss 4.593332
+INFO 2020-12-07 17:56:03 train.py: 79] Epoch 5, iter 200/6416, lr 0.100000, loss 4.043981
+INFO 2020-12-07 18:03:18 train.py: 79] Epoch 5, iter 400/6416, lr 0.100000, loss 4.023411
+INFO 2020-12-07 18:10:34 train.py: 79] Epoch 5, iter 600/6416, lr 0.100000, loss 4.102510
+INFO 2020-12-07 18:17:49 train.py: 79] Epoch 5, iter 800/6416, lr 0.100000, loss 4.171846
+INFO 2020-12-07 18:25:05 train.py: 79] Epoch 5, iter 1000/6416, lr 0.100000, loss 4.227398
+INFO 2020-12-07 18:32:19 train.py: 79] Epoch 5, iter 1200/6416, lr 0.100000, loss 4.255724
+INFO 2020-12-07 18:39:35 train.py: 79] Epoch 5, iter 1400/6416, lr 0.100000, loss 4.313682
+INFO 2020-12-07 18:46:49 train.py: 79] Epoch 5, iter 1600/6416, lr 0.100000, loss 4.318452
+INFO 2020-12-07 18:54:04 train.py: 79] Epoch 5, iter 1800/6416, lr 0.100000, loss 4.328173
+INFO 2020-12-07 19:01:19 train.py: 79] Epoch 5, iter 2000/6416, lr 0.100000, loss 4.359922
+INFO 2020-12-07 19:08:34 train.py: 79] Epoch 5, iter 2200/6416, lr 0.100000, loss 4.341381
+INFO 2020-12-07 19:15:49 train.py: 79] Epoch 5, iter 2400/6416, lr 0.100000, loss 4.352831
+INFO 2020-12-07 19:23:04 train.py: 79] Epoch 5, iter 2600/6416, lr 0.100000, loss 4.391685
+INFO 2020-12-07 19:30:18 train.py: 79] Epoch 5, iter 2800/6416, lr 0.100000, loss 4.368737
+INFO 2020-12-07 19:37:31 train.py: 92] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-12-07 19:37:34 train.py: 79] Epoch 5, iter 3000/6416, lr 0.100000, loss 4.380093
+INFO 2020-12-07 19:44:48 train.py: 79] Epoch 5, iter 3200/6416, lr 0.100000, loss 4.357551
+INFO 2020-12-07 19:52:03 train.py: 79] Epoch 5, iter 3400/6416, lr 0.100000, loss 4.368582
+INFO 2020-12-07 19:59:18 train.py: 79] Epoch 5, iter 3600/6416, lr 0.100000, loss 4.379981
+INFO 2020-12-07 20:06:33 train.py: 79] Epoch 5, iter 3800/6416, lr 0.100000, loss 4.343113
+INFO 2020-12-07 20:13:48 train.py: 79] Epoch 5, iter 4000/6416, lr 0.100000, loss 4.360143
+INFO 2020-12-07 20:21:03 train.py: 79] Epoch 5, iter 4200/6416, lr 0.100000, loss 4.329559
+INFO 2020-12-07 20:28:18 train.py: 79] Epoch 5, iter 4400/6416, lr 0.100000, loss 4.363690
+INFO 2020-12-07 20:35:32 train.py: 79] Epoch 5, iter 4600/6416, lr 0.100000, loss 4.378288
+INFO 2020-12-07 20:42:47 train.py: 79] Epoch 5, iter 4800/6416, lr 0.100000, loss 4.332315
+INFO 2020-12-07 20:50:02 train.py: 79] Epoch 5, iter 5000/6416, lr 0.100000, loss 4.354946
+INFO 2020-12-07 20:57:17 train.py: 79] Epoch 5, iter 5200/6416, lr 0.100000, loss 4.329265
+INFO 2020-12-07 21:04:32 train.py: 79] Epoch 5, iter 5400/6416, lr 0.100000, loss 4.330752
+INFO 2020-12-07 21:11:46 train.py: 79] Epoch 5, iter 5600/6416, lr 0.100000, loss 4.310132
+INFO 2020-12-07 21:19:01 train.py: 79] Epoch 5, iter 5800/6416, lr 0.100000, loss 4.276224
+INFO 2020-12-07 21:26:14 train.py: 92] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-12-07 21:26:17 train.py: 79] Epoch 5, iter 6000/6416, lr 0.100000, loss 4.297984
+INFO 2020-12-07 21:33:31 train.py: 79] Epoch 5, iter 6200/6416, lr 0.100000, loss 4.306187
+INFO 2020-12-07 21:40:46 train.py: 79] Epoch 5, iter 6400/6416, lr 0.100000, loss 4.324875
+INFO 2020-12-07 21:41:18 train.py: 97] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-12-07 21:41:21 train.py: 79] Epoch 6, iter 0/6416, lr 0.100000, loss 4.364389
+INFO 2020-12-07 21:48:37 train.py: 79] Epoch 6, iter 200/6416, lr 0.100000, loss 3.830979
+INFO 2020-12-07 21:55:52 train.py: 79] Epoch 6, iter 400/6416, lr 0.100000, loss 3.794176
+INFO 2020-12-07 22:03:08 train.py: 79] Epoch 6, iter 600/6416, lr 0.100000, loss 3.850483
+INFO 2020-12-07 22:10:23 train.py: 79] Epoch 6, iter 800/6416, lr 0.100000, loss 3.934534
+INFO 2020-12-07 22:17:39 train.py: 79] Epoch 6, iter 1000/6416, lr 0.100000, loss 3.989497
+INFO 2020-12-07 22:24:54 train.py: 79] Epoch 6, iter 1200/6416, lr 0.100000, loss 4.021287
+INFO 2020-12-07 22:32:09 train.py: 79] Epoch 6, iter 1400/6416, lr 0.100000, loss 4.039314
+INFO 2020-12-07 22:39:24 train.py: 79] Epoch 6, iter 1600/6416, lr 0.100000, loss 4.080950
+INFO 2020-12-07 22:46:40 train.py: 79] Epoch 6, iter 1800/6416, lr 0.100000, loss 4.105197
+INFO 2020-12-07 22:53:55 train.py: 79] Epoch 6, iter 2000/6416, lr 0.100000, loss 4.098700
+INFO 2020-12-07 23:01:10 train.py: 79] Epoch 6, iter 2200/6416, lr 0.100000, loss 4.132332
+INFO 2020-12-07 23:08:26 train.py: 79] Epoch 6, iter 2400/6416, lr 0.100000, loss 4.165100
+INFO 2020-12-07 23:15:41 train.py: 79] Epoch 6, iter 2600/6416, lr 0.100000, loss 4.130368
+INFO 2020-12-07 23:22:56 train.py: 79] Epoch 6, iter 2800/6416, lr 0.100000, loss 4.162566
+INFO 2020-12-07 23:30:10 train.py: 92] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-12-07 23:30:12 train.py: 79] Epoch 6, iter 3000/6416, lr 0.100000, loss 4.144197
+INFO 2020-12-07 23:37:27 train.py: 79] Epoch 6, iter 3200/6416, lr 0.100000, loss 4.133361
+INFO 2020-12-07 23:44:41 train.py: 79] Epoch 6, iter 3400/6416, lr 0.100000, loss 4.177254
+INFO 2020-12-07 23:51:56 train.py: 79] Epoch 6, iter 3600/6416, lr 0.100000, loss 4.162256
+INFO 2020-12-07 23:59:12 train.py: 79] Epoch 6, iter 3800/6416, lr 0.100000, loss 4.159798
+INFO 2020-12-08 00:06:27 train.py: 79] Epoch 6, iter 4000/6416, lr 0.100000, loss 4.156479
+INFO 2020-12-08 00:13:42 train.py: 79] Epoch 6, iter 4200/6416, lr 0.100000, loss 4.163072
+INFO 2020-12-08 00:20:57 train.py: 79] Epoch 6, iter 4400/6416, lr 0.100000, loss 4.138556
+INFO 2020-12-08 00:28:13 train.py: 79] Epoch 6, iter 4600/6416, lr 0.100000, loss 4.134754
+INFO 2020-12-08 00:35:28 train.py: 79] Epoch 6, iter 4800/6416, lr 0.100000, loss 4.128897
+INFO 2020-12-08 00:42:44 train.py: 79] Epoch 6, iter 5000/6416, lr 0.100000, loss 4.129762
+INFO 2020-12-08 00:49:59 train.py: 79] Epoch 6, iter 5200/6416, lr 0.100000, loss 4.150688
+INFO 2020-12-08 00:57:15 train.py: 79] Epoch 6, iter 5400/6416, lr 0.100000, loss 4.124341
+INFO 2020-12-08 01:04:30 train.py: 79] Epoch 6, iter 5600/6416, lr 0.100000, loss 4.113522
+INFO 2020-12-08 01:11:46 train.py: 79] Epoch 6, iter 5800/6416, lr 0.100000, loss 4.130361
+INFO 2020-12-08 01:19:00 train.py: 92] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-12-08 01:19:02 train.py: 79] Epoch 6, iter 6000/6416, lr 0.100000, loss 4.129368
+INFO 2020-12-08 01:26:17 train.py: 79] Epoch 6, iter 6200/6416, lr 0.100000, loss 4.125476
+INFO 2020-12-08 01:33:32 train.py: 79] Epoch 6, iter 6400/6416, lr 0.100000, loss 4.123388
+INFO 2020-12-08 01:34:03 train.py: 97] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-12-08 01:34:07 train.py: 79] Epoch 7, iter 0/6416, lr 0.100000, loss 4.107350
+INFO 2020-12-08 01:41:22 train.py: 79] Epoch 7, iter 200/6416, lr 0.100000, loss 3.639274
+INFO 2020-12-08 01:48:38 train.py: 79] Epoch 7, iter 400/6416, lr 0.100000, loss 3.595132
+INFO 2020-12-08 01:55:54 train.py: 79] Epoch 7, iter 600/6416, lr 0.100000, loss 3.666903
+INFO 2020-12-08 02:03:09 train.py: 79] Epoch 7, iter 800/6416, lr 0.100000, loss 3.758504
+INFO 2020-12-08 02:10:25 train.py: 79] Epoch 7, iter 1000/6416, lr 0.100000, loss 3.787551
+INFO 2020-12-08 02:17:40 train.py: 79] Epoch 7, iter 1200/6416, lr 0.100000, loss 3.838482
+INFO 2020-12-08 02:24:55 train.py: 79] Epoch 7, iter 1400/6416, lr 0.100000, loss 3.877843
+INFO 2020-12-08 02:32:11 train.py: 79] Epoch 7, iter 1600/6416, lr 0.100000, loss 3.949540
+INFO 2020-12-08 02:39:26 train.py: 79] Epoch 7, iter 1800/6416, lr 0.100000, loss 3.920689
+INFO 2020-12-08 02:46:42 train.py: 79] Epoch 7, iter 2000/6416, lr 0.100000, loss 3.950563
+INFO 2020-12-08 02:53:57 train.py: 79] Epoch 7, iter 2200/6416, lr 0.100000, loss 3.974366
+INFO 2020-12-08 03:01:13 train.py: 79] Epoch 7, iter 2400/6416, lr 0.100000, loss 3.974169
+INFO 2020-12-08 03:08:28 train.py: 79] Epoch 7, iter 2600/6416, lr 0.100000, loss 3.983155
+INFO 2020-12-08 03:15:44 train.py: 79] Epoch 7, iter 2800/6416, lr 0.100000, loss 3.991203
+INFO 2020-12-08 03:22:57 train.py: 92] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-12-08 03:22:59 train.py: 79] Epoch 7, iter 3000/6416, lr 0.100000, loss 3.976505
+INFO 2020-12-08 03:30:14 train.py: 79] Epoch 7, iter 3200/6416, lr 0.100000, loss 4.021536
+INFO 2020-12-08 03:37:29 train.py: 79] Epoch 7, iter 3400/6416, lr 0.100000, loss 4.009361
+INFO 2020-12-08 03:44:44 train.py: 79] Epoch 7, iter 3600/6416, lr 0.100000, loss 3.997068
+INFO 2020-12-08 03:52:00 train.py: 79] Epoch 7, iter 3800/6416, lr 0.100000, loss 4.005455
+INFO 2020-12-08 03:59:15 train.py: 79] Epoch 7, iter 4000/6416, lr 0.100000, loss 3.993581
+INFO 2020-12-08 04:06:30 train.py: 79] Epoch 7, iter 4200/6416, lr 0.100000, loss 3.979407
+INFO 2020-12-08 04:13:45 train.py: 79] Epoch 7, iter 4400/6416, lr 0.100000, loss 4.059376
+INFO 2020-12-08 04:21:01 train.py: 79] Epoch 7, iter 4600/6416, lr 0.100000, loss 4.042143
+INFO 2020-12-08 04:28:16 train.py: 79] Epoch 7, iter 4800/6416, lr 0.100000, loss 4.020873
+INFO 2020-12-08 04:35:31 train.py: 79] Epoch 7, iter 5000/6416, lr 0.100000, loss 4.019922
+INFO 2020-12-08 04:42:47 train.py: 79] Epoch 7, iter 5200/6416, lr 0.100000, loss 4.020736
+INFO 2020-12-08 04:50:02 train.py: 79] Epoch 7, iter 5400/6416, lr 0.100000, loss 4.004260
+INFO 2020-12-08 04:57:18 train.py: 79] Epoch 7, iter 5600/6416, lr 0.100000, loss 3.991970
+INFO 2020-12-08 05:04:34 train.py: 79] Epoch 7, iter 5800/6416, lr 0.100000, loss 3.967162
+INFO 2020-12-08 05:11:48 train.py: 92] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-12-08 05:11:51 train.py: 79] Epoch 7, iter 6000/6416, lr 0.100000, loss 3.965783
+INFO 2020-12-08 05:19:06 train.py: 79] Epoch 7, iter 6200/6416, lr 0.100000, loss 3.968698
+INFO 2020-12-08 05:26:21 train.py: 79] Epoch 7, iter 6400/6416, lr 0.100000, loss 4.231718
+INFO 2020-12-08 05:26:52 train.py: 97] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-12-08 05:26:56 train.py: 79] Epoch 8, iter 0/6416, lr 0.100000, loss 4.895331
+INFO 2020-12-08 05:34:12 train.py: 79] Epoch 8, iter 200/6416, lr 0.100000, loss 3.936440
+INFO 2020-12-08 05:41:27 train.py: 79] Epoch 8, iter 400/6416, lr 0.100000, loss 3.663585
+INFO 2020-12-08 05:48:43 train.py: 79] Epoch 8, iter 600/6416, lr 0.100000, loss 3.645332
+INFO 2020-12-08 05:55:58 train.py: 79] Epoch 8, iter 800/6416, lr 0.100000, loss 3.673107
+INFO 2020-12-08 06:03:13 train.py: 79] Epoch 8, iter 1000/6416, lr 0.100000, loss 3.721533
+INFO 2020-12-08 06:10:29 train.py: 79] Epoch 8, iter 1200/6416, lr 0.100000, loss 3.730598
+INFO 2020-12-08 06:17:44 train.py: 79] Epoch 8, iter 1400/6416, lr 0.100000, loss 3.774283
+INFO 2020-12-08 06:24:59 train.py: 79] Epoch 8, iter 1600/6416, lr 0.100000, loss 3.766718
+INFO 2020-12-08 06:32:14 train.py: 79] Epoch 8, iter 1800/6416, lr 0.100000, loss 3.791457
+INFO 2020-12-08 06:39:30 train.py: 79] Epoch 8, iter 2000/6416, lr 0.100000, loss 3.817394
+INFO 2020-12-08 06:46:45 train.py: 79] Epoch 8, iter 2200/6416, lr 0.100000, loss 3.832890
+INFO 2020-12-08 06:54:00 train.py: 79] Epoch 8, iter 2400/6416, lr 0.100000, loss 3.874387
+INFO 2020-12-08 07:01:16 train.py: 79] Epoch 8, iter 2600/6416, lr 0.100000, loss 3.843073
+INFO 2020-12-08 07:08:31 train.py: 79] Epoch 8, iter 2800/6416, lr 0.100000, loss 3.868332
+INFO 2020-12-08 07:15:44 train.py: 92] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-12-08 07:15:47 train.py: 79] Epoch 8, iter 3000/6416, lr 0.100000, loss 3.871957
+INFO 2020-12-08 07:23:01 train.py: 79] Epoch 8, iter 3200/6416, lr 0.100000, loss 3.861770
+INFO 2020-12-08 07:30:16 train.py: 79] Epoch 8, iter 3400/6416, lr 0.100000, loss 3.844422
+INFO 2020-12-08 07:37:30 train.py: 79] Epoch 8, iter 3600/6416, lr 0.100000, loss 3.855103
+INFO 2020-12-08 07:44:45 train.py: 79] Epoch 8, iter 3800/6416, lr 0.100000, loss 3.868476
+INFO 2020-12-08 07:51:59 train.py: 79] Epoch 8, iter 4000/6416, lr 0.100000, loss 3.867588
+INFO 2020-12-08 07:59:14 train.py: 79] Epoch 8, iter 4200/6416, lr 0.100000, loss 3.854464
+INFO 2020-12-08 08:06:28 train.py: 79] Epoch 8, iter 4400/6416, lr 0.100000, loss 3.857562
+INFO 2020-12-08 08:13:43 train.py: 79] Epoch 8, iter 4600/6416, lr 0.100000, loss 3.883537
+INFO 2020-12-08 08:20:58 train.py: 79] Epoch 8, iter 4800/6416, lr 0.100000, loss 3.866492
+INFO 2020-12-08 08:28:13 train.py: 79] Epoch 8, iter 5000/6416, lr 0.100000, loss 3.876076
+INFO 2020-12-08 08:35:28 train.py: 79] Epoch 8, iter 5200/6416, lr 0.100000, loss 3.886284
+INFO 2020-12-08 08:42:44 train.py: 79] Epoch 8, iter 5400/6416, lr 0.100000, loss 3.867290
+INFO 2020-12-08 08:49:59 train.py: 79] Epoch 8, iter 5600/6416, lr 0.100000, loss 3.885530
+INFO 2020-12-08 08:57:14 train.py: 79] Epoch 8, iter 5800/6416, lr 0.100000, loss 3.887252
+INFO 2020-12-08 09:04:27 train.py: 92] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-12-08 09:04:29 train.py: 79] Epoch 8, iter 6000/6416, lr 0.100000, loss 3.865906
+INFO 2020-12-08 09:11:44 train.py: 79] Epoch 8, iter 6200/6416, lr 0.100000, loss 3.870939
+INFO 2020-12-08 09:18:59 train.py: 79] Epoch 8, iter 6400/6416, lr 0.100000, loss 3.871969
+INFO 2020-12-08 09:19:30 train.py: 97] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-12-08 09:19:33 train.py: 79] Epoch 9, iter 0/6416, lr 0.100000, loss 3.833806
+INFO 2020-12-08 09:26:49 train.py: 79] Epoch 9, iter 200/6416, lr 0.100000, loss 3.428932
+INFO 2020-12-08 09:34:03 train.py: 79] Epoch 9, iter 400/6416, lr 0.100000, loss 3.352119
+INFO 2020-12-08 09:41:18 train.py: 79] Epoch 9, iter 600/6416, lr 0.100000, loss 3.449942
+INFO 2020-12-08 09:48:33 train.py: 79] Epoch 9, iter 800/6416, lr 0.100000, loss 3.493509
+INFO 2020-12-08 09:55:48 train.py: 79] Epoch 9, iter 1000/6416, lr 0.100000, loss 3.562358
+INFO 2020-12-08 10:03:02 train.py: 79] Epoch 9, iter 1200/6416, lr 0.100000, loss 3.615163
+INFO 2020-12-08 10:10:17 train.py: 79] Epoch 9, iter 1400/6416, lr 0.100000, loss 3.640636
+INFO 2020-12-08 10:17:32 train.py: 79] Epoch 9, iter 1600/6416, lr 0.100000, loss 3.668462
+INFO 2020-12-08 10:24:46 train.py: 79] Epoch 9, iter 1800/6416, lr 0.100000, loss 3.671193
+INFO 2020-12-08 10:32:01 train.py: 79] Epoch 9, iter 2000/6416, lr 0.100000, loss 3.750169
+INFO 2020-12-08 10:39:15 train.py: 79] Epoch 9, iter 2200/6416, lr 0.100000, loss 3.759791
+INFO 2020-12-08 10:46:30 train.py: 79] Epoch 9, iter 2400/6416, lr 0.100000, loss 3.757455
+INFO 2020-12-08 10:53:44 train.py: 79] Epoch 9, iter 2600/6416, lr 0.100000, loss 3.772337
+INFO 2020-12-08 11:00:59 train.py: 79] Epoch 9, iter 2800/6416, lr 0.100000, loss 3.764061
+INFO 2020-12-08 11:08:11 train.py: 92] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-12-08 11:08:14 train.py: 79] Epoch 9, iter 3000/6416, lr 0.100000, loss 3.787713
+INFO 2020-12-08 11:15:28 train.py: 79] Epoch 9, iter 3200/6416, lr 0.100000, loss 3.771101
+INFO 2020-12-08 11:22:43 train.py: 79] Epoch 9, iter 3400/6416, lr 0.100000, loss 3.780502
+INFO 2020-12-08 11:29:57 train.py: 79] Epoch 9, iter 3600/6416, lr 0.100000, loss 3.774684
+INFO 2020-12-08 11:37:12 train.py: 79] Epoch 9, iter 3800/6416, lr 0.100000, loss 3.804026
+INFO 2020-12-08 11:44:26 train.py: 79] Epoch 9, iter 4000/6416, lr 0.100000, loss 3.790399
+INFO 2020-12-08 11:51:41 train.py: 79] Epoch 9, iter 4200/6416, lr 0.100000, loss 3.761879
+INFO 2020-12-08 11:58:56 train.py: 79] Epoch 9, iter 4400/6416, lr 0.100000, loss 3.819084
+INFO 2020-12-08 12:06:10 train.py: 79] Epoch 9, iter 4600/6416, lr 0.100000, loss 3.773706
+INFO 2020-12-08 12:13:25 train.py: 79] Epoch 9, iter 4800/6416, lr 0.100000, loss 3.799159
+INFO 2020-12-08 12:20:39 train.py: 79] Epoch 9, iter 5000/6416, lr 0.100000, loss 3.775115
+INFO 2020-12-08 12:27:54 train.py: 79] Epoch 9, iter 5200/6416, lr 0.100000, loss 3.798848
+INFO 2020-12-08 12:35:08 train.py: 79] Epoch 9, iter 5400/6416, lr 0.100000, loss 3.777364
+INFO 2020-12-08 12:42:23 train.py: 79] Epoch 9, iter 5600/6416, lr 0.100000, loss 3.740130
+INFO 2020-12-08 12:49:37 train.py: 79] Epoch 9, iter 5800/6416, lr 0.100000, loss 3.754799
+INFO 2020-12-08 12:56:50 train.py: 92] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-12-08 12:56:52 train.py: 79] Epoch 9, iter 6000/6416, lr 0.100000, loss 3.763572
+INFO 2020-12-08 13:04:06 train.py: 79] Epoch 9, iter 6200/6416, lr 0.100000, loss 3.744781
+INFO 2020-12-08 13:11:21 train.py: 79] Epoch 9, iter 6400/6416, lr 0.100000, loss 3.760781
+INFO 2020-12-08 13:11:53 train.py: 97] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-12-08 13:11:56 train.py: 79] Epoch 10, iter 0/6416, lr 0.010000, loss 3.800927
+INFO 2020-12-08 13:19:11 train.py: 79] Epoch 10, iter 200/6416, lr 0.010000, loss 2.698821
+INFO 2020-12-08 13:26:25 train.py: 79] Epoch 10, iter 400/6416, lr 0.010000, loss 2.456147
+INFO 2020-12-08 13:33:40 train.py: 79] Epoch 10, iter 600/6416, lr 0.010000, loss 2.348257
+INFO 2020-12-08 13:40:55 train.py: 79] Epoch 10, iter 800/6416, lr 0.010000, loss 2.277365
+INFO 2020-12-08 13:48:10 train.py: 79] Epoch 10, iter 1000/6416, lr 0.010000, loss 2.235313
+INFO 2020-12-08 13:55:25 train.py: 79] Epoch 10, iter 1200/6416, lr 0.010000, loss 2.199413
+INFO 2020-12-08 14:02:41 train.py: 79] Epoch 10, iter 1400/6416, lr 0.010000, loss 2.163125
+INFO 2020-12-08 14:09:56 train.py: 79] Epoch 10, iter 1600/6416, lr 0.010000, loss 2.135294
+INFO 2020-12-08 14:17:11 train.py: 79] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.099613
+INFO 2020-12-08 14:24:26 train.py: 79] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.080774
+INFO 2020-12-08 14:31:41 train.py: 79] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.066044
+INFO 2020-12-08 14:38:56 train.py: 79] Epoch 10, iter 2400/6416, lr 0.010000, loss 2.021690
+INFO 2020-12-08 14:46:11 train.py: 79] Epoch 10, iter 2600/6416, lr 0.010000, loss 2.005259
+INFO 2020-12-08 14:53:26 train.py: 79] Epoch 10, iter 2800/6416, lr 0.010000, loss 1.986953
+INFO 2020-12-08 15:00:40 train.py: 92] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-12-08 15:00:42 train.py: 79] Epoch 10, iter 3000/6416, lr 0.010000, loss 1.952786
+INFO 2020-12-08 15:07:56 train.py: 79] Epoch 10, iter 3200/6416, lr 0.010000, loss 1.947060
+INFO 2020-12-08 15:15:11 train.py: 79] Epoch 10, iter 3400/6416, lr 0.010000, loss 1.927209
+INFO 2020-12-08 15:22:25 train.py: 79] Epoch 10, iter 3600/6416, lr 0.010000, loss 1.930897
+INFO 2020-12-08 15:29:40 train.py: 79] Epoch 10, iter 3800/6416, lr 0.010000, loss 1.910759
+INFO 2020-12-08 15:36:55 train.py: 79] Epoch 10, iter 4000/6416, lr 0.010000, loss 1.896521
+INFO 2020-12-08 15:44:10 train.py: 79] Epoch 10, iter 4200/6416, lr 0.010000, loss 1.884202
+INFO 2020-12-08 15:51:25 train.py: 79] Epoch 10, iter 4400/6416, lr 0.010000, loss 1.866119
+INFO 2020-12-08 15:58:40 train.py: 79] Epoch 10, iter 4600/6416, lr 0.010000, loss 1.842797
+INFO 2020-12-08 16:05:55 train.py: 79] Epoch 10, iter 4800/6416, lr 0.010000, loss 1.842283
+INFO 2020-12-08 16:13:10 train.py: 79] Epoch 10, iter 5000/6416, lr 0.010000, loss 1.836908
+INFO 2020-12-08 16:20:26 train.py: 79] Epoch 10, iter 5200/6416, lr 0.010000, loss 1.820286
+INFO 2020-12-08 16:27:41 train.py: 79] Epoch 10, iter 5400/6416, lr 0.010000, loss 1.813036
+INFO 2020-12-08 16:34:56 train.py: 79] Epoch 10, iter 5600/6416, lr 0.010000, loss 1.813383
+INFO 2020-12-08 16:42:11 train.py: 79] Epoch 10, iter 5800/6416, lr 0.010000, loss 1.810559
+INFO 2020-12-08 16:49:25 train.py: 92] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-12-08 16:49:27 train.py: 79] Epoch 10, iter 6000/6416, lr 0.010000, loss 1.782043
+INFO 2020-12-08 16:56:41 train.py: 79] Epoch 10, iter 6200/6416, lr 0.010000, loss 1.783661
+INFO 2020-12-08 17:03:56 train.py: 79] Epoch 10, iter 6400/6416, lr 0.010000, loss 1.758968
+INFO 2020-12-08 17:04:28 train.py: 97] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-12-08 17:04:31 train.py: 79] Epoch 11, iter 0/6416, lr 0.010000, loss 1.754459
+INFO 2020-12-08 17:11:46 train.py: 79] Epoch 11, iter 200/6416, lr 0.010000, loss 1.490951
+INFO 2020-12-08 17:19:02 train.py: 79] Epoch 11, iter 400/6416, lr 0.010000, loss 1.503394
+INFO 2020-12-08 17:26:17 train.py: 79] Epoch 11, iter 600/6416, lr 0.010000, loss 1.510283
+INFO 2020-12-08 17:33:32 train.py: 79] Epoch 11, iter 800/6416, lr 0.010000, loss 1.467629
+INFO 2020-12-08 17:40:47 train.py: 79] Epoch 11, iter 1000/6416, lr 0.010000, loss 1.475949
+INFO 2020-12-08 17:48:02 train.py: 79] Epoch 11, iter 1200/6416, lr 0.010000, loss 1.488209
+INFO 2020-12-08 17:55:17 train.py: 79] Epoch 11, iter 1400/6416, lr 0.010000, loss 1.509780
+INFO 2020-12-08 18:02:32 train.py: 79] Epoch 11, iter 1600/6416, lr 0.010000, loss 1.491372
+INFO 2020-12-08 18:09:48 train.py: 79] Epoch 11, iter 1800/6416, lr 0.010000, loss 1.483094
+INFO 2020-12-08 18:17:02 train.py: 79] Epoch 11, iter 2000/6416, lr 0.010000, loss 1.478413
+INFO 2020-12-08 18:24:17 train.py: 79] Epoch 11, iter 2200/6416, lr 0.010000, loss 1.485048
+INFO 2020-12-08 18:31:32 train.py: 79] Epoch 11, iter 2400/6416, lr 0.010000, loss 1.476659
+INFO 2020-12-08 18:38:47 train.py: 79] Epoch 11, iter 2600/6416, lr 0.010000, loss 1.477922
+INFO 2020-12-08 18:46:02 train.py: 79] Epoch 11, iter 2800/6416, lr 0.010000, loss 1.486131
+INFO 2020-12-08 18:53:15 train.py: 92] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-12-08 18:53:18 train.py: 79] Epoch 11, iter 3000/6416, lr 0.010000, loss 1.464989
+INFO 2020-12-08 19:00:32 train.py: 79] Epoch 11, iter 3200/6416, lr 0.010000, loss 1.474209
+INFO 2020-12-08 19:07:46 train.py: 79] Epoch 11, iter 3400/6416, lr 0.010000, loss 1.482722
+INFO 2020-12-08 19:15:01 train.py: 79] Epoch 11, iter 3600/6416, lr 0.010000, loss 1.478906
+INFO 2020-12-08 19:22:15 train.py: 79] Epoch 11, iter 3800/6416, lr 0.010000, loss 1.486639
+INFO 2020-12-08 19:29:29 train.py: 79] Epoch 11, iter 4000/6416, lr 0.010000, loss 1.474517
+INFO 2020-12-08 19:36:43 train.py: 79] Epoch 11, iter 4200/6416, lr 0.010000, loss 1.476221
+INFO 2020-12-08 19:43:58 train.py: 79] Epoch 11, iter 4400/6416, lr 0.010000, loss 1.470594
+INFO 2020-12-08 19:51:12 train.py: 79] Epoch 11, iter 4600/6416, lr 0.010000, loss 1.490632
+INFO 2020-12-08 19:58:26 train.py: 79] Epoch 11, iter 4800/6416, lr 0.010000, loss 1.487515
+INFO 2020-12-08 20:05:41 train.py: 79] Epoch 11, iter 5000/6416, lr 0.010000, loss 1.473472
+INFO 2020-12-08 20:12:55 train.py: 79] Epoch 11, iter 5200/6416, lr 0.010000, loss 1.470557
+INFO 2020-12-08 20:20:09 train.py: 79] Epoch 11, iter 5400/6416, lr 0.010000, loss 1.467415
+INFO 2020-12-08 20:27:24 train.py: 79] Epoch 11, iter 5600/6416, lr 0.010000, loss 1.478022
+INFO 2020-12-08 20:34:39 train.py: 79] Epoch 11, iter 5800/6416, lr 0.010000, loss 1.455494
+INFO 2020-12-08 20:41:52 train.py: 92] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-12-08 20:41:54 train.py: 79] Epoch 11, iter 6000/6416, lr 0.010000, loss 1.478754
+INFO 2020-12-08 20:49:09 train.py: 79] Epoch 11, iter 6200/6416, lr 0.010000, loss 1.454890
+INFO 2020-12-08 20:56:23 train.py: 79] Epoch 11, iter 6400/6416, lr 0.010000, loss 1.465753
+INFO 2020-12-08 20:56:55 train.py: 97] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-12-08 20:56:58 train.py: 79] Epoch 12, iter 0/6416, lr 0.010000, loss 1.484015
+INFO 2020-12-08 21:04:13 train.py: 79] Epoch 12, iter 200/6416, lr 0.010000, loss 1.233789
+INFO 2020-12-08 21:11:29 train.py: 79] Epoch 12, iter 400/6416, lr 0.010000, loss 1.230974
+INFO 2020-12-08 21:18:44 train.py: 79] Epoch 12, iter 600/6416, lr 0.010000, loss 1.221417
+INFO 2020-12-08 21:25:59 train.py: 79] Epoch 12, iter 800/6416, lr 0.010000, loss 1.247169
+INFO 2020-12-08 21:33:14 train.py: 79] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.238401
+INFO 2020-12-08 21:40:29 train.py: 79] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.254406
+INFO 2020-12-08 21:47:44 train.py: 79] Epoch 12, iter 1400/6416, lr 0.010000, loss 1.242645
+INFO 2020-12-08 21:54:59 train.py: 79] Epoch 12, iter 1600/6416, lr 0.010000, loss 1.251614
+INFO 2020-12-08 22:02:14 train.py: 79] Epoch 12, iter 1800/6416, lr 0.010000, loss 1.255217
+INFO 2020-12-08 22:09:29 train.py: 79] Epoch 12, iter 2000/6416, lr 0.010000, loss 1.267736
+INFO 2020-12-08 22:16:44 train.py: 79] Epoch 12, iter 2200/6416, lr 0.010000, loss 1.267111
+INFO 2020-12-08 22:23:59 train.py: 79] Epoch 12, iter 2400/6416, lr 0.010000, loss 1.263658
+INFO 2020-12-08 22:31:14 train.py: 79] Epoch 12, iter 2600/6416, lr 0.010000, loss 1.282978
+INFO 2020-12-08 22:38:29 train.py: 79] Epoch 12, iter 2800/6416, lr 0.010000, loss 1.279125
+INFO 2020-12-08 22:45:42 train.py: 92] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-12-08 22:45:44 train.py: 79] Epoch 12, iter 3000/6416, lr 0.010000, loss 1.264927
+INFO 2020-12-08 22:52:58 train.py: 79] Epoch 12, iter 3200/6416, lr 0.010000, loss 1.287299
+INFO 2020-12-08 23:00:13 train.py: 79] Epoch 12, iter 3400/6416, lr 0.010000, loss 1.274587
+INFO 2020-12-08 23:07:27 train.py: 79] Epoch 12, iter 3600/6416, lr 0.010000, loss 1.302100
+INFO 2020-12-08 23:14:41 train.py: 79] Epoch 12, iter 3800/6416, lr 0.010000, loss 1.283710
+INFO 2020-12-08 23:21:55 train.py: 79] Epoch 12, iter 4000/6416, lr 0.010000, loss 1.283689
+INFO 2020-12-08 23:29:10 train.py: 79] Epoch 12, iter 4200/6416, lr 0.010000, loss 1.292065
+INFO 2020-12-08 23:36:24 train.py: 79] Epoch 12, iter 4400/6416, lr 0.010000, loss 1.310679
+INFO 2020-12-08 23:43:38 train.py: 79] Epoch 12, iter 4600/6416, lr 0.010000, loss 1.294607
+INFO 2020-12-08 23:50:53 train.py: 79] Epoch 12, iter 4800/6416, lr 0.010000, loss 1.292776
+INFO 2020-12-08 23:58:07 train.py: 79] Epoch 12, iter 5000/6416, lr 0.010000, loss 1.301286
+INFO 2020-12-09 00:05:22 train.py: 79] Epoch 12, iter 5200/6416, lr 0.010000, loss 1.322327
+INFO 2020-12-09 00:12:37 train.py: 79] Epoch 12, iter 5400/6416, lr 0.010000, loss 1.323649
+INFO 2020-12-09 00:19:52 train.py: 79] Epoch 12, iter 5600/6416, lr 0.010000, loss 1.319972
+INFO 2020-12-09 00:27:07 train.py: 79] Epoch 12, iter 5800/6416, lr 0.010000, loss 1.311442
+INFO 2020-12-09 00:34:20 train.py: 92] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-12-09 00:34:22 train.py: 79] Epoch 12, iter 6000/6416, lr 0.010000, loss 1.319461
+INFO 2020-12-09 00:41:37 train.py: 79] Epoch 12, iter 6200/6416, lr 0.010000, loss 1.315675
+INFO 2020-12-09 00:48:51 train.py: 79] Epoch 12, iter 6400/6416, lr 0.010000, loss 1.320610
+INFO 2020-12-09 00:49:22 train.py: 97] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-12-09 00:49:26 train.py: 79] Epoch 13, iter 0/6416, lr 0.001000, loss 1.363125
+INFO 2020-12-09 00:56:41 train.py: 79] Epoch 13, iter 200/6416, lr 0.001000, loss 1.054507
+INFO 2020-12-09 01:03:55 train.py: 79] Epoch 13, iter 400/6416, lr 0.001000, loss 1.028544
+INFO 2020-12-09 01:11:10 train.py: 79] Epoch 13, iter 600/6416, lr 0.001000, loss 1.029790
+INFO 2020-12-09 01:18:24 train.py: 79] Epoch 13, iter 800/6416, lr 0.001000, loss 1.025129
+INFO 2020-12-09 01:25:38 train.py: 79] Epoch 13, iter 1000/6416, lr 0.001000, loss 1.007467
+INFO 2020-12-09 01:32:52 train.py: 79] Epoch 13, iter 1200/6416, lr 0.001000, loss 1.010181
+INFO 2020-12-09 01:40:07 train.py: 79] Epoch 13, iter 1400/6416, lr 0.001000, loss 1.009079
+INFO 2020-12-09 01:47:21 train.py: 79] Epoch 13, iter 1600/6416, lr 0.001000, loss 1.019332
+INFO 2020-12-09 01:54:35 train.py: 79] Epoch 13, iter 1800/6416, lr 0.001000, loss 1.006859
+INFO 2020-12-09 02:01:49 train.py: 79] Epoch 13, iter 2000/6416, lr 0.001000, loss 1.011971
+INFO 2020-12-09 02:09:04 train.py: 79] Epoch 13, iter 2200/6416, lr 0.001000, loss 1.015564
+INFO 2020-12-09 02:16:18 train.py: 79] Epoch 13, iter 2400/6416, lr 0.001000, loss 1.014718
+INFO 2020-12-09 02:23:32 train.py: 79] Epoch 13, iter 2600/6416, lr 0.001000, loss 1.008467
+INFO 2020-12-09 02:30:46 train.py: 79] Epoch 13, iter 2800/6416, lr 0.001000, loss 1.011032
+INFO 2020-12-09 02:37:59 train.py: 92] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-12-09 02:38:01 train.py: 79] Epoch 13, iter 3000/6416, lr 0.001000, loss 0.999696
+INFO 2020-12-09 02:45:16 train.py: 79] Epoch 13, iter 3200/6416, lr 0.001000, loss 1.014576
+INFO 2020-12-09 02:52:30 train.py: 79] Epoch 13, iter 3400/6416, lr 0.001000, loss 1.002025
+INFO 2020-12-09 02:59:44 train.py: 79] Epoch 13, iter 3600/6416, lr 0.001000, loss 1.006575
+INFO 2020-12-09 03:06:58 train.py: 79] Epoch 13, iter 3800/6416, lr 0.001000, loss 1.015797
+INFO 2020-12-09 03:14:13 train.py: 79] Epoch 13, iter 4000/6416, lr 0.001000, loss 1.018409
+INFO 2020-12-09 03:21:27 train.py: 79] Epoch 13, iter 4200/6416, lr 0.001000, loss 0.999023
+INFO 2020-12-09 03:28:42 train.py: 79] Epoch 13, iter 4400/6416, lr 0.001000, loss 1.010068
+INFO 2020-12-09 03:35:56 train.py: 79] Epoch 13, iter 4600/6416, lr 0.001000, loss 1.005167
+INFO 2020-12-09 03:43:10 train.py: 79] Epoch 13, iter 4800/6416, lr 0.001000, loss 1.004244
+INFO 2020-12-09 03:50:25 train.py: 79] Epoch 13, iter 5000/6416, lr 0.001000, loss 0.999028
+INFO 2020-12-09 03:57:39 train.py: 79] Epoch 13, iter 5200/6416, lr 0.001000, loss 1.005059
+INFO 2020-12-09 04:04:54 train.py: 79] Epoch 13, iter 5400/6416, lr 0.001000, loss 0.998904
+INFO 2020-12-09 04:12:08 train.py: 79] Epoch 13, iter 5600/6416, lr 0.001000, loss 1.017914
+INFO 2020-12-09 04:19:23 train.py: 79] Epoch 13, iter 5800/6416, lr 0.001000, loss 1.002128
+INFO 2020-12-09 04:26:36 train.py: 92] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-12-09 04:26:38 train.py: 79] Epoch 13, iter 6000/6416, lr 0.001000, loss 1.005669
+INFO 2020-12-09 04:33:52 train.py: 79] Epoch 13, iter 6200/6416, lr 0.001000, loss 1.007808
+INFO 2020-12-09 04:41:07 train.py: 79] Epoch 13, iter 6400/6416, lr 0.001000, loss 1.015425
+INFO 2020-12-09 04:41:39 train.py: 97] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-12-09 04:41:42 train.py: 79] Epoch 14, iter 0/6416, lr 0.001000, loss 0.995089
+INFO 2020-12-09 04:48:57 train.py: 79] Epoch 14, iter 200/6416, lr 0.001000, loss 0.969068
+INFO 2020-12-09 04:56:11 train.py: 79] Epoch 14, iter 400/6416, lr 0.001000, loss 0.977145
+INFO 2020-12-09 05:03:26 train.py: 79] Epoch 14, iter 600/6416, lr 0.001000, loss 0.970506
+INFO 2020-12-09 05:10:40 train.py: 79] Epoch 14, iter 800/6416, lr 0.001000, loss 0.973748
+INFO 2020-12-09 05:17:54 train.py: 79] Epoch 14, iter 1000/6416, lr 0.001000, loss 0.967163
+INFO 2020-12-09 05:25:09 train.py: 79] Epoch 14, iter 1200/6416, lr 0.001000, loss 0.975248
+INFO 2020-12-09 05:32:23 train.py: 79] Epoch 14, iter 1400/6416, lr 0.001000, loss 0.979443
+INFO 2020-12-09 05:39:38 train.py: 79] Epoch 14, iter 1600/6416, lr 0.001000, loss 0.981217
+INFO 2020-12-09 05:46:53 train.py: 79] Epoch 14, iter 1800/6416, lr 0.001000, loss 0.972316
+INFO 2020-12-09 05:54:08 train.py: 79] Epoch 14, iter 2000/6416, lr 0.001000, loss 0.971438
+INFO 2020-12-09 06:01:23 train.py: 79] Epoch 14, iter 2200/6416, lr 0.001000, loss 0.979107
+INFO 2020-12-09 06:08:38 train.py: 79] Epoch 14, iter 2400/6416, lr 0.001000, loss 0.971288
+INFO 2020-12-09 06:15:52 train.py: 79] Epoch 14, iter 2600/6416, lr 0.001000, loss 0.968721
+INFO 2020-12-09 06:23:07 train.py: 79] Epoch 14, iter 2800/6416, lr 0.001000, loss 0.976437
+INFO 2020-12-09 06:30:20 train.py: 92] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-12-09 06:30:23 train.py: 79] Epoch 14, iter 3000/6416, lr 0.001000, loss 0.977988
+INFO 2020-12-09 06:37:37 train.py: 79] Epoch 14, iter 3200/6416, lr 0.001000, loss 0.976456
+INFO 2020-12-09 06:44:51 train.py: 79] Epoch 14, iter 3400/6416, lr 0.001000, loss 0.971258
+INFO 2020-12-09 06:52:05 train.py: 79] Epoch 14, iter 3600/6416, lr 0.001000, loss 0.974655
+INFO 2020-12-09 06:59:20 train.py: 79] Epoch 14, iter 3800/6416, lr 0.001000, loss 0.969142
+INFO 2020-12-09 07:06:34 train.py: 79] Epoch 14, iter 4000/6416, lr 0.001000, loss 0.978815
+INFO 2020-12-09 07:13:48 train.py: 79] Epoch 14, iter 4200/6416, lr 0.001000, loss 0.976680
+INFO 2020-12-09 07:21:02 train.py: 79] Epoch 14, iter 4400/6416, lr 0.001000, loss 0.975473
+INFO 2020-12-09 07:28:17 train.py: 79] Epoch 14, iter 4600/6416, lr 0.001000, loss 0.981068
+INFO 2020-12-09 07:35:32 train.py: 79] Epoch 14, iter 4800/6416, lr 0.001000, loss 0.970727
+INFO 2020-12-09 07:42:47 train.py: 79] Epoch 14, iter 5000/6416, lr 0.001000, loss 0.986718
+INFO 2020-12-09 07:50:02 train.py: 79] Epoch 14, iter 5200/6416, lr 0.001000, loss 0.982766
+INFO 2020-12-09 07:57:17 train.py: 79] Epoch 14, iter 5400/6416, lr 0.001000, loss 0.980030
+INFO 2020-12-09 08:04:32 train.py: 79] Epoch 14, iter 5600/6416, lr 0.001000, loss 0.983091
+INFO 2020-12-09 08:11:47 train.py: 79] Epoch 14, iter 5800/6416, lr 0.001000, loss 0.985169
+INFO 2020-12-09 08:19:00 train.py: 92] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-12-09 08:19:02 train.py: 79] Epoch 14, iter 6000/6416, lr 0.001000, loss 0.991201
+INFO 2020-12-09 08:26:16 train.py: 79] Epoch 14, iter 6200/6416, lr 0.001000, loss 0.968671
+INFO 2020-12-09 08:33:31 train.py: 79] Epoch 14, iter 6400/6416, lr 0.001000, loss 0.975769
+INFO 2020-12-09 08:34:02 train.py: 97] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-12-09 08:34:06 train.py: 79] Epoch 15, iter 0/6416, lr 0.001000, loss 1.011397
+INFO 2020-12-09 08:41:21 train.py: 79] Epoch 15, iter 200/6416, lr 0.001000, loss 0.948173
+INFO 2020-12-09 08:48:36 train.py: 79] Epoch 15, iter 400/6416, lr 0.001000, loss 0.955711
+INFO 2020-12-09 08:55:51 train.py: 79] Epoch 15, iter 600/6416, lr 0.001000, loss 0.946194
+INFO 2020-12-09 09:03:06 train.py: 79] Epoch 15, iter 800/6416, lr 0.001000, loss 0.945617
+INFO 2020-12-09 09:10:21 train.py: 79] Epoch 15, iter 1000/6416, lr 0.001000, loss 0.944328
+INFO 2020-12-09 09:17:36 train.py: 79] Epoch 15, iter 1200/6416, lr 0.001000, loss 0.955536
+INFO 2020-12-09 09:24:50 train.py: 79] Epoch 15, iter 1400/6416, lr 0.001000, loss 0.947339
+INFO 2020-12-09 09:32:05 train.py: 79] Epoch 15, iter 1600/6416, lr 0.001000, loss 0.950708
+INFO 2020-12-09 09:39:20 train.py: 79] Epoch 15, iter 1800/6416, lr 0.001000, loss 0.957520
+INFO 2020-12-09 09:46:35 train.py: 79] Epoch 15, iter 2000/6416, lr 0.001000, loss 0.958083
+INFO 2020-12-09 09:53:49 train.py: 79] Epoch 15, iter 2200/6416, lr 0.001000, loss 0.955983
+INFO 2020-12-09 10:01:04 train.py: 79] Epoch 15, iter 2400/6416, lr 0.001000, loss 0.963296
+INFO 2020-12-09 10:08:19 train.py: 79] Epoch 15, iter 2600/6416, lr 0.001000, loss 0.957618
+INFO 2020-12-09 10:15:34 train.py: 79] Epoch 15, iter 2800/6416, lr 0.001000, loss 0.956104
+INFO 2020-12-09 10:22:47 train.py: 92] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-12-09 10:22:49 train.py: 79] Epoch 15, iter 3000/6416, lr 0.001000, loss 0.964389
+INFO 2020-12-09 10:30:03 train.py: 79] Epoch 15, iter 3200/6416, lr 0.001000, loss 0.950037
+INFO 2020-12-09 10:37:17 train.py: 79] Epoch 15, iter 3400/6416, lr 0.001000, loss 0.956096
+INFO 2020-12-09 10:44:32 train.py: 79] Epoch 15, iter 3600/6416, lr 0.001000, loss 0.953946
+INFO 2020-12-09 10:51:46 train.py: 79] Epoch 15, iter 3800/6416, lr 0.001000, loss 0.955661
+INFO 2020-12-09 10:59:01 train.py: 79] Epoch 15, iter 4000/6416, lr 0.001000, loss 0.957458
+INFO 2020-12-09 11:06:16 train.py: 79] Epoch 15, iter 4200/6416, lr 0.001000, loss 0.954589
+INFO 2020-12-09 11:13:31 train.py: 79] Epoch 15, iter 4400/6416, lr 0.001000, loss 0.952797
+INFO 2020-12-09 11:20:46 train.py: 79] Epoch 15, iter 4600/6416, lr 0.001000, loss 0.956928
+INFO 2020-12-09 11:28:00 train.py: 79] Epoch 15, iter 4800/6416, lr 0.001000, loss 0.961252
+INFO 2020-12-09 11:35:15 train.py: 79] Epoch 15, iter 5000/6416, lr 0.001000, loss 0.964729
+INFO 2020-12-09 11:42:31 train.py: 79] Epoch 15, iter 5200/6416, lr 0.001000, loss 0.936394
+INFO 2020-12-09 11:49:46 train.py: 79] Epoch 15, iter 5400/6416, lr 0.001000, loss 0.959212
+INFO 2020-12-09 11:57:02 train.py: 79] Epoch 15, iter 5600/6416, lr 0.001000, loss 0.955976
+INFO 2020-12-09 12:04:18 train.py: 79] Epoch 15, iter 5800/6416, lr 0.001000, loss 0.958369
+INFO 2020-12-09 12:11:32 train.py: 92] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-12-09 12:11:34 train.py: 79] Epoch 15, iter 6000/6416, lr 0.001000, loss 0.952067
+INFO 2020-12-09 12:18:49 train.py: 79] Epoch 15, iter 6200/6416, lr 0.001000, loss 0.952807
+INFO 2020-12-09 12:26:04 train.py: 79] Epoch 15, iter 6400/6416, lr 0.001000, loss 0.962677
+INFO 2020-12-09 12:26:36 train.py: 97] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-12-09 12:26:39 train.py: 79] Epoch 16, iter 0/6416, lr 0.000100, loss 0.954858
+INFO 2020-12-09 12:33:54 train.py: 79] Epoch 16, iter 200/6416, lr 0.000100, loss 0.926850
+INFO 2020-12-09 12:41:09 train.py: 79] Epoch 16, iter 400/6416, lr 0.000100, loss 0.920335
+INFO 2020-12-09 12:48:24 train.py: 79] Epoch 16, iter 600/6416, lr 0.000100, loss 0.926612
+INFO 2020-12-09 12:55:39 train.py: 79] Epoch 16, iter 800/6416, lr 0.000100, loss 0.910835
+INFO 2020-12-09 13:02:54 train.py: 79] Epoch 16, iter 1000/6416, lr 0.000100, loss 0.927042
+INFO 2020-12-09 13:10:09 train.py: 79] Epoch 16, iter 1200/6416, lr 0.000100, loss 0.922178
+INFO 2020-12-09 13:17:23 train.py: 79] Epoch 16, iter 1400/6416, lr 0.000100, loss 0.919548
+INFO 2020-12-09 13:24:38 train.py: 79] Epoch 16, iter 1600/6416, lr 0.000100, loss 0.931134
+INFO 2020-12-09 13:31:53 train.py: 79] Epoch 16, iter 1800/6416, lr 0.000100, loss 0.928784
+INFO 2020-12-09 13:39:07 train.py: 79] Epoch 16, iter 2000/6416, lr 0.000100, loss 0.925331
+INFO 2020-12-09 13:46:22 train.py: 79] Epoch 16, iter 2200/6416, lr 0.000100, loss 0.914949
+INFO 2020-12-09 13:53:36 train.py: 79] Epoch 16, iter 2400/6416, lr 0.000100, loss 0.924347
+INFO 2020-12-09 14:00:51 train.py: 79] Epoch 16, iter 2600/6416, lr 0.000100, loss 0.923306
+INFO 2020-12-09 14:08:06 train.py: 79] Epoch 16, iter 2800/6416, lr 0.000100, loss 0.925473
+INFO 2020-12-09 14:15:19 train.py: 92] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-12-09 14:15:21 train.py: 79] Epoch 16, iter 3000/6416, lr 0.000100, loss 0.918383
+INFO 2020-12-09 14:22:35 train.py: 79] Epoch 16, iter 3200/6416, lr 0.000100, loss 0.918482
+INFO 2020-12-09 14:29:49 train.py: 79] Epoch 16, iter 3400/6416, lr 0.000100, loss 0.923224
+INFO 2020-12-09 14:37:03 train.py: 79] Epoch 16, iter 3600/6416, lr 0.000100, loss 0.922412
+INFO 2020-12-09 14:44:17 train.py: 79] Epoch 16, iter 3800/6416, lr 0.000100, loss 0.909996
+INFO 2020-12-09 14:51:31 train.py: 79] Epoch 16, iter 4000/6416, lr 0.000100, loss 0.918164
+INFO 2020-12-09 14:58:45 train.py: 79] Epoch 16, iter 4200/6416, lr 0.000100, loss 0.916819
+INFO 2020-12-09 15:05:59 train.py: 79] Epoch 16, iter 4400/6416, lr 0.000100, loss 0.930096
+INFO 2020-12-09 15:13:14 train.py: 79] Epoch 16, iter 4600/6416, lr 0.000100, loss 0.923030
+INFO 2020-12-09 15:20:28 train.py: 79] Epoch 16, iter 4800/6416, lr 0.000100, loss 0.922276
+INFO 2020-12-09 15:27:42 train.py: 79] Epoch 16, iter 5000/6416, lr 0.000100, loss 0.927248
+INFO 2020-12-09 15:34:56 train.py: 79] Epoch 16, iter 5200/6416, lr 0.000100, loss 0.926332
+INFO 2020-12-09 15:42:11 train.py: 79] Epoch 16, iter 5400/6416, lr 0.000100, loss 0.915357
+INFO 2020-12-09 15:49:26 train.py: 79] Epoch 16, iter 5600/6416, lr 0.000100, loss 0.927079
+INFO 2020-12-09 15:56:40 train.py: 79] Epoch 16, iter 5800/6416, lr 0.000100, loss 0.920080
+INFO 2020-12-09 16:03:54 train.py: 92] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-12-09 16:03:56 train.py: 79] Epoch 16, iter 6000/6416, lr 0.000100, loss 0.923692
+INFO 2020-12-09 16:11:10 train.py: 79] Epoch 16, iter 6200/6416, lr 0.000100, loss 0.924486
+INFO 2020-12-09 16:18:25 train.py: 79] Epoch 16, iter 6400/6416, lr 0.000100, loss 0.925978
+INFO 2020-12-09 16:18:56 train.py: 97] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-12-09 16:19:00 train.py: 79] Epoch 17, iter 0/6416, lr 0.000100, loss 0.924411
+INFO 2020-12-09 16:26:14 train.py: 79] Epoch 17, iter 200/6416, lr 0.000100, loss 0.926630
+INFO 2020-12-09 16:33:29 train.py: 79] Epoch 17, iter 400/6416, lr 0.000100, loss 0.922395
+INFO 2020-12-09 16:40:43 train.py: 79] Epoch 17, iter 600/6416, lr 0.000100, loss 0.905566
+INFO 2020-12-09 16:47:57 train.py: 79] Epoch 17, iter 800/6416, lr 0.000100, loss 0.909944
+INFO 2020-12-09 16:55:11 train.py: 79] Epoch 17, iter 1000/6416, lr 0.000100, loss 0.917820
+INFO 2020-12-09 17:02:25 train.py: 79] Epoch 17, iter 1200/6416, lr 0.000100, loss 0.912816
+INFO 2020-12-09 17:09:39 train.py: 79] Epoch 17, iter 1400/6416, lr 0.000100, loss 0.918477
+INFO 2020-12-09 17:16:53 train.py: 79] Epoch 17, iter 1600/6416, lr 0.000100, loss 0.917358
+INFO 2020-12-09 17:24:07 train.py: 79] Epoch 17, iter 1800/6416, lr 0.000100, loss 0.915639
+INFO 2020-12-09 17:31:20 train.py: 79] Epoch 17, iter 2000/6416, lr 0.000100, loss 0.918681
+INFO 2020-12-09 17:38:34 train.py: 79] Epoch 17, iter 2200/6416, lr 0.000100, loss 0.919049
+INFO 2020-12-09 17:45:48 train.py: 79] Epoch 17, iter 2400/6416, lr 0.000100, loss 0.921275
+INFO 2020-12-09 17:53:02 train.py: 79] Epoch 17, iter 2600/6416, lr 0.000100, loss 0.925036
+INFO 2020-12-09 18:00:16 train.py: 79] Epoch 17, iter 2800/6416, lr 0.000100, loss 0.915752
+INFO 2020-12-09 18:07:29 train.py: 92] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-12-09 18:07:31 train.py: 79] Epoch 17, iter 3000/6416, lr 0.000100, loss 0.918238
+INFO 2020-12-09 18:14:45 train.py: 79] Epoch 17, iter 3200/6416, lr 0.000100, loss 0.915892
+INFO 2020-12-09 18:21:59 train.py: 79] Epoch 17, iter 3400/6416, lr 0.000100, loss 0.919799
+INFO 2020-12-09 18:29:13 train.py: 79] Epoch 17, iter 3600/6416, lr 0.000100, loss 0.920145
+INFO 2020-12-09 18:36:27 train.py: 79] Epoch 17, iter 3800/6416, lr 0.000100, loss 0.919046
+INFO 2020-12-09 18:43:41 train.py: 79] Epoch 17, iter 4000/6416, lr 0.000100, loss 0.918274
+INFO 2020-12-09 18:50:55 train.py: 79] Epoch 17, iter 4200/6416, lr 0.000100, loss 0.914328
+INFO 2020-12-09 18:58:09 train.py: 79] Epoch 17, iter 4400/6416, lr 0.000100, loss 0.910781
+INFO 2020-12-09 19:05:23 train.py: 79] Epoch 17, iter 4600/6416, lr 0.000100, loss 0.920605
+INFO 2020-12-09 19:12:37 train.py: 79] Epoch 17, iter 4800/6416, lr 0.000100, loss 0.930098
+INFO 2020-12-09 19:19:51 train.py: 79] Epoch 17, iter 5000/6416, lr 0.000100, loss 0.920116
+INFO 2020-12-09 19:27:05 train.py: 79] Epoch 17, iter 5200/6416, lr 0.000100, loss 0.920246
+INFO 2020-12-09 19:34:20 train.py: 79] Epoch 17, iter 5400/6416, lr 0.000100, loss 0.930487
+INFO 2020-12-09 19:41:34 train.py: 79] Epoch 17, iter 5600/6416, lr 0.000100, loss 0.926128
+INFO 2020-12-09 19:48:49 train.py: 79] Epoch 17, iter 5800/6416, lr 0.000100, loss 0.920784
+INFO 2020-12-09 19:56:01 train.py: 92] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-12-09 19:56:03 train.py: 79] Epoch 17, iter 6000/6416, lr 0.000100, loss 0.915791
+INFO 2020-12-09 20:03:17 train.py: 79] Epoch 17, iter 6200/6416, lr 0.000100, loss 0.923196
+INFO 2020-12-09 20:10:32 train.py: 79] Epoch 17, iter 6400/6416, lr 0.000100, loss 0.923991
+INFO 2020-12-09 20:11:04 train.py: 97] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-12-09 20:11:04 train.py: 180] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/.gitkeep b/bob/bio/facexzoo/models/backbones/AttentionNet92/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_African.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6456b781dcdb61ca532934ee0c646389b3de9632
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_African.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9705 | 0.0015325682471259722 |
+| Epoch_14_batch_2999.pt | 0.9698333333333334 | 0.001979057014506322 |
+| Epoch_16_batch_2999.pt | 0.9696666666666666 | 0.001606314699422325 |
+| Epoch_17_batch_2999.pt | 0.9691666666666668 | 0.0013888888888888883 |
+| Epoch_17.pt | 0.9691666666666666 | 0.0017258027296676735 |
+| Epoch_13_batch_5999.pt | 0.9691666666666666 | 0.001991494258770543 |
+| Epoch_16_batch_5999.pt | 0.9690000000000001 | 0.0016887426837300754 |
+| Epoch_14_batch_5999.pt | 0.9683333333333334 | 0.001610152971798828 |
+| Epoch_16.pt | 0.9683333333333334 | 0.001756820922315767 |
+| Epoch_15_batch_5999.pt | 0.9681666666666668 | 0.0017993483045224145 |
+| Epoch_14.pt | 0.9681666666666666 | 0.0017293758240303758 |
+| Epoch_17_batch_5999.pt | 0.9676666666666666 | 0.0021401511426953615 |
+| Epoch_15.pt | 0.9673333333333334 | 0.0017249082995844493 |
+| Epoch_13.pt | 0.9666666666666668 | 0.0023306863292670067 |
+| Epoch_13_batch_2999.pt | 0.9666666666666666 | 0.0015713484026367705 |
+| Epoch_12.pt | 0.9665000000000001 | 0.002040485296911887 |
+| Epoch_11_batch_5999.pt | 0.9664999999999999 | 0.0019790570145063195 |
+| Epoch_12_batch_2999.pt | 0.9654999999999999 | 0.002222916558193627 |
+| Epoch_11.pt | 0.9646666666666667 | 0.001855921454276675 |
+| Epoch_10.pt | 0.9636666666666667 | 0.0018888888888888887 |
+| Epoch_11_batch_2999.pt | 0.9633333333333333 | 0.0016850834320114552 |
+| Epoch_12_batch_5999.pt | 0.9629999999999999 | 0.001625415426480867 |
+| Epoch_10_batch_5999.pt | 0.9603333333333332 | 0.001937288419351413 |
+| Epoch_10_batch_2999.pt | 0.9596666666666666 | 0.0022054925823643584 |
+| Epoch_7.pt | 0.9405000000000001 | 0.0036519063176766167 |
+| Epoch_9_batch_2999.pt | 0.9404999999999999 | 0.0016111111111111087 |
+| Epoch_6.pt | 0.9391666666666666 | 0.0029107081994288273 |
+| Epoch_9_batch_5999.pt | 0.9355000000000002 | 0.003152991916369489 |
+| Epoch_9.pt | 0.9333333333333333 | 0.0031720227608044833 |
+| Epoch_3.pt | 0.9326666666666666 | 0.002878185299330807 |
+| Epoch_5.pt | 0.9308333333333334 | 0.003977172517576763 |
+| Epoch_4.pt | 0.9298333333333332 | 0.0037222222222222157 |
+| Epoch_8.pt | 0.9265000000000001 | 0.004175546094247459 |
+| Epoch_2.pt | 0.9161666666666666 | 0.004588296975545775 |
+| Epoch_1.pt | 0.8939999999999999 | 0.003115076837528037 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Asian.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..87aef30cf98c6f266482a0d7cdf6f2327db7484f
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Asian.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9568333333333333 | 0.0031175528547996026 |
+| Epoch_16.pt | 0.9566666666666667 | 0.002876039801232177 |
+| Epoch_15.pt | 0.9563333333333335 | 0.0024695678634325457 |
+| Epoch_17.pt | 0.9563333333333333 | 0.0032183885239911664 |
+| Epoch_13.pt | 0.9563333333333333 | 0.0027307123838765626 |
+| Epoch_14.pt | 0.9558333333333333 | 0.0029943362173804134 |
+| Epoch_14_batch_5999.pt | 0.9555 | 0.0029402485827553847 |
+| Epoch_12.pt | 0.9555 | 0.0027335365778094556 |
+| Epoch_17_batch_2999.pt | 0.9555 | 0.0030837086858617555 |
+| Epoch_15_batch_5999.pt | 0.9555 | 0.0027672020900916124 |
+| Epoch_14_batch_2999.pt | 0.9555 | 0.002811462428640002 |
+| Epoch_16_batch_2999.pt | 0.9553333333333335 | 0.0030102704853653484 |
+| Epoch_16_batch_5999.pt | 0.9553333333333333 | 0.002830608711746007 |
+| Epoch_13_batch_5999.pt | 0.9553333333333333 | 0.0029689753813083126 |
+| Epoch_15_batch_2999.pt | 0.9550000000000001 | 0.0031622776601683772 |
+| Epoch_11_batch_2999.pt | 0.9548333333333334 | 0.002255993389360776 |
+| Epoch_17_batch_5999.pt | 0.9548333333333334 | 0.0028485430540653904 |
+| Epoch_12_batch_5999.pt | 0.9546666666666667 | 0.002696591355447022 |
+| Epoch_11.pt | 0.9544999999999998 | 0.002108916988455811 |
+| Epoch_10_batch_2999.pt | 0.9541666666666666 | 0.0021264065787782956 |
+| Epoch_10_batch_5999.pt | 0.9536666666666667 | 0.0029481109247603532 |
+| Epoch_12_batch_2999.pt | 0.9531666666666666 | 0.0021437534868416377 |
+| Epoch_11_batch_5999.pt | 0.9528333333333332 | 0.0025706078447242818 |
+| Epoch_10.pt | 0.9521666666666666 | 0.0021949718072376014 |
+| Epoch_7.pt | 0.9321666666666667 | 0.0036094013046179493 |
+| Epoch_9_batch_2999.pt | 0.9313333333333332 | 0.0029896942326830423 |
+| Epoch_9.pt | 0.9296666666666666 | 0.0031308895119123094 |
+| Epoch_6.pt | 0.9269999999999999 | 0.0026851213274654635 |
+| Epoch_9_batch_5999.pt | 0.9248333333333333 | 0.004749593874775885 |
+| Epoch_5.pt | 0.9225000000000001 | 0.002952818281315182 |
+| Epoch_8.pt | 0.9216666666666666 | 0.002732971972499743 |
+| Epoch_3.pt | 0.9188333333333333 | 0.0028442057191489425 |
+| Epoch_4.pt | 0.9178333333333333 | 0.003230353724468141 |
+| Epoch_2.pt | 0.9076666666666668 | 0.0018459164139817896 |
+| Epoch_1.pt | 0.8913333333333334 | 0.004126098806139365 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Caucasian.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..42c1f5f862db26a292be0ce9231f0d06fc53d774
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Caucasian.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.993 | 0.0012862041003100268 |
+| Epoch_16.pt | 0.9928333333333332 | 0.0013619611857923655 |
+| Epoch_15.pt | 0.9928333333333332 | 0.0013158576980363368 |
+| Epoch_13_batch_5999.pt | 0.9926666666666666 | 0.0012957670877434013 |
+| Epoch_13.pt | 0.9925 | 0.0013437096247164275 |
+| Epoch_11_batch_2999.pt | 0.9924999999999999 | 0.0014326441064697415 |
+| Epoch_16_batch_5999.pt | 0.9923333333333332 | 0.0013425606637327329 |
+| Epoch_14.pt | 0.9923333333333332 | 0.0013877773329774256 |
+| Epoch_12_batch_2999.pt | 0.9923333333333332 | 0.001342560663732732 |
+| Epoch_17_batch_5999.pt | 0.9921666666666666 | 0.0014065543223524663 |
+| Epoch_17_batch_2999.pt | 0.9920000000000002 | 0.001527525231651945 |
+| Epoch_14_batch_5999.pt | 0.992 | 0.0013562839573037482 |
+| Epoch_12_batch_5999.pt | 0.992 | 0.001333333333333336 |
+| Epoch_14_batch_2999.pt | 0.992 | 0.001333333333333336 |
+| Epoch_10_batch_5999.pt | 0.9918333333333335 | 0.0017471316881684917 |
+| Epoch_17.pt | 0.9918333333333333 | 0.0014153043558729993 |
+| Epoch_15_batch_2999.pt | 0.9918333333333333 | 0.0013709958532503398 |
+| Epoch_11_batch_5999.pt | 0.9918333333333333 | 0.0014792807728549332 |
+| Epoch_16_batch_2999.pt | 0.9915 | 0.0014999999999999985 |
+| Epoch_10_batch_2999.pt | 0.9914999999999999 | 0.0015204369092671154 |
+| Epoch_10.pt | 0.9913333333333334 | 0.0014656562175858778 |
+| Epoch_11.pt | 0.9913333333333334 | 0.001644294287438749 |
+| Epoch_13_batch_2999.pt | 0.9913333333333332 | 0.0014229164972073011 |
+| Epoch_12.pt | 0.991 | 0.0014315665251916822 |
+| Epoch_7.pt | 0.9799999999999999 | 0.0017568209223157553 |
+| Epoch_9.pt | 0.9791666666666666 | 0.0020824072015545817 |
+| Epoch_9_batch_2999.pt | 0.9771666666666668 | 0.0013158576980363298 |
+| Epoch_9_batch_5999.pt | 0.9766666666666666 | 0.0016850834320114542 |
+| Epoch_6.pt | 0.9766666666666666 | 0.0014272480642961195 |
+| Epoch_5.pt | 0.9758333333333333 | 0.002386303510546058 |
+| Epoch_4.pt | 0.9748333333333334 | 0.0020705161281760324 |
+| Epoch_8.pt | 0.9746666666666666 | 0.0018559214542766735 |
+| Epoch_3.pt | 0.9724999999999999 | 0.002225691736001119 |
+| Epoch_2.pt | 0.9671666666666667 | 0.0020038543107634933 |
+| Epoch_1.pt | 0.9556666666666667 | 0.0019116278371205894 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Indian.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f82602507899239eb887155fee258cab3eb8ed4c
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_RFW_Indian.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9678333333333333 | 0.0016489802310728759 |
+| Epoch_14_batch_5999.pt | 0.967 | 0.0019051586888313668 |
+| Epoch_15_batch_2999.pt | 0.9668333333333333 | 0.0020853693754581677 |
+| Epoch_11_batch_5999.pt | 0.9665000000000001 | 0.0021865187393504907 |
+| Epoch_17_batch_5999.pt | 0.9664999999999999 | 0.0018500917561496382 |
+| Epoch_13_batch_5999.pt | 0.9664999999999999 | 0.0021293075440818716 |
+| Epoch_10_batch_5999.pt | 0.9664999999999999 | 0.002465189748037093 |
+| Epoch_15_batch_5999.pt | 0.9663333333333334 | 0.0018223577185396407 |
+| Epoch_16.pt | 0.9663333333333333 | 0.0018559214542766735 |
+| Epoch_17_batch_2999.pt | 0.9663333333333333 | 0.0020608041101101613 |
+| Epoch_13.pt | 0.9661666666666667 | 0.002180865158487834 |
+| Epoch_16_batch_5999.pt | 0.9656666666666667 | 0.001862561623804466 |
+| Epoch_12.pt | 0.9656666666666667 | 0.0018790593916986442 |
+| Epoch_10.pt | 0.9655000000000001 | 0.002383715328102766 |
+| Epoch_14_batch_2999.pt | 0.9653333333333334 | 0.002260776661041757 |
+| Epoch_15.pt | 0.9653333333333333 | 0.0017533037597843863 |
+| Epoch_14.pt | 0.9651666666666667 | 0.0020853693754581673 |
+| Epoch_17.pt | 0.9649999999999999 | 0.0020487876571761983 |
+| Epoch_13_batch_2999.pt | 0.9646666666666667 | 0.0021773242158072675 |
+| Epoch_11.pt | 0.9644999999999999 | 0.0024349563334234554 |
+| Epoch_11_batch_2999.pt | 0.9638333333333333 | 0.002108916988455817 |
+| Epoch_12_batch_2999.pt | 0.9623333333333333 | 0.002320068113091234 |
+| Epoch_12_batch_5999.pt | 0.9613333333333334 | 0.0018223577185396392 |
+| Epoch_10_batch_2999.pt | 0.961 | 0.002619961360567026 |
+| Epoch_9_batch_5999.pt | 0.9501666666666667 | 0.0021147629234082622 |
+| Epoch_9_batch_2999.pt | 0.9463333333333335 | 0.0020905430802474227 |
+| Epoch_6.pt | 0.9455 | 0.0030636257064360515 |
+| Epoch_7.pt | 0.9448333333333334 | 0.0026579719364234816 |
+| Epoch_8.pt | 0.9446666666666665 | 0.0029059326290271125 |
+| Epoch_9.pt | 0.9410000000000001 | 0.002802115602870776 |
+| Epoch_5.pt | 0.9391666666666666 | 0.0020971762320196492 |
+| Epoch_3.pt | 0.9384999999999998 | 0.001994591452335139 |
+| Epoch_4.pt | 0.9366666666666668 | 0.003990729999126216 |
+| Epoch_2.pt | 0.9281666666666666 | 0.003262677080005783 |
+| Epoch_1.pt | 0.9175000000000001 | 0.0034627648982772058 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_agedb30.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_agedb30.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e69997667d4808e56d0694b5792b37b7deefce9b
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_agedb30.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.9808333333333333 | 0.0024999999999999953 |
+| Epoch_11_batch_2999.pt | 0.9806666666666667 | 0.0024113927126900737 |
+| Epoch_15.pt | 0.9804999999999999 | 0.002664929990046986 |
+| Epoch_10_batch_5999.pt | 0.9803333333333333 | 0.0027977062915587 |
+| Epoch_13_batch_5999.pt | 0.9801666666666667 | 0.0026463345252921195 |
+| Epoch_13_batch_2999.pt | 0.9801666666666666 | 0.00244002125066641 |
+| Epoch_14.pt | 0.9799999999999999 | 0.0025458753860865746 |
+| Epoch_16_batch_2999.pt | 0.9799999999999999 | 0.002484519974999761 |
+| Epoch_11_batch_5999.pt | 0.9798333333333333 | 0.0025992639033976784 |
+| Epoch_16_batch_5999.pt | 0.9796666666666667 | 0.0025067809272618807 |
+| Epoch_17_batch_5999.pt | 0.9795000000000001 | 0.002594509873074578 |
+| Epoch_10_batch_2999.pt | 0.9795 | 0.002582586510926547 |
+| Epoch_10.pt | 0.9795 | 0.002676486548987977 |
+| Epoch_16.pt | 0.9793333333333333 | 0.002572408200620048 |
+| Epoch_12_batch_5999.pt | 0.9791666666666666 | 0.0026787918780535928 |
+| Epoch_15_batch_2999.pt | 0.9791666666666666 | 0.0024501196745777714 |
+| Epoch_12_batch_2999.pt | 0.9791666666666664 | 0.002952818281315175 |
+| Epoch_12.pt | 0.9789999999999999 | 0.002723922371584722 |
+| Epoch_13.pt | 0.9789999999999999 | 0.0026666666666666644 |
+| Epoch_17.pt | 0.9788333333333334 | 0.0027783332777888818 |
+| Epoch_11.pt | 0.9788333333333334 | 0.0027222222222222257 |
+| Epoch_17_batch_2999.pt | 0.9786666666666667 | 0.002457038265277331 |
+| Epoch_14_batch_5999.pt | 0.9785 | 0.0027605018330677383 |
+| Epoch_15_batch_5999.pt | 0.9781666666666666 | 0.0026579719364234807 |
+| Epoch_6.pt | 0.9723333333333335 | 0.002523959264800124 |
+| Epoch_9.pt | 0.9723333333333333 | 0.0030751894451328987 |
+| Epoch_9_batch_2999.pt | 0.9716666666666667 | 0.003370166864022908 |
+| Epoch_7.pt | 0.9711666666666666 | 0.002855036701673244 |
+| Epoch_8.pt | 0.9711666666666666 | 0.002344549760667683 |
+| Epoch_5.pt | 0.9710000000000001 | 0.0031642290813106958 |
+| Epoch_9_batch_5999.pt | 0.9695 | 0.002357677241546996 |
+| Epoch_4.pt | 0.9680000000000002 | 0.0033314809667922074 |
+| Epoch_3.pt | 0.9666666666666668 | 0.002421610524189264 |
+| Epoch_2.pt | 0.9661666666666665 | 0.00252212432507026 |
+| Epoch_1.pt | 0.9526666666666668 | 0.0031150768375280418 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..fe01984681abfcf7e6e62f317c221d8bd0d165c7
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_calfw.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.9588333333333333 | 0.003849402711404444 |
+| Epoch_14_batch_5999.pt | 0.9586666666666666 | 0.0038151743807531982 |
+| Epoch_16_batch_5999.pt | 0.9584999999999999 | 0.003947573094109004 |
+| Epoch_15_batch_5999.pt | 0.9584999999999999 | 0.003924047419202289 |
+| Epoch_17_batch_5999.pt | 0.9584999999999999 | 0.003844588950399101 |
+| Epoch_13.pt | 0.9583333333333333 | 0.003967460238079361 |
+| Epoch_16_batch_2999.pt | 0.9581666666666665 | 0.004047938051543746 |
+| Epoch_15_batch_2999.pt | 0.9581666666666665 | 0.00395538389140612 |
+| Epoch_14_batch_2999.pt | 0.9581666666666665 | 0.004032659876435441 |
+| Epoch_17_batch_2999.pt | 0.9579999999999999 | 0.003981438414926422 |
+| Epoch_13_batch_2999.pt | 0.9578333333333333 | 0.004202071799411263 |
+| Epoch_16.pt | 0.9578333333333333 | 0.0038972133127323557 |
+| Epoch_11.pt | 0.9578333333333333 | 0.004082860894816251 |
+| Epoch_12_batch_2999.pt | 0.9578333333333333 | 0.00415776827601504 |
+| Epoch_15.pt | 0.9576666666666667 | 0.004158881616047309 |
+| Epoch_17.pt | 0.9574999999999999 | 0.004084372505713951 |
+| Epoch_12.pt | 0.9574999999999999 | 0.004084372505713957 |
+| Epoch_11_batch_5999.pt | 0.9574999999999998 | 0.003859012219291619 |
+| Epoch_12_batch_5999.pt | 0.9571666666666665 | 0.003913020367320073 |
+| Epoch_10.pt | 0.9571666666666665 | 0.004105476619298294 |
+| Epoch_14.pt | 0.9569999999999999 | 0.004050605807457372 |
+| Epoch_11_batch_2999.pt | 0.9566666666666667 | 0.0038409746705430252 |
+| Epoch_10_batch_5999.pt | 0.9566666666666664 | 0.004172218523448578 |
+| Epoch_10_batch_2999.pt | 0.9555 | 0.004037249399783144 |
+| Epoch_9_batch_2999.pt | 0.9513333333333331 | 0.004185512933741636 |
+| Epoch_9_batch_5999.pt | 0.9511666666666667 | 0.0038813418832685715 |
+| Epoch_7.pt | 0.951 | 0.0037859388972001883 |
+| Epoch_8.pt | 0.9501666666666667 | 0.0038042374035044367 |
+| Epoch_9.pt | 0.9486666666666667 | 0.004214905941864593 |
+| Epoch_3.pt | 0.9481666666666666 | 0.004017323597731317 |
+| Epoch_6.pt | 0.9480000000000001 | 0.003485419364746254 |
+| Epoch_5.pt | 0.9471666666666666 | 0.004150338376804387 |
+| Epoch_2.pt | 0.9446666666666668 | 0.0032375116187407693 |
+| Epoch_4.pt | 0.9428333333333334 | 0.0038654052859553255 |
+| Epoch_1.pt | 0.9356666666666668 | 0.004715354483093983 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f248914235c29d0f898a1b71ac33825c6fe71bec
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_cplfw.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.9033333333333333 | 0.005533288710217213 |
+| Epoch_13.pt | 0.9031666666666667 | 0.005112620550059319 |
+| Epoch_14_batch_5999.pt | 0.9023333333333333 | 0.004920353294552114 |
+| Epoch_16.pt | 0.9018333333333335 | 0.004965622559901973 |
+| Epoch_15_batch_5999.pt | 0.9013333333333333 | 0.005029542354558341 |
+| Epoch_17_batch_5999.pt | 0.9011666666666667 | 0.005420867317056913 |
+| Epoch_15.pt | 0.901 | 0.005528824579833208 |
+| Epoch_16_batch_5999.pt | 0.9006666666666666 | 0.005528824579833213 |
+| Epoch_13_batch_2999.pt | 0.9006666666666666 | 0.0048444954125759195 |
+| Epoch_17_batch_2999.pt | 0.9006666666666666 | 0.0054614247677531175 |
+| Epoch_15_batch_2999.pt | 0.9004999999999999 | 0.00483077837962853 |
+| Epoch_12_batch_5999.pt | 0.9003333333333332 | 0.005269291422835276 |
+| Epoch_17.pt | 0.9 | 0.005235208439728779 |
+| Epoch_16_batch_2999.pt | 0.8998333333333333 | 0.005411749905255874 |
+| Epoch_14.pt | 0.8993333333333332 | 0.004957847003849572 |
+| Epoch_13_batch_5999.pt | 0.8985 | 0.004896773944315621 |
+| Epoch_10_batch_5999.pt | 0.8985 | 0.006375347969200976 |
+| Epoch_12_batch_2999.pt | 0.8978333333333332 | 0.0046417988211787335 |
+| Epoch_10.pt | 0.8968333333333334 | 0.0055691500337812245 |
+| Epoch_11_batch_5999.pt | 0.8968333333333334 | 0.005928847659360227 |
+| Epoch_11.pt | 0.8965 | 0.005722222222222225 |
+| Epoch_10_batch_2999.pt | 0.8961666666666666 | 0.006030017505041844 |
+| Epoch_11_batch_2999.pt | 0.8939999999999999 | 0.005545546539319011 |
+| Epoch_12.pt | 0.8938333333333333 | 0.004925682255181749 |
+| Epoch_9_batch_5999.pt | 0.866 | 0.006246974576386931 |
+| Epoch_5.pt | 0.866 | 0.00622222222222222 |
+| Epoch_7.pt | 0.8651666666666665 | 0.007346999051431023 |
+| Epoch_6.pt | 0.8641666666666667 | 0.007670892555145885 |
+| Epoch_9.pt | 0.8634999999999999 | 0.006297412755045308 |
+| Epoch_4.pt | 0.8606666666666667 | 0.005937950762616616 |
+| Epoch_9_batch_2999.pt | 0.8598333333333332 | 0.004965622559901967 |
+| Epoch_3.pt | 0.8591666666666666 | 0.005617707893376858 |
+| Epoch_2.pt | 0.8588333333333333 | 0.006070826819437769 |
+| Epoch_8.pt | 0.8583333333333334 | 0.005538863813606443 |
+| Epoch_1.pt | 0.841 | 0.006147367197165804 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..30984d13f38f3f23aff2ac4e67507437e9d7290b
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_lfw.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.9981666666666668 | 0.0006309898162000297 |
+| Epoch_15_batch_5999.pt | 0.9981666666666668 | 0.0006781419786518711 |
+| Epoch_17_batch_5999.pt | 0.9981666666666668 | 0.0006309898162000297 |
+| Epoch_13.pt | 0.9981666666666668 | 0.0006309898162000297 |
+| Epoch_10.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_13_batch_2999.pt | 0.9979999999999999 | 0.0005983516452371659 |
+| Epoch_13_batch_5999.pt | 0.9979999999999999 | 0.0005983516452371659 |
+| Epoch_12_batch_2999.pt | 0.9979999999999999 | 0.0005983516452371659 |
+| Epoch_17.pt | 0.9978333333333333 | 0.0007049209744694192 |
+| Epoch_12_batch_5999.pt | 0.9978333333333333 | 0.0007049209744694192 |
+| Epoch_15_batch_2999.pt | 0.9978333333333333 | 0.0007049209744694192 |
+| Epoch_10_batch_5999.pt | 0.9978333333333333 | 0.0007474235581707629 |
+| Epoch_16_batch_5999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_11_batch_2999.pt | 0.9976666666666667 | 0.0008314794192831015 |
+| Epoch_16.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_17_batch_2999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_9.pt | 0.9976666666666667 | 0.0008678055195451802 |
+| Epoch_16_batch_2999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_14_batch_2999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_5.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_11_batch_5999.pt | 0.9976666666666667 | 0.0008678055195451803 |
+| Epoch_11.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_14_batch_5999.pt | 0.9974999999999999 | 0.000833333333333329 |
+| Epoch_10_batch_2999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_9_batch_5999.pt | 0.9974999999999999 | 0.000668977476599572 |
+| Epoch_15.pt | 0.9974999999999999 | 0.0007954345035153557 |
+| Epoch_12.pt | 0.9973333333333333 | 0.0011439589045541142 |
+| Epoch_6.pt | 0.9973333333333333 | 0.0009686442096757043 |
+| Epoch_4.pt | 0.9971666666666668 | 0.0008258927081843642 |
+| Epoch_7.pt | 0.9970000000000001 | 0.000853460638652065 |
+| Epoch_3.pt | 0.9966666666666667 | 0.0008958064164776142 |
+| Epoch_2.pt | 0.9966666666666667 | 0.0008958064164776142 |
+| Epoch_8.pt | 0.9964999999999999 | 0.0009111788592698182 |
+| Epoch_9_batch_2999.pt | 0.9961666666666666 | 0.0013158576980363342 |
+| Epoch_1.pt | 0.9951666666666666 | 0.0010671873729054782 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..924d78a28fd28f5aac20a47e6b0ab855589fd0cb
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.9808831378470911 |
+| 2 | 0.9867867789681971 |
+| 3 | 0.9886092922139631 |
+| 4 | 0.989644219378523 |
+| 5 | 0.9904187875079735 |
+| 6 | 0.9909785594334587 |
+| 7 | 0.9914537146725334 |
+| 8 | 0.9919288699116081 |
+| 9 | 0.9922868635848836 |
+| 10 | 0.9925081687647266 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/AttentionNet92/log.log b/bob/bio/facexzoo/models/backbones/AttentionNet92/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..f1ac01ef49f7d7a549b7af02af1b2a9dd6e9130d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/AttentionNet92/log.log
@@ -0,0 +1,441 @@
+INFO 2021-01-22 11:27:19 train_amp.py: 180] Start optimization.
+INFO 2021-01-22 11:27:19 train_amp.py: 181] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='AttentionNet', batch_size=512, data_root='/home/wangjun492/wj_data/facex-zoo/private_file/train_data/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-hrnet', train_file='/home/wangjun492/wj_data/facex-zoo/private_file/train_data/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f9c284d07f0>)
+backbone param:
+{'stage1_modules': 1, 'stage2_modules': 2, 'stage3_modules': 3, 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+Selected optimization level O1: Insert automatic casts around Pytorch functions and Tensor methods.
+
+Defaults for this optimization level are:
+enabled : True
+opt_level : O1
+cast_model_type : None
+patch_torch_functions : True
+keep_batchnorm_fp32 : None
+master_weights : None
+loss_scale : dynamic
+Processing user overrides (additional kwargs that are not None)...
+After processing overrides, optimization options are:
+enabled : True
+opt_level : O1
+cast_model_type : None
+patch_torch_functions : True
+keep_batchnorm_fp32 : None
+master_weights : None
+loss_scale : dynamic
+INFO 2021-01-22 11:27:47 train_amp.py: 80] Epoch 0, iter 0/6416, lr 0.100000, loss 16.200769
+INFO 2021-01-22 11:36:01 train_amp.py: 80] Epoch 0, iter 200/6416, lr 0.100000, loss 15.650791
+INFO 2021-01-22 11:44:16 train_amp.py: 80] Epoch 0, iter 400/6416, lr 0.100000, loss 15.396689
+INFO 2021-01-22 11:52:31 train_amp.py: 80] Epoch 0, iter 600/6416, lr 0.100000, loss 15.359234
+INFO 2021-01-22 12:00:46 train_amp.py: 80] Epoch 0, iter 800/6416, lr 0.100000, loss 15.313069
+INFO 2021-01-22 12:09:02 train_amp.py: 80] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.244729
+INFO 2021-01-22 12:17:18 train_amp.py: 80] Epoch 0, iter 1200/6416, lr 0.100000, loss 15.122207
+INFO 2021-01-22 12:25:33 train_amp.py: 80] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.887803
+INFO 2021-01-22 12:33:49 train_amp.py: 80] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.569927
+INFO 2021-01-22 12:42:04 train_amp.py: 80] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.213781
+INFO 2021-01-22 12:50:20 train_amp.py: 80] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.832066
+INFO 2021-01-22 12:58:36 train_amp.py: 80] Epoch 0, iter 2200/6416, lr 0.100000, loss 13.460406
+INFO 2021-01-22 13:06:53 train_amp.py: 80] Epoch 0, iter 2400/6416, lr 0.100000, loss 13.025999
+INFO 2021-01-22 13:15:09 train_amp.py: 80] Epoch 0, iter 2600/6416, lr 0.100000, loss 12.596452
+INFO 2021-01-22 13:23:26 train_amp.py: 80] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.186990
+INFO 2021-01-22 13:31:40 train_amp.py: 93] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-01-22 13:31:42 train_amp.py: 80] Epoch 0, iter 3000/6416, lr 0.100000, loss 11.875443
+INFO 2021-01-22 13:39:57 train_amp.py: 80] Epoch 0, iter 3200/6416, lr 0.100000, loss 11.691376
+INFO 2021-01-22 13:48:12 train_amp.py: 80] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.708680
+INFO 2021-01-22 13:56:26 train_amp.py: 80] Epoch 0, iter 3600/6416, lr 0.100000, loss 11.784122
+INFO 2021-01-22 14:04:39 train_amp.py: 80] Epoch 0, iter 3800/6416, lr 0.100000, loss 12.037083
+INFO 2021-01-22 14:12:51 train_amp.py: 80] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.344254
+INFO 2021-01-22 14:21:03 train_amp.py: 80] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.699137
+INFO 2021-01-22 14:29:15 train_amp.py: 80] Epoch 0, iter 4400/6416, lr 0.100000, loss 12.954928
+INFO 2021-01-22 14:37:25 train_amp.py: 80] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.173759
+INFO 2021-01-22 14:45:35 train_amp.py: 80] Epoch 0, iter 4800/6416, lr 0.100000, loss 13.346399
+INFO 2021-01-22 14:53:44 train_amp.py: 80] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.424897
+INFO 2021-01-22 15:01:54 train_amp.py: 80] Epoch 0, iter 5200/6416, lr 0.100000, loss 13.438772
+INFO 2021-01-22 15:10:03 train_amp.py: 80] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.295961
+INFO 2021-01-22 15:18:12 train_amp.py: 80] Epoch 0, iter 5600/6416, lr 0.100000, loss 13.154845
+INFO 2021-01-22 15:26:20 train_amp.py: 80] Epoch 0, iter 5800/6416, lr 0.100000, loss 12.944470
+INFO 2021-01-22 15:34:27 train_amp.py: 93] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-01-22 15:34:29 train_amp.py: 80] Epoch 0, iter 6000/6416, lr 0.100000, loss 12.652119
+INFO 2021-01-22 15:42:37 train_amp.py: 80] Epoch 0, iter 6200/6416, lr 0.100000, loss 12.393544
+INFO 2021-01-22 15:50:46 train_amp.py: 80] Epoch 0, iter 6400/6416, lr 0.100000, loss 12.058863
+INFO 2021-01-22 15:51:21 train_amp.py: 98] Save checkpoint Epoch_0.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-01-22 15:51:25 train_amp.py: 80] Epoch 1, iter 0/6416, lr 0.100000, loss 11.820808
+INFO 2021-01-22 15:59:31 train_amp.py: 80] Epoch 1, iter 200/6416, lr 0.100000, loss 11.347225
+INFO 2021-01-22 16:07:37 train_amp.py: 80] Epoch 1, iter 400/6416, lr 0.100000, loss 11.088671
+INFO 2021-01-22 16:15:44 train_amp.py: 80] Epoch 1, iter 600/6416, lr 0.100000, loss 10.792624
+INFO 2021-01-22 16:23:50 train_amp.py: 80] Epoch 1, iter 800/6416, lr 0.100000, loss 10.556611
+INFO 2021-01-22 16:31:56 train_amp.py: 80] Epoch 1, iter 1000/6416, lr 0.100000, loss 10.294086
+INFO 2021-01-22 16:40:02 train_amp.py: 80] Epoch 1, iter 1200/6416, lr 0.100000, loss 9.994435
+INFO 2021-01-22 16:48:08 train_amp.py: 80] Epoch 1, iter 1400/6416, lr 0.100000, loss 9.727042
+INFO 2021-01-22 16:56:15 train_amp.py: 80] Epoch 1, iter 1600/6416, lr 0.100000, loss 9.491597
+INFO 2021-01-22 17:04:21 train_amp.py: 80] Epoch 1, iter 1800/6416, lr 0.100000, loss 9.267840
+INFO 2021-01-22 17:12:27 train_amp.py: 80] Epoch 1, iter 2000/6416, lr 0.100000, loss 9.047816
+INFO 2021-01-22 17:20:34 train_amp.py: 80] Epoch 1, iter 2200/6416, lr 0.100000, loss 8.831102
+INFO 2021-01-22 17:28:41 train_amp.py: 80] Epoch 1, iter 2400/6416, lr 0.100000, loss 8.630591
+INFO 2021-01-22 17:36:47 train_amp.py: 80] Epoch 1, iter 2600/6416, lr 0.100000, loss 8.447793
+INFO 2021-01-22 17:44:53 train_amp.py: 80] Epoch 1, iter 2800/6416, lr 0.100000, loss 8.307894
+INFO 2021-01-22 17:52:58 train_amp.py: 93] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-01-22 17:53:01 train_amp.py: 80] Epoch 1, iter 3000/6416, lr 0.100000, loss 8.133036
+INFO 2021-01-22 18:01:07 train_amp.py: 80] Epoch 1, iter 3200/6416, lr 0.100000, loss 7.956565
+INFO 2021-01-22 18:09:14 train_amp.py: 80] Epoch 1, iter 3400/6416, lr 0.100000, loss 7.832110
+INFO 2021-01-22 18:17:20 train_amp.py: 80] Epoch 1, iter 3600/6416, lr 0.100000, loss 7.698390
+INFO 2021-01-22 18:25:27 train_amp.py: 80] Epoch 1, iter 3800/6416, lr 0.100000, loss 7.564181
+INFO 2021-01-22 18:33:34 train_amp.py: 80] Epoch 1, iter 4000/6416, lr 0.100000, loss 7.421879
+INFO 2021-01-22 18:41:40 train_amp.py: 80] Epoch 1, iter 4200/6416, lr 0.100000, loss 7.337094
+INFO 2021-01-22 18:49:47 train_amp.py: 80] Epoch 1, iter 4400/6416, lr 0.100000, loss 7.203218
+INFO 2021-01-22 18:57:54 train_amp.py: 80] Epoch 1, iter 4600/6416, lr 0.100000, loss 7.126578
+INFO 2021-01-22 19:06:00 train_amp.py: 80] Epoch 1, iter 4800/6416, lr 0.100000, loss 6.999827
+INFO 2021-01-22 19:14:07 train_amp.py: 80] Epoch 1, iter 5000/6416, lr 0.100000, loss 6.957920
+INFO 2021-01-22 19:22:14 train_amp.py: 80] Epoch 1, iter 5200/6416, lr 0.100000, loss 6.833013
+INFO 2021-01-22 19:30:20 train_amp.py: 80] Epoch 1, iter 5400/6416, lr 0.100000, loss 6.763639
+INFO 2021-01-22 19:38:27 train_amp.py: 80] Epoch 1, iter 5600/6416, lr 0.100000, loss 6.675965
+INFO 2021-01-22 19:46:33 train_amp.py: 80] Epoch 1, iter 5800/6416, lr 0.100000, loss 6.589048
+INFO 2021-01-22 19:54:38 train_amp.py: 93] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-01-22 19:54:40 train_amp.py: 80] Epoch 1, iter 6000/6416, lr 0.100000, loss 6.542221
+INFO 2021-01-22 20:02:46 train_amp.py: 80] Epoch 1, iter 6200/6416, lr 0.100000, loss 6.475078
+INFO 2021-01-22 20:10:53 train_amp.py: 80] Epoch 1, iter 6400/6416, lr 0.100000, loss 6.411744
+INFO 2021-01-22 20:11:28 train_amp.py: 98] Save checkpoint Epoch_1.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-01-22 20:11:32 train_amp.py: 80] Epoch 2, iter 0/6416, lr 0.100000, loss 6.430531
+INFO 2021-01-22 20:19:38 train_amp.py: 80] Epoch 2, iter 200/6416, lr 0.100000, loss 5.644422
+INFO 2021-01-22 20:27:44 train_amp.py: 80] Epoch 2, iter 400/6416, lr 0.100000, loss 5.725728
+INFO 2021-01-22 20:35:51 train_amp.py: 80] Epoch 2, iter 600/6416, lr 0.100000, loss 5.765044
+INFO 2021-01-22 20:43:57 train_amp.py: 80] Epoch 2, iter 800/6416, lr 0.100000, loss 5.819584
+INFO 2021-01-22 20:52:04 train_amp.py: 80] Epoch 2, iter 1000/6416, lr 0.100000, loss 5.813375
+INFO 2021-01-22 21:00:10 train_amp.py: 80] Epoch 2, iter 1200/6416, lr 0.100000, loss 5.825755
+INFO 2021-01-22 21:08:17 train_amp.py: 80] Epoch 2, iter 1400/6416, lr 0.100000, loss 5.800301
+INFO 2021-01-22 21:16:23 train_amp.py: 80] Epoch 2, iter 1600/6416, lr 0.100000, loss 5.800774
+INFO 2021-01-22 21:24:30 train_amp.py: 80] Epoch 2, iter 1800/6416, lr 0.100000, loss 5.803499
+INFO 2021-01-22 21:32:37 train_amp.py: 80] Epoch 2, iter 2000/6416, lr 0.100000, loss 5.793053
+INFO 2021-01-22 21:40:44 train_amp.py: 80] Epoch 2, iter 2200/6416, lr 0.100000, loss 5.752284
+INFO 2021-01-22 21:48:51 train_amp.py: 80] Epoch 2, iter 2400/6416, lr 0.100000, loss 5.708198
+INFO 2021-01-22 21:56:57 train_amp.py: 80] Epoch 2, iter 2600/6416, lr 0.100000, loss 5.727209
+INFO 2021-01-22 22:05:04 train_amp.py: 80] Epoch 2, iter 2800/6416, lr 0.100000, loss 5.657213
+INFO 2021-01-22 22:13:09 train_amp.py: 93] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-01-22 22:13:12 train_amp.py: 80] Epoch 2, iter 3000/6416, lr 0.100000, loss 5.644669
+INFO 2021-01-22 22:21:19 train_amp.py: 80] Epoch 2, iter 3200/6416, lr 0.100000, loss 5.608233
+INFO 2021-01-22 22:29:26 train_amp.py: 80] Epoch 2, iter 3400/6416, lr 0.100000, loss 5.576870
+INFO 2021-01-22 22:37:33 train_amp.py: 80] Epoch 2, iter 3600/6416, lr 0.100000, loss 5.532724
+INFO 2021-01-22 22:45:40 train_amp.py: 80] Epoch 2, iter 3800/6416, lr 0.100000, loss 5.531711
+INFO 2021-01-22 22:53:47 train_amp.py: 80] Epoch 2, iter 4000/6416, lr 0.100000, loss 5.492676
+INFO 2021-01-22 23:01:54 train_amp.py: 80] Epoch 2, iter 4200/6416, lr 0.100000, loss 5.450874
+INFO 2021-01-22 23:10:01 train_amp.py: 80] Epoch 2, iter 4400/6416, lr 0.100000, loss 5.415592
+INFO 2021-01-22 23:18:09 train_amp.py: 80] Epoch 2, iter 4600/6416, lr 0.100000, loss 5.451228
+INFO 2021-01-22 23:26:16 train_amp.py: 80] Epoch 2, iter 4800/6416, lr 0.100000, loss 5.393117
+INFO 2021-01-22 23:34:24 train_amp.py: 80] Epoch 2, iter 5000/6416, lr 0.100000, loss 5.343950
+INFO 2021-01-22 23:42:31 train_amp.py: 80] Epoch 2, iter 5200/6416, lr 0.100000, loss 5.354063
+INFO 2021-01-22 23:50:39 train_amp.py: 80] Epoch 2, iter 5400/6416, lr 0.100000, loss 5.321516
+INFO 2021-01-22 23:58:47 train_amp.py: 80] Epoch 2, iter 5600/6416, lr 0.100000, loss 5.272619
+INFO 2021-01-23 00:06:54 train_amp.py: 80] Epoch 2, iter 5800/6416, lr 0.100000, loss 5.256074
+INFO 2021-01-23 00:14:59 train_amp.py: 93] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-01-23 00:15:02 train_amp.py: 80] Epoch 2, iter 6000/6416, lr 0.100000, loss 5.242422
+INFO 2021-01-23 00:23:09 train_amp.py: 80] Epoch 2, iter 6200/6416, lr 0.100000, loss 5.222620
+INFO 2021-01-23 00:31:16 train_amp.py: 80] Epoch 2, iter 6400/6416, lr 0.100000, loss 5.177877
+INFO 2021-01-23 00:31:52 train_amp.py: 98] Save checkpoint Epoch_2.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-01-23 00:31:55 train_amp.py: 80] Epoch 3, iter 0/6416, lr 0.100000, loss 5.053004
+INFO 2021-01-23 00:40:02 train_amp.py: 80] Epoch 3, iter 200/6416, lr 0.100000, loss 4.496916
+INFO 2021-01-23 00:48:08 train_amp.py: 80] Epoch 3, iter 400/6416, lr 0.100000, loss 4.586180
+INFO 2021-01-23 00:56:14 train_amp.py: 80] Epoch 3, iter 600/6416, lr 0.100000, loss 4.670625
+INFO 2021-01-23 01:04:21 train_amp.py: 80] Epoch 3, iter 800/6416, lr 0.100000, loss 4.706992
+INFO 2021-01-23 01:12:27 train_amp.py: 80] Epoch 3, iter 1000/6416, lr 0.100000, loss 4.776439
+INFO 2021-01-23 01:20:33 train_amp.py: 80] Epoch 3, iter 1200/6416, lr 0.100000, loss 4.773361
+INFO 2021-01-23 01:28:40 train_amp.py: 80] Epoch 3, iter 1400/6416, lr 0.100000, loss 4.837910
+INFO 2021-01-23 01:36:46 train_amp.py: 80] Epoch 3, iter 1600/6416, lr 0.100000, loss 4.836994
+INFO 2021-01-23 01:44:53 train_amp.py: 80] Epoch 3, iter 1800/6416, lr 0.100000, loss 4.864613
+INFO 2021-01-23 01:53:00 train_amp.py: 80] Epoch 3, iter 2000/6416, lr 0.100000, loss 4.810300
+INFO 2021-01-23 02:01:07 train_amp.py: 80] Epoch 3, iter 2200/6416, lr 0.100000, loss 4.842055
+INFO 2021-01-23 02:09:13 train_amp.py: 80] Epoch 3, iter 2400/6416, lr 0.100000, loss 4.846513
+INFO 2021-01-23 02:17:20 train_amp.py: 80] Epoch 3, iter 2600/6416, lr 0.100000, loss 4.828449
+INFO 2021-01-23 02:25:27 train_amp.py: 80] Epoch 3, iter 2800/6416, lr 0.100000, loss 4.795540
+INFO 2021-01-23 02:33:32 train_amp.py: 93] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-01-23 02:33:35 train_amp.py: 80] Epoch 3, iter 3000/6416, lr 0.100000, loss 4.791901
+INFO 2021-01-23 02:41:42 train_amp.py: 80] Epoch 3, iter 3200/6416, lr 0.100000, loss 4.817109
+INFO 2021-01-23 02:49:49 train_amp.py: 80] Epoch 3, iter 3400/6416, lr 0.100000, loss 4.805560
+INFO 2021-01-23 02:57:56 train_amp.py: 80] Epoch 3, iter 3600/6416, lr 0.100000, loss 4.740652
+INFO 2021-01-23 03:06:03 train_amp.py: 80] Epoch 3, iter 3800/6416, lr 0.100000, loss 4.779515
+INFO 2021-01-23 03:14:10 train_amp.py: 80] Epoch 3, iter 4000/6416, lr 0.100000, loss 4.769814
+INFO 2021-01-23 03:22:17 train_amp.py: 80] Epoch 3, iter 4200/6416, lr 0.100000, loss 4.716411
+INFO 2021-01-23 03:30:24 train_amp.py: 80] Epoch 3, iter 4400/6416, lr 0.100000, loss 4.723036
+INFO 2021-01-23 03:38:31 train_amp.py: 80] Epoch 3, iter 4600/6416, lr 0.100000, loss 4.743889
+INFO 2021-01-23 03:46:37 train_amp.py: 80] Epoch 3, iter 4800/6416, lr 0.100000, loss 4.711898
+INFO 2021-01-23 03:54:44 train_amp.py: 80] Epoch 3, iter 5000/6416, lr 0.100000, loss 4.676345
+INFO 2021-01-23 04:02:51 train_amp.py: 80] Epoch 3, iter 5200/6416, lr 0.100000, loss 4.682190
+INFO 2021-01-23 04:10:58 train_amp.py: 80] Epoch 3, iter 5400/6416, lr 0.100000, loss 4.627394
+INFO 2021-01-23 04:19:06 train_amp.py: 80] Epoch 3, iter 5600/6416, lr 0.100000, loss 4.623215
+INFO 2021-01-23 04:27:13 train_amp.py: 80] Epoch 3, iter 5800/6416, lr 0.100000, loss 4.642187
+INFO 2021-01-23 04:35:18 train_amp.py: 93] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-01-23 04:35:20 train_amp.py: 80] Epoch 3, iter 6000/6416, lr 0.100000, loss 4.630307
+INFO 2021-01-23 04:43:26 train_amp.py: 80] Epoch 3, iter 6200/6416, lr 0.100000, loss 4.629063
+INFO 2021-01-23 04:51:33 train_amp.py: 80] Epoch 3, iter 6400/6416, lr 0.100000, loss 4.587063
+INFO 2021-01-23 04:52:08 train_amp.py: 98] Save checkpoint Epoch_3.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+INFO 2021-01-23 04:52:12 train_amp.py: 80] Epoch 4, iter 0/6416, lr 0.100000, loss 4.536706
+INFO 2021-01-23 05:00:17 train_amp.py: 80] Epoch 4, iter 200/6416, lr 0.100000, loss 3.937602
+INFO 2021-01-23 05:08:23 train_amp.py: 80] Epoch 4, iter 400/6416, lr 0.100000, loss 4.016549
+INFO 2021-01-23 05:16:29 train_amp.py: 80] Epoch 4, iter 600/6416, lr 0.100000, loss 4.092552
+INFO 2021-01-23 05:24:35 train_amp.py: 80] Epoch 4, iter 800/6416, lr 0.100000, loss 4.187135
+INFO 2021-01-23 05:32:40 train_amp.py: 80] Epoch 4, iter 1000/6416, lr 0.100000, loss 4.228737
+INFO 2021-01-23 05:40:46 train_amp.py: 80] Epoch 4, iter 1200/6416, lr 0.100000, loss 4.270726
+INFO 2021-01-23 05:48:52 train_amp.py: 80] Epoch 4, iter 1400/6416, lr 0.100000, loss 4.313067
+INFO 2021-01-23 05:56:58 train_amp.py: 80] Epoch 4, iter 1600/6416, lr 0.100000, loss 4.297603
+INFO 2021-01-23 06:05:04 train_amp.py: 80] Epoch 4, iter 1800/6416, lr 0.100000, loss 4.343759
+INFO 2021-01-23 06:13:10 train_amp.py: 80] Epoch 4, iter 2000/6416, lr 0.100000, loss 4.329446
+INFO 2021-01-23 06:21:15 train_amp.py: 80] Epoch 4, iter 2200/6416, lr 0.100000, loss 4.371943
+INFO 2021-01-23 06:29:22 train_amp.py: 80] Epoch 4, iter 2400/6416, lr 0.100000, loss 4.353969
+INFO 2021-01-23 06:37:28 train_amp.py: 80] Epoch 4, iter 2600/6416, lr 0.100000, loss 4.375121
+INFO 2021-01-23 06:45:34 train_amp.py: 80] Epoch 4, iter 2800/6416, lr 0.100000, loss 4.366602
+INFO 2021-01-23 06:53:38 train_amp.py: 93] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-01-23 06:53:41 train_amp.py: 80] Epoch 4, iter 3000/6416, lr 0.100000, loss 4.327982
+INFO 2021-01-23 07:01:47 train_amp.py: 80] Epoch 4, iter 3200/6416, lr 0.100000, loss 4.371157
+INFO 2021-01-23 07:09:53 train_amp.py: 80] Epoch 4, iter 3400/6416, lr 0.100000, loss 4.326776
+INFO 2021-01-23 07:17:59 train_amp.py: 80] Epoch 4, iter 3600/6416, lr 0.100000, loss 4.325711
+INFO 2021-01-23 07:26:05 train_amp.py: 80] Epoch 4, iter 3800/6416, lr 0.100000, loss 4.324026
+INFO 2021-01-23 07:34:12 train_amp.py: 80] Epoch 4, iter 4000/6416, lr 0.100000, loss 4.341099
+INFO 2021-01-23 07:42:18 train_amp.py: 80] Epoch 4, iter 4200/6416, lr 0.100000, loss 4.367066
+INFO 2021-01-23 07:50:24 train_amp.py: 80] Epoch 4, iter 4400/6416, lr 0.100000, loss 4.339646
+INFO 2021-01-23 07:58:31 train_amp.py: 80] Epoch 4, iter 4600/6416, lr 0.100000, loss 4.344855
+INFO 2021-01-23 08:06:37 train_amp.py: 80] Epoch 4, iter 4800/6416, lr 0.100000, loss 4.316831
+INFO 2021-01-23 08:14:43 train_amp.py: 80] Epoch 4, iter 5000/6416, lr 0.100000, loss 4.285109
+INFO 2021-01-23 08:22:50 train_amp.py: 80] Epoch 4, iter 5200/6416, lr 0.100000, loss 4.293263
+INFO 2021-01-23 08:30:56 train_amp.py: 80] Epoch 4, iter 5400/6416, lr 0.100000, loss 4.305744
+INFO 2021-01-23 08:39:03 train_amp.py: 80] Epoch 4, iter 5600/6416, lr 0.100000, loss 4.295299
+INFO 2021-01-23 08:47:09 train_amp.py: 80] Epoch 4, iter 5800/6416, lr 0.100000, loss 4.273463
+INFO 2021-01-23 08:55:14 train_amp.py: 93] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-01-23 08:55:16 train_amp.py: 80] Epoch 4, iter 6000/6416, lr 0.100000, loss 4.270777
+INFO 2021-01-23 09:03:24 train_amp.py: 80] Epoch 4, iter 6200/6416, lr 0.100000, loss 4.282375
+INFO 2021-01-23 09:11:31 train_amp.py: 80] Epoch 4, iter 6400/6416, lr 0.100000, loss 4.245695
+INFO 2021-01-23 09:12:07 train_amp.py: 98] Save checkpoint Epoch_4.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 262144.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-01-23 09:12:10 train_amp.py: 80] Epoch 5, iter 0/6416, lr 0.100000, loss 4.193446
+INFO 2021-01-23 09:20:16 train_amp.py: 80] Epoch 5, iter 200/6416, lr 0.100000, loss 3.615340
+INFO 2021-01-23 09:28:22 train_amp.py: 80] Epoch 5, iter 400/6416, lr 0.100000, loss 3.705306
+INFO 2021-01-23 09:36:28 train_amp.py: 80] Epoch 5, iter 600/6416, lr 0.100000, loss 3.771797
+INFO 2021-01-23 09:44:34 train_amp.py: 80] Epoch 5, iter 800/6416, lr 0.100000, loss 3.832407
+INFO 2021-01-23 09:52:40 train_amp.py: 80] Epoch 5, iter 1000/6416, lr 0.100000, loss 3.905287
+INFO 2021-01-23 10:00:47 train_amp.py: 80] Epoch 5, iter 1200/6416, lr 0.100000, loss 3.948975
+INFO 2021-01-23 10:08:53 train_amp.py: 80] Epoch 5, iter 1400/6416, lr 0.100000, loss 3.995046
+INFO 2021-01-23 10:17:00 train_amp.py: 80] Epoch 5, iter 1600/6416, lr 0.100000, loss 4.005833
+INFO 2021-01-23 10:25:06 train_amp.py: 80] Epoch 5, iter 1800/6416, lr 0.100000, loss 4.016435
+INFO 2021-01-23 10:33:13 train_amp.py: 80] Epoch 5, iter 2000/6416, lr 0.100000, loss 4.020118
+INFO 2021-01-23 10:41:20 train_amp.py: 80] Epoch 5, iter 2200/6416, lr 0.100000, loss 4.049165
+INFO 2021-01-23 10:49:27 train_amp.py: 80] Epoch 5, iter 2400/6416, lr 0.100000, loss 4.064548
+INFO 2021-01-23 10:57:34 train_amp.py: 80] Epoch 5, iter 2600/6416, lr 0.100000, loss 4.068893
+INFO 2021-01-23 11:05:41 train_amp.py: 80] Epoch 5, iter 2800/6416, lr 0.100000, loss 4.445986
+INFO 2021-01-23 11:13:46 train_amp.py: 93] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-01-23 11:13:49 train_amp.py: 80] Epoch 5, iter 3000/6416, lr 0.100000, loss 4.139378
+INFO 2021-01-23 11:21:56 train_amp.py: 80] Epoch 5, iter 3200/6416, lr 0.100000, loss 4.113524
+INFO 2021-01-23 11:30:03 train_amp.py: 80] Epoch 5, iter 3400/6416, lr 0.100000, loss 4.125536
+INFO 2021-01-23 11:38:10 train_amp.py: 80] Epoch 5, iter 3600/6416, lr 0.100000, loss 4.081725
+INFO 2021-01-23 11:46:17 train_amp.py: 80] Epoch 5, iter 3800/6416, lr 0.100000, loss 4.092283
+INFO 2021-01-23 11:54:24 train_amp.py: 80] Epoch 5, iter 4000/6416, lr 0.100000, loss 4.112386
+INFO 2021-01-23 12:02:31 train_amp.py: 80] Epoch 5, iter 4200/6416, lr 0.100000, loss 4.087631
+INFO 2021-01-23 12:10:38 train_amp.py: 80] Epoch 5, iter 4400/6416, lr 0.100000, loss 4.075816
+INFO 2021-01-23 12:18:44 train_amp.py: 80] Epoch 5, iter 4600/6416, lr 0.100000, loss 4.083638
+INFO 2021-01-23 12:26:52 train_amp.py: 80] Epoch 5, iter 4800/6416, lr 0.100000, loss 4.073061
+INFO 2021-01-23 12:34:58 train_amp.py: 80] Epoch 5, iter 5000/6416, lr 0.100000, loss 4.086428
+INFO 2021-01-23 12:43:05 train_amp.py: 80] Epoch 5, iter 5200/6416, lr 0.100000, loss 4.040670
+INFO 2021-01-23 12:51:12 train_amp.py: 80] Epoch 5, iter 5400/6416, lr 0.100000, loss 4.063632
+INFO 2021-01-23 12:59:19 train_amp.py: 80] Epoch 5, iter 5600/6416, lr 0.100000, loss 4.040343
+INFO 2021-01-23 13:07:26 train_amp.py: 80] Epoch 5, iter 5800/6416, lr 0.100000, loss 4.016819
+INFO 2021-01-23 13:15:31 train_amp.py: 93] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-01-23 13:15:34 train_amp.py: 80] Epoch 5, iter 6000/6416, lr 0.100000, loss 4.039247
+INFO 2021-01-23 13:23:40 train_amp.py: 80] Epoch 5, iter 6200/6416, lr 0.100000, loss 3.971773
+INFO 2021-01-23 13:31:47 train_amp.py: 80] Epoch 5, iter 6400/6416, lr 0.100000, loss 3.991577
+INFO 2021-01-23 13:32:22 train_amp.py: 98] Save checkpoint Epoch_5.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+INFO 2021-01-23 13:32:26 train_amp.py: 80] Epoch 6, iter 0/6416, lr 0.100000, loss 3.981857
+INFO 2021-01-23 13:40:31 train_amp.py: 80] Epoch 6, iter 200/6416, lr 0.100000, loss 3.516364
+INFO 2021-01-23 13:48:36 train_amp.py: 80] Epoch 6, iter 400/6416, lr 0.100000, loss 3.493047
+INFO 2021-01-23 13:56:42 train_amp.py: 80] Epoch 6, iter 600/6416, lr 0.100000, loss 3.576897
+INFO 2021-01-23 14:04:47 train_amp.py: 80] Epoch 6, iter 800/6416, lr 0.100000, loss 3.664588
+INFO 2021-01-23 14:12:52 train_amp.py: 80] Epoch 6, iter 1000/6416, lr 0.100000, loss 3.700190
+INFO 2021-01-23 14:20:58 train_amp.py: 80] Epoch 6, iter 1200/6416, lr 0.100000, loss 3.727872
+INFO 2021-01-23 14:29:03 train_amp.py: 80] Epoch 6, iter 1400/6416, lr 0.100000, loss 3.748572
+INFO 2021-01-23 14:37:09 train_amp.py: 80] Epoch 6, iter 1600/6416, lr 0.100000, loss 3.791810
+INFO 2021-01-23 14:45:15 train_amp.py: 80] Epoch 6, iter 1800/6416, lr 0.100000, loss 3.818300
+INFO 2021-01-23 14:53:21 train_amp.py: 80] Epoch 6, iter 2000/6416, lr 0.100000, loss 3.808685
+INFO 2021-01-23 15:01:27 train_amp.py: 80] Epoch 6, iter 2200/6416, lr 0.100000, loss 3.854625
+INFO 2021-01-23 15:09:33 train_amp.py: 80] Epoch 6, iter 2400/6416, lr 0.100000, loss 3.866210
+INFO 2021-01-23 15:17:38 train_amp.py: 80] Epoch 6, iter 2600/6416, lr 0.100000, loss 3.853472
+INFO 2021-01-23 15:25:44 train_amp.py: 80] Epoch 6, iter 2800/6416, lr 0.100000, loss 3.871405
+INFO 2021-01-23 15:33:48 train_amp.py: 93] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-01-23 15:33:51 train_amp.py: 80] Epoch 6, iter 3000/6416, lr 0.100000, loss 3.855416
+INFO 2021-01-23 15:41:57 train_amp.py: 80] Epoch 6, iter 3200/6416, lr 0.100000, loss 3.884241
+INFO 2021-01-23 15:50:03 train_amp.py: 80] Epoch 6, iter 3400/6416, lr 0.100000, loss 3.908138
+INFO 2021-01-23 15:58:09 train_amp.py: 80] Epoch 6, iter 3600/6416, lr 0.100000, loss 3.886681
+INFO 2021-01-23 16:06:16 train_amp.py: 80] Epoch 6, iter 3800/6416, lr 0.100000, loss 3.887693
+INFO 2021-01-23 16:14:22 train_amp.py: 80] Epoch 6, iter 4000/6416, lr 0.100000, loss 3.887921
+INFO 2021-01-23 16:22:28 train_amp.py: 80] Epoch 6, iter 4200/6416, lr 0.100000, loss 3.892143
+INFO 2021-01-23 16:30:34 train_amp.py: 80] Epoch 6, iter 4400/6416, lr 0.100000, loss 3.874170
+INFO 2021-01-23 16:38:41 train_amp.py: 80] Epoch 6, iter 4600/6416, lr 0.100000, loss 3.874878
+INFO 2021-01-23 16:46:47 train_amp.py: 80] Epoch 6, iter 4800/6416, lr 0.100000, loss 3.862447
+INFO 2021-01-23 16:54:53 train_amp.py: 80] Epoch 6, iter 5000/6416, lr 0.100000, loss 3.856809
+INFO 2021-01-23 17:02:59 train_amp.py: 80] Epoch 6, iter 5200/6416, lr 0.100000, loss 3.843367
+INFO 2021-01-23 17:11:05 train_amp.py: 80] Epoch 6, iter 5400/6416, lr 0.100000, loss 3.860189
+INFO 2021-01-23 17:19:11 train_amp.py: 80] Epoch 6, iter 5600/6416, lr 0.100000, loss 3.888652
+INFO 2021-01-23 17:27:18 train_amp.py: 80] Epoch 6, iter 5800/6416, lr 0.100000, loss 3.844141
+INFO 2021-01-23 17:35:22 train_amp.py: 93] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-01-23 17:35:25 train_amp.py: 80] Epoch 6, iter 6000/6416, lr 0.100000, loss 3.815569
+INFO 2021-01-23 17:43:32 train_amp.py: 80] Epoch 6, iter 6200/6416, lr 0.100000, loss 3.845630
+INFO 2021-01-23 17:51:39 train_amp.py: 80] Epoch 6, iter 6400/6416, lr 0.100000, loss 3.844138
+INFO 2021-01-23 17:52:14 train_amp.py: 98] Save checkpoint Epoch_6.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+INFO 2021-01-23 17:52:18 train_amp.py: 80] Epoch 7, iter 0/6416, lr 0.100000, loss 3.852761
+INFO 2021-01-23 18:00:24 train_amp.py: 80] Epoch 7, iter 200/6416, lr 0.100000, loss 3.233932
+INFO 2021-01-23 18:08:30 train_amp.py: 80] Epoch 7, iter 400/6416, lr 0.100000, loss 3.333622
+INFO 2021-01-23 18:16:36 train_amp.py: 80] Epoch 7, iter 600/6416, lr 0.100000, loss 3.372904
+INFO 2021-01-23 18:24:42 train_amp.py: 80] Epoch 7, iter 800/6416, lr 0.100000, loss 3.467978
+INFO 2021-01-23 18:32:48 train_amp.py: 80] Epoch 7, iter 1000/6416, lr 0.100000, loss 3.515557
+INFO 2021-01-23 18:40:54 train_amp.py: 80] Epoch 7, iter 1200/6416, lr 0.100000, loss 3.548216
+INFO 2021-01-23 18:49:00 train_amp.py: 80] Epoch 7, iter 1400/6416, lr 0.100000, loss 3.605942
+INFO 2021-01-23 18:57:06 train_amp.py: 80] Epoch 7, iter 1600/6416, lr 0.100000, loss 3.625488
+INFO 2021-01-23 19:05:13 train_amp.py: 80] Epoch 7, iter 1800/6416, lr 0.100000, loss 3.661789
+INFO 2021-01-23 19:13:19 train_amp.py: 80] Epoch 7, iter 2000/6416, lr 0.100000, loss 3.696077
+INFO 2021-01-23 19:21:26 train_amp.py: 80] Epoch 7, iter 2200/6416, lr 0.100000, loss 3.688675
+INFO 2021-01-23 19:29:32 train_amp.py: 80] Epoch 7, iter 2400/6416, lr 0.100000, loss 3.675459
+INFO 2021-01-23 19:37:39 train_amp.py: 80] Epoch 7, iter 2600/6416, lr 0.100000, loss 3.684645
+INFO 2021-01-23 19:45:46 train_amp.py: 80] Epoch 7, iter 2800/6416, lr 0.100000, loss 3.725261
+INFO 2021-01-23 19:53:51 train_amp.py: 93] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-01-23 19:53:54 train_amp.py: 80] Epoch 7, iter 3000/6416, lr 0.100000, loss 3.713844
+INFO 2021-01-23 20:02:01 train_amp.py: 80] Epoch 7, iter 3200/6416, lr 0.100000, loss 3.748073
+INFO 2021-01-23 20:10:08 train_amp.py: 80] Epoch 7, iter 3400/6416, lr 0.100000, loss 3.712234
+INFO 2021-01-23 20:18:15 train_amp.py: 80] Epoch 7, iter 3600/6416, lr 0.100000, loss 3.741482
+INFO 2021-01-23 20:26:22 train_amp.py: 80] Epoch 7, iter 3800/6416, lr 0.100000, loss 3.737261
+INFO 2021-01-23 20:34:29 train_amp.py: 80] Epoch 7, iter 4000/6416, lr 0.100000, loss 3.729734
+INFO 2021-01-23 20:42:36 train_amp.py: 80] Epoch 7, iter 4200/6416, lr 0.100000, loss 3.723875
+INFO 2021-01-23 20:50:42 train_amp.py: 80] Epoch 7, iter 4400/6416, lr 0.100000, loss 3.694521
+INFO 2021-01-23 20:58:49 train_amp.py: 80] Epoch 7, iter 4600/6416, lr 0.100000, loss 3.726618
+INFO 2021-01-23 21:06:56 train_amp.py: 80] Epoch 7, iter 4800/6416, lr 0.100000, loss 3.707411
+INFO 2021-01-23 21:15:03 train_amp.py: 80] Epoch 7, iter 5000/6416, lr 0.100000, loss 3.758164
+INFO 2021-01-23 21:23:10 train_amp.py: 80] Epoch 7, iter 5200/6416, lr 0.100000, loss 3.713144
+INFO 2021-01-23 21:31:17 train_amp.py: 80] Epoch 7, iter 5400/6416, lr 0.100000, loss 3.727918
+INFO 2021-01-23 21:39:24 train_amp.py: 80] Epoch 7, iter 5600/6416, lr 0.100000, loss 3.740217
+INFO 2021-01-23 21:47:31 train_amp.py: 80] Epoch 7, iter 5800/6416, lr 0.100000, loss 3.695104
+INFO 2021-01-23 21:55:37 train_amp.py: 93] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-01-23 21:55:39 train_amp.py: 80] Epoch 7, iter 6000/6416, lr 0.100000, loss 3.708433
+INFO 2021-01-23 22:03:47 train_amp.py: 80] Epoch 7, iter 6200/6416, lr 0.100000, loss 3.690604
+INFO 2021-01-23 22:11:54 train_amp.py: 80] Epoch 7, iter 6400/6416, lr 0.100000, loss 3.702294
+INFO 2021-01-23 22:12:30 train_amp.py: 98] Save checkpoint Epoch_7.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 262144.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-01-23 22:12:33 train_amp.py: 80] Epoch 8, iter 0/6416, lr 0.100000, loss 3.705586
+INFO 2021-01-23 22:20:39 train_amp.py: 80] Epoch 8, iter 200/6416, lr 0.100000, loss 3.111450
+INFO 2021-01-23 22:28:45 train_amp.py: 80] Epoch 8, iter 400/6416, lr 0.100000, loss 3.183126
+INFO 2021-01-23 22:36:50 train_amp.py: 80] Epoch 8, iter 600/6416, lr 0.100000, loss 3.249723
+INFO 2021-01-23 22:44:56 train_amp.py: 80] Epoch 8, iter 800/6416, lr 0.100000, loss 3.340384
+INFO 2021-01-23 22:53:02 train_amp.py: 80] Epoch 8, iter 1000/6416, lr 0.100000, loss 3.394046
+INFO 2021-01-23 23:01:08 train_amp.py: 80] Epoch 8, iter 1200/6416, lr 0.100000, loss 3.439463
+INFO 2021-01-23 23:09:13 train_amp.py: 80] Epoch 8, iter 1400/6416, lr 0.100000, loss 3.463581
+INFO 2021-01-23 23:17:19 train_amp.py: 80] Epoch 8, iter 1600/6416, lr 0.100000, loss 3.505763
+INFO 2021-01-23 23:25:25 train_amp.py: 80] Epoch 8, iter 1800/6416, lr 0.100000, loss 3.523740
+INFO 2021-01-23 23:33:31 train_amp.py: 80] Epoch 8, iter 2000/6416, lr 0.100000, loss 3.572195
+INFO 2021-01-23 23:41:37 train_amp.py: 80] Epoch 8, iter 2200/6416, lr 0.100000, loss 3.566671
+INFO 2021-01-23 23:49:44 train_amp.py: 80] Epoch 8, iter 2400/6416, lr 0.100000, loss 3.570859
+INFO 2021-01-23 23:57:50 train_amp.py: 80] Epoch 8, iter 2600/6416, lr 0.100000, loss 3.588276
+INFO 2021-01-24 00:05:56 train_amp.py: 80] Epoch 8, iter 2800/6416, lr 0.100000, loss 3.578560
+INFO 2021-01-24 00:14:01 train_amp.py: 93] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-01-24 00:14:03 train_amp.py: 80] Epoch 8, iter 3000/6416, lr 0.100000, loss 3.632966
+INFO 2021-01-24 00:22:10 train_amp.py: 80] Epoch 8, iter 3200/6416, lr 0.100000, loss 3.661230
+INFO 2021-01-24 00:30:17 train_amp.py: 80] Epoch 8, iter 3400/6416, lr 0.100000, loss 3.620683
+INFO 2021-01-24 00:38:23 train_amp.py: 80] Epoch 8, iter 3600/6416, lr 0.100000, loss 3.603778
+INFO 2021-01-24 00:46:29 train_amp.py: 80] Epoch 8, iter 3800/6416, lr 0.100000, loss 3.626872
+INFO 2021-01-24 00:54:36 train_amp.py: 80] Epoch 8, iter 4000/6416, lr 0.100000, loss 3.618878
+INFO 2021-01-24 01:02:42 train_amp.py: 80] Epoch 8, iter 4200/6416, lr 0.100000, loss 3.600972
+INFO 2021-01-24 01:10:49 train_amp.py: 80] Epoch 8, iter 4400/6416, lr 0.100000, loss 3.599495
+INFO 2021-01-24 01:18:55 train_amp.py: 80] Epoch 8, iter 4600/6416, lr 0.100000, loss 3.599033
+INFO 2021-01-24 01:27:02 train_amp.py: 80] Epoch 8, iter 4800/6416, lr 0.100000, loss 3.616074
+INFO 2021-01-24 01:35:08 train_amp.py: 80] Epoch 8, iter 5000/6416, lr 0.100000, loss 3.611207
+INFO 2021-01-24 01:43:15 train_amp.py: 80] Epoch 8, iter 5200/6416, lr 0.100000, loss 3.613900
+INFO 2021-01-24 01:51:21 train_amp.py: 80] Epoch 8, iter 5400/6416, lr 0.100000, loss 3.604460
+INFO 2021-01-24 01:59:27 train_amp.py: 80] Epoch 8, iter 5600/6416, lr 0.100000, loss 3.573313
+INFO 2021-01-24 02:07:34 train_amp.py: 80] Epoch 8, iter 5800/6416, lr 0.100000, loss 3.615657
+INFO 2021-01-24 02:15:38 train_amp.py: 93] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-01-24 02:15:41 train_amp.py: 80] Epoch 8, iter 6000/6416, lr 0.100000, loss 3.601553
+INFO 2021-01-24 02:23:47 train_amp.py: 80] Epoch 8, iter 6200/6416, lr 0.100000, loss 3.612821
+INFO 2021-01-24 02:31:53 train_amp.py: 80] Epoch 8, iter 6400/6416, lr 0.100000, loss 3.585719
+INFO 2021-01-24 02:32:28 train_amp.py: 98] Save checkpoint Epoch_8.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-01-24 02:32:32 train_amp.py: 80] Epoch 9, iter 0/6416, lr 0.100000, loss 3.535126
+INFO 2021-01-24 02:40:37 train_amp.py: 80] Epoch 9, iter 200/6416, lr 0.100000, loss 3.000203
+INFO 2021-01-24 02:48:43 train_amp.py: 80] Epoch 9, iter 400/6416, lr 0.100000, loss 3.076744
+INFO 2021-01-24 02:56:49 train_amp.py: 80] Epoch 9, iter 600/6416, lr 0.100000, loss 3.143883
+INFO 2021-01-24 03:04:55 train_amp.py: 80] Epoch 9, iter 800/6416, lr 0.100000, loss 3.222606
+INFO 2021-01-24 03:13:01 train_amp.py: 80] Epoch 9, iter 1000/6416, lr 0.100000, loss 3.259705
+INFO 2021-01-24 03:21:07 train_amp.py: 80] Epoch 9, iter 1200/6416, lr 0.100000, loss 3.338220
+INFO 2021-01-24 03:29:13 train_amp.py: 80] Epoch 9, iter 1400/6416, lr 0.100000, loss 3.380289
+INFO 2021-01-24 03:37:20 train_amp.py: 80] Epoch 9, iter 1600/6416, lr 0.100000, loss 3.384573
+INFO 2021-01-24 03:45:27 train_amp.py: 80] Epoch 9, iter 1800/6416, lr 0.100000, loss 3.427571
+INFO 2021-01-24 03:53:33 train_amp.py: 80] Epoch 9, iter 2000/6416, lr 0.100000, loss 3.481041
+INFO 2021-01-24 04:01:40 train_amp.py: 80] Epoch 9, iter 2200/6416, lr 0.100000, loss 3.472408
+INFO 2021-01-24 04:09:46 train_amp.py: 80] Epoch 9, iter 2400/6416, lr 0.100000, loss 3.480687
+INFO 2021-01-24 04:17:53 train_amp.py: 80] Epoch 9, iter 2600/6416, lr 0.100000, loss 3.479921
+INFO 2021-01-24 04:25:59 train_amp.py: 80] Epoch 9, iter 2800/6416, lr 0.100000, loss 3.487379
+INFO 2021-01-24 04:34:05 train_amp.py: 93] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-01-24 04:34:07 train_amp.py: 80] Epoch 9, iter 3000/6416, lr 0.100000, loss 3.479804
+INFO 2021-01-24 04:42:14 train_amp.py: 80] Epoch 9, iter 3200/6416, lr 0.100000, loss 3.510940
+INFO 2021-01-24 04:50:21 train_amp.py: 80] Epoch 9, iter 3400/6416, lr 0.100000, loss 3.503377
+INFO 2021-01-24 04:58:28 train_amp.py: 80] Epoch 9, iter 3600/6416, lr 0.100000, loss 3.521613
+INFO 2021-01-24 05:06:35 train_amp.py: 80] Epoch 9, iter 3800/6416, lr 0.100000, loss 3.500217
+INFO 2021-01-24 05:14:42 train_amp.py: 80] Epoch 9, iter 4000/6416, lr 0.100000, loss 3.527514
+INFO 2021-01-24 05:22:49 train_amp.py: 80] Epoch 9, iter 4200/6416, lr 0.100000, loss 3.539997
+INFO 2021-01-24 05:30:56 train_amp.py: 80] Epoch 9, iter 4400/6416, lr 0.100000, loss 3.511874
+INFO 2021-01-24 05:39:03 train_amp.py: 80] Epoch 9, iter 4600/6416, lr 0.100000, loss 3.500803
+INFO 2021-01-24 05:47:10 train_amp.py: 80] Epoch 9, iter 4800/6416, lr 0.100000, loss 3.516222
+INFO 2021-01-24 05:55:17 train_amp.py: 80] Epoch 9, iter 5000/6416, lr 0.100000, loss 3.520138
+INFO 2021-01-24 06:03:24 train_amp.py: 80] Epoch 9, iter 5200/6416, lr 0.100000, loss 3.503154
+INFO 2021-01-24 06:11:31 train_amp.py: 80] Epoch 9, iter 5400/6416, lr 0.100000, loss 3.531918
+INFO 2021-01-24 06:19:38 train_amp.py: 80] Epoch 9, iter 5600/6416, lr 0.100000, loss 3.527579
+INFO 2021-01-24 06:27:44 train_amp.py: 80] Epoch 9, iter 5800/6416, lr 0.100000, loss 3.507871
+INFO 2021-01-24 06:35:49 train_amp.py: 93] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-01-24 06:35:52 train_amp.py: 80] Epoch 9, iter 6000/6416, lr 0.100000, loss 3.534637
+INFO 2021-01-24 06:43:59 train_amp.py: 80] Epoch 9, iter 6200/6416, lr 0.100000, loss 3.502730
+INFO 2021-01-24 06:52:06 train_amp.py: 80] Epoch 9, iter 6400/6416, lr 0.100000, loss 3.515443
+INFO 2021-01-24 06:52:41 train_amp.py: 98] Save checkpoint Epoch_9.pt to disk...
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 131072.0
+INFO 2021-01-24 06:52:45 train_amp.py: 80] Epoch 10, iter 0/6416, lr 0.010000, loss 3.468381
+INFO 2021-01-24 07:00:51 train_amp.py: 80] Epoch 10, iter 200/6416, lr 0.010000, loss 2.396712
+INFO 2021-01-24 07:08:56 train_amp.py: 80] Epoch 10, iter 400/6416, lr 0.010000, loss 2.183902
+INFO 2021-01-24 07:17:02 train_amp.py: 80] Epoch 10, iter 600/6416, lr 0.010000, loss 2.073261
+INFO 2021-01-24 07:25:08 train_amp.py: 80] Epoch 10, iter 800/6416, lr 0.010000, loss 2.029956
+INFO 2021-01-24 07:33:14 train_amp.py: 80] Epoch 10, iter 1000/6416, lr 0.010000, loss 1.977752
+INFO 2021-01-24 07:41:20 train_amp.py: 80] Epoch 10, iter 1200/6416, lr 0.010000, loss 1.921100
+INFO 2021-01-24 07:49:26 train_amp.py: 80] Epoch 10, iter 1400/6416, lr 0.010000, loss 1.907342
+INFO 2021-01-24 07:57:32 train_amp.py: 80] Epoch 10, iter 1600/6416, lr 0.010000, loss 1.889862
+INFO 2021-01-24 08:05:39 train_amp.py: 80] Epoch 10, iter 1800/6416, lr 0.010000, loss 1.855511
+INFO 2021-01-24 08:13:45 train_amp.py: 80] Epoch 10, iter 2000/6416, lr 0.010000, loss 1.824386
+INFO 2021-01-24 08:21:52 train_amp.py: 80] Epoch 10, iter 2200/6416, lr 0.010000, loss 1.809564
+INFO 2021-01-24 08:29:58 train_amp.py: 80] Epoch 10, iter 2400/6416, lr 0.010000, loss 1.774494
+INFO 2021-01-24 08:38:05 train_amp.py: 80] Epoch 10, iter 2600/6416, lr 0.010000, loss 1.762994
+INFO 2021-01-24 08:46:12 train_amp.py: 80] Epoch 10, iter 2800/6416, lr 0.010000, loss 1.723699
+INFO 2021-01-24 08:54:17 train_amp.py: 93] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-01-24 08:54:19 train_amp.py: 80] Epoch 10, iter 3000/6416, lr 0.010000, loss 1.706979
+INFO 2021-01-24 09:02:26 train_amp.py: 80] Epoch 10, iter 3200/6416, lr 0.010000, loss 1.696589
+INFO 2021-01-24 09:10:32 train_amp.py: 80] Epoch 10, iter 3400/6416, lr 0.010000, loss 1.695312
+INFO 2021-01-24 09:18:39 train_amp.py: 80] Epoch 10, iter 3600/6416, lr 0.010000, loss 1.675947
+INFO 2021-01-24 09:26:46 train_amp.py: 80] Epoch 10, iter 3800/6416, lr 0.010000, loss 1.644826
+INFO 2021-01-24 09:34:52 train_amp.py: 80] Epoch 10, iter 4000/6416, lr 0.010000, loss 1.638687
+INFO 2021-01-24 09:42:59 train_amp.py: 80] Epoch 10, iter 4200/6416, lr 0.010000, loss 1.627776
+INFO 2021-01-24 09:51:05 train_amp.py: 80] Epoch 10, iter 4400/6416, lr 0.010000, loss 1.612081
+INFO 2021-01-24 09:59:12 train_amp.py: 80] Epoch 10, iter 4600/6416, lr 0.010000, loss 1.604629
+INFO 2021-01-24 10:07:19 train_amp.py: 80] Epoch 10, iter 4800/6416, lr 0.010000, loss 1.592123
+INFO 2021-01-24 10:15:26 train_amp.py: 80] Epoch 10, iter 5000/6416, lr 0.010000, loss 1.577276
+INFO 2021-01-24 10:23:33 train_amp.py: 80] Epoch 10, iter 5200/6416, lr 0.010000, loss 1.578025
+INFO 2021-01-24 10:31:39 train_amp.py: 80] Epoch 10, iter 5400/6416, lr 0.010000, loss 1.547893
+INFO 2021-01-24 10:39:46 train_amp.py: 80] Epoch 10, iter 5600/6416, lr 0.010000, loss 1.549067
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/.gitkeep b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d89bfbb72d988cda38e27e6e56cbc1837eb0c511
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9663333333333334 | 0.0031894889098682973 |
+| Epoch_14_batch_5999.pt | 0.9660000000000002 | 0.00339934634239519 |
+| Epoch_16_batch_5999.pt | 0.9656666666666668 | 0.003850605211369666 |
+| Epoch_17.pt | 0.9653333333333334 | 0.00329421490674969 |
+| Epoch_17_batch_2999.pt | 0.9651666666666667 | 0.0038924586790229725 |
+| Epoch_15_batch_5999.pt | 0.9650000000000001 | 0.0036430214023900004 |
+| Epoch_13.pt | 0.9648333333333333 | 0.0032815420945827515 |
+| Epoch_15.pt | 0.9645000000000001 | 0.0038333333333333383 |
+| Epoch_14.pt | 0.9643333333333335 | 0.0036868133384526853 |
+| Epoch_13_batch_5999.pt | 0.9641666666666667 | 0.0035853784000583503 |
+| Epoch_16.pt | 0.9638333333333333 | 0.0037847158511669188 |
+| Epoch_16_batch_2999.pt | 0.9633333333333335 | 0.0033884334848837583 |
+| Epoch_14_batch_2999.pt | 0.9633333333333333 | 0.0037679611017362607 |
+| Epoch_15_batch_2999.pt | 0.9631666666666667 | 0.003578485092263985 |
+| Epoch_17_batch_5999.pt | 0.9630000000000001 | 0.0036497928234793015 |
+| Epoch_12.pt | 0.9626666666666667 | 0.0035329140212410353 |
+| Epoch_11_batch_5999.pt | 0.9626666666666667 | 0.003325917677132393 |
+| Epoch_12_batch_5999.pt | 0.9621666666666666 | 0.0036349639562123517 |
+| Epoch_12_batch_2999.pt | 0.9616666666666667 | 0.003951870943061946 |
+| Epoch_10_batch_2999.pt | 0.9603333333333335 | 0.003581502546952484 |
+| Epoch_10_batch_5999.pt | 0.9601666666666666 | 0.003374285475346716 |
+| Epoch_10.pt | 0.9598333333333333 | 0.0033096380019125545 |
+| Epoch_11.pt | 0.9593333333333334 | 0.0040383959656413925 |
+| Epoch_11_batch_2999.pt | 0.958 | 0.0034676636742949447 |
+| Epoch_9_batch_5999.pt | 0.9498333333333333 | 0.004570774038239255 |
+| Epoch_7_batch_5999.pt | 0.9484999999999999 | 0.0035956935833965364 |
+| Epoch_9_batch_2999.pt | 0.9468333333333332 | 0.003300299275617587 |
+| Epoch_8_batch_2999.pt | 0.9461666666666666 | 0.004360668767375644 |
+| Epoch_8_batch_5999.pt | 0.9458333333333334 | 0.0027470522922391346 |
+| Epoch_7.pt | 0.9453333333333334 | 0.0041484788227987655 |
+| Epoch_9.pt | 0.9433333333333334 | 0.004209043761083513 |
+| Epoch_6_batch_2999.pt | 0.9411666666666667 | 0.004231349783184352 |
+| Epoch_7_batch_2999.pt | 0.9410000000000001 | 0.0042903854614017405 |
+| Epoch_6_batch_5999.pt | 0.9404999999999999 | 0.004142895152778258 |
+| Epoch_6.pt | 0.9395 | 0.004493136054576833 |
+| Epoch_8.pt | 0.9391666666666666 | 0.004362084109434132 |
+| Epoch_5.pt | 0.9384999999999998 | 0.00375524324802693 |
+| Epoch_5_batch_2999.pt | 0.9375 | 0.00409945795874962 |
+| Epoch_5_batch_5999.pt | 0.9365 | 0.0036383587400073184 |
+| Epoch_4_batch_5999.pt | 0.9348333333333333 | 0.004730058856940736 |
+| Epoch_4_batch_2999.pt | 0.932 | 0.004751867728965126 |
+| Epoch_4.pt | 0.9316666666666666 | 0.0040368671387966525 |
+| Epoch_3.pt | 0.9293333333333333 | 0.004173697771331763 |
+| Epoch_3_batch_2999.pt | 0.9273333333333333 | 0.005031996388544109 |
+| Epoch_3_batch_5999.pt | 0.9238333333333332 | 0.0055168093301459315 |
+| Epoch_2.pt | 0.9113333333333333 | 0.004309048762147854 |
+| Epoch_2_batch_5999.pt | 0.9056666666666666 | 0.005013561854523765 |
+| Epoch_2_batch_2999.pt | 0.9019999999999999 | 0.005413745651146596 |
+| Epoch_1_batch_5999.pt | 0.8671666666666666 | 0.004931944248881377 |
+| Epoch_1.pt | 0.857 | 0.005909816065320982 |
+| Epoch_1_batch_2999.pt | 0.8113333333333334 | 0.006942444156361474 |
+| Epoch_0_batch_5999.pt | 0.6745000000000001 | 0.004628481339075794 |
+| Epoch_0.pt | 0.6385000000000001 | 0.005094477765552 |
+| Epoch_0_batch_2999.pt | 0.5281666666666667 | 0.0028267898296018964 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..adea368b5e99f84e3cf9a3af1b70c75ee1e7b75f
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_calfw.txt
@@ -0,0 +1,21 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.9436666666666665 | 0.002979352817287095 |
+| Epoch_14.pt | 0.9433333333333334 | 0.0034783279649996685 |
+| Epoch_17.pt | 0.9431666666666667 | 0.0035698497392038513 |
+| Epoch_16_batch_5999.pt | 0.9424999999999999 | 0.0035681201607403157 |
+| Epoch_17_batch_5999.pt | 0.9423333333333332 | 0.00363623737154524 |
+| Epoch_16_batch_2999.pt | 0.9423333333333332 | 0.003627739492736554 |
+| Epoch_16.pt | 0.9421666666666665 | 0.0035403319925232943 |
+| Epoch_15_batch_2999.pt | 0.9421666666666665 | 0.003928763824052697 |
+| Epoch_10_batch_5999.pt | 0.942 | 0.0037908271353848813 |
+| Epoch_15_batch_5999.pt | 0.942 | 0.0038473977095957288 |
+| Epoch_13.pt | 0.9419999999999998 | 0.0036750745352313843 |
+| Epoch_11.pt | 0.9418333333333335 | 0.0034106767729268515 |
+| Epoch_17_batch_2999.pt | 0.9418333333333333 | 0.003655285366576887 |
+| Epoch_13_batch_2999.pt | 0.9418333333333331 | 0.0036637193541772337 |
+| Epoch_14_batch_2999.pt | 0.9416666666666667 | 0.0036514837167011035 |
+| Epoch_13_batch_5999.pt | 0.9411666666666667 | 0.0032207851201412705 |
+| Epoch_15.pt | 0.9403333333333332 | 0.003590109871423003 |
+| Epoch_12_batch_5999.pt | 0.9398333333333335 | 0.004093430448841924 |
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6e8f60b62b3d58a6245839edc39da2538e61b29d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_cplfw.txt
@@ -0,0 +1,20 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8471666666666667 | 0.00594236723832024 |
+| Epoch_16_batch_2999.pt | 0.8455000000000001 | 0.005736228812351742 |
+| Epoch_14_batch_2999.pt | 0.8455 | 0.005589065603568065 |
+| Epoch_16.pt | 0.8451666666666668 | 0.006188649506641124 |
+| Epoch_15_batch_2999.pt | 0.8451666666666666 | 0.0067515430335097016 |
+| Epoch_15_batch_5999.pt | 0.8451666666666666 | 0.006113383415926358 |
+| Epoch_13_batch_2999.pt | 0.8451666666666666 | 0.005791916461439711 |
+| Epoch_17_batch_2999.pt | 0.8448333333333334 | 0.006037179048545562 |
+| Epoch_17_batch_5999.pt | 0.8445 | 0.006216515239939106 |
+| Epoch_13_batch_5999.pt | 0.844 | 0.006227180564089805 |
+| Epoch_13.pt | 0.8436666666666666 | 0.006506407098647713 |
+| Epoch_16_batch_5999.pt | 0.8435 | 0.006341369171004238 |
+| Epoch_15.pt | 0.8428333333333334 | 0.006070826819437778 |
+| Epoch_11.pt | 0.8421666666666667 | 0.0058639408650783455 |
+| Epoch_12.pt | 0.8420000000000002 | 0.006259807120445905 |
+| Epoch_12_batch_2999.pt | 0.8413333333333334 | 0.006033855103282431 |
+| Epoch_12_batch_5999.pt | 0.8411666666666665 | 0.006459990826733539 |
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b1529b73de0037fc5dc43a5756696c2811598c00
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11_batch_2999.pt | 0.9955 | 0.0010555555555555596 |
+| Epoch_12.pt | 0.9954999999999998 | 0.000963852865160968 |
+| Epoch_11.pt | 0.9953333333333335 | 0.0011055415967851348 |
+| Epoch_10_batch_5999.pt | 0.9951666666666668 | 0.0010378634273483006 |
+| Epoch_15_batch_5999.pt | 0.9951666666666666 | 0.0010378634273483065 |
+| Epoch_10.pt | 0.9951666666666666 | 0.0010957268290731133 |
+| Epoch_12_batch_5999.pt | 0.9950000000000001 | 0.0010829771494232222 |
+| Epoch_10_batch_2999.pt | 0.9948333333333335 | 0.000976577546180387 |
+| Epoch_9.pt | 0.9946666666666667 | 0.0013562839573037487 |
+| Epoch_17_batch_2999.pt | 0.9946666666666667 | 0.0012120791238484135 |
+| Epoch_7.pt | 0.9945 | 0.0011666666666666696 |
+| Epoch_17.pt | 0.9945 | 0.0011399046960379597 |
+| Epoch_15_batch_2999.pt | 0.9945 | 0.0011399046960379542 |
+| Epoch_13_batch_5999.pt | 0.9943333333333333 | 0.001143958904554113 |
+| Epoch_16_batch_5999.pt | 0.9943333333333333 | 0.0011967032904743344 |
+| Epoch_17_batch_5999.pt | 0.9941666666666669 | 0.001118033988749893 |
+| Epoch_14_batch_2999.pt | 0.9941666666666666 | 0.0010613873985857113 |
+| Epoch_12_batch_2999.pt | 0.9941666666666666 | 0.0011719457283182776 |
+| Epoch_14.pt | 0.9940000000000001 | 0.0010599324460188288 |
+| Epoch_15.pt | 0.9940000000000001 | 0.0010599324460188288 |
+| Epoch_14_batch_5999.pt | 0.9940000000000001 | 0.0010599324460188288 |
+| Epoch_11_batch_5999.pt | 0.9940000000000001 | 0.0011166528467912136 |
+| Epoch_16.pt | 0.9938333333333335 | 0.0011399046960379573 |
+| Epoch_9_batch_5999.pt | 0.9938333333333335 | 0.0011666666666666709 |
+| Epoch_9_batch_2999.pt | 0.9938333333333335 | 0.001112499133027825 |
+| Epoch_13.pt | 0.9938333333333335 | 0.0009953596037316098 |
+| Epoch_8_batch_5999.pt | 0.9938333333333335 | 0.001084401183107948 |
+| Epoch_16_batch_2999.pt | 0.9938333333333335 | 0.0010844011831079511 |
+| Epoch_6_batch_5999.pt | 0.9936666666666667 | 0.0013099806802835086 |
+| Epoch_13_batch_2999.pt | 0.9936666666666667 | 0.0010183501544346332 |
+| Epoch_7_batch_2999.pt | 0.9933333333333334 | 0.001427248064296131 |
+| Epoch_5_batch_5999.pt | 0.9931666666666666 | 0.0012777777777777852 |
+| Epoch_5_batch_2999.pt | 0.9931666666666666 | 0.0014582671942674158 |
+| Epoch_6_batch_2999.pt | 0.993 | 0.0012619796324000575 |
+| Epoch_4_batch_5999.pt | 0.993 | 0.0014444444444444448 |
+| Epoch_8.pt | 0.9928333333333335 | 0.001361961185792361 |
+| Epoch_5.pt | 0.9928333333333332 | 0.0015723301886761047 |
+| Epoch_8_batch_2999.pt | 0.9925 | 0.0013888888888888883 |
+| Epoch_7_batch_5999.pt | 0.9921666666666666 | 0.001055555555555557 |
+| Epoch_4_batch_2999.pt | 0.9921666666666666 | 0.0009953596037316124 |
+| Epoch_6.pt | 0.992 | 0.0011600340565456216 |
+| Epoch_4.pt | 0.9918333333333333 | 0.0011235415786753757 |
+| Epoch_3_batch_5999.pt | 0.991 | 0.001409841948938838 |
+| Epoch_3_batch_2999.pt | 0.9896666666666667 | 0.0017881641043812299 |
+| Epoch_3.pt | 0.9895000000000002 | 0.0016489802310728672 |
+| Epoch_2.pt | 0.9886666666666665 | 0.0013562839573037452 |
+| Epoch_2_batch_5999.pt | 0.9883333333333333 | 0.0011653431646335003 |
+| Epoch_2_batch_2999.pt | 0.9863333333333333 | 0.001048220125784065 |
+| Epoch_1_batch_5999.pt | 0.9784999999999998 | 0.0017821127702606022 |
+| Epoch_1.pt | 0.9775 | 0.001536590742882148 |
+| Epoch_1_batch_2999.pt | 0.9585000000000001 | 0.001979057014506319 |
+| Epoch_0.pt | 0.9030000000000001 | 0.004456926915584795 |
+| Epoch_0_batch_5999.pt | 0.8895 | 0.004588296975545772 |
+| Epoch_0_batch_2999.pt | 0.583 | 0.007555555555555552 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b32c1e2bed36691d5c5d5d5097ff54bbfb8f34b1
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.8966666666666667 | 0.003548604316149175 |
+| Epoch_13_batch_5999.pt | 0.8958333333333333 | 0.003670452590924201 |
+| Epoch_14.pt | 0.8953333333333333 | 0.003581502546952488 |
+| Epoch_13.pt | 0.8951666666666668 | 0.0035870996571906507 |
+| Epoch_15.pt | 0.8945000000000001 | 0.0034251250315107413 |
+| Epoch_17_batch_5999.pt | 0.8945000000000001 | 0.0035490391674854282 |
+| Epoch_17.pt | 0.8941666666666667 | 0.003542075139843345 |
+| Epoch_16_batch_2999.pt | 0.8934999999999998 | 0.003578485092263977 |
+| Epoch_15_batch_5999.pt | 0.893 | 0.0040582183039443685 |
+| Epoch_16.pt | 0.8928333333333335 | 0.003333796264150549 |
+| Epoch_11_batch_5999.pt | 0.8928333333333333 | 0.0036771734985273138 |
+| Epoch_15_batch_2999.pt | 0.8926666666666667 | 0.004151453709393197 |
+| Epoch_10.pt | 0.8926666666666667 | 0.004812535072823936 |
+| Epoch_17_batch_2999.pt | 0.8923333333333334 | 0.0033902547338649607 |
+| Epoch_14_batch_2999.pt | 0.8918333333333335 | 0.004024999041484435 |
+| Epoch_12_batch_5999.pt | 0.8916666666666668 | 0.00253372316688697 |
+| Epoch_11.pt | 0.8916666666666666 | 0.0037433067839828947 |
+| Epoch_10_batch_5999.pt | 0.8911666666666667 | 0.003540331992523296 |
+| Epoch_16_batch_5999.pt | 0.8911666666666667 | 0.0036771734985273107 |
+| Epoch_14_batch_5999.pt | 0.8901666666666666 | 0.0036213529972738286 |
+| Epoch_11_batch_2999.pt | 0.8881666666666668 | 0.0038123418809550575 |
+| Epoch_10_batch_2999.pt | 0.8881666666666665 | 0.003595693583396534 |
+| Epoch_12.pt | 0.8863333333333333 | 0.00405821830394437 |
+| Epoch_12_batch_2999.pt | 0.8843333333333334 | 0.003890475866669243 |
+| Epoch_9_batch_5999.pt | 0.8601666666666666 | 0.004677566635509284 |
+| Epoch_9_batch_2999.pt | 0.8585 | 0.005406043714397372 |
+| Epoch_7.pt | 0.8584999999999999 | 0.005094477765552003 |
+| Epoch_8_batch_5999.pt | 0.8578333333333333 | 0.004956913117487634 |
+| Epoch_7_batch_5999.pt | 0.8558333333333333 | 0.004319422114983366 |
+| Epoch_8_batch_2999.pt | 0.8541666666666666 | 0.0044531165394114895 |
+| Epoch_6_batch_5999.pt | 0.8538333333333334 | 0.005483139138742954 |
+| Epoch_9.pt | 0.8528333333333334 | 0.005061656879952483 |
+| Epoch_5_batch_5999.pt | 0.8513333333333334 | 0.006069555681665629 |
+| Epoch_5.pt | 0.8493333333333333 | 0.0052363874079769685 |
+| Epoch_7_batch_2999.pt | 0.8491666666666667 | 0.006607375228289023 |
+| Epoch_6_batch_2999.pt | 0.8461666666666666 | 0.004317992789294009 |
+| Epoch_5_batch_2999.pt | 0.8453333333333333 | 0.006155395104206458 |
+| Epoch_8.pt | 0.8443333333333334 | 0.004920353294552119 |
+| Epoch_4_batch_2999.pt | 0.842 | 0.005615235008219233 |
+| Epoch_4.pt | 0.841 | 0.005381724905916597 |
+| Epoch_6.pt | 0.8385 | 0.006828815127778958 |
+| Epoch_4_batch_5999.pt | 0.8353333333333332 | 0.00515081199399967 |
+| Epoch_3.pt | 0.8306666666666667 | 0.004038395965641387 |
+| Epoch_3_batch_2999.pt | 0.8283333333333334 | 0.005762801102173307 |
+| Epoch_3_batch_5999.pt | 0.8278333333333334 | 0.005837564602968194 |
+| Epoch_2_batch_5999.pt | 0.8101666666666668 | 0.00715546066189666 |
+| Epoch_2.pt | 0.8021666666666667 | 0.005122270426382005 |
+| Epoch_2_batch_2999.pt | 0.7916666666666666 | 0.004520187912624002 |
+| Epoch_1_batch_5999.pt | 0.7523333333333333 | 0.007132779486979569 |
+| Epoch_1.pt | 0.7516666666666667 | 0.005957669608202005 |
+| Epoch_1_batch_2999.pt | 0.692 | 0.0043233503240763796 |
+| Epoch_0.pt | 0.6098333333333333 | 0.006163662932115997 |
+| Epoch_0_batch_5999.pt | 0.5868333333333333 | 0.008470901536407215 |
+| Epoch_0_batch_2999.pt | 0.5143333333333333 | 0.006051221690751157 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6937c6c5192059ad646a4e80f6910618b0edecff
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.8931666666666664 | 0.005178600316163422 |
+| Epoch_13_batch_2999.pt | 0.8923333333333332 | 0.0055176484524156145 |
+| Epoch_14.pt | 0.8918333333333333 | 0.005400331494169 |
+| Epoch_15.pt | 0.8915 | 0.005196449405098935 |
+| Epoch_14_batch_5999.pt | 0.8915 | 0.006057593948462085 |
+| Epoch_17_batch_5999.pt | 0.8913333333333334 | 0.005017254179944731 |
+| Epoch_13_batch_5999.pt | 0.8911666666666667 | 0.005386597450967118 |
+| Epoch_17_batch_2999.pt | 0.8908333333333335 | 0.005189317136741494 |
+| Epoch_16_batch_5999.pt | 0.8901666666666668 | 0.0052555085739032295 |
+| Epoch_13.pt | 0.8895 | 0.005403759550900713 |
+| Epoch_16_batch_2999.pt | 0.8893333333333333 | 0.005099019513592788 |
+| Epoch_17.pt | 0.8888333333333334 | 0.005049446858839779 |
+| Epoch_15_batch_2999.pt | 0.8888333333333334 | 0.0056875975096782935 |
+| Epoch_15_batch_5999.pt | 0.8888333333333334 | 0.005116241386596798 |
+| Epoch_16.pt | 0.8886666666666667 | 0.004546060565661954 |
+| Epoch_10_batch_5999.pt | 0.8886666666666665 | 0.004949123878071136 |
+| Epoch_11_batch_2999.pt | 0.8876666666666667 | 0.004793256580027328 |
+| Epoch_11.pt | 0.8871666666666667 | 0.0045950187737227146 |
+| Epoch_12_batch_2999.pt | 0.8871666666666667 | 0.005223699554629561 |
+| Epoch_12_batch_5999.pt | 0.8860000000000001 | 0.0049453806853928675 |
+| Epoch_11_batch_5999.pt | 0.8858333333333335 | 0.005147515282177112 |
+| Epoch_10_batch_2999.pt | 0.8831666666666667 | 0.004946940693218612 |
+| Epoch_12.pt | 0.883 | 0.004222222222222221 |
+| Epoch_10.pt | 0.8826666666666666 | 0.005038126243890947 |
+| Epoch_8_batch_5999.pt | 0.8629999999999999 | 0.005174724898753341 |
+| Epoch_9_batch_2999.pt | 0.8615 | 0.005045778090959789 |
+| Epoch_7_batch_5999.pt | 0.8540000000000001 | 0.004728427295767375 |
+| Epoch_6_batch_2999.pt | 0.8496666666666666 | 0.005222222222222215 |
+| Epoch_8_batch_2999.pt | 0.8495000000000001 | 0.004641798821178732 |
+| Epoch_7_batch_2999.pt | 0.8493333333333333 | 0.006359594676112972 |
+| Epoch_6_batch_5999.pt | 0.8485000000000001 | 0.0051606894936028404 |
+| Epoch_9.pt | 0.8481666666666667 | 0.003836552593471013 |
+| Epoch_9_batch_5999.pt | 0.8456666666666667 | 0.005922336876105079 |
+| Epoch_4.pt | 0.8455 | 0.005516809330145931 |
+| Epoch_5_batch_2999.pt | 0.8448333333333332 | 0.005388888888888897 |
+| Epoch_5.pt | 0.8443333333333334 | 0.005300826863056258 |
+| Epoch_5_batch_5999.pt | 0.844 | 0.00460943744726479 |
+| Epoch_7.pt | 0.8433333333333332 | 0.005036900869619174 |
+| Epoch_8.pt | 0.8421666666666667 | 0.005067750858658319 |
+| Epoch_4_batch_5999.pt | 0.837 | 0.005920251913225434 |
+| Epoch_6.pt | 0.8365 | 0.005445861493593779 |
+| Epoch_4_batch_2999.pt | 0.8316666666666667 | 0.004906533814626588 |
+| Epoch_3_batch_2999.pt | 0.826 | 0.004754465087852151 |
+| Epoch_3_batch_5999.pt | 0.8256666666666665 | 0.004562325592810619 |
+| Epoch_3.pt | 0.819 | 0.00429757324573638 |
+| Epoch_2_batch_5999.pt | 0.8161666666666667 | 0.004029597290175235 |
+| Epoch_2.pt | 0.8145 | 0.003865405285955327 |
+| Epoch_2_batch_2999.pt | 0.7981666666666667 | 0.004984234403857092 |
+| Epoch_1.pt | 0.7721666666666668 | 0.005217787715303805 |
+| Epoch_1_batch_5999.pt | 0.7670000000000001 | 0.006069555681665625 |
+| Epoch_1_batch_2999.pt | 0.7223333333333334 | 0.005111111111111112 |
+| Epoch_0_batch_5999.pt | 0.6678333333333333 | 0.006540708233634249 |
+| Epoch_0.pt | 0.661 | 0.006607141665474894 |
+| Epoch_0_batch_2999.pt | 0.5161666666666667 | 0.005884956758497271 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..394c17e62fc1a7ca17d48f21e46d1c2687f7bf74
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.9610000000000001 | 0.001990719207463218 |
+| Epoch_17_batch_2999.pt | 0.9598333333333333 | 0.00212930754408187 |
+| Epoch_16_batch_5999.pt | 0.9593333333333331 | 0.0021545243810739186 |
+| Epoch_13_batch_2999.pt | 0.9591666666666667 | 0.00248762368635979 |
+| Epoch_15_batch_5999.pt | 0.9590000000000002 | 0.002398559238324773 |
+| Epoch_15_batch_2999.pt | 0.959 | 0.0023333333333333322 |
+| Epoch_14_batch_5999.pt | 0.9588333333333334 | 0.0022505143445032353 |
+| Epoch_17.pt | 0.9588333333333333 | 0.0025945098730745793 |
+| Epoch_16_batch_2999.pt | 0.9588333333333333 | 0.0020645449084513334 |
+| Epoch_17_batch_5999.pt | 0.9584999999999999 | 0.0021001175746039823 |
+| Epoch_15.pt | 0.9583333333333334 | 0.001791612832955231 |
+| Epoch_13.pt | 0.9578333333333333 | 0.002049540749521857 |
+| Epoch_14_batch_2999.pt | 0.9576666666666668 | 0.001812167381144451 |
+| Epoch_16.pt | 0.9573333333333333 | 0.0019116278371205792 |
+| Epoch_13_batch_5999.pt | 0.9566666666666667 | 0.0023959842947608684 |
+| Epoch_10.pt | 0.9561666666666667 | 0.001957101661534281 |
+| Epoch_12_batch_5999.pt | 0.9560000000000001 | 0.0016329931618554486 |
+| Epoch_12_batch_2999.pt | 0.9556666666666667 | 0.002306726610225191 |
+| Epoch_11_batch_2999.pt | 0.9555 | 0.0018765939726727238 |
+| Epoch_12.pt | 0.9553333333333333 | 0.00248203421152097 |
+| Epoch_11.pt | 0.9546666666666666 | 0.002457038265277336 |
+| Epoch_10_batch_5999.pt | 0.9525 | 0.0019914942587705518 |
+| Epoch_11_batch_5999.pt | 0.9521666666666666 | 0.002064544908451343 |
+| Epoch_10_batch_2999.pt | 0.9489999999999998 | 0.0012472191289246474 |
+| Epoch_9_batch_5999.pt | 0.9341666666666667 | 0.0018299632324445083 |
+| Epoch_7_batch_2999.pt | 0.9334999999999999 | 0.003057575090181559 |
+| Epoch_9_batch_2999.pt | 0.9316666666666666 | 0.0027777777777777688 |
+| Epoch_7_batch_5999.pt | 0.9305000000000001 | 0.004029597290175244 |
+| Epoch_8_batch_2999.pt | 0.9303333333333332 | 0.003546864377669408 |
+| Epoch_7.pt | 0.9301666666666666 | 0.0038924586790229643 |
+| Epoch_6_batch_5999.pt | 0.9288333333333334 | 0.0034341243245612366 |
+| Epoch_9.pt | 0.9276666666666668 | 0.0030449310235654875 |
+| Epoch_8_batch_5999.pt | 0.9276666666666665 | 0.003371997978998558 |
+| Epoch_5_batch_5999.pt | 0.9271666666666667 | 0.003343041418517864 |
+| Epoch_5_batch_2999.pt | 0.9218333333333334 | 0.00341971408811386 |
+| Epoch_6_batch_2999.pt | 0.9216666666666666 | 0.0024845199749997694 |
+| Epoch_6.pt | 0.9209999999999999 | 0.0028349668493718072 |
+| Epoch_8.pt | 0.9208333333333334 | 0.0034805455794838045 |
+| Epoch_5.pt | 0.9206666666666667 | 0.0036192216610894078 |
+| Epoch_4_batch_5999.pt | 0.9176666666666666 | 0.003711842908553351 |
+| Epoch_4_batch_2999.pt | 0.9163333333333334 | 0.0034228715112776427 |
+| Epoch_4.pt | 0.9161666666666667 | 0.0037354656608920593 |
+| Epoch_3_batch_5999.pt | 0.9036666666666667 | 0.003666666666666667 |
+| Epoch_3.pt | 0.8985 | 0.0029444444444444487 |
+| Epoch_3_batch_2999.pt | 0.8955000000000002 | 0.0028658267503397653 |
+| Epoch_2_batch_5999.pt | 0.8921666666666667 | 0.0035577250325745232 |
+| Epoch_2.pt | 0.8891666666666665 | 0.004629814811111254 |
+| Epoch_2_batch_2999.pt | 0.8781666666666668 | 0.0029128281619818473 |
+| Epoch_1_batch_5999.pt | 0.8511666666666666 | 0.0036349639562123512 |
+| Epoch_1.pt | 0.8476666666666667 | 0.005147215476400212 |
+| Epoch_1_batch_2999.pt | 0.8071666666666666 | 0.004919412290502854 |
+| Epoch_0.pt | 0.7268333333333333 | 0.004306541105840776 |
+| Epoch_0_batch_5999.pt | 0.7035000000000001 | 0.0043351135944220895 |
+| Epoch_0_batch_2999.pt | 0.5223333333333333 | 0.005859465277082314 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e533e28dee3644365f26b9924945cbed8d2262a1
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.9193333333333333 | 0.004114113018127475 |
+| Epoch_12.pt | 0.9186666666666667 | 0.004784233364802442 |
+| Epoch_13_batch_2999.pt | 0.9186666666666667 | 0.003831319922125935 |
+| Epoch_13_batch_5999.pt | 0.9183333333333333 | 0.004186987484759282 |
+| Epoch_17.pt | 0.9181666666666667 | 0.0037552432480269364 |
+| Epoch_15_batch_5999.pt | 0.9181666666666667 | 0.0038924586790229686 |
+| Epoch_13.pt | 0.9179999999999999 | 0.003965904066127726 |
+| Epoch_17_batch_2999.pt | 0.9179999999999999 | 0.004103596736137643 |
+| Epoch_15.pt | 0.9178333333333335 | 0.003967849185704242 |
+| Epoch_14.pt | 0.9175000000000001 | 0.004305107503568754 |
+| Epoch_15_batch_2999.pt | 0.9174999999999999 | 0.0039145975622117395 |
+| Epoch_16.pt | 0.917 | 0.0041260988061393705 |
+| Epoch_16_batch_5999.pt | 0.9168333333333333 | 0.0038845213569807416 |
+| Epoch_16_batch_2999.pt | 0.9166666666666666 | 0.003857012212824401 |
+| Epoch_10_batch_5999.pt | 0.9164999999999999 | 0.0037222222222222275 |
+| Epoch_14_batch_5999.pt | 0.9161666666666667 | 0.004021930621719612 |
+| Epoch_14_batch_2999.pt | 0.9156666666666669 | 0.0045215533220835154 |
+| Epoch_11.pt | 0.9138333333333334 | 0.004067713890663439 |
+| Epoch_11_batch_5999.pt | 0.9138333333333334 | 0.003296556377588176 |
+| Epoch_12_batch_5999.pt | 0.9133333333333334 | 0.003591828861165473 |
+| Epoch_11_batch_2999.pt | 0.913 | 0.004163331998932264 |
+| Epoch_10_batch_2999.pt | 0.9128333333333332 | 0.0034964708839021275 |
+| Epoch_10.pt | 0.9118333333333333 | 0.003262677080005785 |
+| Epoch_12_batch_2999.pt | 0.9096666666666667 | 0.004273085449562091 |
+| Epoch_7_batch_5999.pt | 0.8903333333333334 | 0.0032659863237108934 |
+| Epoch_8_batch_5999.pt | 0.8883333333333333 | 0.004785523437190676 |
+| Epoch_6_batch_5999.pt | 0.8876666666666667 | 0.004210510071443643 |
+| Epoch_8_batch_2999.pt | 0.8865000000000001 | 0.0053831584652254474 |
+| Epoch_9_batch_5999.pt | 0.8863333333333333 | 0.004294699575575043 |
+| Epoch_7_batch_2999.pt | 0.8846666666666667 | 0.004579880854561103 |
+| Epoch_7.pt | 0.8846666666666666 | 0.003413842554608272 |
+| Epoch_9_batch_2999.pt | 0.883 | 0.004065816547451549 |
+| Epoch_5_batch_2999.pt | 0.8795 | 0.003897213312732355 |
+| Epoch_6.pt | 0.8785000000000001 | 0.005337093581832214 |
+| Epoch_5_batch_5999.pt | 0.8781666666666667 | 0.004820545023230008 |
+| Epoch_8.pt | 0.8780000000000001 | 0.0053679432574138515 |
+| Epoch_6_batch_2999.pt | 0.8773333333333333 | 0.004225145187073673 |
+| Epoch_5.pt | 0.8768333333333335 | 0.0044752405273658525 |
+| Epoch_9.pt | 0.876 | 0.0029938207967349973 |
+| Epoch_4_batch_2999.pt | 0.8748333333333334 | 0.005866045842572359 |
+| Epoch_4.pt | 0.8748333333333334 | 0.004475240527365849 |
+| Epoch_4_batch_5999.pt | 0.8696666666666666 | 0.004764840364975905 |
+| Epoch_3_batch_5999.pt | 0.8661666666666668 | 0.005264896561421621 |
+| Epoch_3_batch_2999.pt | 0.8603333333333334 | 0.0034228715112776345 |
+| Epoch_3.pt | 0.859 | 0.00382648341282723 |
+| Epoch_2_batch_5999.pt | 0.8526666666666667 | 0.004582575694955836 |
+| Epoch_2.pt | 0.8431666666666666 | 0.006042289239953614 |
+| Epoch_2_batch_2999.pt | 0.8318333333333332 | 0.005094477765551996 |
+| Epoch_1_batch_5999.pt | 0.8045 | 0.006086059764477351 |
+| Epoch_1.pt | 0.8024999999999999 | 0.0077190241179396065 |
+| Epoch_1_batch_2999.pt | 0.7508333333333334 | 0.0037944891814509297 |
+| Epoch_0.pt | 0.6948333333333332 | 0.003215030288051173 |
+| Epoch_0_batch_5999.pt | 0.6801666666666666 | 0.0038526085439079577 |
+| Epoch_0_batch_2999.pt | 0.5255 | 0.005006477285958142 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/EfficientNet_B0/log.log b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..4444bfe2a4b523de8f00872dc7c8a53a8fcf1a3b
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/EfficientNet_B0/log.log
@@ -0,0 +1,657 @@
+INFO 2020-12-04 15:43:22 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2020-12-04 15:43:22 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2020-12-04 15:43:22 train.py: 177] Start optimization.
+INFO 2020-12-04 15:43:22 train.py: 178] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='EfficientNet', batch_size=512, data_root='/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-effi', train_file='/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f9ae71fce80>)
+backbone param:
+{'width': 1.0, 'depth': 1.0, 'image_size': 112, 'drop_ratio': 0.2, 'out_h': 7, 'out_w': 7, 'feat_dim': 512}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-12-04 15:43:46 train.py: 79] Epoch 0, iter 0/6416, lr 0.100000, loss 16.337477
+INFO 2020-12-04 15:47:20 train.py: 79] Epoch 0, iter 200/6416, lr 0.100000, loss 15.642263
+INFO 2020-12-04 15:50:53 train.py: 79] Epoch 0, iter 400/6416, lr 0.100000, loss 15.402893
+INFO 2020-12-04 15:54:27 train.py: 79] Epoch 0, iter 600/6416, lr 0.100000, loss 15.359507
+INFO 2020-12-04 15:58:01 train.py: 79] Epoch 0, iter 800/6416, lr 0.100000, loss 15.329064
+INFO 2020-12-04 16:01:35 train.py: 79] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.299876
+INFO 2020-12-04 16:05:08 train.py: 79] Epoch 0, iter 1200/6416, lr 0.100000, loss 15.263649
+INFO 2020-12-04 16:08:42 train.py: 79] Epoch 0, iter 1400/6416, lr 0.100000, loss 15.239414
+INFO 2020-12-04 16:12:16 train.py: 79] Epoch 0, iter 1600/6416, lr 0.100000, loss 15.198864
+INFO 2020-12-04 16:15:50 train.py: 79] Epoch 0, iter 1800/6416, lr 0.100000, loss 15.156965
+INFO 2020-12-04 16:19:23 train.py: 79] Epoch 0, iter 2000/6416, lr 0.100000, loss 15.117281
+INFO 2020-12-04 16:22:57 train.py: 79] Epoch 0, iter 2200/6416, lr 0.100000, loss 15.080080
+INFO 2020-12-04 16:26:31 train.py: 79] Epoch 0, iter 2400/6416, lr 0.100000, loss 15.038412
+INFO 2020-12-04 16:30:04 train.py: 79] Epoch 0, iter 2600/6416, lr 0.100000, loss 14.995138
+INFO 2020-12-04 16:33:38 train.py: 79] Epoch 0, iter 2800/6416, lr 0.100000, loss 14.940116
+INFO 2020-12-04 16:37:11 train.py: 92] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-12-04 16:37:12 train.py: 79] Epoch 0, iter 3000/6416, lr 0.100000, loss 14.870924
+INFO 2020-12-04 16:40:46 train.py: 79] Epoch 0, iter 3200/6416, lr 0.100000, loss 14.767175
+INFO 2020-12-04 16:44:20 train.py: 79] Epoch 0, iter 3400/6416, lr 0.100000, loss 14.647236
+INFO 2020-12-04 16:47:53 train.py: 79] Epoch 0, iter 3600/6416, lr 0.100000, loss 14.528890
+INFO 2020-12-04 16:51:27 train.py: 79] Epoch 0, iter 3800/6416, lr 0.100000, loss 14.379911
+INFO 2020-12-04 16:55:01 train.py: 79] Epoch 0, iter 4000/6416, lr 0.100000, loss 14.205200
+INFO 2020-12-04 16:58:35 train.py: 79] Epoch 0, iter 4200/6416, lr 0.100000, loss 14.002715
+INFO 2020-12-04 17:02:09 train.py: 79] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.768731
+INFO 2020-12-04 17:05:43 train.py: 79] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.555681
+INFO 2020-12-04 17:09:16 train.py: 79] Epoch 0, iter 4800/6416, lr 0.100000, loss 13.300397
+INFO 2020-12-04 17:12:50 train.py: 79] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.058598
+INFO 2020-12-04 17:16:24 train.py: 79] Epoch 0, iter 5200/6416, lr 0.100000, loss 12.808439
+INFO 2020-12-04 17:19:58 train.py: 79] Epoch 0, iter 5400/6416, lr 0.100000, loss 12.547190
+INFO 2020-12-04 17:23:32 train.py: 79] Epoch 0, iter 5600/6416, lr 0.100000, loss 12.235485
+INFO 2020-12-04 17:27:06 train.py: 79] Epoch 0, iter 5800/6416, lr 0.100000, loss 11.904796
+INFO 2020-12-04 17:30:39 train.py: 92] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-12-04 17:30:40 train.py: 79] Epoch 0, iter 6000/6416, lr 0.100000, loss 11.575368
+INFO 2020-12-04 17:34:13 train.py: 79] Epoch 0, iter 6200/6416, lr 0.100000, loss 11.263143
+INFO 2020-12-04 17:37:47 train.py: 79] Epoch 0, iter 6400/6416, lr 0.100000, loss 10.939203
+INFO 2020-12-04 17:38:02 train.py: 97] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-12-04 17:38:05 train.py: 79] Epoch 1, iter 0/6416, lr 0.100000, loss 10.780567
+INFO 2020-12-04 17:41:38 train.py: 79] Epoch 1, iter 200/6416, lr 0.100000, loss 10.553443
+INFO 2020-12-04 17:45:10 train.py: 79] Epoch 1, iter 400/6416, lr 0.100000, loss 10.256111
+INFO 2020-12-04 17:48:42 train.py: 79] Epoch 1, iter 600/6416, lr 0.100000, loss 10.118830
+INFO 2020-12-04 17:52:14 train.py: 79] Epoch 1, iter 800/6416, lr 0.100000, loss 10.001785
+INFO 2020-12-04 17:55:45 train.py: 79] Epoch 1, iter 1000/6416, lr 0.100000, loss 9.989004
+INFO 2020-12-04 17:59:16 train.py: 79] Epoch 1, iter 1200/6416, lr 0.100000, loss 10.098805
+INFO 2020-12-04 18:02:47 train.py: 79] Epoch 1, iter 1400/6416, lr 0.100000, loss 10.293372
+INFO 2020-12-04 18:06:18 train.py: 79] Epoch 1, iter 1600/6416, lr 0.100000, loss 10.591329
+INFO 2020-12-04 18:09:48 train.py: 79] Epoch 1, iter 1800/6416, lr 0.100000, loss 10.944686
+INFO 2020-12-04 18:13:17 train.py: 79] Epoch 1, iter 2000/6416, lr 0.100000, loss 11.360150
+INFO 2020-12-04 18:16:47 train.py: 79] Epoch 1, iter 2200/6416, lr 0.100000, loss 11.743507
+INFO 2020-12-04 18:20:16 train.py: 79] Epoch 1, iter 2400/6416, lr 0.100000, loss 12.141191
+INFO 2020-12-04 18:23:46 train.py: 79] Epoch 1, iter 2600/6416, lr 0.100000, loss 12.527966
+INFO 2020-12-04 18:27:15 train.py: 79] Epoch 1, iter 2800/6416, lr 0.100000, loss 12.834362
+INFO 2020-12-04 18:30:43 train.py: 92] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-12-04 18:30:44 train.py: 79] Epoch 1, iter 3000/6416, lr 0.100000, loss 13.081664
+INFO 2020-12-04 18:34:13 train.py: 79] Epoch 1, iter 3200/6416, lr 0.100000, loss 13.314389
+INFO 2020-12-04 18:37:42 train.py: 79] Epoch 1, iter 3400/6416, lr 0.100000, loss 13.413196
+INFO 2020-12-04 18:41:10 train.py: 79] Epoch 1, iter 3600/6416, lr 0.100000, loss 13.508192
+INFO 2020-12-04 18:44:39 train.py: 79] Epoch 1, iter 3800/6416, lr 0.100000, loss 13.529262
+INFO 2020-12-04 18:48:08 train.py: 79] Epoch 1, iter 4000/6416, lr 0.100000, loss 13.488719
+INFO 2020-12-04 18:51:36 train.py: 79] Epoch 1, iter 4200/6416, lr 0.100000, loss 13.417769
+INFO 2020-12-04 18:55:05 train.py: 79] Epoch 1, iter 4400/6416, lr 0.100000, loss 13.342571
+INFO 2020-12-04 18:58:33 train.py: 79] Epoch 1, iter 4600/6416, lr 0.100000, loss 13.225288
+INFO 2020-12-04 19:02:02 train.py: 79] Epoch 1, iter 4800/6416, lr 0.100000, loss 13.083499
+INFO 2020-12-04 19:05:30 train.py: 79] Epoch 1, iter 5000/6416, lr 0.100000, loss 12.969828
+INFO 2020-12-04 19:08:59 train.py: 79] Epoch 1, iter 5200/6416, lr 0.100000, loss 12.814550
+INFO 2020-12-04 19:12:27 train.py: 79] Epoch 1, iter 5400/6416, lr 0.100000, loss 12.689435
+INFO 2020-12-04 19:15:55 train.py: 79] Epoch 1, iter 5600/6416, lr 0.100000, loss 12.555713
+INFO 2020-12-04 19:19:24 train.py: 79] Epoch 1, iter 5800/6416, lr 0.100000, loss 12.389005
+INFO 2020-12-04 19:22:52 train.py: 92] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-12-04 19:22:53 train.py: 79] Epoch 1, iter 6000/6416, lr 0.100000, loss 12.243248
+INFO 2020-12-04 19:26:22 train.py: 79] Epoch 1, iter 6200/6416, lr 0.100000, loss 12.139172
+INFO 2020-12-04 19:29:51 train.py: 79] Epoch 1, iter 6400/6416, lr 0.100000, loss 11.995660
+INFO 2020-12-04 19:30:06 train.py: 97] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-12-04 19:30:08 train.py: 79] Epoch 2, iter 0/6416, lr 0.100000, loss 11.869083
+INFO 2020-12-04 19:33:37 train.py: 79] Epoch 2, iter 200/6416, lr 0.100000, loss 11.396316
+INFO 2020-12-04 19:37:04 train.py: 79] Epoch 2, iter 400/6416, lr 0.100000, loss 11.301600
+INFO 2020-12-04 19:40:32 train.py: 79] Epoch 2, iter 600/6416, lr 0.100000, loss 11.337017
+INFO 2020-12-04 19:44:00 train.py: 79] Epoch 2, iter 800/6416, lr 0.100000, loss 11.255260
+INFO 2020-12-04 19:47:28 train.py: 79] Epoch 2, iter 1000/6416, lr 0.100000, loss 11.191981
+INFO 2020-12-04 19:50:56 train.py: 79] Epoch 2, iter 1200/6416, lr 0.100000, loss 11.137195
+INFO 2020-12-04 19:54:24 train.py: 79] Epoch 2, iter 1400/6416, lr 0.100000, loss 11.056727
+INFO 2020-12-04 19:57:52 train.py: 79] Epoch 2, iter 1600/6416, lr 0.100000, loss 10.999593
+INFO 2020-12-04 20:01:20 train.py: 79] Epoch 2, iter 1800/6416, lr 0.100000, loss 10.909856
+INFO 2020-12-04 20:04:48 train.py: 79] Epoch 2, iter 2000/6416, lr 0.100000, loss 10.802479
+INFO 2020-12-04 20:08:16 train.py: 79] Epoch 2, iter 2200/6416, lr 0.100000, loss 10.733868
+INFO 2020-12-04 20:11:44 train.py: 79] Epoch 2, iter 2400/6416, lr 0.100000, loss 10.669921
+INFO 2020-12-04 20:15:12 train.py: 79] Epoch 2, iter 2600/6416, lr 0.100000, loss 10.629951
+INFO 2020-12-04 20:18:40 train.py: 79] Epoch 2, iter 2800/6416, lr 0.100000, loss 10.516694
+INFO 2020-12-04 20:22:07 train.py: 92] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-12-04 20:22:08 train.py: 79] Epoch 2, iter 3000/6416, lr 0.100000, loss 10.495982
+INFO 2020-12-04 20:25:36 train.py: 79] Epoch 2, iter 3200/6416, lr 0.100000, loss 10.373624
+INFO 2020-12-04 20:29:04 train.py: 79] Epoch 2, iter 3400/6416, lr 0.100000, loss 10.340557
+INFO 2020-12-04 20:32:33 train.py: 79] Epoch 2, iter 3600/6416, lr 0.100000, loss 10.252735
+INFO 2020-12-04 20:36:01 train.py: 79] Epoch 2, iter 3800/6416, lr 0.100000, loss 10.198890
+INFO 2020-12-04 20:39:29 train.py: 79] Epoch 2, iter 4000/6416, lr 0.100000, loss 10.113380
+INFO 2020-12-04 20:42:57 train.py: 79] Epoch 2, iter 4200/6416, lr 0.100000, loss 10.069223
+INFO 2020-12-04 20:46:26 train.py: 79] Epoch 2, iter 4400/6416, lr 0.100000, loss 9.975065
+INFO 2020-12-04 20:49:54 train.py: 79] Epoch 2, iter 4600/6416, lr 0.100000, loss 9.942636
+INFO 2020-12-04 20:53:22 train.py: 79] Epoch 2, iter 4800/6416, lr 0.100000, loss 9.886392
+INFO 2020-12-04 20:56:51 train.py: 79] Epoch 2, iter 5000/6416, lr 0.100000, loss 9.821362
+INFO 2020-12-04 21:00:19 train.py: 79] Epoch 2, iter 5200/6416, lr 0.100000, loss 9.751015
+INFO 2020-12-04 21:03:48 train.py: 79] Epoch 2, iter 5400/6416, lr 0.100000, loss 9.727385
+INFO 2020-12-04 21:07:16 train.py: 79] Epoch 2, iter 5600/6416, lr 0.100000, loss 9.650240
+INFO 2020-12-04 21:10:45 train.py: 79] Epoch 2, iter 5800/6416, lr 0.100000, loss 9.607512
+INFO 2020-12-04 21:14:13 train.py: 92] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-12-04 21:14:14 train.py: 79] Epoch 2, iter 6000/6416, lr 0.100000, loss 9.560512
+INFO 2020-12-04 21:17:42 train.py: 79] Epoch 2, iter 6200/6416, lr 0.100000, loss 9.496267
+INFO 2020-12-04 21:21:10 train.py: 79] Epoch 2, iter 6400/6416, lr 0.100000, loss 9.444142
+INFO 2020-12-04 21:21:26 train.py: 97] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-12-04 21:21:28 train.py: 79] Epoch 3, iter 0/6416, lr 0.100000, loss 9.351611
+INFO 2020-12-04 21:24:56 train.py: 79] Epoch 3, iter 200/6416, lr 0.100000, loss 8.886291
+INFO 2020-12-04 21:28:23 train.py: 79] Epoch 3, iter 400/6416, lr 0.100000, loss 8.856868
+INFO 2020-12-04 21:31:51 train.py: 79] Epoch 3, iter 600/6416, lr 0.100000, loss 8.932419
+INFO 2020-12-04 21:35:18 train.py: 79] Epoch 3, iter 800/6416, lr 0.100000, loss 8.954494
+INFO 2020-12-04 21:38:46 train.py: 79] Epoch 3, iter 1000/6416, lr 0.100000, loss 9.012721
+INFO 2020-12-04 21:42:13 train.py: 79] Epoch 3, iter 1200/6416, lr 0.100000, loss 9.015286
+INFO 2020-12-04 21:45:41 train.py: 79] Epoch 3, iter 1400/6416, lr 0.100000, loss 8.978294
+INFO 2020-12-04 21:49:08 train.py: 79] Epoch 3, iter 1600/6416, lr 0.100000, loss 8.984096
+INFO 2020-12-04 21:52:36 train.py: 79] Epoch 3, iter 1800/6416, lr 0.100000, loss 8.929338
+INFO 2020-12-04 21:56:03 train.py: 79] Epoch 3, iter 2000/6416, lr 0.100000, loss 8.930364
+INFO 2020-12-04 21:59:31 train.py: 79] Epoch 3, iter 2200/6416, lr 0.100000, loss 8.922982
+INFO 2020-12-04 22:02:59 train.py: 79] Epoch 3, iter 2400/6416, lr 0.100000, loss 8.888956
+INFO 2020-12-04 22:06:26 train.py: 79] Epoch 3, iter 2600/6416, lr 0.100000, loss 8.841022
+INFO 2020-12-04 22:09:54 train.py: 79] Epoch 3, iter 2800/6416, lr 0.100000, loss 8.831702
+INFO 2020-12-04 22:13:21 train.py: 92] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-12-04 22:13:22 train.py: 79] Epoch 3, iter 3000/6416, lr 0.100000, loss 8.784187
+INFO 2020-12-04 22:16:50 train.py: 79] Epoch 3, iter 3200/6416, lr 0.100000, loss 8.809937
+INFO 2020-12-04 22:20:18 train.py: 79] Epoch 3, iter 3400/6416, lr 0.100000, loss 8.764032
+INFO 2020-12-04 22:23:46 train.py: 79] Epoch 3, iter 3600/6416, lr 0.100000, loss 8.720683
+INFO 2020-12-04 22:27:14 train.py: 79] Epoch 3, iter 3800/6416, lr 0.100000, loss 8.690780
+INFO 2020-12-04 22:30:41 train.py: 79] Epoch 3, iter 4000/6416, lr 0.100000, loss 8.663415
+INFO 2020-12-04 22:34:09 train.py: 79] Epoch 3, iter 4200/6416, lr 0.100000, loss 8.598696
+INFO 2020-12-04 22:37:37 train.py: 79] Epoch 3, iter 4400/6416, lr 0.100000, loss 8.581060
+INFO 2020-12-04 22:41:05 train.py: 79] Epoch 3, iter 4600/6416, lr 0.100000, loss 8.598070
+INFO 2020-12-04 22:44:33 train.py: 79] Epoch 3, iter 4800/6416, lr 0.100000, loss 8.551843
+INFO 2020-12-04 22:48:01 train.py: 79] Epoch 3, iter 5000/6416, lr 0.100000, loss 8.496203
+INFO 2020-12-04 22:51:29 train.py: 79] Epoch 3, iter 5200/6416, lr 0.100000, loss 8.500524
+INFO 2020-12-04 22:54:57 train.py: 79] Epoch 3, iter 5400/6416, lr 0.100000, loss 8.467890
+INFO 2020-12-04 22:58:26 train.py: 79] Epoch 3, iter 5600/6416, lr 0.100000, loss 8.427388
+INFO 2020-12-04 23:01:54 train.py: 79] Epoch 3, iter 5800/6416, lr 0.100000, loss 8.392213
+INFO 2020-12-04 23:05:22 train.py: 92] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-12-04 23:05:23 train.py: 79] Epoch 3, iter 6000/6416, lr 0.100000, loss 8.376185
+INFO 2020-12-04 23:08:51 train.py: 79] Epoch 3, iter 6200/6416, lr 0.100000, loss 8.376378
+INFO 2020-12-04 23:12:20 train.py: 79] Epoch 3, iter 6400/6416, lr 0.100000, loss 8.333553
+INFO 2020-12-04 23:12:35 train.py: 97] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-12-04 23:12:37 train.py: 79] Epoch 4, iter 0/6416, lr 0.100000, loss 8.324027
+INFO 2020-12-04 23:16:05 train.py: 79] Epoch 4, iter 200/6416, lr 0.100000, loss 7.799595
+INFO 2020-12-04 23:19:33 train.py: 79] Epoch 4, iter 400/6416, lr 0.100000, loss 7.770641
+INFO 2020-12-04 23:23:01 train.py: 79] Epoch 4, iter 600/6416, lr 0.100000, loss 7.842314
+INFO 2020-12-04 23:26:29 train.py: 79] Epoch 4, iter 800/6416, lr 0.100000, loss 7.927975
+INFO 2020-12-04 23:29:56 train.py: 79] Epoch 4, iter 1000/6416, lr 0.100000, loss 7.971727
+INFO 2020-12-04 23:33:24 train.py: 79] Epoch 4, iter 1200/6416, lr 0.100000, loss 7.991209
+INFO 2020-12-04 23:36:52 train.py: 79] Epoch 4, iter 1400/6416, lr 0.100000, loss 8.028504
+INFO 2020-12-04 23:40:20 train.py: 79] Epoch 4, iter 1600/6416, lr 0.100000, loss 8.006097
+INFO 2020-12-04 23:43:47 train.py: 79] Epoch 4, iter 1800/6416, lr 0.100000, loss 7.999764
+INFO 2020-12-04 23:47:15 train.py: 79] Epoch 4, iter 2000/6416, lr 0.100000, loss 7.988119
+INFO 2020-12-04 23:50:43 train.py: 79] Epoch 4, iter 2200/6416, lr 0.100000, loss 7.981561
+INFO 2020-12-04 23:54:11 train.py: 79] Epoch 4, iter 2400/6416, lr 0.100000, loss 7.943922
+INFO 2020-12-04 23:57:39 train.py: 79] Epoch 4, iter 2600/6416, lr 0.100000, loss 7.992059
+INFO 2020-12-05 00:01:07 train.py: 79] Epoch 4, iter 2800/6416, lr 0.100000, loss 7.957175
+INFO 2020-12-05 00:04:34 train.py: 92] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-12-05 00:04:35 train.py: 79] Epoch 4, iter 3000/6416, lr 0.100000, loss 7.946537
+INFO 2020-12-05 00:08:03 train.py: 79] Epoch 4, iter 3200/6416, lr 0.100000, loss 7.963321
+INFO 2020-12-05 00:11:31 train.py: 79] Epoch 4, iter 3400/6416, lr 0.100000, loss 7.934362
+INFO 2020-12-05 00:14:59 train.py: 79] Epoch 4, iter 3600/6416, lr 0.100000, loss 7.910169
+INFO 2020-12-05 00:18:27 train.py: 79] Epoch 4, iter 3800/6416, lr 0.100000, loss 7.893520
+INFO 2020-12-05 00:21:55 train.py: 79] Epoch 4, iter 4000/6416, lr 0.100000, loss 7.861820
+INFO 2020-12-05 00:25:24 train.py: 79] Epoch 4, iter 4200/6416, lr 0.100000, loss 7.854077
+INFO 2020-12-05 00:28:52 train.py: 79] Epoch 4, iter 4400/6416, lr 0.100000, loss 7.842075
+INFO 2020-12-05 00:32:20 train.py: 79] Epoch 4, iter 4600/6416, lr 0.100000, loss 7.816697
+INFO 2020-12-05 00:35:48 train.py: 79] Epoch 4, iter 4800/6416, lr 0.100000, loss 7.801469
+INFO 2020-12-05 00:39:17 train.py: 79] Epoch 4, iter 5000/6416, lr 0.100000, loss 7.793809
+INFO 2020-12-05 00:42:45 train.py: 79] Epoch 4, iter 5200/6416, lr 0.100000, loss 7.778498
+INFO 2020-12-05 00:46:13 train.py: 79] Epoch 4, iter 5400/6416, lr 0.100000, loss 7.765427
+INFO 2020-12-05 00:49:42 train.py: 79] Epoch 4, iter 5600/6416, lr 0.100000, loss 7.735386
+INFO 2020-12-05 00:53:10 train.py: 79] Epoch 4, iter 5800/6416, lr 0.100000, loss 7.705831
+INFO 2020-12-05 00:56:38 train.py: 92] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-12-05 00:56:39 train.py: 79] Epoch 4, iter 6000/6416, lr 0.100000, loss 7.682049
+INFO 2020-12-05 01:00:08 train.py: 79] Epoch 4, iter 6200/6416, lr 0.100000, loss 7.694603
+INFO 2020-12-05 01:03:36 train.py: 79] Epoch 4, iter 6400/6416, lr 0.100000, loss 7.685364
+INFO 2020-12-05 01:03:52 train.py: 97] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-12-05 01:03:54 train.py: 79] Epoch 5, iter 0/6416, lr 0.100000, loss 7.716165
+INFO 2020-12-05 01:07:22 train.py: 79] Epoch 5, iter 200/6416, lr 0.100000, loss 7.153336
+INFO 2020-12-05 01:10:50 train.py: 79] Epoch 5, iter 400/6416, lr 0.100000, loss 7.126025
+INFO 2020-12-05 01:14:17 train.py: 79] Epoch 5, iter 600/6416, lr 0.100000, loss 7.225863
+INFO 2020-12-05 01:17:45 train.py: 79] Epoch 5, iter 800/6416, lr 0.100000, loss 7.277958
+INFO 2020-12-05 01:21:13 train.py: 79] Epoch 5, iter 1000/6416, lr 0.100000, loss 7.334002
+INFO 2020-12-05 01:24:40 train.py: 79] Epoch 5, iter 1200/6416, lr 0.100000, loss 7.341764
+INFO 2020-12-05 01:28:08 train.py: 79] Epoch 5, iter 1400/6416, lr 0.100000, loss 7.411059
+INFO 2020-12-05 01:31:36 train.py: 79] Epoch 5, iter 1600/6416, lr 0.100000, loss 7.412400
+INFO 2020-12-05 01:35:04 train.py: 79] Epoch 5, iter 1800/6416, lr 0.100000, loss 7.455484
+INFO 2020-12-05 01:38:32 train.py: 79] Epoch 5, iter 2000/6416, lr 0.100000, loss 7.427902
+INFO 2020-12-05 01:41:59 train.py: 79] Epoch 5, iter 2200/6416, lr 0.100000, loss 7.423447
+INFO 2020-12-05 01:45:27 train.py: 79] Epoch 5, iter 2400/6416, lr 0.100000, loss 7.421161
+INFO 2020-12-05 01:48:55 train.py: 79] Epoch 5, iter 2600/6416, lr 0.100000, loss 7.422858
+INFO 2020-12-05 01:52:23 train.py: 79] Epoch 5, iter 2800/6416, lr 0.100000, loss 7.421201
+INFO 2020-12-05 01:55:50 train.py: 92] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-12-05 01:55:51 train.py: 79] Epoch 5, iter 3000/6416, lr 0.100000, loss 7.422262
+INFO 2020-12-05 01:59:20 train.py: 79] Epoch 5, iter 3200/6416, lr 0.100000, loss 7.413162
+INFO 2020-12-05 02:02:48 train.py: 79] Epoch 5, iter 3400/6416, lr 0.100000, loss 7.389495
+INFO 2020-12-05 02:06:16 train.py: 79] Epoch 5, iter 3600/6416, lr 0.100000, loss 7.362105
+INFO 2020-12-05 02:09:44 train.py: 79] Epoch 5, iter 3800/6416, lr 0.100000, loss 7.405298
+INFO 2020-12-05 02:13:12 train.py: 79] Epoch 5, iter 4000/6416, lr 0.100000, loss 7.404870
+INFO 2020-12-05 02:16:40 train.py: 79] Epoch 5, iter 4200/6416, lr 0.100000, loss 7.356420
+INFO 2020-12-05 02:20:08 train.py: 79] Epoch 5, iter 4400/6416, lr 0.100000, loss 7.366529
+INFO 2020-12-05 02:23:36 train.py: 79] Epoch 5, iter 4600/6416, lr 0.100000, loss 7.322106
+INFO 2020-12-05 02:27:05 train.py: 79] Epoch 5, iter 4800/6416, lr 0.100000, loss 7.370430
+INFO 2020-12-05 02:30:33 train.py: 79] Epoch 5, iter 5000/6416, lr 0.100000, loss 7.328980
+INFO 2020-12-05 02:34:01 train.py: 79] Epoch 5, iter 5200/6416, lr 0.100000, loss 7.298283
+INFO 2020-12-05 02:37:30 train.py: 79] Epoch 5, iter 5400/6416, lr 0.100000, loss 7.266570
+INFO 2020-12-05 02:40:58 train.py: 79] Epoch 5, iter 5600/6416, lr 0.100000, loss 7.309567
+INFO 2020-12-05 02:44:26 train.py: 79] Epoch 5, iter 5800/6416, lr 0.100000, loss 7.280971
+INFO 2020-12-05 02:47:54 train.py: 92] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-12-05 02:47:55 train.py: 79] Epoch 5, iter 6000/6416, lr 0.100000, loss 7.274906
+INFO 2020-12-05 02:51:23 train.py: 79] Epoch 5, iter 6200/6416, lr 0.100000, loss 7.251253
+INFO 2020-12-05 02:54:51 train.py: 79] Epoch 5, iter 6400/6416, lr 0.100000, loss 7.247135
+INFO 2020-12-05 02:55:06 train.py: 97] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-12-05 02:55:09 train.py: 79] Epoch 6, iter 0/6416, lr 0.100000, loss 7.152089
+INFO 2020-12-05 02:58:36 train.py: 79] Epoch 6, iter 200/6416, lr 0.100000, loss 6.677138
+INFO 2020-12-05 03:02:04 train.py: 79] Epoch 6, iter 400/6416, lr 0.100000, loss 6.716689
+INFO 2020-12-05 03:05:31 train.py: 79] Epoch 6, iter 600/6416, lr 0.100000, loss 6.827116
+INFO 2020-12-05 03:08:59 train.py: 79] Epoch 6, iter 800/6416, lr 0.100000, loss 6.880670
+INFO 2020-12-05 03:12:26 train.py: 79] Epoch 6, iter 1000/6416, lr 0.100000, loss 6.925452
+INFO 2020-12-05 03:15:54 train.py: 79] Epoch 6, iter 1200/6416, lr 0.100000, loss 6.990694
+INFO 2020-12-05 03:19:21 train.py: 79] Epoch 6, iter 1400/6416, lr 0.100000, loss 6.995743
+INFO 2020-12-05 03:22:49 train.py: 79] Epoch 6, iter 1600/6416, lr 0.100000, loss 7.024012
+INFO 2020-12-05 03:26:16 train.py: 79] Epoch 6, iter 1800/6416, lr 0.100000, loss 7.045314
+INFO 2020-12-05 03:29:44 train.py: 79] Epoch 6, iter 2000/6416, lr 0.100000, loss 7.062760
+INFO 2020-12-05 03:33:11 train.py: 79] Epoch 6, iter 2200/6416, lr 0.100000, loss 7.032008
+INFO 2020-12-05 03:36:39 train.py: 79] Epoch 6, iter 2400/6416, lr 0.100000, loss 7.061680
+INFO 2020-12-05 03:40:07 train.py: 79] Epoch 6, iter 2600/6416, lr 0.100000, loss 7.048713
+INFO 2020-12-05 03:43:34 train.py: 79] Epoch 6, iter 2800/6416, lr 0.100000, loss 7.054901
+INFO 2020-12-05 03:47:01 train.py: 92] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-12-05 03:47:02 train.py: 79] Epoch 6, iter 3000/6416, lr 0.100000, loss 7.054258
+INFO 2020-12-05 03:50:30 train.py: 79] Epoch 6, iter 3200/6416, lr 0.100000, loss 7.048441
+INFO 2020-12-05 03:53:58 train.py: 79] Epoch 6, iter 3400/6416, lr 0.100000, loss 7.035745
+INFO 2020-12-05 03:57:27 train.py: 79] Epoch 6, iter 3600/6416, lr 0.100000, loss 7.040420
+INFO 2020-12-05 04:00:55 train.py: 79] Epoch 6, iter 3800/6416, lr 0.100000, loss 7.047064
+INFO 2020-12-05 04:04:23 train.py: 79] Epoch 6, iter 4000/6416, lr 0.100000, loss 7.030522
+INFO 2020-12-05 04:07:51 train.py: 79] Epoch 6, iter 4200/6416, lr 0.100000, loss 7.000274
+INFO 2020-12-05 04:11:20 train.py: 79] Epoch 6, iter 4400/6416, lr 0.100000, loss 6.978023
+INFO 2020-12-05 04:14:48 train.py: 79] Epoch 6, iter 4600/6416, lr 0.100000, loss 7.004088
+INFO 2020-12-05 04:18:16 train.py: 79] Epoch 6, iter 4800/6416, lr 0.100000, loss 6.994035
+INFO 2020-12-05 04:21:44 train.py: 79] Epoch 6, iter 5000/6416, lr 0.100000, loss 6.966543
+INFO 2020-12-05 04:25:13 train.py: 79] Epoch 6, iter 5200/6416, lr 0.100000, loss 6.992030
+INFO 2020-12-05 04:28:41 train.py: 79] Epoch 6, iter 5400/6416, lr 0.100000, loss 6.966049
+INFO 2020-12-05 04:32:09 train.py: 79] Epoch 6, iter 5600/6416, lr 0.100000, loss 6.966196
+INFO 2020-12-05 04:35:38 train.py: 79] Epoch 6, iter 5800/6416, lr 0.100000, loss 6.980391
+INFO 2020-12-05 04:39:05 train.py: 92] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-12-05 04:39:06 train.py: 79] Epoch 6, iter 6000/6416, lr 0.100000, loss 6.939235
+INFO 2020-12-05 04:42:34 train.py: 79] Epoch 6, iter 6200/6416, lr 0.100000, loss 6.932310
+INFO 2020-12-05 04:46:02 train.py: 79] Epoch 6, iter 6400/6416, lr 0.100000, loss 6.920093
+INFO 2020-12-05 04:46:18 train.py: 97] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-12-05 04:46:20 train.py: 79] Epoch 7, iter 0/6416, lr 0.100000, loss 6.918180
+INFO 2020-12-05 04:49:48 train.py: 79] Epoch 7, iter 200/6416, lr 0.100000, loss 6.396653
+INFO 2020-12-05 04:53:15 train.py: 79] Epoch 7, iter 400/6416, lr 0.100000, loss 6.413509
+INFO 2020-12-05 04:56:43 train.py: 79] Epoch 7, iter 600/6416, lr 0.100000, loss 6.514897
+INFO 2020-12-05 05:00:10 train.py: 79] Epoch 7, iter 800/6416, lr 0.100000, loss 6.566967
+INFO 2020-12-05 05:03:38 train.py: 79] Epoch 7, iter 1000/6416, lr 0.100000, loss 6.618975
+INFO 2020-12-05 05:07:05 train.py: 79] Epoch 7, iter 1200/6416, lr 0.100000, loss 6.687779
+INFO 2020-12-05 05:10:33 train.py: 79] Epoch 7, iter 1400/6416, lr 0.100000, loss 6.724841
+INFO 2020-12-05 05:14:00 train.py: 79] Epoch 7, iter 1600/6416, lr 0.100000, loss 6.741530
+INFO 2020-12-05 05:17:28 train.py: 79] Epoch 7, iter 1800/6416, lr 0.100000, loss 6.729090
+INFO 2020-12-05 05:20:55 train.py: 79] Epoch 7, iter 2000/6416, lr 0.100000, loss 6.782476
+INFO 2020-12-05 05:24:23 train.py: 79] Epoch 7, iter 2200/6416, lr 0.100000, loss 6.763539
+INFO 2020-12-05 05:27:50 train.py: 79] Epoch 7, iter 2400/6416, lr 0.100000, loss 6.769798
+INFO 2020-12-05 05:31:18 train.py: 79] Epoch 7, iter 2600/6416, lr 0.100000, loss 6.760862
+INFO 2020-12-05 05:34:45 train.py: 79] Epoch 7, iter 2800/6416, lr 0.100000, loss 6.796310
+INFO 2020-12-05 05:38:12 train.py: 92] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-12-05 05:38:13 train.py: 79] Epoch 7, iter 3000/6416, lr 0.100000, loss 6.765018
+INFO 2020-12-05 05:41:41 train.py: 79] Epoch 7, iter 3200/6416, lr 0.100000, loss 6.790659
+INFO 2020-12-05 05:45:09 train.py: 79] Epoch 7, iter 3400/6416, lr 0.100000, loss 6.779472
+INFO 2020-12-05 05:48:37 train.py: 79] Epoch 7, iter 3600/6416, lr 0.100000, loss 6.754166
+INFO 2020-12-05 05:52:04 train.py: 79] Epoch 7, iter 3800/6416, lr 0.100000, loss 6.741950
+INFO 2020-12-05 05:55:32 train.py: 79] Epoch 7, iter 4000/6416, lr 0.100000, loss 6.796215
+INFO 2020-12-05 05:59:00 train.py: 79] Epoch 7, iter 4200/6416, lr 0.100000, loss 6.730734
+INFO 2020-12-05 06:02:28 train.py: 79] Epoch 7, iter 4400/6416, lr 0.100000, loss 6.720681
+INFO 2020-12-05 06:05:56 train.py: 79] Epoch 7, iter 4600/6416, lr 0.100000, loss 6.737967
+INFO 2020-12-05 06:09:24 train.py: 79] Epoch 7, iter 4800/6416, lr 0.100000, loss 6.723388
+INFO 2020-12-05 06:12:52 train.py: 79] Epoch 7, iter 5000/6416, lr 0.100000, loss 6.755905
+INFO 2020-12-05 06:16:20 train.py: 79] Epoch 7, iter 5200/6416, lr 0.100000, loss 6.756323
+INFO 2020-12-05 06:19:48 train.py: 79] Epoch 7, iter 5400/6416, lr 0.100000, loss 6.730527
+INFO 2020-12-05 06:23:16 train.py: 79] Epoch 7, iter 5600/6416, lr 0.100000, loss 6.712811
+INFO 2020-12-05 06:26:44 train.py: 79] Epoch 7, iter 5800/6416, lr 0.100000, loss 6.676762
+INFO 2020-12-05 06:30:12 train.py: 92] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-12-05 06:30:13 train.py: 79] Epoch 7, iter 6000/6416, lr 0.100000, loss 6.709044
+INFO 2020-12-05 06:33:42 train.py: 79] Epoch 7, iter 6200/6416, lr 0.100000, loss 6.735850
+INFO 2020-12-05 06:37:10 train.py: 79] Epoch 7, iter 6400/6416, lr 0.100000, loss 6.704747
+INFO 2020-12-05 06:37:26 train.py: 97] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-12-05 06:37:28 train.py: 79] Epoch 8, iter 0/6416, lr 0.100000, loss 6.655569
+INFO 2020-12-05 06:40:56 train.py: 79] Epoch 8, iter 200/6416, lr 0.100000, loss 6.200425
+INFO 2020-12-05 06:44:23 train.py: 79] Epoch 8, iter 400/6416, lr 0.100000, loss 6.194023
+INFO 2020-12-05 06:47:50 train.py: 79] Epoch 8, iter 600/6416, lr 0.100000, loss 6.280918
+INFO 2020-12-05 06:51:18 train.py: 79] Epoch 8, iter 800/6416, lr 0.100000, loss 6.356980
+INFO 2020-12-05 06:54:45 train.py: 79] Epoch 8, iter 1000/6416, lr 0.100000, loss 6.396135
+INFO 2020-12-05 06:58:13 train.py: 79] Epoch 8, iter 1200/6416, lr 0.100000, loss 6.428632
+INFO 2020-12-05 07:01:40 train.py: 79] Epoch 8, iter 1400/6416, lr 0.100000, loss 6.503153
+INFO 2020-12-05 07:05:08 train.py: 79] Epoch 8, iter 1600/6416, lr 0.100000, loss 6.500470
+INFO 2020-12-05 07:08:35 train.py: 79] Epoch 8, iter 1800/6416, lr 0.100000, loss 6.529297
+INFO 2020-12-05 07:12:03 train.py: 79] Epoch 8, iter 2000/6416, lr 0.100000, loss 6.530378
+INFO 2020-12-05 07:15:30 train.py: 79] Epoch 8, iter 2200/6416, lr 0.100000, loss 6.522300
+INFO 2020-12-05 07:18:58 train.py: 79] Epoch 8, iter 2400/6416, lr 0.100000, loss 6.557148
+INFO 2020-12-05 07:22:25 train.py: 79] Epoch 8, iter 2600/6416, lr 0.100000, loss 6.568744
+INFO 2020-12-05 07:25:53 train.py: 79] Epoch 8, iter 2800/6416, lr 0.100000, loss 6.559994
+INFO 2020-12-05 07:29:20 train.py: 92] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-12-05 07:29:21 train.py: 79] Epoch 8, iter 3000/6416, lr 0.100000, loss 6.589960
+INFO 2020-12-05 07:32:48 train.py: 79] Epoch 8, iter 3200/6416, lr 0.100000, loss 6.560607
+INFO 2020-12-05 07:36:16 train.py: 79] Epoch 8, iter 3400/6416, lr 0.100000, loss 6.576931
+INFO 2020-12-05 07:39:44 train.py: 79] Epoch 8, iter 3600/6416, lr 0.100000, loss 6.571948
+INFO 2020-12-05 07:43:12 train.py: 79] Epoch 8, iter 3800/6416, lr 0.100000, loss 6.568509
+INFO 2020-12-05 07:46:39 train.py: 79] Epoch 8, iter 4000/6416, lr 0.100000, loss 6.534901
+INFO 2020-12-05 07:50:07 train.py: 79] Epoch 8, iter 4200/6416, lr 0.100000, loss 6.555075
+INFO 2020-12-05 07:53:35 train.py: 79] Epoch 8, iter 4400/6416, lr 0.100000, loss 6.562300
+INFO 2020-12-05 07:57:03 train.py: 79] Epoch 8, iter 4600/6416, lr 0.100000, loss 6.534326
+INFO 2020-12-05 08:00:31 train.py: 79] Epoch 8, iter 4800/6416, lr 0.100000, loss 6.543559
+INFO 2020-12-05 08:03:59 train.py: 79] Epoch 8, iter 5000/6416, lr 0.100000, loss 6.543015
+INFO 2020-12-05 08:07:27 train.py: 79] Epoch 8, iter 5200/6416, lr 0.100000, loss 6.498902
+INFO 2020-12-05 08:10:55 train.py: 79] Epoch 8, iter 5400/6416, lr 0.100000, loss 6.530536
+INFO 2020-12-05 08:14:23 train.py: 79] Epoch 8, iter 5600/6416, lr 0.100000, loss 6.513833
+INFO 2020-12-05 08:17:51 train.py: 79] Epoch 8, iter 5800/6416, lr 0.100000, loss 6.507943
+INFO 2020-12-05 08:21:19 train.py: 92] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-12-05 08:21:20 train.py: 79] Epoch 8, iter 6000/6416, lr 0.100000, loss 6.484468
+INFO 2020-12-05 08:24:48 train.py: 79] Epoch 8, iter 6200/6416, lr 0.100000, loss 6.504925
+INFO 2020-12-05 08:28:17 train.py: 79] Epoch 8, iter 6400/6416, lr 0.100000, loss 6.526360
+INFO 2020-12-05 08:28:32 train.py: 97] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-12-05 08:28:35 train.py: 79] Epoch 9, iter 0/6416, lr 0.100000, loss 6.546625
+INFO 2020-12-05 08:32:03 train.py: 79] Epoch 9, iter 200/6416, lr 0.100000, loss 5.988418
+INFO 2020-12-05 08:35:30 train.py: 79] Epoch 9, iter 400/6416, lr 0.100000, loss 6.013575
+INFO 2020-12-05 08:38:58 train.py: 79] Epoch 9, iter 600/6416, lr 0.100000, loss 6.090373
+INFO 2020-12-05 08:42:26 train.py: 79] Epoch 9, iter 800/6416, lr 0.100000, loss 6.155948
+INFO 2020-12-05 08:45:53 train.py: 79] Epoch 9, iter 1000/6416, lr 0.100000, loss 6.204142
+INFO 2020-12-05 08:49:21 train.py: 79] Epoch 9, iter 1200/6416, lr 0.100000, loss 6.273956
+INFO 2020-12-05 08:52:49 train.py: 79] Epoch 9, iter 1400/6416, lr 0.100000, loss 6.308727
+INFO 2020-12-05 08:56:17 train.py: 79] Epoch 9, iter 1600/6416, lr 0.100000, loss 6.318594
+INFO 2020-12-05 08:59:44 train.py: 79] Epoch 9, iter 1800/6416, lr 0.100000, loss 6.349388
+INFO 2020-12-05 09:03:12 train.py: 79] Epoch 9, iter 2000/6416, lr 0.100000, loss 6.348874
+INFO 2020-12-05 09:06:40 train.py: 79] Epoch 9, iter 2200/6416, lr 0.100000, loss 6.340981
+INFO 2020-12-05 09:10:08 train.py: 79] Epoch 9, iter 2400/6416, lr 0.100000, loss 6.333893
+INFO 2020-12-05 09:13:36 train.py: 79] Epoch 9, iter 2600/6416, lr 0.100000, loss 6.407534
+INFO 2020-12-05 09:17:04 train.py: 79] Epoch 9, iter 2800/6416, lr 0.100000, loss 6.404898
+INFO 2020-12-05 09:20:31 train.py: 92] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-12-05 09:20:32 train.py: 79] Epoch 9, iter 3000/6416, lr 0.100000, loss 6.430646
+INFO 2020-12-05 09:24:00 train.py: 79] Epoch 9, iter 3200/6416, lr 0.100000, loss 6.422098
+INFO 2020-12-05 09:27:28 train.py: 79] Epoch 9, iter 3400/6416, lr 0.100000, loss 6.392207
+INFO 2020-12-05 09:30:56 train.py: 79] Epoch 9, iter 3600/6416, lr 0.100000, loss 6.416785
+INFO 2020-12-05 09:34:24 train.py: 79] Epoch 9, iter 3800/6416, lr 0.100000, loss 6.398945
+INFO 2020-12-05 09:37:52 train.py: 79] Epoch 9, iter 4000/6416, lr 0.100000, loss 6.388043
+INFO 2020-12-05 09:41:20 train.py: 79] Epoch 9, iter 4200/6416, lr 0.100000, loss 6.380904
+INFO 2020-12-05 09:44:49 train.py: 79] Epoch 9, iter 4400/6416, lr 0.100000, loss 6.393300
+INFO 2020-12-05 09:48:17 train.py: 79] Epoch 9, iter 4600/6416, lr 0.100000, loss 6.420892
+INFO 2020-12-05 09:51:45 train.py: 79] Epoch 9, iter 4800/6416, lr 0.100000, loss 6.401572
+INFO 2020-12-05 09:55:13 train.py: 79] Epoch 9, iter 5000/6416, lr 0.100000, loss 6.367910
+INFO 2020-12-05 09:58:42 train.py: 79] Epoch 9, iter 5200/6416, lr 0.100000, loss 6.354099
+INFO 2020-12-05 10:02:10 train.py: 79] Epoch 9, iter 5400/6416, lr 0.100000, loss 6.391977
+INFO 2020-12-05 10:05:38 train.py: 79] Epoch 9, iter 5600/6416, lr 0.100000, loss 6.369676
+INFO 2020-12-05 10:09:07 train.py: 79] Epoch 9, iter 5800/6416, lr 0.100000, loss 6.352410
+INFO 2020-12-05 10:12:34 train.py: 92] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-12-05 10:12:35 train.py: 79] Epoch 9, iter 6000/6416, lr 0.100000, loss 6.361461
+INFO 2020-12-05 10:16:04 train.py: 79] Epoch 9, iter 6200/6416, lr 0.100000, loss 6.336442
+INFO 2020-12-05 10:19:32 train.py: 79] Epoch 9, iter 6400/6416, lr 0.100000, loss 6.363317
+INFO 2020-12-05 10:19:48 train.py: 97] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-12-05 10:19:50 train.py: 79] Epoch 10, iter 0/6416, lr 0.010000, loss 6.385041
+INFO 2020-12-05 10:23:17 train.py: 79] Epoch 10, iter 200/6416, lr 0.010000, loss 5.172713
+INFO 2020-12-05 10:26:45 train.py: 79] Epoch 10, iter 400/6416, lr 0.010000, loss 4.925981
+INFO 2020-12-05 10:30:12 train.py: 79] Epoch 10, iter 600/6416, lr 0.010000, loss 4.823651
+INFO 2020-12-05 10:33:40 train.py: 79] Epoch 10, iter 800/6416, lr 0.010000, loss 4.755726
+INFO 2020-12-05 10:37:07 train.py: 79] Epoch 10, iter 1000/6416, lr 0.010000, loss 4.711808
+INFO 2020-12-05 10:40:34 train.py: 79] Epoch 10, iter 1200/6416, lr 0.010000, loss 4.656096
+INFO 2020-12-05 10:44:02 train.py: 79] Epoch 10, iter 1400/6416, lr 0.010000, loss 4.620695
+INFO 2020-12-05 10:47:29 train.py: 79] Epoch 10, iter 1600/6416, lr 0.010000, loss 4.570408
+INFO 2020-12-05 10:50:56 train.py: 79] Epoch 10, iter 1800/6416, lr 0.010000, loss 4.555992
+INFO 2020-12-05 10:54:24 train.py: 79] Epoch 10, iter 2000/6416, lr 0.010000, loss 4.514627
+INFO 2020-12-05 10:57:51 train.py: 79] Epoch 10, iter 2200/6416, lr 0.010000, loss 4.499152
+INFO 2020-12-05 11:01:18 train.py: 79] Epoch 10, iter 2400/6416, lr 0.010000, loss 4.501562
+INFO 2020-12-05 11:04:46 train.py: 79] Epoch 10, iter 2600/6416, lr 0.010000, loss 4.447203
+INFO 2020-12-05 11:08:13 train.py: 79] Epoch 10, iter 2800/6416, lr 0.010000, loss 4.449026
+INFO 2020-12-05 11:11:40 train.py: 92] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-12-05 11:11:41 train.py: 79] Epoch 10, iter 3000/6416, lr 0.010000, loss 4.419977
+INFO 2020-12-05 11:15:09 train.py: 79] Epoch 10, iter 3200/6416, lr 0.010000, loss 4.400013
+INFO 2020-12-05 11:18:36 train.py: 79] Epoch 10, iter 3400/6416, lr 0.010000, loss 4.355941
+INFO 2020-12-05 11:22:04 train.py: 79] Epoch 10, iter 3600/6416, lr 0.010000, loss 4.361823
+INFO 2020-12-05 11:25:32 train.py: 79] Epoch 10, iter 3800/6416, lr 0.010000, loss 4.329315
+INFO 2020-12-05 11:28:59 train.py: 79] Epoch 10, iter 4000/6416, lr 0.010000, loss 4.302991
+INFO 2020-12-05 11:32:27 train.py: 79] Epoch 10, iter 4200/6416, lr 0.010000, loss 4.288169
+INFO 2020-12-05 11:35:55 train.py: 79] Epoch 10, iter 4400/6416, lr 0.010000, loss 4.288176
+INFO 2020-12-05 11:39:23 train.py: 79] Epoch 10, iter 4600/6416, lr 0.010000, loss 4.257224
+INFO 2020-12-05 11:42:51 train.py: 79] Epoch 10, iter 4800/6416, lr 0.010000, loss 4.265672
+INFO 2020-12-05 11:46:18 train.py: 79] Epoch 10, iter 5000/6416, lr 0.010000, loss 4.237838
+INFO 2020-12-05 11:49:46 train.py: 79] Epoch 10, iter 5200/6416, lr 0.010000, loss 4.239720
+INFO 2020-12-05 11:53:14 train.py: 79] Epoch 10, iter 5400/6416, lr 0.010000, loss 4.222967
+INFO 2020-12-05 11:56:42 train.py: 79] Epoch 10, iter 5600/6416, lr 0.010000, loss 4.225243
+INFO 2020-12-05 12:00:10 train.py: 79] Epoch 10, iter 5800/6416, lr 0.010000, loss 4.196087
+INFO 2020-12-05 12:03:37 train.py: 92] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-12-05 12:03:38 train.py: 79] Epoch 10, iter 6000/6416, lr 0.010000, loss 4.195523
+INFO 2020-12-05 12:07:07 train.py: 79] Epoch 10, iter 6200/6416, lr 0.010000, loss 4.190834
+INFO 2020-12-05 12:10:35 train.py: 79] Epoch 10, iter 6400/6416, lr 0.010000, loss 4.175358
+INFO 2020-12-05 12:10:51 train.py: 97] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-12-05 12:10:53 train.py: 79] Epoch 11, iter 0/6416, lr 0.010000, loss 4.148418
+INFO 2020-12-05 12:14:21 train.py: 79] Epoch 11, iter 200/6416, lr 0.010000, loss 3.855050
+INFO 2020-12-05 12:17:48 train.py: 79] Epoch 11, iter 400/6416, lr 0.010000, loss 3.864178
+INFO 2020-12-05 12:21:16 train.py: 79] Epoch 11, iter 600/6416, lr 0.010000, loss 3.853740
+INFO 2020-12-05 12:24:44 train.py: 79] Epoch 11, iter 800/6416, lr 0.010000, loss 3.849989
+INFO 2020-12-05 12:28:11 train.py: 79] Epoch 11, iter 1000/6416, lr 0.010000, loss 3.855244
+INFO 2020-12-05 12:31:39 train.py: 79] Epoch 11, iter 1200/6416, lr 0.010000, loss 3.837125
+INFO 2020-12-05 12:35:06 train.py: 79] Epoch 11, iter 1400/6416, lr 0.010000, loss 3.864083
+INFO 2020-12-05 12:38:34 train.py: 79] Epoch 11, iter 1600/6416, lr 0.010000, loss 3.834262
+INFO 2020-12-05 12:42:02 train.py: 79] Epoch 11, iter 1800/6416, lr 0.010000, loss 3.877466
+INFO 2020-12-05 12:45:29 train.py: 79] Epoch 11, iter 2000/6416, lr 0.010000, loss 3.848900
+INFO 2020-12-05 12:48:57 train.py: 79] Epoch 11, iter 2200/6416, lr 0.010000, loss 3.851610
+INFO 2020-12-05 12:52:25 train.py: 79] Epoch 11, iter 2400/6416, lr 0.010000, loss 3.841831
+INFO 2020-12-05 12:55:53 train.py: 79] Epoch 11, iter 2600/6416, lr 0.010000, loss 3.858794
+INFO 2020-12-05 12:59:20 train.py: 79] Epoch 11, iter 2800/6416, lr 0.010000, loss 3.857770
+INFO 2020-12-05 13:02:47 train.py: 92] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-12-05 13:02:48 train.py: 79] Epoch 11, iter 3000/6416, lr 0.010000, loss 3.851545
+INFO 2020-12-05 13:06:16 train.py: 79] Epoch 11, iter 3200/6416, lr 0.010000, loss 3.849674
+INFO 2020-12-05 13:09:44 train.py: 79] Epoch 11, iter 3400/6416, lr 0.010000, loss 3.872307
+INFO 2020-12-05 13:13:12 train.py: 79] Epoch 11, iter 3600/6416, lr 0.010000, loss 3.860439
+INFO 2020-12-05 13:16:40 train.py: 79] Epoch 11, iter 3800/6416, lr 0.010000, loss 3.881402
+INFO 2020-12-05 13:20:08 train.py: 79] Epoch 11, iter 4000/6416, lr 0.010000, loss 3.852261
+INFO 2020-12-05 13:23:36 train.py: 79] Epoch 11, iter 4200/6416, lr 0.010000, loss 3.874088
+INFO 2020-12-05 13:27:04 train.py: 79] Epoch 11, iter 4400/6416, lr 0.010000, loss 3.888295
+INFO 2020-12-05 13:30:32 train.py: 79] Epoch 11, iter 4600/6416, lr 0.010000, loss 3.883151
+INFO 2020-12-05 13:34:00 train.py: 79] Epoch 11, iter 4800/6416, lr 0.010000, loss 3.837264
+INFO 2020-12-05 13:37:29 train.py: 79] Epoch 11, iter 5000/6416, lr 0.010000, loss 3.858073
+INFO 2020-12-05 13:40:57 train.py: 79] Epoch 11, iter 5200/6416, lr 0.010000, loss 3.867437
+INFO 2020-12-05 13:44:25 train.py: 79] Epoch 11, iter 5400/6416, lr 0.010000, loss 3.870497
+INFO 2020-12-05 13:47:53 train.py: 79] Epoch 11, iter 5600/6416, lr 0.010000, loss 3.880058
+INFO 2020-12-05 13:51:21 train.py: 79] Epoch 11, iter 5800/6416, lr 0.010000, loss 3.858043
+INFO 2020-12-05 13:54:49 train.py: 92] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-12-05 13:54:50 train.py: 79] Epoch 11, iter 6000/6416, lr 0.010000, loss 3.861370
+INFO 2020-12-05 13:58:18 train.py: 79] Epoch 11, iter 6200/6416, lr 0.010000, loss 3.863755
+INFO 2020-12-05 14:01:46 train.py: 79] Epoch 11, iter 6400/6416, lr 0.010000, loss 3.876178
+INFO 2020-12-05 14:02:01 train.py: 97] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-12-05 14:02:04 train.py: 79] Epoch 12, iter 0/6416, lr 0.010000, loss 3.795015
+INFO 2020-12-05 14:05:31 train.py: 79] Epoch 12, iter 200/6416, lr 0.010000, loss 3.565641
+INFO 2020-12-05 14:08:59 train.py: 79] Epoch 12, iter 400/6416, lr 0.010000, loss 3.555529
+INFO 2020-12-05 14:12:27 train.py: 79] Epoch 12, iter 600/6416, lr 0.010000, loss 3.565018
+INFO 2020-12-05 14:15:55 train.py: 79] Epoch 12, iter 800/6416, lr 0.010000, loss 3.605192
+INFO 2020-12-05 14:19:22 train.py: 79] Epoch 12, iter 1000/6416, lr 0.010000, loss 3.586902
+INFO 2020-12-05 14:22:50 train.py: 79] Epoch 12, iter 1200/6416, lr 0.010000, loss 3.585633
+INFO 2020-12-05 14:26:17 train.py: 79] Epoch 12, iter 1400/6416, lr 0.010000, loss 3.617846
+INFO 2020-12-05 14:29:45 train.py: 79] Epoch 12, iter 1600/6416, lr 0.010000, loss 3.610591
+INFO 2020-12-05 14:33:12 train.py: 79] Epoch 12, iter 1800/6416, lr 0.010000, loss 3.611443
+INFO 2020-12-05 14:36:40 train.py: 79] Epoch 12, iter 2000/6416, lr 0.010000, loss 3.624212
+INFO 2020-12-05 14:40:08 train.py: 79] Epoch 12, iter 2200/6416, lr 0.010000, loss 3.625046
+INFO 2020-12-05 14:43:35 train.py: 79] Epoch 12, iter 2400/6416, lr 0.010000, loss 3.653922
+INFO 2020-12-05 14:47:03 train.py: 79] Epoch 12, iter 2600/6416, lr 0.010000, loss 3.662666
+INFO 2020-12-05 14:50:31 train.py: 79] Epoch 12, iter 2800/6416, lr 0.010000, loss 3.666486
+INFO 2020-12-05 14:53:58 train.py: 92] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-12-05 14:53:59 train.py: 79] Epoch 12, iter 3000/6416, lr 0.010000, loss 3.677757
+INFO 2020-12-05 14:57:27 train.py: 79] Epoch 12, iter 3200/6416, lr 0.010000, loss 3.679683
+INFO 2020-12-05 15:00:55 train.py: 79] Epoch 12, iter 3400/6416, lr 0.010000, loss 3.698292
+INFO 2020-12-05 15:04:23 train.py: 79] Epoch 12, iter 3600/6416, lr 0.010000, loss 3.692556
+INFO 2020-12-05 15:07:51 train.py: 79] Epoch 12, iter 3800/6416, lr 0.010000, loss 3.697790
+INFO 2020-12-05 15:11:19 train.py: 79] Epoch 12, iter 4000/6416, lr 0.010000, loss 3.699502
+INFO 2020-12-05 15:14:47 train.py: 79] Epoch 12, iter 4200/6416, lr 0.010000, loss 3.705153
+INFO 2020-12-05 15:18:15 train.py: 79] Epoch 12, iter 4400/6416, lr 0.010000, loss 3.735897
+INFO 2020-12-05 15:21:43 train.py: 79] Epoch 12, iter 4600/6416, lr 0.010000, loss 3.722900
+INFO 2020-12-05 15:25:11 train.py: 79] Epoch 12, iter 4800/6416, lr 0.010000, loss 3.724094
+INFO 2020-12-05 15:28:39 train.py: 79] Epoch 12, iter 5000/6416, lr 0.010000, loss 3.745422
+INFO 2020-12-05 15:32:07 train.py: 79] Epoch 12, iter 5200/6416, lr 0.010000, loss 3.746051
+INFO 2020-12-05 15:35:35 train.py: 79] Epoch 12, iter 5400/6416, lr 0.010000, loss 3.729365
+INFO 2020-12-05 15:39:04 train.py: 79] Epoch 12, iter 5600/6416, lr 0.010000, loss 3.745029
+INFO 2020-12-05 15:42:32 train.py: 79] Epoch 12, iter 5800/6416, lr 0.010000, loss 3.755959
+INFO 2020-12-05 15:46:00 train.py: 92] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-12-05 15:46:01 train.py: 79] Epoch 12, iter 6000/6416, lr 0.010000, loss 3.720827
+INFO 2020-12-05 15:49:29 train.py: 79] Epoch 12, iter 6200/6416, lr 0.010000, loss 3.757693
+INFO 2020-12-05 15:52:57 train.py: 79] Epoch 12, iter 6400/6416, lr 0.010000, loss 3.760766
+INFO 2020-12-05 15:53:13 train.py: 97] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-12-05 15:53:15 train.py: 79] Epoch 13, iter 0/6416, lr 0.001000, loss 3.819709
+INFO 2020-12-05 15:56:42 train.py: 79] Epoch 13, iter 200/6416, lr 0.001000, loss 3.375830
+INFO 2020-12-05 16:00:10 train.py: 79] Epoch 13, iter 400/6416, lr 0.001000, loss 3.318987
+INFO 2020-12-05 16:03:37 train.py: 79] Epoch 13, iter 600/6416, lr 0.001000, loss 3.315471
+INFO 2020-12-05 16:07:04 train.py: 79] Epoch 13, iter 800/6416, lr 0.001000, loss 3.302148
+INFO 2020-12-05 16:10:32 train.py: 79] Epoch 13, iter 1000/6416, lr 0.001000, loss 3.296407
+INFO 2020-12-05 16:13:59 train.py: 79] Epoch 13, iter 1200/6416, lr 0.001000, loss 3.294618
+INFO 2020-12-05 16:17:26 train.py: 79] Epoch 13, iter 1400/6416, lr 0.001000, loss 3.303942
+INFO 2020-12-05 16:20:54 train.py: 79] Epoch 13, iter 1600/6416, lr 0.001000, loss 3.284822
+INFO 2020-12-05 16:24:21 train.py: 79] Epoch 13, iter 1800/6416, lr 0.001000, loss 3.292487
+INFO 2020-12-05 16:27:48 train.py: 79] Epoch 13, iter 2000/6416, lr 0.001000, loss 3.281661
+INFO 2020-12-05 16:31:16 train.py: 79] Epoch 13, iter 2200/6416, lr 0.001000, loss 3.294156
+INFO 2020-12-05 16:34:43 train.py: 79] Epoch 13, iter 2400/6416, lr 0.001000, loss 3.296767
+INFO 2020-12-05 16:38:10 train.py: 79] Epoch 13, iter 2600/6416, lr 0.001000, loss 3.306324
+INFO 2020-12-05 16:41:38 train.py: 79] Epoch 13, iter 2800/6416, lr 0.001000, loss 3.280782
+INFO 2020-12-05 16:45:05 train.py: 92] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-12-05 16:45:06 train.py: 79] Epoch 13, iter 3000/6416, lr 0.001000, loss 3.293406
+INFO 2020-12-05 16:48:33 train.py: 79] Epoch 13, iter 3200/6416, lr 0.001000, loss 3.288028
+INFO 2020-12-05 16:52:01 train.py: 79] Epoch 13, iter 3400/6416, lr 0.001000, loss 3.301002
+INFO 2020-12-05 16:55:29 train.py: 79] Epoch 13, iter 3600/6416, lr 0.001000, loss 3.289780
+INFO 2020-12-05 16:58:56 train.py: 79] Epoch 13, iter 3800/6416, lr 0.001000, loss 3.290293
+INFO 2020-12-05 17:02:24 train.py: 79] Epoch 13, iter 4000/6416, lr 0.001000, loss 3.290176
+INFO 2020-12-05 17:05:52 train.py: 79] Epoch 13, iter 4200/6416, lr 0.001000, loss 3.299990
+INFO 2020-12-05 17:09:19 train.py: 79] Epoch 13, iter 4400/6416, lr 0.001000, loss 3.277490
+INFO 2020-12-05 17:12:47 train.py: 79] Epoch 13, iter 4600/6416, lr 0.001000, loss 3.306302
+INFO 2020-12-05 17:16:15 train.py: 79] Epoch 13, iter 4800/6416, lr 0.001000, loss 3.305148
+INFO 2020-12-05 17:19:43 train.py: 79] Epoch 13, iter 5000/6416, lr 0.001000, loss 3.283517
+INFO 2020-12-05 17:23:10 train.py: 79] Epoch 13, iter 5200/6416, lr 0.001000, loss 3.281036
+INFO 2020-12-05 17:26:38 train.py: 79] Epoch 13, iter 5400/6416, lr 0.001000, loss 3.287068
+INFO 2020-12-05 17:30:06 train.py: 79] Epoch 13, iter 5600/6416, lr 0.001000, loss 3.285729
+INFO 2020-12-05 17:33:34 train.py: 79] Epoch 13, iter 5800/6416, lr 0.001000, loss 3.287088
+INFO 2020-12-05 17:37:01 train.py: 92] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-12-05 17:37:02 train.py: 79] Epoch 13, iter 6000/6416, lr 0.001000, loss 3.285475
+INFO 2020-12-05 17:40:30 train.py: 79] Epoch 13, iter 6200/6416, lr 0.001000, loss 3.308075
+INFO 2020-12-05 17:43:59 train.py: 79] Epoch 13, iter 6400/6416, lr 0.001000, loss 3.281814
+INFO 2020-12-05 17:44:14 train.py: 97] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-12-05 17:44:16 train.py: 79] Epoch 14, iter 0/6416, lr 0.001000, loss 3.289893
+INFO 2020-12-05 17:47:44 train.py: 79] Epoch 14, iter 200/6416, lr 0.001000, loss 3.253529
+INFO 2020-12-05 17:51:12 train.py: 79] Epoch 14, iter 400/6416, lr 0.001000, loss 3.232601
+INFO 2020-12-05 17:54:40 train.py: 79] Epoch 14, iter 600/6416, lr 0.001000, loss 3.233960
+INFO 2020-12-05 17:58:07 train.py: 79] Epoch 14, iter 800/6416, lr 0.001000, loss 3.233396
+INFO 2020-12-05 18:01:35 train.py: 79] Epoch 14, iter 1000/6416, lr 0.001000, loss 3.222918
+INFO 2020-12-05 18:05:02 train.py: 79] Epoch 14, iter 1200/6416, lr 0.001000, loss 3.249890
+INFO 2020-12-05 18:08:30 train.py: 79] Epoch 14, iter 1400/6416, lr 0.001000, loss 3.252724
+INFO 2020-12-05 18:11:57 train.py: 79] Epoch 14, iter 1600/6416, lr 0.001000, loss 3.250110
+INFO 2020-12-05 18:15:25 train.py: 79] Epoch 14, iter 1800/6416, lr 0.001000, loss 3.257493
+INFO 2020-12-05 18:18:53 train.py: 79] Epoch 14, iter 2000/6416, lr 0.001000, loss 3.271572
+INFO 2020-12-05 18:22:20 train.py: 79] Epoch 14, iter 2200/6416, lr 0.001000, loss 3.276865
+INFO 2020-12-05 18:25:48 train.py: 79] Epoch 14, iter 2400/6416, lr 0.001000, loss 3.249085
+INFO 2020-12-05 18:29:16 train.py: 79] Epoch 14, iter 2600/6416, lr 0.001000, loss 3.256323
+INFO 2020-12-05 18:32:43 train.py: 79] Epoch 14, iter 2800/6416, lr 0.001000, loss 3.233061
+INFO 2020-12-05 18:36:10 train.py: 92] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-12-05 18:36:11 train.py: 79] Epoch 14, iter 3000/6416, lr 0.001000, loss 3.240132
+INFO 2020-12-05 18:39:39 train.py: 79] Epoch 14, iter 3200/6416, lr 0.001000, loss 3.276102
+INFO 2020-12-05 18:43:07 train.py: 79] Epoch 14, iter 3400/6416, lr 0.001000, loss 3.255621
+INFO 2020-12-05 18:46:35 train.py: 79] Epoch 14, iter 3600/6416, lr 0.001000, loss 3.256739
+INFO 2020-12-05 18:50:03 train.py: 79] Epoch 14, iter 3800/6416, lr 0.001000, loss 3.258472
+INFO 2020-12-05 18:53:31 train.py: 79] Epoch 14, iter 4000/6416, lr 0.001000, loss 3.240507
+INFO 2020-12-05 18:56:59 train.py: 79] Epoch 14, iter 4200/6416, lr 0.001000, loss 3.260787
+INFO 2020-12-05 19:00:27 train.py: 79] Epoch 14, iter 4400/6416, lr 0.001000, loss 3.276287
+INFO 2020-12-05 19:03:55 train.py: 79] Epoch 14, iter 4600/6416, lr 0.001000, loss 3.269647
+INFO 2020-12-05 19:07:23 train.py: 79] Epoch 14, iter 4800/6416, lr 0.001000, loss 3.248438
+INFO 2020-12-05 19:10:51 train.py: 79] Epoch 14, iter 5000/6416, lr 0.001000, loss 3.242763
+INFO 2020-12-05 19:14:19 train.py: 79] Epoch 14, iter 5200/6416, lr 0.001000, loss 3.272917
+INFO 2020-12-05 19:17:48 train.py: 79] Epoch 14, iter 5400/6416, lr 0.001000, loss 3.270104
+INFO 2020-12-05 19:21:16 train.py: 79] Epoch 14, iter 5600/6416, lr 0.001000, loss 3.246871
+INFO 2020-12-05 19:24:44 train.py: 79] Epoch 14, iter 5800/6416, lr 0.001000, loss 3.283965
+INFO 2020-12-05 19:28:12 train.py: 92] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-12-05 19:28:13 train.py: 79] Epoch 14, iter 6000/6416, lr 0.001000, loss 3.269561
+INFO 2020-12-05 19:31:41 train.py: 79] Epoch 14, iter 6200/6416, lr 0.001000, loss 3.265996
+INFO 2020-12-05 19:35:09 train.py: 79] Epoch 14, iter 6400/6416, lr 0.001000, loss 3.297731
+INFO 2020-12-05 19:35:24 train.py: 97] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-12-05 19:35:27 train.py: 79] Epoch 15, iter 0/6416, lr 0.001000, loss 3.274030
+INFO 2020-12-05 19:38:54 train.py: 79] Epoch 15, iter 200/6416, lr 0.001000, loss 3.202996
+INFO 2020-12-05 19:42:22 train.py: 79] Epoch 15, iter 400/6416, lr 0.001000, loss 3.218926
+INFO 2020-12-05 19:45:50 train.py: 79] Epoch 15, iter 600/6416, lr 0.001000, loss 3.221272
+INFO 2020-12-05 19:49:17 train.py: 79] Epoch 15, iter 800/6416, lr 0.001000, loss 3.229126
+INFO 2020-12-05 19:52:45 train.py: 79] Epoch 15, iter 1000/6416, lr 0.001000, loss 3.203917
+INFO 2020-12-05 19:56:12 train.py: 79] Epoch 15, iter 1200/6416, lr 0.001000, loss 3.220771
+INFO 2020-12-05 19:59:40 train.py: 79] Epoch 15, iter 1400/6416, lr 0.001000, loss 3.237727
+INFO 2020-12-05 20:03:08 train.py: 79] Epoch 15, iter 1600/6416, lr 0.001000, loss 3.207903
+INFO 2020-12-05 20:06:35 train.py: 79] Epoch 15, iter 1800/6416, lr 0.001000, loss 3.238711
+INFO 2020-12-05 20:10:03 train.py: 79] Epoch 15, iter 2000/6416, lr 0.001000, loss 3.208129
+INFO 2020-12-05 20:13:30 train.py: 79] Epoch 15, iter 2200/6416, lr 0.001000, loss 3.235893
+INFO 2020-12-05 20:16:58 train.py: 79] Epoch 15, iter 2400/6416, lr 0.001000, loss 3.236612
+INFO 2020-12-05 20:20:26 train.py: 79] Epoch 15, iter 2600/6416, lr 0.001000, loss 3.227373
+INFO 2020-12-05 20:23:54 train.py: 79] Epoch 15, iter 2800/6416, lr 0.001000, loss 3.240951
+INFO 2020-12-05 20:27:21 train.py: 92] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-12-05 20:27:22 train.py: 79] Epoch 15, iter 3000/6416, lr 0.001000, loss 3.238829
+INFO 2020-12-05 20:30:49 train.py: 79] Epoch 15, iter 3200/6416, lr 0.001000, loss 3.225033
+INFO 2020-12-05 20:34:17 train.py: 79] Epoch 15, iter 3400/6416, lr 0.001000, loss 3.232243
+INFO 2020-12-05 20:37:45 train.py: 79] Epoch 15, iter 3600/6416, lr 0.001000, loss 3.249092
+INFO 2020-12-05 20:41:13 train.py: 79] Epoch 15, iter 3800/6416, lr 0.001000, loss 3.243252
+INFO 2020-12-05 20:44:41 train.py: 79] Epoch 15, iter 4000/6416, lr 0.001000, loss 3.235035
+INFO 2020-12-05 20:48:09 train.py: 79] Epoch 15, iter 4200/6416, lr 0.001000, loss 3.274263
+INFO 2020-12-05 20:51:37 train.py: 79] Epoch 15, iter 4400/6416, lr 0.001000, loss 3.233355
+INFO 2020-12-05 20:55:05 train.py: 79] Epoch 15, iter 4600/6416, lr 0.001000, loss 3.219492
+INFO 2020-12-05 20:58:33 train.py: 79] Epoch 15, iter 4800/6416, lr 0.001000, loss 3.229777
+INFO 2020-12-05 21:02:01 train.py: 79] Epoch 15, iter 5000/6416, lr 0.001000, loss 3.255224
+INFO 2020-12-05 21:05:30 train.py: 79] Epoch 15, iter 5200/6416, lr 0.001000, loss 3.230233
+INFO 2020-12-05 21:08:58 train.py: 79] Epoch 15, iter 5400/6416, lr 0.001000, loss 3.263279
+INFO 2020-12-05 21:12:26 train.py: 79] Epoch 15, iter 5600/6416, lr 0.001000, loss 3.246884
+INFO 2020-12-05 21:15:54 train.py: 79] Epoch 15, iter 5800/6416, lr 0.001000, loss 3.228916
+INFO 2020-12-05 21:19:22 train.py: 92] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-12-05 21:19:23 train.py: 79] Epoch 15, iter 6000/6416, lr 0.001000, loss 3.242338
+INFO 2020-12-05 21:22:51 train.py: 79] Epoch 15, iter 6200/6416, lr 0.001000, loss 3.253669
+INFO 2020-12-05 21:26:19 train.py: 79] Epoch 15, iter 6400/6416, lr 0.001000, loss 3.258864
+INFO 2020-12-05 21:26:35 train.py: 97] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-12-05 21:26:37 train.py: 79] Epoch 16, iter 0/6416, lr 0.000100, loss 3.292948
+INFO 2020-12-05 21:30:04 train.py: 79] Epoch 16, iter 200/6416, lr 0.000100, loss 3.183219
+INFO 2020-12-05 21:33:32 train.py: 79] Epoch 16, iter 400/6416, lr 0.000100, loss 3.197252
+INFO 2020-12-05 21:36:59 train.py: 79] Epoch 16, iter 600/6416, lr 0.000100, loss 3.181016
+INFO 2020-12-05 21:40:26 train.py: 79] Epoch 16, iter 800/6416, lr 0.000100, loss 3.192165
+INFO 2020-12-05 21:43:54 train.py: 79] Epoch 16, iter 1000/6416, lr 0.000100, loss 3.184664
+INFO 2020-12-05 21:47:21 train.py: 79] Epoch 16, iter 1200/6416, lr 0.000100, loss 3.198980
+INFO 2020-12-05 21:50:48 train.py: 79] Epoch 16, iter 1400/6416, lr 0.000100, loss 3.192734
+INFO 2020-12-05 21:54:15 train.py: 79] Epoch 16, iter 1600/6416, lr 0.000100, loss 3.191035
+INFO 2020-12-05 21:57:42 train.py: 79] Epoch 16, iter 1800/6416, lr 0.000100, loss 3.173339
+INFO 2020-12-05 22:01:10 train.py: 79] Epoch 16, iter 2000/6416, lr 0.000100, loss 3.204828
+INFO 2020-12-05 22:04:37 train.py: 79] Epoch 16, iter 2200/6416, lr 0.000100, loss 3.194531
+INFO 2020-12-05 22:08:04 train.py: 79] Epoch 16, iter 2400/6416, lr 0.000100, loss 3.175848
+INFO 2020-12-05 22:11:31 train.py: 79] Epoch 16, iter 2600/6416, lr 0.000100, loss 3.189726
+INFO 2020-12-05 22:14:58 train.py: 79] Epoch 16, iter 2800/6416, lr 0.000100, loss 3.177625
+INFO 2020-12-05 22:18:25 train.py: 92] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-12-05 22:18:26 train.py: 79] Epoch 16, iter 3000/6416, lr 0.000100, loss 3.170502
+INFO 2020-12-05 22:21:53 train.py: 79] Epoch 16, iter 3200/6416, lr 0.000100, loss 3.186420
+INFO 2020-12-05 22:25:20 train.py: 79] Epoch 16, iter 3400/6416, lr 0.000100, loss 3.185336
+INFO 2020-12-05 22:28:48 train.py: 79] Epoch 16, iter 3600/6416, lr 0.000100, loss 3.188959
+INFO 2020-12-05 22:32:15 train.py: 79] Epoch 16, iter 3800/6416, lr 0.000100, loss 3.207331
+INFO 2020-12-05 22:35:43 train.py: 79] Epoch 16, iter 4000/6416, lr 0.000100, loss 3.175016
+INFO 2020-12-05 22:39:10 train.py: 79] Epoch 16, iter 4200/6416, lr 0.000100, loss 3.202153
+INFO 2020-12-05 22:42:38 train.py: 79] Epoch 16, iter 4400/6416, lr 0.000100, loss 3.175727
+INFO 2020-12-05 22:46:05 train.py: 79] Epoch 16, iter 4600/6416, lr 0.000100, loss 3.197430
+INFO 2020-12-05 22:49:33 train.py: 79] Epoch 16, iter 4800/6416, lr 0.000100, loss 3.209093
+INFO 2020-12-05 22:53:00 train.py: 79] Epoch 16, iter 5000/6416, lr 0.000100, loss 3.200509
+INFO 2020-12-05 22:56:27 train.py: 79] Epoch 16, iter 5200/6416, lr 0.000100, loss 3.166731
+INFO 2020-12-05 22:59:55 train.py: 79] Epoch 16, iter 5400/6416, lr 0.000100, loss 3.189580
+INFO 2020-12-05 23:03:23 train.py: 79] Epoch 16, iter 5600/6416, lr 0.000100, loss 3.176489
+INFO 2020-12-05 23:06:50 train.py: 79] Epoch 16, iter 5800/6416, lr 0.000100, loss 3.185435
+INFO 2020-12-05 23:10:18 train.py: 92] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-12-05 23:10:19 train.py: 79] Epoch 16, iter 6000/6416, lr 0.000100, loss 3.183122
+INFO 2020-12-05 23:13:47 train.py: 79] Epoch 16, iter 6200/6416, lr 0.000100, loss 3.173565
+INFO 2020-12-05 23:17:15 train.py: 79] Epoch 16, iter 6400/6416, lr 0.000100, loss 3.172358
+INFO 2020-12-05 23:17:31 train.py: 97] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-12-05 23:17:33 train.py: 79] Epoch 17, iter 0/6416, lr 0.000100, loss 3.189538
+INFO 2020-12-05 23:21:01 train.py: 79] Epoch 17, iter 200/6416, lr 0.000100, loss 3.189161
+INFO 2020-12-05 23:24:29 train.py: 79] Epoch 17, iter 400/6416, lr 0.000100, loss 3.188404
+INFO 2020-12-05 23:27:56 train.py: 79] Epoch 17, iter 600/6416, lr 0.000100, loss 3.164539
+INFO 2020-12-05 23:31:24 train.py: 79] Epoch 17, iter 800/6416, lr 0.000100, loss 3.182564
+INFO 2020-12-05 23:34:51 train.py: 79] Epoch 17, iter 1000/6416, lr 0.000100, loss 3.187699
+INFO 2020-12-05 23:38:19 train.py: 79] Epoch 17, iter 1200/6416, lr 0.000100, loss 3.190330
+INFO 2020-12-05 23:41:46 train.py: 79] Epoch 17, iter 1400/6416, lr 0.000100, loss 3.179394
+INFO 2020-12-05 23:45:14 train.py: 79] Epoch 17, iter 1600/6416, lr 0.000100, loss 3.190443
+INFO 2020-12-05 23:48:41 train.py: 79] Epoch 17, iter 1800/6416, lr 0.000100, loss 3.182039
+INFO 2020-12-05 23:52:09 train.py: 79] Epoch 17, iter 2000/6416, lr 0.000100, loss 3.194719
+INFO 2020-12-05 23:55:36 train.py: 79] Epoch 17, iter 2200/6416, lr 0.000100, loss 3.172530
+INFO 2020-12-05 23:59:04 train.py: 79] Epoch 17, iter 2400/6416, lr 0.000100, loss 3.193706
+INFO 2020-12-06 00:02:31 train.py: 79] Epoch 17, iter 2600/6416, lr 0.000100, loss 3.183820
+INFO 2020-12-06 00:05:59 train.py: 79] Epoch 17, iter 2800/6416, lr 0.000100, loss 3.187860
+INFO 2020-12-06 00:09:26 train.py: 92] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-12-06 00:09:27 train.py: 79] Epoch 17, iter 3000/6416, lr 0.000100, loss 3.198377
+INFO 2020-12-06 00:12:54 train.py: 79] Epoch 17, iter 3200/6416, lr 0.000100, loss 3.143905
+INFO 2020-12-06 00:16:22 train.py: 79] Epoch 17, iter 3400/6416, lr 0.000100, loss 3.172025
+INFO 2020-12-06 00:19:49 train.py: 79] Epoch 17, iter 3600/6416, lr 0.000100, loss 3.192636
+INFO 2020-12-06 00:23:17 train.py: 79] Epoch 17, iter 3800/6416, lr 0.000100, loss 3.177560
+INFO 2020-12-06 00:26:45 train.py: 79] Epoch 17, iter 4000/6416, lr 0.000100, loss 3.188232
+INFO 2020-12-06 00:30:12 train.py: 79] Epoch 17, iter 4200/6416, lr 0.000100, loss 3.191702
+INFO 2020-12-06 00:33:40 train.py: 79] Epoch 17, iter 4400/6416, lr 0.000100, loss 3.177440
+INFO 2020-12-06 00:37:08 train.py: 79] Epoch 17, iter 4600/6416, lr 0.000100, loss 3.184048
+INFO 2020-12-06 00:40:35 train.py: 79] Epoch 17, iter 4800/6416, lr 0.000100, loss 3.184476
+INFO 2020-12-06 00:44:03 train.py: 79] Epoch 17, iter 5000/6416, lr 0.000100, loss 3.183765
+INFO 2020-12-06 00:47:31 train.py: 79] Epoch 17, iter 5200/6416, lr 0.000100, loss 3.175542
+INFO 2020-12-06 00:50:59 train.py: 79] Epoch 17, iter 5400/6416, lr 0.000100, loss 3.190452
+INFO 2020-12-06 00:54:26 train.py: 79] Epoch 17, iter 5600/6416, lr 0.000100, loss 3.204129
+INFO 2020-12-06 00:57:55 train.py: 79] Epoch 17, iter 5800/6416, lr 0.000100, loss 3.163028
+INFO 2020-12-06 01:01:22 train.py: 92] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-12-06 01:01:23 train.py: 79] Epoch 17, iter 6000/6416, lr 0.000100, loss 3.187762
+INFO 2020-12-06 01:04:51 train.py: 79] Epoch 17, iter 6200/6416, lr 0.000100, loss 3.182176
+INFO 2020-12-06 01:08:19 train.py: 79] Epoch 17, iter 6400/6416, lr 0.000100, loss 3.163325
+INFO 2020-12-06 01:08:35 train.py: 97] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-12-06 01:08:35 train.py: 180] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/.gitkeep b/bob/bio/facexzoo/models/backbones/GhostNet/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..57e1f56b3c7524906bf628e03959d797d6a80bff
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_agedb.txt
@@ -0,0 +1,45 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9570000000000001 | 0.0032848323331321023 |
+| Epoch_17_batch_5999.pt | 0.9568333333333333 | 0.0033467323292953426 |
+| Epoch_13.pt | 0.9563333333333333 | 0.0036243347622889155 |
+| Epoch_15_batch_5999.pt | 0.9563333333333333 | 0.0031991511219751087 |
+| Epoch_15.pt | 0.9561666666666667 | 0.00324940640353106 |
+| Epoch_14.pt | 0.9555 | 0.0033706247360261103 |
+| Epoch_16_batch_5999.pt | 0.9555 | 0.0035490391674854248 |
+| Epoch_12_batch_5999.pt | 0.9553333333333333 | 0.0031700761372638027 |
+| Epoch_10.pt | 0.9551666666666666 | 0.002986078811194823 |
+| Epoch_15_batch_2999.pt | 0.9551666666666666 | 0.003309638001912554 |
+| Epoch_13_batch_5999.pt | 0.9548333333333334 | 0.0029860788111948163 |
+| Epoch_14_batch_2999.pt | 0.9548333333333334 | 0.003107636949099334 |
+| Epoch_13_batch_2999.pt | 0.9548333333333332 | 0.003047464033756106 |
+| Epoch_16.pt | 0.9548333333333332 | 0.0035088072610306984 |
+| Epoch_12.pt | 0.9545 | 0.0037437190197403204 |
+| Epoch_12_batch_2999.pt | 0.9543333333333333 | 0.0032603112780269315 |
+| Epoch_14_batch_5999.pt | 0.9541666666666668 | 0.00307569123014809 |
+| Epoch_17_batch_2999.pt | 0.9541666666666668 | 0.003203489609630777 |
+| Epoch_16_batch_2999.pt | 0.9541666666666666 | 0.0035939764421413028 |
+| Epoch_10_batch_5999.pt | 0.9533333333333334 | 0.0030530292586742436 |
+| Epoch_11_batch_2999.pt | 0.9533333333333334 | 0.003676753801727618 |
+| Epoch_10_batch_2999.pt | 0.9528333333333334 | 0.0030332519321597104 |
+| Epoch_11_batch_5999.pt | 0.9526666666666668 | 0.003506607519568787 |
+| Epoch_11.pt | 0.9491666666666667 | 0.0032796604762545395 |
+| Epoch_8_batch_5999.pt | 0.9399999999999998 | 0.003975231959999623 |
+| Epoch_9_batch_2999.pt | 0.9391666666666667 | 0.003125462928674486 |
+| Epoch_6_batch_2999.pt | 0.9381666666666668 | 0.00428498671072748 |
+| Epoch_7_batch_2999.pt | 0.9361666666666666 | 0.005299953412327276 |
+| Epoch_7_batch_5999.pt | 0.9351666666666667 | 0.0043776226706870956 |
+| Epoch_8_batch_2999.pt | 0.9351666666666667 | 0.004814138129086651 |
+| Epoch_5_batch_2999.pt | 0.9341666666666667 | 0.004261875203113481 |
+| Epoch_6_batch_5999.pt | 0.9333333333333332 | 0.0037350525142159086 |
+| Epoch_7.pt | 0.9333333333333332 | 0.004245549594342846 |
+| Epoch_9_batch_5999.pt | 0.9331666666666667 | 0.0039003798485484635 |
+| Epoch_8.pt | 0.932 | 0.0046067583203617535 |
+| Epoch_5_batch_5999.pt | 0.9318333333333333 | 0.005954819606269348 |
+| Epoch_4_batch_2999.pt | 0.9306666666666666 | 0.004682512779045661 |
+| Epoch_6.pt | 0.9306666666666666 | 0.004932882862316247 |
+| Epoch_9.pt | 0.9306666666666666 | 0.003532914021241036 |
+| Epoch_5.pt | 0.9288333333333334 | 0.003760171390892831 |
+| Epoch_3_batch_5999.pt | 0.9276666666666665 | 0.004786813161897336 |
+| Epoch_4_batch_5999.pt | 0.9253333333333333 | 0.0040809705938237685 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0c86c978f4ef9c9d63b730ac2599db3c790ef5a2
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_calfw.txt
@@ -0,0 +1,46 @@
++------------------------+--------------------+-----------------------+ [13/206]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.9393333333333335 | 0.003055050463303892 |
+| Epoch_16_batch_2999.pt | 0.9386666666666666 | 0.003340732528527317 |
+| Epoch_16_batch_5999.pt | 0.9386666666666666 | 0.0028523328117763197 |
+| Epoch_16.pt | 0.9385 | 0.00274929845147969 |
+| Epoch_14_batch_2999.pt | 0.9381666666666668 | 0.002902212863690886 |
+| Epoch_14.pt | 0.9381666666666666 | 0.0036298658275376984 |
+| Epoch_15_batch_2999.pt | 0.9381666666666666 | 0.0027938424357066977 |
+| Epoch_12_batch_2999.pt | 0.938 | 0.0034047896546765674 |
+| Epoch_14_batch_5999.pt | 0.9378333333333334 | 0.0030837086858617594 |
+| Epoch_13_batch_5999.pt | 0.9373333333333334 | 0.003279189902971932 |
+| Epoch_13.pt | 0.9373333333333334 | 0.0028888888888888844 |
+| Epoch_15.pt | 0.9371666666666666 | 0.0030332519321597095 |
+| Epoch_17_batch_2999.pt | 0.9369999999999999 | 0.002948110924760356 |
+| Epoch_12.pt | 0.9368333333333334 | 0.0026925824035672562 |
+| Epoch_17.pt | 0.9368333333333334 | 0.0030169275516412184 |
+| Epoch_12_batch_5999.pt | 0.9366666666666668 | 0.0030123203803835365 |
+| Epoch_13_batch_2999.pt | 0.9366666666666668 | 0.00255796987404918 |
+| Epoch_11.pt | 0.9361666666666666 | 0.0030025709148603684 |
+| Epoch_11_batch_2999.pt | 0.9358333333333334 | 0.003154949081000144 |
+| Epoch_10_batch_5999.pt | 0.9356666666666665 | 0.003506607519568778 |
+| Epoch_10.pt | 0.9353333333333331 | 0.003368334753605355 |
+| Epoch_15_batch_5999.pt | 0.9351666666666667 | 0.0031076369490993356 |
+| Epoch_11_batch_5999.pt | 0.9333333333333333 | 0.0036851386559504434 |
+| Epoch_10_batch_2999.pt | 0.9326666666666666 | 0.0037367048268444995 |
+| Epoch_7_batch_5999.pt | 0.9268333333333333 | 0.0038445889503991018 |
+| Epoch_7_batch_2999.pt | 0.9245000000000001 | 0.004346489998681923 |
+| Epoch_8_batch_5999.pt | 0.924 | 0.004030746032714904 |
+| Epoch_8_batch_2999.pt | 0.9224999999999998 | 0.00431942211498337 |
+| Epoch_6_batch_2999.pt | 0.9221666666666666 | 0.004701260666231114 |
+| Epoch_9.pt | 0.9209999999999999 | 0.0043047490284023335 |
+| Epoch_5_batch_5999.pt | 0.9206666666666667 | 0.006475023237481541 |
+| Epoch_6_batch_5999.pt | 0.9203333333333334 | 0.005090538306202354 |
+| Epoch_9_batch_5999.pt | 0.9201666666666666 | 0.004342227339926471 |
+| Epoch_7.pt | 0.9199999999999999 | 0.0041573970964154895 |
+| Epoch_4_batch_2999.pt | 0.9186666666666665 | 0.004626813958590444 |
+| Epoch_9_batch_2999.pt | 0.9183333333333333 | 0.0035486043161491836 |
+| Epoch_5_batch_2999.pt | 0.9161666666666667 | 0.004444791653104356 |
+| Epoch_6.pt | 0.916 | 0.003866602809178964 |
+| Epoch_5.pt | 0.9153333333333334 | 0.004790680252735582 |
+| Epoch_4_batch_5999.pt | 0.9149999999999998 | 0.003920506392083858 |
+| Epoch_8.pt | 0.9145000000000001 | 0.003425125031510741 |
+| Epoch_3_batch_5999.pt | 0.9108333333333333 | 0.00583862194124945 |
+| Epoch_4.pt | 0.9099999999999999 | 0.004157397096415489 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b3fadf49108c3354924d656119472963a5c48d47
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_cplfw.txt
@@ -0,0 +1,44 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.8351666666666666 | 0.00529645816898448 |
+| Epoch_14.pt | 0.8351666666666666 | 0.005502244211049226 |
+| Epoch_15.pt | 0.8346666666666666 | 0.006308919807368171 |
+| Epoch_13_batch_5999.pt | 0.8343333333333334 | 0.005606433694019055 |
+| Epoch_17_batch_5999.pt | 0.8338333333333333 | 0.005346338310781819 |
+| Epoch_14_batch_2999.pt | 0.8335000000000001 | 0.005406043714397371 |
+| Epoch_17.pt | 0.8334999999999999 | 0.005148714330710073 |
+| Epoch_16_batch_2999.pt | 0.8333333333333334 | 0.006382847385042249 |
+| Epoch_15_batch_5999.pt | 0.833 | 0.00634307243386315 |
+| Epoch_16_batch_5999.pt | 0.8328333333333333 | 0.00574697988318889 |
+| Epoch_17_batch_2999.pt | 0.8328333333333333 | 0.005863940865078348 |
+| Epoch_16.pt | 0.8323333333333334 | 0.005731653500679523 |
+| Epoch_13_batch_2999.pt | 0.8314999999999999 | 0.005985837194042308 |
+| Epoch_15_batch_2999.pt | 0.8308333333333333 | 0.005969314537648581 |
+| Epoch_12_batch_2999.pt | 0.8298333333333334 | 0.005700606422149049 |
+| Epoch_11_batch_5999.pt | 0.8296666666666666 | 0.005436502143433363 |
+| Epoch_12.pt | 0.8295 | 0.005488765180792179 |
+| Epoch_13.pt | 0.8295 | 0.006014642626617643 |
+| Epoch_11_batch_2999.pt | 0.829 | 0.005628411174183974 |
+| Epoch_11.pt | 0.8281666666666666 | 0.0058184996581253584 |
+| Epoch_10.pt | 0.827 | 0.005430821959933297 |
+| Epoch_10_batch_2999.pt | 0.8266666666666668 | 0.006521569151479632 |
+| Epoch_10_batch_5999.pt | 0.8265 | 0.006118429960797555 |
+| Epoch_12_batch_5999.pt | 0.8251666666666667 | 0.0063413691710042415 |
+| Epoch_8_batch_5999.pt | 0.8028333333333333 | 0.006161659628924338 |
+| Epoch_9_batch_5999.pt | 0.8006666666666667 | 0.005890984951916251 |
+| Epoch_8_batch_2999.pt | 0.8003333333333333 | 0.0070088127946267365 |
+| Epoch_7.pt | 0.7993333333333333 | 0.006513044839890153 |
+| Epoch_6_batch_2999.pt | 0.7986666666666667 | 0.006084791808930259 |
+| Epoch_5_batch_2999.pt | 0.7985 | 0.006032064527932023 |
+| Epoch_7_batch_5999.pt | 0.7983333333333335 | 0.004733646312031285 |
+| Epoch_7_batch_2999.pt | 0.7971666666666666 | 0.005018791847141928 |
+| Epoch_9.pt | 0.7968333333333334 | 0.008094541675615219 |
+| Epoch_8.pt | 0.7965 | 0.005255508573903227 |
+| Epoch_9_batch_2999.pt | 0.7955 | 0.004987948438950589 |
+| Epoch_6_batch_5999.pt | 0.7933333333333333 | 0.005698169513314707 |
+| Epoch_4_batch_2999.pt | 0.7928333333333334 | 0.007345318488922299 |
+| Epoch_3_batch_5999.pt | 0.7928333333333333 | 0.006479073626817992 |
+| Epoch_5_batch_5999.pt | 0.7925 | 0.0049206669225789835 |
+| Epoch_4_batch_5999.pt | 0.7914999999999999 | 0.006737814790282384 |
+| Epoch_3_batch_2999.pt | 0.7888333333333333 | 0.005409468151215264 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..1347bab37bfb5cf96ee39964bc2b3e2fee54c585
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_lfw.txt
@@ -0,0 +1,46 @@
++------------------------+--------------------+-----------------------+ [20/79]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11_batch_2999.pt | 0.9964999999999999 | 0.0009444444444444436 |
+| Epoch_13_batch_2999.pt | 0.9964999999999999 | 0.0009444444444444436 |
+| Epoch_13_batch_5999.pt | 0.9963333333333333 | 0.0010772621905369643 |
+| Epoch_14_batch_2999.pt | 0.9963333333333333 | 0.0008164965809277232 |
+| Epoch_13.pt | 0.9961666666666668 | 0.0007474235581707571 |
+| Epoch_17_batch_5999.pt | 0.9960000000000001 | 0.0009362388636862611 |
+| Epoch_14.pt | 0.9958333333333333 | 0.0009043789220055396 |
+| Epoch_16.pt | 0.9958333333333333 | 0.0009043789220055396 |
+| Epoch_15_batch_2999.pt | 0.9956666666666667 | 0.0008314794192830969 |
+| Epoch_15.pt | 0.9956666666666667 | 0.0007535922203472557 |
+| Epoch_16_batch_2999.pt | 0.9956666666666667 | 0.0008314794192830969 |
+| Epoch_16_batch_5999.pt | 0.9956666666666667 | 0.0010304020550550815 |
+| Epoch_17.pt | 0.9956666666666667 | 0.000968644209675708 |
+| Epoch_10.pt | 0.9955 | 0.0007876359377087687 |
+| Epoch_12.pt | 0.9955 | 0.0007876359377087669 |
+| Epoch_14_batch_5999.pt | 0.9955 | 0.000963852865160973 |
+| Epoch_15_batch_5999.pt | 0.9955 | 0.0009312808119022384 |
+| Epoch_17_batch_2999.pt | 0.9955 | 0.000963852865160973 |
+| Epoch_11_batch_5999.pt | 0.9951666666666666 | 0.0008766518798922007 |
+| Epoch_11.pt | 0.9951666666666666 | 0.0011772011166898341 |
+| Epoch_12_batch_5999.pt | 0.9951666666666666 | 0.0009765775461803897 |
+| Epoch_10_batch_5999.pt | 0.9949999999999999 | 0.000785674201318386 |
+| Epoch_12_batch_2999.pt | 0.9949999999999999 | 0.0011111111111111072 |
+| Epoch_10_batch_2999.pt | 0.9943333333333332 | 0.0011706281947614185 |
+| Epoch_4_batch_2999.pt | 0.994 | 0.0010599324460188319 |
+| Epoch_7.pt | 0.994 | 0.0007114582486036464 |
+| Epoch_7_batch_5999.pt | 0.9933333333333334 | 0.001531560972454468 |
+| Epoch_8_batch_5999.pt | 0.9933333333333334 | 0.0011385500851066193 |
+| Epoch_9.pt | 0.9933333333333332 | 0.0013146843962443628 |
+| Epoch_5_batch_2999.pt | 0.9931666666666666 | 0.001228519132638661 |
+| Epoch_9_batch_5999.pt | 0.9931666666666666 | 0.001203133768205985 |
+| Epoch_5_batch_5999.pt | 0.9926666666666668 | 0.0010599324460188304 |
+| Epoch_7_batch_2999.pt | 0.9926666666666668 | 0.0009026709338484449 |
+| Epoch_8_batch_2999.pt | 0.9926666666666666 | 0.0008314794192831023 |
+| Epoch_9_batch_2999.pt | 0.9924999999999999 | 0.0010015420209622196 |
+| Epoch_3.pt | 0.9923333333333334 | 0.0010000000000000015 |
+| Epoch_5.pt | 0.9921666666666666 | 0.0011124991330278197 |
+| Epoch_8.pt | 0.9921666666666666 | 0.0015918387535438186 |
+| Epoch_2_batch_5999.pt | 0.9918333333333333 | 0.0010957268290731155 |
+| Epoch_6_batch_5999.pt | 0.9918333333333333 | 0.001037863427348303 |
+| Epoch_3_batch_5999.pt | 0.9916666666666668 | 0.001405456737852609 |
+| Epoch_4_batch_5999.pt | 0.9914999999999999 | 0.0009444444444444503 |
+| Epoch_6.pt | 0.9913333333333334 | 0.0010772621905369587 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f2b20c203bf358e6a66f75d35e9d4b73624b0b65
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.8941770701797779 |
+| 2 | 0.915207571240741 |
+| 3 | 0.9250100889126105 |
+| 4 | 0.9306598799744848 |
+| 5 | 0.934838642487991 |
+| 6 | 0.9378457893435047 |
+| 7 | 0.9401434578283453 |
+| 8 | 0.9422979288438758 |
+| 9 | 0.9442050587760522 |
+| 10 | 0.9458062668419751 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b60d5fae88eb704a591935ad8a08d25781f8d383
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_African.txt
@@ -0,0 +1,44 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.8865000000000001 | 0.006471447259278996 |
+| Epoch_16_batch_5999.pt | 0.8848333333333335 | 0.00664556846679903 |
+| Epoch_14_batch_5999.pt | 0.8848333333333332 | 0.005797242852542748 |
+| Epoch_13_batch_5999.pt | 0.8846666666666666 | 0.006906786785742248 |
+| Epoch_16_batch_2999.pt | 0.8846666666666666 | 0.006094928067512211 |
+| Epoch_15_batch_2999.pt | 0.8843333333333334 | 0.006667592528301109 |
+| Epoch_17.pt | 0.8843333333333332 | 0.0066024686741091455 |
+| Epoch_16.pt | 0.8841666666666667 | 0.006369535905725081 |
+| Epoch_14.pt | 0.8835 | 0.006088087944076928 |
+| Epoch_17_batch_2999.pt | 0.8833333333333334 | 0.00622123007963085 |
+| Epoch_15_batch_5999.pt | 0.8818333333333334 | 0.006263011148159904 |
+| Epoch_12.pt | 0.8815000000000002 | 0.006243267979330092 |
+| Epoch_13_batch_2999.pt | 0.8808333333333334 | 0.00626202546808642 |
+| Epoch_13.pt | 0.8808333333333334 | 0.006446599256608759 |
+| Epoch_14_batch_2999.pt | 0.8808333333333334 | 0.006522752202582612 |
+| Epoch_12_batch_2999.pt | 0.8805 | 0.005725457553552456 |
+| Epoch_12_batch_5999.pt | 0.8801666666666665 | 0.005939250066971811 |
+| Epoch_15.pt | 0.8799999999999999 | 0.0064788354387170025 |
+| Epoch_11_batch_2999.pt | 0.8795 | 0.006334551539760064 |
+| Epoch_10_batch_2999.pt | 0.8783333333333333 | 0.005895174842169915 |
+| Epoch_11_batch_5999.pt | 0.8775000000000001 | 0.0050077101048111 |
+| Epoch_11.pt | 0.8761666666666666 | 0.005730845713556816 |
+| Epoch_10.pt | 0.8736666666666666 | 0.0047842333648024345 |
+| Epoch_10_batch_5999.pt | 0.8711666666666666 | 0.005505608812389576 |
+| Epoch_8_batch_5999.pt | 0.8514999999999999 | 0.005860781981280213 |
+| Epoch_9_batch_5999.pt | 0.8476666666666667 | 0.006310876364961303 |
+| Epoch_8_batch_2999.pt | 0.8466666666666667 | 0.005687326174253215 |
+| Epoch_9_batch_2999.pt | 0.845 | 0.00411261233851594 |
+| Epoch_7_batch_5999.pt | 0.8428333333333333 | 0.006045353281347278 |
+| Epoch_9.pt | 0.8413333333333334 | 0.006969067458029653 |
+| Epoch_5_batch_5999.pt | 0.841 | 0.005769224438109846 |
+| Epoch_7.pt | 0.8385 | 0.004877828397611296 |
+| Epoch_4_batch_5999.pt | 0.8366666666666667 | 0.005024630690931122 |
+| Epoch_6_batch_2999.pt | 0.8354999999999999 | 0.0065830403122787 |
+| Epoch_8.pt | 0.8353333333333334 | 0.004620138419175955 |
+| Epoch_5_batch_2999.pt | 0.8348333333333333 | 0.0057060180489820235 |
+| Epoch_4_batch_2999.pt | 0.8335000000000001 | 0.00538888888888889 |
+| Epoch_7_batch_2999.pt | 0.8331666666666667 | 0.005348646996330754 |
+| Epoch_6_batch_5999.pt | 0.8295 | 0.005805754914174107 |
+| Epoch_3_batch_5999.pt | 0.8290000000000001 | 0.005864730320176291 |
+| Epoch_6.pt | 0.8233333333333335 | 0.006004113816047236 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b9d2f7df9a81a11f7c127d6305fdb89c61bb0f25
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,44 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.8856666666666667 | 0.004548775442324154 |
+| Epoch_17.pt | 0.8850000000000001 | 0.004919098582484149 |
+| Epoch_15_batch_2999.pt | 0.8848333333333332 | 0.004839714816616041 |
+| Epoch_15_batch_5999.pt | 0.8843333333333334 | 0.004340449996583247 |
+| Epoch_14_batch_5999.pt | 0.8841666666666667 | 0.004643128468533754 |
+| Epoch_11_batch_5999.pt | 0.8838333333333335 | 0.004919412290502857 |
+| Epoch_13_batch_5999.pt | 0.8836666666666666 | 0.004955356249106163 |
+| Epoch_17_batch_2999.pt | 0.883 | 0.0048419463487779845 |
+| Epoch_13.pt | 0.8826666666666666 | 0.005171145012542263 |
+| Epoch_12_batch_5999.pt | 0.8825 | 0.0035246048723470893 |
+| Epoch_17_batch_5999.pt | 0.8823333333333334 | 0.004368800722519442 |
+| Epoch_16_batch_2999.pt | 0.8821666666666668 | 0.005079916884709394 |
+| Epoch_12_batch_2999.pt | 0.8818333333333335 | 0.005640735877043689 |
+| Epoch_16.pt | 0.8815 | 0.004996603784843864 |
+| Epoch_16_batch_5999.pt | 0.8811666666666668 | 0.004900554264444525 |
+| Epoch_15.pt | 0.8801666666666668 | 0.0045368858629387335 |
+| Epoch_14.pt | 0.8801666666666665 | 0.004597704741377902 |
+| Epoch_13_batch_2999.pt | 0.8800000000000001 | 0.004733646312031285 |
+| Epoch_11_batch_2999.pt | 0.8791666666666667 | 0.004844813951249541 |
+| Epoch_12.pt | 0.8786666666666667 | 0.004579880854561098 |
+| Epoch_10_batch_5999.pt | 0.8781666666666667 | 0.004342227339926473 |
+| Epoch_11.pt | 0.8779999999999999 | 0.0060542812114778915 |
+| Epoch_10.pt | 0.8766666666666666 | 0.004238273582835707 |
+| Epoch_10_batch_2999.pt | 0.875 | 0.004975247372719576 |
+| Epoch_9_batch_5999.pt | 0.8550000000000001 | 0.0042236839574995985 |
+| Epoch_8_batch_2999.pt | 0.8523333333333334 | 0.005774571760082798 |
+| Epoch_6_batch_2999.pt | 0.8488333333333333 | 0.004267664813801922 |
+| Epoch_9_batch_2999.pt | 0.8486666666666667 | 0.004660048216780167 |
+| Epoch_8_batch_5999.pt | 0.8481666666666665 | 0.0052378607451119925 |
+| Epoch_7_batch_2999.pt | 0.8458333333333334 | 0.002973649709127253 |
+| Epoch_7_batch_5999.pt | 0.8416666666666666 | 0.004296136650929158 |
+| Epoch_8.pt | 0.8415000000000001 | 0.004277777777777781 |
+| Epoch_7.pt | 0.8408333333333333 | 0.00513550939928294 |
+| Epoch_4_batch_2999.pt | 0.8404999999999999 | 0.005264896561421614 |
+| Epoch_9.pt | 0.8400000000000001 | 0.0066989956866590535 |
+| Epoch_4_batch_5999.pt | 0.8398333333333332 | 0.004040306185683717 |
+| Epoch_5_batch_2999.pt | 0.8385 | 0.002349809553617403 |
+| Epoch_6_batch_5999.pt | 0.8361666666666666 | 0.004245913067618993 |
+| Epoch_3_batch_5999.pt | 0.8338333333333333 | 0.004727448088778849 |
+| Epoch_6.pt | 0.8320000000000001 | 0.003458751648060755 |
+| Epoch_5_batch_5999.pt | 0.8313333333333333 | 0.003950308629918044 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..51ecc8a2529a37c55ee030db0a1f68eaca602f8a
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,45 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.9536666666666667 | 0.0026851213274654675 |
+| Epoch_15_batch_5999.pt | 0.9536666666666667 | 0.002808716591058781 |
+| Epoch_16_batch_5999.pt | 0.9535 | 0.0027716599296147585 |
+| Epoch_17_batch_5999.pt | 0.9535 | 0.0027827732858596013 |
+| Epoch_17.pt | 0.9533333333333334 | 0.00234389145663656 |
+| Epoch_17_batch_2999.pt | 0.9523333333333334 | 0.002746490465401846 |
+| Epoch_16_batch_2999.pt | 0.952 | 0.0028952920490656403 |
+| Epoch_15_batch_2999.pt | 0.9518333333333334 | 0.002575405996999832 |
+| Epoch_15.pt | 0.9516666666666668 | 0.0030731814857642955 |
+| Epoch_13_batch_2999.pt | 0.9515 | 0.0028915585954829084 |
+| Epoch_13_batch_5999.pt | 0.9515 | 0.0024017740356908107 |
+| Epoch_13.pt | 0.9513333333333334 | 0.002650413431528128 |
+| Epoch_14.pt | 0.9513333333333334 | 0.002852332811776319 |
+| Epoch_16.pt | 0.9513333333333334 | 0.002905932629027119 |
+| Epoch_14_batch_2999.pt | 0.9511666666666667 | 0.0036434449849043595 |
+| Epoch_12.pt | 0.9493333333333334 | 0.003115076837528045 |
+| Epoch_11_batch_5999.pt | 0.9483333333333335 | 0.0028436630871266086 |
+| Epoch_10.pt | 0.9478333333333333 | 0.002699451247390294 |
+| Epoch_12_batch_2999.pt | 0.9476666666666667 | 0.002813108644704928 |
+| Epoch_10_batch_2999.pt | 0.9469999999999998 | 0.0031308895119123016 |
+| Epoch_11_batch_2999.pt | 0.9469999999999998 | 0.0032659863237109008 |
+| Epoch_11.pt | 0.946 | 0.0032413226990699617 |
+| Epoch_12_batch_5999.pt | 0.9456666666666667 | 0.0021401511426953554 |
+| Epoch_10_batch_5999.pt | 0.9445 | 0.0024222477062879567 |
+| Epoch_7_batch_2999.pt | 0.9259999999999999 | 0.003335184671067475 |
+| Epoch_9_batch_2999.pt | 0.9255000000000001 | 0.00310366171632521 |
+| Epoch_8_batch_5999.pt | 0.925 | 0.002545875386086577 |
+| Epoch_7.pt | 0.9246666666666667 | 0.003179797338056482 |
+| Epoch_5_batch_2999.pt | 0.9233333333333335 | 0.004395564543300551 |
+| Epoch_9_batch_5999.pt | 0.9224999999999998 | 0.001595711846260564 |
+| Epoch_6_batch_2999.pt | 0.9223333333333332 | 0.0032979604621457396 |
+| Epoch_8_batch_2999.pt | 0.9213333333333333 | 0.003237511618740773 |
+| Epoch_7_batch_5999.pt | 0.9189999999999999 | 0.003426476432246508 |
+| Epoch_8.pt | 0.9164999999999999 | 0.002551325000712229 |
+| Epoch_5_batch_5999.pt | 0.9156666666666666 | 0.0028131086447049218 |
+| Epoch_4_batch_5999.pt | 0.9153333333333332 | 0.0038151743807531995 |
+| Epoch_6_batch_5999.pt | 0.9148333333333334 | 0.0033742854753467133 |
+| Epoch_6.pt | 0.9143333333333332 | 0.0030550504633038962 |
+| Epoch_9.pt | 0.9138333333333334 | 0.004759979769642663 |
+| Epoch_3_batch_5999.pt | 0.9125 | 0.002724488852275389 |
+| Epoch_4_batch_2999.pt | 0.9103333333333333 | 0.003071172213574503 |
+| Epoch_5.pt | 0.9101666666666667 | 0.0031957726707265944 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3c61e4910568a001f8c32b6993f4d38a3de477e5
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,44 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.908 | 0.004163331998932263 |
+| Epoch_13.pt | 0.9078333333333333 | 0.0038091021592048776 |
+| Epoch_14_batch_2999.pt | 0.9073333333333334 | 0.003818408950542129 |
+| Epoch_15_batch_2999.pt | 0.9073333333333332 | 0.003644715437079271 |
+| Epoch_16.pt | 0.907 | 0.003440858348267113 |
+| Epoch_14.pt | 0.9066666666666666 | 0.003522414996477193 |
+| Epoch_12_batch_5999.pt | 0.9056666666666668 | 0.004982686072173039 |
+| Epoch_14_batch_5999.pt | 0.9056666666666666 | 0.003653173827283017 |
+| Epoch_16_batch_2999.pt | 0.9055000000000002 | 0.00426042657111952 |
+| Epoch_13_batch_5999.pt | 0.9055 | 0.004090413363155548 |
+| Epoch_12_batch_2999.pt | 0.9053333333333334 | 0.0034047896546765713 |
+| Epoch_15_batch_5999.pt | 0.9048333333333334 | 0.00425607771673218 |
+| Epoch_17_batch_2999.pt | 0.9046666666666667 | 0.004118611776189025 |
+| Epoch_11_batch_2999.pt | 0.9045 | 0.0029191788123154767 |
+| Epoch_17_batch_5999.pt | 0.9045 | 0.0037022682403435874 |
+| Epoch_10_batch_2999.pt | 0.9043333333333333 | 0.0036951753662924254 |
+| Epoch_10.pt | 0.9041666666666668 | 0.003164716748583581 |
+| Epoch_16_batch_5999.pt | 0.9041666666666666 | 0.004046412829917809 |
+| Epoch_11.pt | 0.9034999999999999 | 0.003621352997273831 |
+| Epoch_13_batch_2999.pt | 0.9030000000000001 | 0.0043659739343085546 |
+| Epoch_12.pt | 0.9026666666666665 | 0.003753187945345448 |
+| Epoch_10_batch_5999.pt | 0.9019999999999999 | 0.003749897117930264 |
+| Epoch_11_batch_5999.pt | 0.9018333333333333 | 0.002986078811194815 |
+| Epoch_15.pt | 0.9001666666666667 | 0.004017323597731312 |
+| Epoch_8_batch_2999.pt | 0.8823333333333332 | 0.001613982116259329 |
+| Epoch_5_batch_2999.pt | 0.8813333333333333 | 0.0035294178165041303 |
+| Epoch_7_batch_2999.pt | 0.8806666666666667 | 0.004575835618216443 |
+| Epoch_6_batch_5999.pt | 0.8805 | 0.003967849185704245 |
+| Epoch_9_batch_5999.pt | 0.8803333333333333 | 0.003431876713662331 |
+| Epoch_5_batch_5999.pt | 0.8781666666666667 | 0.002214570158666091 |
+| Epoch_7_batch_5999.pt | 0.8779999999999999 | 0.003725123247608937 |
+| Epoch_6_batch_2999.pt | 0.8763333333333332 | 0.004373037478700985 |
+| Epoch_7.pt | 0.8753333333333334 | 0.003431876713662335 |
+| Epoch_8_batch_5999.pt | 0.8746666666666666 | 0.004155912046503341 |
+| Epoch_9.pt | 0.8741666666666665 | 0.004008093663428393 |
+| Epoch_9_batch_2999.pt | 0.8709999999999999 | 0.004982686072173035 |
+| Epoch_6.pt | 0.8700000000000001 | 0.004029214303215277 |
+| Epoch_3_batch_5999.pt | 0.8699999999999999 | 0.005055250296034369 |
+| Epoch_5.pt | 0.8695 | 0.004621808207954014 |
+| Epoch_4_batch_2999.pt | 0.8678333333333335 | 0.0046084329571978895 |
+| Epoch_4_batch_5999.pt | 0.8643333333333333 | 0.0024993826398226654 |
diff --git a/bob/bio/facexzoo/models/backbones/GhostNet/log.log b/bob/bio/facexzoo/models/backbones/GhostNet/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..ef2803a45b7380fd147a60ce716c7ee82e000182
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/GhostNet/log.log
@@ -0,0 +1,655 @@
+INFO 2021-01-21 20:12:59 train.py: 176] Start optimization.
+INFO 2021-01-21 20:12:59 train.py: 177] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='GhostNet', batch_size=512, data_root='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-resnet', train_file='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msceleb_deepglint_train_file.txt1', writer=<tensorboardX.writer.SummaryWriter object at 0x7f35ace577f0>)
+backbone param:
+{'width': 1.0, 'drop_ratio': 0.2, 'out_h': 7, 'out_w': 7, 'feat_dim': 512}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2021-01-21 20:13:32 train.py: 78] Epoch 0, iter 0/6416, lr 0.100000, loss 16.250774
+INFO 2021-01-21 20:25:19 train.py: 78] Epoch 0, iter 200/6416, lr 0.100000, loss 15.764360
+INFO 2021-01-21 20:36:01 train.py: 78] Epoch 0, iter 400/6416, lr 0.100000, loss 15.331061
+INFO 2021-01-21 20:46:33 train.py: 78] Epoch 0, iter 600/6416, lr 0.100000, loss 15.235012
+INFO 2021-01-21 20:55:56 train.py: 78] Epoch 0, iter 800/6416, lr 0.100000, loss 15.091175
+INFO 2021-01-21 21:05:38 train.py: 78] Epoch 0, iter 1000/6416, lr 0.100000, loss 14.853463
+INFO 2021-01-21 21:15:24 train.py: 78] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.560143
+INFO 2021-01-21 21:24:43 train.py: 78] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.265671
+INFO 2021-01-21 21:34:31 train.py: 78] Epoch 0, iter 1600/6416, lr 0.100000, loss 13.921368
+INFO 2021-01-21 21:44:24 train.py: 78] Epoch 0, iter 1800/6416, lr 0.100000, loss 13.560656
+INFO 2021-01-21 21:53:54 train.py: 78] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.195586
+INFO 2021-01-21 22:03:30 train.py: 78] Epoch 0, iter 2200/6416, lr 0.100000, loss 12.816410
+INFO 2021-01-21 22:12:59 train.py: 78] Epoch 0, iter 2400/6416, lr 0.100000, loss 12.464761
+INFO 2021-01-21 22:22:24 train.py: 78] Epoch 0, iter 2600/6416, lr 0.100000, loss 12.151867
+INFO 2021-01-21 22:31:43 train.py: 78] Epoch 0, iter 2800/6416, lr 0.100000, loss 11.947212
+INFO 2021-01-21 22:41:37 train.py: 91] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-01-21 22:41:38 train.py: 78] Epoch 0, iter 3000/6416, lr 0.100000, loss 11.872883
+INFO 2021-01-21 22:51:55 train.py: 78] Epoch 0, iter 3200/6416, lr 0.100000, loss 11.948144
+INFO 2021-01-21 23:01:23 train.py: 78] Epoch 0, iter 3400/6416, lr 0.100000, loss 12.136700
+INFO 2021-01-21 23:11:01 train.py: 78] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.427660
+INFO 2021-01-21 23:20:49 train.py: 78] Epoch 0, iter 3800/6416, lr 0.100000, loss 12.762948
+INFO 2021-01-21 23:29:55 train.py: 78] Epoch 0, iter 4000/6416, lr 0.100000, loss 13.112913
+INFO 2021-01-21 23:39:00 train.py: 78] Epoch 0, iter 4200/6416, lr 0.100000, loss 13.417453
+INFO 2021-01-21 23:48:48 train.py: 78] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.704103
+INFO 2021-01-21 23:57:52 train.py: 78] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.941063
+INFO 2021-01-22 00:07:10 train.py: 78] Epoch 0, iter 4800/6416, lr 0.100000, loss 14.112939
+INFO 2021-01-22 00:15:55 train.py: 78] Epoch 0, iter 5000/6416, lr 0.100000, loss 14.235073
+INFO 2021-01-22 00:23:58 train.py: 78] Epoch 0, iter 5200/6416, lr 0.100000, loss 14.305190
+INFO 2021-01-22 00:33:30 train.py: 78] Epoch 0, iter 5400/6416, lr 0.100000, loss 14.280206
+INFO 2021-01-22 00:42:12 train.py: 78] Epoch 0, iter 5600/6416, lr 0.100000, loss 14.278833
+INFO 2021-01-22 00:52:18 train.py: 78] Epoch 0, iter 5800/6416, lr 0.100000, loss 14.196568
+INFO 2021-01-22 01:01:32 train.py: 91] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-01-22 01:01:33 train.py: 78] Epoch 0, iter 6000/6416, lr 0.100000, loss 14.100001
+INFO 2021-01-22 01:10:58 train.py: 78] Epoch 0, iter 6200/6416, lr 0.100000, loss 13.986286
+INFO 2021-01-22 01:19:56 train.py: 78] Epoch 0, iter 6400/6416, lr 0.100000, loss 13.823606
+INFO 2021-01-22 01:20:28 train.py: 96] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-01-22 01:20:30 train.py: 78] Epoch 1, iter 0/6416, lr 0.100000, loss 13.770247
+INFO 2021-01-22 01:22:35 train.py: 78] Epoch 1, iter 200/6416, lr 0.100000, loss 13.470351
+INFO 2021-01-22 01:24:38 train.py: 78] Epoch 1, iter 400/6416, lr 0.100000, loss 13.320353
+INFO 2021-01-22 01:26:42 train.py: 78] Epoch 1, iter 600/6416, lr 0.100000, loss 13.157153
+INFO 2021-01-22 01:28:46 train.py: 78] Epoch 1, iter 800/6416, lr 0.100000, loss 12.991815
+INFO 2021-01-22 01:30:49 train.py: 78] Epoch 1, iter 1000/6416, lr 0.100000, loss 12.854444
+INFO 2021-01-22 01:32:53 train.py: 78] Epoch 1, iter 1200/6416, lr 0.100000, loss 12.679881
+INFO 2021-01-22 01:34:56 train.py: 78] Epoch 1, iter 1400/6416, lr 0.100000, loss 12.565455
+INFO 2021-01-22 01:36:59 train.py: 78] Epoch 1, iter 1600/6416, lr 0.100000, loss 12.393085
+INFO 2021-01-22 01:39:03 train.py: 78] Epoch 1, iter 1800/6416, lr 0.100000, loss 12.235998
+INFO 2021-01-22 01:41:06 train.py: 78] Epoch 1, iter 2000/6416, lr 0.100000, loss 12.078922
+INFO 2021-01-22 01:43:09 train.py: 78] Epoch 1, iter 2200/6416, lr 0.100000, loss 11.955585
+INFO 2021-01-22 01:45:13 train.py: 78] Epoch 1, iter 2400/6416, lr 0.100000, loss 11.815526
+INFO 2021-01-22 01:47:16 train.py: 78] Epoch 1, iter 2600/6416, lr 0.100000, loss 11.653279
+INFO 2021-01-22 01:49:19 train.py: 78] Epoch 1, iter 2800/6416, lr 0.100000, loss 11.547822
+INFO 2021-01-22 01:51:22 train.py: 91] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-01-22 01:51:23 train.py: 78] Epoch 1, iter 3000/6416, lr 0.100000, loss 11.399225
+INFO 2021-01-22 01:53:26 train.py: 78] Epoch 1, iter 3200/6416, lr 0.100000, loss 11.332256
+INFO 2021-01-22 01:55:29 train.py: 78] Epoch 1, iter 3400/6416, lr 0.100000, loss 11.179480
+INFO 2021-01-22 01:57:33 train.py: 78] Epoch 1, iter 3600/6416, lr 0.100000, loss 11.079146
+INFO 2021-01-22 01:59:37 train.py: 78] Epoch 1, iter 3800/6416, lr 0.100000, loss 11.000047
+INFO 2021-01-22 02:01:40 train.py: 78] Epoch 1, iter 4000/6416, lr 0.100000, loss 10.885364
+INFO 2021-01-22 02:03:44 train.py: 78] Epoch 1, iter 4200/6416, lr 0.100000, loss 10.799798
+INFO 2021-01-22 02:05:47 train.py: 78] Epoch 1, iter 4400/6416, lr 0.100000, loss 10.700391
+INFO 2021-01-22 02:07:51 train.py: 78] Epoch 1, iter 4600/6416, lr 0.100000, loss 10.611290
+INFO 2021-01-22 02:09:54 train.py: 78] Epoch 1, iter 4800/6416, lr 0.100000, loss 10.554503
+INFO 2021-01-22 02:11:58 train.py: 78] Epoch 1, iter 5000/6416, lr 0.100000, loss 10.446805
+INFO 2021-01-22 02:14:01 train.py: 78] Epoch 1, iter 5200/6416, lr 0.100000, loss 10.398536
+INFO 2021-01-22 02:16:05 train.py: 78] Epoch 1, iter 5400/6416, lr 0.100000, loss 10.320798
+INFO 2021-01-22 02:18:08 train.py: 78] Epoch 1, iter 5600/6416, lr 0.100000, loss 10.229470
+INFO 2021-01-22 02:20:12 train.py: 78] Epoch 1, iter 5800/6416, lr 0.100000, loss 10.170319
+INFO 2021-01-22 02:22:15 train.py: 91] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-01-22 02:22:16 train.py: 78] Epoch 1, iter 6000/6416, lr 0.100000, loss 10.089223
+INFO 2021-01-22 02:24:20 train.py: 78] Epoch 1, iter 6200/6416, lr 0.100000, loss 10.078914
+INFO 2021-01-22 02:26:23 train.py: 78] Epoch 1, iter 6400/6416, lr 0.100000, loss 10.004797
+INFO 2021-01-22 02:26:33 train.py: 96] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-01-22 02:26:34 train.py: 78] Epoch 2, iter 0/6416, lr 0.100000, loss 10.012711
+INFO 2021-01-22 02:28:38 train.py: 78] Epoch 2, iter 200/6416, lr 0.100000, loss 9.402851
+INFO 2021-01-22 02:30:42 train.py: 78] Epoch 2, iter 400/6416, lr 0.100000, loss 9.393676
+INFO 2021-01-22 02:32:45 train.py: 78] Epoch 2, iter 600/6416, lr 0.100000, loss 9.449439
+INFO 2021-01-22 02:34:48 train.py: 78] Epoch 2, iter 800/6416, lr 0.100000, loss 9.473360
+INFO 2021-01-22 02:36:52 train.py: 78] Epoch 2, iter 1000/6416, lr 0.100000, loss 9.482772
+INFO 2021-01-22 02:38:55 train.py: 78] Epoch 2, iter 1200/6416, lr 0.100000, loss 9.489313
+INFO 2021-01-22 02:40:58 train.py: 78] Epoch 2, iter 1400/6416, lr 0.100000, loss 9.495777
+INFO 2021-01-22 02:43:01 train.py: 78] Epoch 2, iter 1600/6416, lr 0.100000, loss 9.471175
+INFO 2021-01-22 02:45:05 train.py: 78] Epoch 2, iter 1800/6416, lr 0.100000, loss 9.406645
+INFO 2021-01-22 02:47:08 train.py: 78] Epoch 2, iter 2000/6416, lr 0.100000, loss 9.389534
+INFO 2021-01-22 02:49:11 train.py: 78] Epoch 2, iter 2200/6416, lr 0.100000, loss 9.421353
+INFO 2021-01-22 02:51:14 train.py: 78] Epoch 2, iter 2400/6416, lr 0.100000, loss 9.368639
+INFO 2021-01-22 02:53:17 train.py: 78] Epoch 2, iter 2600/6416, lr 0.100000, loss 9.315423
+INFO 2021-01-22 02:55:21 train.py: 78] Epoch 2, iter 2800/6416, lr 0.100000, loss 9.281135
+INFO 2021-01-22 02:57:24 train.py: 91] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-01-22 02:57:25 train.py: 78] Epoch 2, iter 3000/6416, lr 0.100000, loss 9.246886
+INFO 2021-01-22 02:59:28 train.py: 78] Epoch 2, iter 3200/6416, lr 0.100000, loss 9.243442
+INFO 2021-01-22 03:01:31 train.py: 78] Epoch 2, iter 3400/6416, lr 0.100000, loss 9.188717
+INFO 2021-01-22 03:03:35 train.py: 78] Epoch 2, iter 3600/6416, lr 0.100000, loss 9.148053
+INFO 2021-01-22 03:05:38 train.py: 78] Epoch 2, iter 3800/6416, lr 0.100000, loss 9.152092
+INFO 2021-01-22 03:07:42 train.py: 78] Epoch 2, iter 4000/6416, lr 0.100000, loss 9.113306
+INFO 2021-01-22 03:09:45 train.py: 78] Epoch 2, iter 4200/6416, lr 0.100000, loss 9.050850
+INFO 2021-01-22 03:11:48 train.py: 78] Epoch 2, iter 4400/6416, lr 0.100000, loss 9.076974
+INFO 2021-01-22 03:13:52 train.py: 78] Epoch 2, iter 4600/6416, lr 0.100000, loss 9.038068
+INFO 2021-01-22 03:15:55 train.py: 78] Epoch 2, iter 4800/6416, lr 0.100000, loss 8.987925
+INFO 2021-01-22 03:17:59 train.py: 78] Epoch 2, iter 5000/6416, lr 0.100000, loss 8.973293
+INFO 2021-01-22 03:20:02 train.py: 78] Epoch 2, iter 5200/6416, lr 0.100000, loss 8.931205
+INFO 2021-01-22 03:22:06 train.py: 78] Epoch 2, iter 5400/6416, lr 0.100000, loss 8.921868
+INFO 2021-01-22 03:24:09 train.py: 78] Epoch 2, iter 5600/6416, lr 0.100000, loss 8.861114
+INFO 2021-01-22 03:26:12 train.py: 78] Epoch 2, iter 5800/6416, lr 0.100000, loss 8.842090
+INFO 2021-01-22 03:28:15 train.py: 91] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-01-22 03:28:16 train.py: 78] Epoch 2, iter 6000/6416, lr 0.100000, loss 8.846736
+INFO 2021-01-22 03:30:19 train.py: 78] Epoch 2, iter 6200/6416, lr 0.100000, loss 8.801316
+INFO 2021-01-22 03:32:23 train.py: 78] Epoch 2, iter 6400/6416, lr 0.100000, loss 8.772921
+INFO 2021-01-22 03:32:33 train.py: 96] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-01-22 03:32:34 train.py: 78] Epoch 3, iter 0/6416, lr 0.100000, loss 8.785591
+INFO 2021-01-22 03:34:38 train.py: 78] Epoch 3, iter 200/6416, lr 0.100000, loss 8.233217
+INFO 2021-01-22 03:36:42 train.py: 78] Epoch 3, iter 400/6416, lr 0.100000, loss 8.247045
+INFO 2021-01-22 03:38:45 train.py: 78] Epoch 3, iter 600/6416, lr 0.100000, loss 8.277921
+INFO 2021-01-22 03:40:48 train.py: 78] Epoch 3, iter 800/6416, lr 0.100000, loss 8.360723
+INFO 2021-01-22 03:42:51 train.py: 78] Epoch 3, iter 1000/6416, lr 0.100000, loss 8.408463
+INFO 2021-01-22 03:44:54 train.py: 78] Epoch 3, iter 1200/6416, lr 0.100000, loss 8.444841
+INFO 2021-01-22 03:46:57 train.py: 78] Epoch 3, iter 1400/6416, lr 0.100000, loss 8.453999
+INFO 2021-01-22 03:49:00 train.py: 78] Epoch 3, iter 1600/6416, lr 0.100000, loss 8.438502
+INFO 2021-01-22 03:51:03 train.py: 78] Epoch 3, iter 1800/6416, lr 0.100000, loss 8.429155
+INFO 2021-01-22 03:53:06 train.py: 78] Epoch 3, iter 2000/6416, lr 0.100000, loss 8.441304
+INFO 2021-01-22 03:55:09 train.py: 78] Epoch 3, iter 2200/6416, lr 0.100000, loss 8.450672
+INFO 2021-01-22 03:57:12 train.py: 78] Epoch 3, iter 2400/6416, lr 0.100000, loss 8.463969
+INFO 2021-01-22 03:59:15 train.py: 78] Epoch 3, iter 2600/6416, lr 0.100000, loss 8.451363
+INFO 2021-01-22 04:01:19 train.py: 78] Epoch 3, iter 2800/6416, lr 0.100000, loss 8.440834
+INFO 2021-01-22 04:03:21 train.py: 91] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-01-22 04:03:22 train.py: 78] Epoch 3, iter 3000/6416, lr 0.100000, loss 8.449899
+INFO 2021-01-22 04:05:25 train.py: 78] Epoch 3, iter 3200/6416, lr 0.100000, loss 8.414726
+INFO 2021-01-22 04:07:28 train.py: 78] Epoch 3, iter 3400/6416, lr 0.100000, loss 8.380796
+INFO 2021-01-22 04:09:31 train.py: 78] Epoch 3, iter 3600/6416, lr 0.100000, loss 8.378931
+INFO 2021-01-22 04:11:35 train.py: 78] Epoch 3, iter 3800/6416, lr 0.100000, loss 8.358698
+INFO 2021-01-22 04:13:38 train.py: 78] Epoch 3, iter 4000/6416, lr 0.100000, loss 8.365002
+INFO 2021-01-22 04:15:41 train.py: 78] Epoch 3, iter 4200/6416, lr 0.100000, loss 8.348931
+INFO 2021-01-22 04:17:44 train.py: 78] Epoch 3, iter 4400/6416, lr 0.100000, loss 8.329774
+INFO 2021-01-22 04:19:47 train.py: 78] Epoch 3, iter 4600/6416, lr 0.100000, loss 8.308526
+INFO 2021-01-22 04:21:50 train.py: 78] Epoch 3, iter 4800/6416, lr 0.100000, loss 8.303235
+INFO 2021-01-22 04:23:53 train.py: 78] Epoch 3, iter 5000/6416, lr 0.100000, loss 8.269777
+INFO 2021-01-22 04:25:57 train.py: 78] Epoch 3, iter 5200/6416, lr 0.100000, loss 8.280975
+INFO 2021-01-22 04:28:00 train.py: 78] Epoch 3, iter 5400/6416, lr 0.100000, loss 8.265018
+INFO 2021-01-22 04:30:03 train.py: 78] Epoch 3, iter 5600/6416, lr 0.100000, loss 8.241350
+INFO 2021-01-22 04:32:06 train.py: 78] Epoch 3, iter 5800/6416, lr 0.100000, loss 8.202742
+INFO 2021-01-22 04:34:09 train.py: 91] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-01-22 04:34:09 train.py: 78] Epoch 3, iter 6000/6416, lr 0.100000, loss 8.204603
+INFO 2021-01-22 04:36:12 train.py: 78] Epoch 3, iter 6200/6416, lr 0.100000, loss 8.211601
+INFO 2021-01-22 04:38:15 train.py: 78] Epoch 3, iter 6400/6416, lr 0.100000, loss 8.189393
+INFO 2021-01-22 04:38:25 train.py: 96] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-01-22 04:38:27 train.py: 78] Epoch 4, iter 0/6416, lr 0.100000, loss 8.182916
+INFO 2021-01-22 04:40:30 train.py: 78] Epoch 4, iter 200/6416, lr 0.100000, loss 7.598081
+INFO 2021-01-22 04:42:34 train.py: 78] Epoch 4, iter 400/6416, lr 0.100000, loss 7.644563
+INFO 2021-01-22 04:44:37 train.py: 78] Epoch 4, iter 600/6416, lr 0.100000, loss 7.719511
+INFO 2021-01-22 04:46:40 train.py: 78] Epoch 4, iter 800/6416, lr 0.100000, loss 7.801908
+INFO 2021-01-22 04:48:43 train.py: 78] Epoch 4, iter 1000/6416, lr 0.100000, loss 7.892504
+INFO 2021-01-22 04:50:46 train.py: 78] Epoch 4, iter 1200/6416, lr 0.100000, loss 7.906984
+INFO 2021-01-22 04:52:49 train.py: 78] Epoch 4, iter 1400/6416, lr 0.100000, loss 7.933156
+INFO 2021-01-22 04:54:53 train.py: 78] Epoch 4, iter 1600/6416, lr 0.100000, loss 7.944929
+INFO 2021-01-22 04:56:55 train.py: 78] Epoch 4, iter 1800/6416, lr 0.100000, loss 7.953894
+INFO 2021-01-22 04:58:58 train.py: 78] Epoch 4, iter 2000/6416, lr 0.100000, loss 7.948719
+INFO 2021-01-22 05:01:02 train.py: 78] Epoch 4, iter 2200/6416, lr 0.100000, loss 7.990220
+INFO 2021-01-22 05:03:05 train.py: 78] Epoch 4, iter 2400/6416, lr 0.100000, loss 7.968251
+INFO 2021-01-22 05:05:08 train.py: 78] Epoch 4, iter 2600/6416, lr 0.100000, loss 7.954469
+INFO 2021-01-22 05:07:11 train.py: 78] Epoch 4, iter 2800/6416, lr 0.100000, loss 7.949867
+INFO 2021-01-22 05:09:14 train.py: 91] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-01-22 05:09:15 train.py: 78] Epoch 4, iter 3000/6416, lr 0.100000, loss 7.924289
+INFO 2021-01-22 05:11:18 train.py: 78] Epoch 4, iter 3200/6416, lr 0.100000, loss 7.950249
+INFO 2021-01-22 05:13:21 train.py: 78] Epoch 4, iter 3400/6416, lr 0.100000, loss 7.953858
+INFO 2021-01-22 05:15:25 train.py: 78] Epoch 4, iter 3600/6416, lr 0.100000, loss 7.936049
+INFO 2021-01-22 05:17:28 train.py: 78] Epoch 4, iter 3800/6416, lr 0.100000, loss 7.930690
+INFO 2021-01-22 05:19:31 train.py: 78] Epoch 4, iter 4000/6416, lr 0.100000, loss 7.934540
+INFO 2021-01-22 05:21:35 train.py: 78] Epoch 4, iter 4200/6416, lr 0.100000, loss 7.934243
+INFO 2021-01-22 05:23:38 train.py: 78] Epoch 4, iter 4400/6416, lr 0.100000, loss 7.884119
+INFO 2021-01-22 05:25:41 train.py: 78] Epoch 4, iter 4600/6416, lr 0.100000, loss 7.863164
+INFO 2021-01-22 05:27:44 train.py: 78] Epoch 4, iter 4800/6416, lr 0.100000, loss 7.864212
+INFO 2021-01-22 05:29:48 train.py: 78] Epoch 4, iter 5000/6416, lr 0.100000, loss 7.869936
+INFO 2021-01-22 05:31:51 train.py: 78] Epoch 4, iter 5200/6416, lr 0.100000, loss 7.847592
+INFO 2021-01-22 05:33:54 train.py: 78] Epoch 4, iter 5400/6416, lr 0.100000, loss 7.844493
+INFO 2021-01-22 05:35:58 train.py: 78] Epoch 4, iter 5600/6416, lr 0.100000, loss 7.818415
+INFO 2021-01-22 05:38:01 train.py: 78] Epoch 4, iter 5800/6416, lr 0.100000, loss 7.784891
+INFO 2021-01-22 05:40:04 train.py: 91] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-01-22 05:40:04 train.py: 78] Epoch 4, iter 6000/6416, lr 0.100000, loss 7.840141
+INFO 2021-01-22 05:42:08 train.py: 78] Epoch 4, iter 6200/6416, lr 0.100000, loss 7.777532
+INFO 2021-01-22 05:44:11 train.py: 78] Epoch 4, iter 6400/6416, lr 0.100000, loss 7.802946
+INFO 2021-01-22 05:44:21 train.py: 96] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-01-22 05:44:22 train.py: 78] Epoch 5, iter 0/6416, lr 0.100000, loss 7.888285
+INFO 2021-01-22 05:46:26 train.py: 78] Epoch 5, iter 200/6416, lr 0.100000, loss 7.297671
+INFO 2021-01-22 05:48:29 train.py: 78] Epoch 5, iter 400/6416, lr 0.100000, loss 7.319428
+INFO 2021-01-22 05:50:32 train.py: 78] Epoch 5, iter 600/6416, lr 0.100000, loss 7.379807
+INFO 2021-01-22 05:52:35 train.py: 78] Epoch 5, iter 800/6416, lr 0.100000, loss 7.438146
+INFO 2021-01-22 05:54:38 train.py: 78] Epoch 5, iter 1000/6416, lr 0.100000, loss 7.508583
+INFO 2021-01-22 05:56:41 train.py: 78] Epoch 5, iter 1200/6416, lr 0.100000, loss 7.532811
+INFO 2021-01-22 05:58:44 train.py: 78] Epoch 5, iter 1400/6416, lr 0.100000, loss 7.546482
+INFO 2021-01-22 06:00:47 train.py: 78] Epoch 5, iter 1600/6416, lr 0.100000, loss 7.591708
+INFO 2021-01-22 06:02:50 train.py: 78] Epoch 5, iter 1800/6416, lr 0.100000, loss 7.594268
+INFO 2021-01-22 06:04:53 train.py: 78] Epoch 5, iter 2000/6416, lr 0.100000, loss 7.642017
+INFO 2021-01-22 06:06:56 train.py: 78] Epoch 5, iter 2200/6416, lr 0.100000, loss 7.606971
+INFO 2021-01-22 06:08:59 train.py: 78] Epoch 5, iter 2400/6416, lr 0.100000, loss 7.602692
+INFO 2021-01-22 06:11:02 train.py: 78] Epoch 5, iter 2600/6416, lr 0.100000, loss 7.594610
+INFO 2021-01-22 06:13:05 train.py: 78] Epoch 5, iter 2800/6416, lr 0.100000, loss 7.628082
+INFO 2021-01-22 06:15:08 train.py: 91] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-01-22 06:15:09 train.py: 78] Epoch 5, iter 3000/6416, lr 0.100000, loss 7.634071
+INFO 2021-01-22 06:17:11 train.py: 78] Epoch 5, iter 3200/6416, lr 0.100000, loss 7.652492
+INFO 2021-01-22 06:19:15 train.py: 78] Epoch 5, iter 3400/6416, lr 0.100000, loss 7.627753
+INFO 2021-01-22 06:21:18 train.py: 78] Epoch 5, iter 3600/6416, lr 0.100000, loss 7.628190
+INFO 2021-01-22 06:23:21 train.py: 78] Epoch 5, iter 3800/6416, lr 0.100000, loss 7.608053
+INFO 2021-01-22 06:25:24 train.py: 78] Epoch 5, iter 4000/6416, lr 0.100000, loss 7.611481
+INFO 2021-01-22 06:27:27 train.py: 78] Epoch 5, iter 4200/6416, lr 0.100000, loss 7.582783
+INFO 2021-01-22 06:29:30 train.py: 78] Epoch 5, iter 4400/6416, lr 0.100000, loss 7.607936
+INFO 2021-01-22 06:31:33 train.py: 78] Epoch 5, iter 4600/6416, lr 0.100000, loss 7.568005
+INFO 2021-01-22 06:33:36 train.py: 78] Epoch 5, iter 4800/6416, lr 0.100000, loss 7.613409
+INFO 2021-01-22 06:35:39 train.py: 78] Epoch 5, iter 5000/6416, lr 0.100000, loss 7.579131
+INFO 2021-01-22 06:37:42 train.py: 78] Epoch 5, iter 5200/6416, lr 0.100000, loss 7.563879
+INFO 2021-01-22 06:39:45 train.py: 78] Epoch 5, iter 5400/6416, lr 0.100000, loss 7.588154
+INFO 2021-01-22 06:41:48 train.py: 78] Epoch 5, iter 5600/6416, lr 0.100000, loss 7.568004
+INFO 2021-01-22 06:43:51 train.py: 78] Epoch 5, iter 5800/6416, lr 0.100000, loss 7.600856
+INFO 2021-01-22 06:45:54 train.py: 91] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-01-22 06:45:55 train.py: 78] Epoch 5, iter 6000/6416, lr 0.100000, loss 7.565823
+INFO 2021-01-22 06:47:58 train.py: 78] Epoch 5, iter 6200/6416, lr 0.100000, loss 7.553977
+INFO 2021-01-22 06:50:01 train.py: 78] Epoch 5, iter 6400/6416, lr 0.100000, loss 7.539934
+INFO 2021-01-22 06:50:10 train.py: 96] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-01-22 06:50:12 train.py: 78] Epoch 6, iter 0/6416, lr 0.100000, loss 7.630221
+INFO 2021-01-22 06:52:16 train.py: 78] Epoch 6, iter 200/6416, lr 0.100000, loss 7.020682
+INFO 2021-01-22 06:54:18 train.py: 78] Epoch 6, iter 400/6416, lr 0.100000, loss 7.017615
+INFO 2021-01-22 06:56:21 train.py: 78] Epoch 6, iter 600/6416, lr 0.100000, loss 7.125110
+INFO 2021-01-22 06:58:24 train.py: 78] Epoch 6, iter 800/6416, lr 0.100000, loss 7.177220
+INFO 2021-01-22 07:00:28 train.py: 78] Epoch 6, iter 1000/6416, lr 0.100000, loss 7.237925
+INFO 2021-01-22 07:02:31 train.py: 78] Epoch 6, iter 1200/6416, lr 0.100000, loss 7.311243
+INFO 2021-01-22 07:04:34 train.py: 78] Epoch 6, iter 1400/6416, lr 0.100000, loss 7.315162
+INFO 2021-01-22 07:06:37 train.py: 78] Epoch 6, iter 1600/6416, lr 0.100000, loss 7.330540
+INFO 2021-01-22 07:08:40 train.py: 78] Epoch 6, iter 1800/6416, lr 0.100000, loss 7.378689
+INFO 2021-01-22 07:10:43 train.py: 78] Epoch 6, iter 2000/6416, lr 0.100000, loss 7.363947
+INFO 2021-01-22 07:12:46 train.py: 78] Epoch 6, iter 2200/6416, lr 0.100000, loss 7.388986
+INFO 2021-01-22 07:14:49 train.py: 78] Epoch 6, iter 2400/6416, lr 0.100000, loss 7.385360
+INFO 2021-01-22 07:16:51 train.py: 78] Epoch 6, iter 2600/6416, lr 0.100000, loss 7.364944
+INFO 2021-01-22 07:18:54 train.py: 78] Epoch 6, iter 2800/6416, lr 0.100000, loss 7.404275
+INFO 2021-01-22 07:20:57 train.py: 91] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-01-22 07:20:58 train.py: 78] Epoch 6, iter 3000/6416, lr 0.100000, loss 7.404556
+INFO 2021-01-22 07:23:00 train.py: 78] Epoch 6, iter 3200/6416, lr 0.100000, loss 7.431760
+INFO 2021-01-22 07:25:03 train.py: 78] Epoch 6, iter 3400/6416, lr 0.100000, loss 7.410710
+INFO 2021-01-22 07:27:06 train.py: 78] Epoch 6, iter 3600/6416, lr 0.100000, loss 7.411103
+INFO 2021-01-22 07:29:09 train.py: 78] Epoch 6, iter 3800/6416, lr 0.100000, loss 7.401692
+INFO 2021-01-22 07:31:12 train.py: 78] Epoch 6, iter 4000/6416, lr 0.100000, loss 7.382090
+INFO 2021-01-22 07:33:15 train.py: 78] Epoch 6, iter 4200/6416, lr 0.100000, loss 7.397586
+INFO 2021-01-22 07:35:18 train.py: 78] Epoch 6, iter 4400/6416, lr 0.100000, loss 7.387713
+INFO 2021-01-22 07:37:21 train.py: 78] Epoch 6, iter 4600/6416, lr 0.100000, loss 7.387556
+INFO 2021-01-22 07:39:24 train.py: 78] Epoch 6, iter 4800/6416, lr 0.100000, loss 7.406343
+INFO 2021-01-22 07:41:27 train.py: 78] Epoch 6, iter 5000/6416, lr 0.100000, loss 7.386101
+INFO 2021-01-22 07:43:29 train.py: 78] Epoch 6, iter 5200/6416, lr 0.100000, loss 7.406947
+INFO 2021-01-22 07:45:32 train.py: 78] Epoch 6, iter 5400/6416, lr 0.100000, loss 7.365710
+INFO 2021-01-22 07:47:35 train.py: 78] Epoch 6, iter 5600/6416, lr 0.100000, loss 7.355272
+INFO 2021-01-22 07:49:38 train.py: 78] Epoch 6, iter 5800/6416, lr 0.100000, loss 7.353986
+INFO 2021-01-22 07:51:41 train.py: 91] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-01-22 07:51:41 train.py: 78] Epoch 6, iter 6000/6416, lr 0.100000, loss 7.374732
+INFO 2021-01-22 07:53:44 train.py: 78] Epoch 6, iter 6200/6416, lr 0.100000, loss 7.370151
+INFO 2021-01-22 07:55:47 train.py: 78] Epoch 6, iter 6400/6416, lr 0.100000, loss 7.330472
+INFO 2021-01-22 07:55:57 train.py: 96] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-01-22 07:55:58 train.py: 78] Epoch 7, iter 0/6416, lr 0.100000, loss 7.303059
+INFO 2021-01-22 07:58:02 train.py: 78] Epoch 7, iter 200/6416, lr 0.100000, loss 6.846964
+INFO 2021-01-22 08:00:05 train.py: 78] Epoch 7, iter 400/6416, lr 0.100000, loss 6.844414
+INFO 2021-01-22 08:02:08 train.py: 78] Epoch 7, iter 600/6416, lr 0.100000, loss 6.915306
+INFO 2021-01-22 08:04:10 train.py: 78] Epoch 7, iter 800/6416, lr 0.100000, loss 6.987713
+INFO 2021-01-22 08:06:14 train.py: 78] Epoch 7, iter 1000/6416, lr 0.100000, loss 7.036404
+INFO 2021-01-22 08:08:16 train.py: 78] Epoch 7, iter 1200/6416, lr 0.100000, loss 7.103609
+INFO 2021-01-22 08:10:19 train.py: 78] Epoch 7, iter 1400/6416, lr 0.100000, loss 7.162308
+INFO 2021-01-22 08:12:22 train.py: 78] Epoch 7, iter 1600/6416, lr 0.100000, loss 7.186649
+INFO 2021-01-22 08:14:25 train.py: 78] Epoch 7, iter 1800/6416, lr 0.100000, loss 7.188459
+INFO 2021-01-22 08:16:28 train.py: 78] Epoch 7, iter 2000/6416, lr 0.100000, loss 7.189919
+INFO 2021-01-22 08:18:31 train.py: 78] Epoch 7, iter 2200/6416, lr 0.100000, loss 7.201545
+INFO 2021-01-22 08:20:34 train.py: 78] Epoch 7, iter 2400/6416, lr 0.100000, loss 7.237635
+INFO 2021-01-22 08:22:37 train.py: 78] Epoch 7, iter 2600/6416, lr 0.100000, loss 7.207448
+INFO 2021-01-22 08:24:40 train.py: 78] Epoch 7, iter 2800/6416, lr 0.100000, loss 7.245673
+INFO 2021-01-22 08:26:43 train.py: 91] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-01-22 08:26:43 train.py: 78] Epoch 7, iter 3000/6416, lr 0.100000, loss 7.234166
+INFO 2021-01-22 08:28:47 train.py: 78] Epoch 7, iter 3200/6416, lr 0.100000, loss 7.247087
+INFO 2021-01-22 08:30:49 train.py: 78] Epoch 7, iter 3400/6416, lr 0.100000, loss 7.243866
+INFO 2021-01-22 08:32:52 train.py: 78] Epoch 7, iter 3600/6416, lr 0.100000, loss 7.232373
+INFO 2021-01-22 08:34:55 train.py: 78] Epoch 7, iter 3800/6416, lr 0.100000, loss 7.211621
+INFO 2021-01-22 08:36:58 train.py: 78] Epoch 7, iter 4000/6416, lr 0.100000, loss 7.213890
+INFO 2021-01-22 08:39:01 train.py: 78] Epoch 7, iter 4200/6416, lr 0.100000, loss 7.190381
+INFO 2021-01-22 08:41:05 train.py: 78] Epoch 7, iter 4400/6416, lr 0.100000, loss 7.219901
+INFO 2021-01-22 08:43:07 train.py: 78] Epoch 7, iter 4600/6416, lr 0.100000, loss 7.233510
+INFO 2021-01-22 08:45:11 train.py: 78] Epoch 7, iter 4800/6416, lr 0.100000, loss 7.247017
+INFO 2021-01-22 08:47:14 train.py: 78] Epoch 7, iter 5000/6416, lr 0.100000, loss 7.222211
+INFO 2021-01-22 08:49:17 train.py: 78] Epoch 7, iter 5200/6416, lr 0.100000, loss 7.201700
+INFO 2021-01-22 08:51:19 train.py: 78] Epoch 7, iter 5400/6416, lr 0.100000, loss 7.179126
+INFO 2021-01-22 08:53:23 train.py: 78] Epoch 7, iter 5600/6416, lr 0.100000, loss 7.222539
+INFO 2021-01-22 08:55:26 train.py: 78] Epoch 7, iter 5800/6416, lr 0.100000, loss 7.203858
+INFO 2021-01-22 08:57:28 train.py: 91] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-01-22 08:57:29 train.py: 78] Epoch 7, iter 6000/6416, lr 0.100000, loss 7.188316
+INFO 2021-01-22 08:59:32 train.py: 78] Epoch 7, iter 6200/6416, lr 0.100000, loss 7.184431
+INFO 2021-01-22 09:01:35 train.py: 78] Epoch 7, iter 6400/6416, lr 0.100000, loss 7.207356
+INFO 2021-01-22 09:01:44 train.py: 96] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-01-22 09:01:46 train.py: 78] Epoch 8, iter 0/6416, lr 0.100000, loss 7.145245
+INFO 2021-01-22 09:03:49 train.py: 78] Epoch 8, iter 200/6416, lr 0.100000, loss 6.656028
+INFO 2021-01-22 09:05:52 train.py: 78] Epoch 8, iter 400/6416, lr 0.100000, loss 6.703750
+INFO 2021-01-22 09:07:55 train.py: 78] Epoch 8, iter 600/6416, lr 0.100000, loss 6.789069
+INFO 2021-01-22 09:09:58 train.py: 78] Epoch 8, iter 800/6416, lr 0.100000, loss 6.858890
+INFO 2021-01-22 09:12:01 train.py: 78] Epoch 8, iter 1000/6416, lr 0.100000, loss 6.921686
+INFO 2021-01-22 09:14:04 train.py: 78] Epoch 8, iter 1200/6416, lr 0.100000, loss 6.932669
+INFO 2021-01-22 09:16:07 train.py: 78] Epoch 8, iter 1400/6416, lr 0.100000, loss 6.984204
+INFO 2021-01-22 09:18:10 train.py: 78] Epoch 8, iter 1600/6416, lr 0.100000, loss 7.011911
+INFO 2021-01-22 09:20:13 train.py: 78] Epoch 8, iter 1800/6416, lr 0.100000, loss 7.052055
+INFO 2021-01-22 09:22:16 train.py: 78] Epoch 8, iter 2000/6416, lr 0.100000, loss 7.089869
+INFO 2021-01-22 09:24:18 train.py: 78] Epoch 8, iter 2200/6416, lr 0.100000, loss 7.046858
+INFO 2021-01-22 09:26:22 train.py: 78] Epoch 8, iter 2400/6416, lr 0.100000, loss 7.085688
+INFO 2021-01-22 09:28:24 train.py: 78] Epoch 8, iter 2600/6416, lr 0.100000, loss 7.074547
+INFO 2021-01-22 09:30:27 train.py: 78] Epoch 8, iter 2800/6416, lr 0.100000, loss 7.071129
+INFO 2021-01-22 09:32:30 train.py: 91] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-01-22 09:32:31 train.py: 78] Epoch 8, iter 3000/6416, lr 0.100000, loss 7.104793
+INFO 2021-01-22 09:34:34 train.py: 78] Epoch 8, iter 3200/6416, lr 0.100000, loss 7.121389
+INFO 2021-01-22 09:36:36 train.py: 78] Epoch 8, iter 3400/6416, lr 0.100000, loss 7.094295
+INFO 2021-01-22 09:38:39 train.py: 78] Epoch 8, iter 3600/6416, lr 0.100000, loss 7.104957
+INFO 2021-01-22 09:40:42 train.py: 78] Epoch 8, iter 3800/6416, lr 0.100000, loss 7.092743
+INFO 2021-01-22 09:42:45 train.py: 78] Epoch 8, iter 4000/6416, lr 0.100000, loss 7.099756
+INFO 2021-01-22 09:44:48 train.py: 78] Epoch 8, iter 4200/6416, lr 0.100000, loss 7.096347
+INFO 2021-01-22 09:46:51 train.py: 78] Epoch 8, iter 4400/6416, lr 0.100000, loss 7.089161
+INFO 2021-01-22 09:48:54 train.py: 78] Epoch 8, iter 4600/6416, lr 0.100000, loss 7.084775
+INFO 2021-01-22 09:50:57 train.py: 78] Epoch 8, iter 4800/6416, lr 0.100000, loss 7.088833
+INFO 2021-01-22 09:53:00 train.py: 78] Epoch 8, iter 5000/6416, lr 0.100000, loss 7.077663
+INFO 2021-01-22 09:55:03 train.py: 78] Epoch 8, iter 5200/6416, lr 0.100000, loss 7.075910
+INFO 2021-01-22 09:57:06 train.py: 78] Epoch 8, iter 5400/6416, lr 0.100000, loss 7.076304
+INFO 2021-01-22 09:59:09 train.py: 78] Epoch 8, iter 5600/6416, lr 0.100000, loss 7.061108
+INFO 2021-01-22 10:01:11 train.py: 78] Epoch 8, iter 5800/6416, lr 0.100000, loss 7.033988
+INFO 2021-01-22 10:03:14 train.py: 91] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-01-22 10:03:15 train.py: 78] Epoch 8, iter 6000/6416, lr 0.100000, loss 7.073775
+INFO 2021-01-22 10:05:17 train.py: 78] Epoch 8, iter 6200/6416, lr 0.100000, loss 7.089445
+INFO 2021-01-22 10:07:20 train.py: 78] Epoch 8, iter 6400/6416, lr 0.100000, loss 7.064998
+INFO 2021-01-22 10:07:30 train.py: 96] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-01-22 10:07:31 train.py: 78] Epoch 9, iter 0/6416, lr 0.100000, loss 7.116372
+INFO 2021-01-22 10:09:34 train.py: 78] Epoch 9, iter 200/6416, lr 0.100000, loss 6.537119
+INFO 2021-01-22 10:11:37 train.py: 78] Epoch 9, iter 400/6416, lr 0.100000, loss 6.545783
+INFO 2021-01-22 10:13:40 train.py: 78] Epoch 9, iter 600/6416, lr 0.100000, loss 6.664490
+INFO 2021-01-22 10:15:43 train.py: 78] Epoch 9, iter 800/6416, lr 0.100000, loss 6.726363
+INFO 2021-01-22 10:17:46 train.py: 78] Epoch 9, iter 1000/6416, lr 0.100000, loss 6.796361
+INFO 2021-01-22 10:19:49 train.py: 78] Epoch 9, iter 1200/6416, lr 0.100000, loss 6.810639
+INFO 2021-01-22 10:21:51 train.py: 78] Epoch 9, iter 1400/6416, lr 0.100000, loss 6.888028
+INFO 2021-01-22 10:23:54 train.py: 78] Epoch 9, iter 1600/6416, lr 0.100000, loss 6.893971
+INFO 2021-01-22 10:25:57 train.py: 78] Epoch 9, iter 1800/6416, lr 0.100000, loss 6.922876
+INFO 2021-01-22 10:28:00 train.py: 78] Epoch 9, iter 2000/6416, lr 0.100000, loss 6.943989
+INFO 2021-01-22 10:30:03 train.py: 78] Epoch 9, iter 2200/6416, lr 0.100000, loss 6.930270
+INFO 2021-01-22 10:32:05 train.py: 78] Epoch 9, iter 2400/6416, lr 0.100000, loss 6.966433
+INFO 2021-01-22 10:34:09 train.py: 78] Epoch 9, iter 2600/6416, lr 0.100000, loss 6.966681
+INFO 2021-01-22 10:36:12 train.py: 78] Epoch 9, iter 2800/6416, lr 0.100000, loss 6.964253
+INFO 2021-01-22 10:38:14 train.py: 91] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-01-22 10:38:15 train.py: 78] Epoch 9, iter 3000/6416, lr 0.100000, loss 6.987408
+INFO 2021-01-22 10:40:18 train.py: 78] Epoch 9, iter 3200/6416, lr 0.100000, loss 6.985979
+INFO 2021-01-22 10:42:21 train.py: 78] Epoch 9, iter 3400/6416, lr 0.100000, loss 6.981175
+INFO 2021-01-22 10:44:24 train.py: 78] Epoch 9, iter 3600/6416, lr 0.100000, loss 6.978908
+INFO 2021-01-22 10:46:27 train.py: 78] Epoch 9, iter 3800/6416, lr 0.100000, loss 6.979999
+INFO 2021-01-22 10:48:30 train.py: 78] Epoch 9, iter 4000/6416, lr 0.100000, loss 6.985364
+INFO 2021-01-22 10:50:33 train.py: 78] Epoch 9, iter 4200/6416, lr 0.100000, loss 6.976088
+INFO 2021-01-22 10:52:36 train.py: 78] Epoch 9, iter 4400/6416, lr 0.100000, loss 6.962269
+INFO 2021-01-22 10:54:39 train.py: 78] Epoch 9, iter 4600/6416, lr 0.100000, loss 6.974390
+INFO 2021-01-22 10:56:42 train.py: 78] Epoch 9, iter 4800/6416, lr 0.100000, loss 6.949354
+INFO 2021-01-22 10:58:45 train.py: 78] Epoch 9, iter 5000/6416, lr 0.100000, loss 7.008740
+INFO 2021-01-22 11:00:48 train.py: 78] Epoch 9, iter 5200/6416, lr 0.100000, loss 7.003042
+INFO 2021-01-22 11:02:51 train.py: 78] Epoch 9, iter 5400/6416, lr 0.100000, loss 6.962869
+INFO 2021-01-22 11:04:54 train.py: 78] Epoch 9, iter 5600/6416, lr 0.100000, loss 6.933342
+INFO 2021-01-22 11:06:57 train.py: 78] Epoch 9, iter 5800/6416, lr 0.100000, loss 6.947052
+INFO 2021-01-22 11:08:59 train.py: 91] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-01-22 11:09:00 train.py: 78] Epoch 9, iter 6000/6416, lr 0.100000, loss 6.963691
+INFO 2021-01-22 11:11:03 train.py: 78] Epoch 9, iter 6200/6416, lr 0.100000, loss 6.968245
+INFO 2021-01-22 11:13:05 train.py: 78] Epoch 9, iter 6400/6416, lr 0.100000, loss 6.993389
+INFO 2021-01-22 11:13:15 train.py: 96] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-01-22 11:13:17 train.py: 78] Epoch 10, iter 0/6416, lr 0.010000, loss 6.902760
+INFO 2021-01-22 11:15:20 train.py: 78] Epoch 10, iter 200/6416, lr 0.010000, loss 5.748563
+INFO 2021-01-22 11:17:22 train.py: 78] Epoch 10, iter 400/6416, lr 0.010000, loss 5.519318
+INFO 2021-01-22 11:19:25 train.py: 78] Epoch 10, iter 600/6416, lr 0.010000, loss 5.457388
+INFO 2021-01-22 11:21:28 train.py: 78] Epoch 10, iter 800/6416, lr 0.010000, loss 5.341965
+INFO 2021-01-22 11:23:30 train.py: 78] Epoch 10, iter 1000/6416, lr 0.010000, loss 5.301196
+INFO 2021-01-22 11:25:33 train.py: 78] Epoch 10, iter 1200/6416, lr 0.010000, loss 5.278854
+INFO 2021-01-22 11:27:36 train.py: 78] Epoch 10, iter 1400/6416, lr 0.010000, loss 5.215471
+INFO 2021-01-22 11:29:39 train.py: 78] Epoch 10, iter 1600/6416, lr 0.010000, loss 5.189592
+INFO 2021-01-22 11:31:42 train.py: 78] Epoch 10, iter 1800/6416, lr 0.010000, loss 5.178332
+INFO 2021-01-22 11:33:44 train.py: 78] Epoch 10, iter 2000/6416, lr 0.010000, loss 5.113112
+INFO 2021-01-22 11:35:47 train.py: 78] Epoch 10, iter 2200/6416, lr 0.010000, loss 5.122206
+INFO 2021-01-22 11:37:50 train.py: 78] Epoch 10, iter 2400/6416, lr 0.010000, loss 5.085119
+INFO 2021-01-22 11:39:53 train.py: 78] Epoch 10, iter 2600/6416, lr 0.010000, loss 5.083674
+INFO 2021-01-22 11:41:56 train.py: 78] Epoch 10, iter 2800/6416, lr 0.010000, loss 5.016466
+INFO 2021-01-22 11:43:59 train.py: 91] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-01-22 11:43:59 train.py: 78] Epoch 10, iter 3000/6416, lr 0.010000, loss 5.036480
+INFO 2021-01-22 11:46:02 train.py: 78] Epoch 10, iter 3200/6416, lr 0.010000, loss 4.996401
+INFO 2021-01-22 11:48:05 train.py: 78] Epoch 10, iter 3400/6416, lr 0.010000, loss 4.995122
+INFO 2021-01-22 11:50:08 train.py: 78] Epoch 10, iter 3600/6416, lr 0.010000, loss 4.961188
+INFO 2021-01-22 11:52:11 train.py: 78] Epoch 10, iter 3800/6416, lr 0.010000, loss 4.936179
+INFO 2021-01-22 11:54:13 train.py: 78] Epoch 10, iter 4000/6416, lr 0.010000, loss 4.932051
+INFO 2021-01-22 11:56:16 train.py: 78] Epoch 10, iter 4200/6416, lr 0.010000, loss 4.928279
+INFO 2021-01-22 11:58:19 train.py: 78] Epoch 10, iter 4400/6416, lr 0.010000, loss 4.902917
+INFO 2021-01-22 12:00:22 train.py: 78] Epoch 10, iter 4600/6416, lr 0.010000, loss 4.865013
+INFO 2021-01-22 12:02:25 train.py: 78] Epoch 10, iter 4800/6416, lr 0.010000, loss 4.857190
+INFO 2021-01-22 12:04:28 train.py: 78] Epoch 10, iter 5000/6416, lr 0.010000, loss 4.863378
+INFO 2021-01-22 12:06:30 train.py: 78] Epoch 10, iter 5200/6416, lr 0.010000, loss 4.856453
+INFO 2021-01-22 12:08:33 train.py: 78] Epoch 10, iter 5400/6416, lr 0.010000, loss 4.829621
+INFO 2021-01-22 12:10:36 train.py: 78] Epoch 10, iter 5600/6416, lr 0.010000, loss 4.798682
+INFO 2021-01-22 12:12:39 train.py: 78] Epoch 10, iter 5800/6416, lr 0.010000, loss 4.840754
+INFO 2021-01-22 12:14:42 train.py: 91] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-01-22 12:14:43 train.py: 78] Epoch 10, iter 6000/6416, lr 0.010000, loss 4.824581
+INFO 2021-01-22 12:16:46 train.py: 78] Epoch 10, iter 6200/6416, lr 0.010000, loss 4.786027
+INFO 2021-01-22 12:18:49 train.py: 78] Epoch 10, iter 6400/6416, lr 0.010000, loss 4.797228
+INFO 2021-01-22 12:18:58 train.py: 96] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-01-22 12:19:00 train.py: 78] Epoch 11, iter 0/6416, lr 0.010000, loss 4.784014
+INFO 2021-01-22 12:21:03 train.py: 78] Epoch 11, iter 200/6416, lr 0.010000, loss 4.470120
+INFO 2021-01-22 12:23:06 train.py: 78] Epoch 11, iter 400/6416, lr 0.010000, loss 4.451558
+INFO 2021-01-22 12:25:08 train.py: 78] Epoch 11, iter 600/6416, lr 0.010000, loss 4.429917
+INFO 2021-01-22 12:27:11 train.py: 78] Epoch 11, iter 800/6416, lr 0.010000, loss 4.436838
+INFO 2021-01-22 12:29:14 train.py: 78] Epoch 11, iter 1000/6416, lr 0.010000, loss 4.467249
+INFO 2021-01-22 12:31:17 train.py: 78] Epoch 11, iter 1200/6416, lr 0.010000, loss 4.462276
+INFO 2021-01-22 12:33:20 train.py: 78] Epoch 11, iter 1400/6416, lr 0.010000, loss 4.441387
+INFO 2021-01-22 12:35:22 train.py: 78] Epoch 11, iter 1600/6416, lr 0.010000, loss 4.458403
+INFO 2021-01-22 12:37:25 train.py: 78] Epoch 11, iter 1800/6416, lr 0.010000, loss 4.457052
+INFO 2021-01-22 12:39:28 train.py: 78] Epoch 11, iter 2000/6416, lr 0.010000, loss 4.471438
+INFO 2021-01-22 12:41:30 train.py: 78] Epoch 11, iter 2200/6416, lr 0.010000, loss 4.469756
+INFO 2021-01-22 12:43:33 train.py: 78] Epoch 11, iter 2400/6416, lr 0.010000, loss 4.486885
+INFO 2021-01-22 12:45:35 train.py: 78] Epoch 11, iter 2600/6416, lr 0.010000, loss 4.446462
+INFO 2021-01-22 12:47:38 train.py: 78] Epoch 11, iter 2800/6416, lr 0.010000, loss 4.466606
+INFO 2021-01-22 12:49:40 train.py: 91] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-01-22 12:49:41 train.py: 78] Epoch 11, iter 3000/6416, lr 0.010000, loss 4.456106
+INFO 2021-01-22 12:51:44 train.py: 78] Epoch 11, iter 3200/6416, lr 0.010000, loss 4.447276
+INFO 2021-01-22 12:53:46 train.py: 78] Epoch 11, iter 3400/6416, lr 0.010000, loss 4.470429
+INFO 2021-01-22 12:55:49 train.py: 78] Epoch 11, iter 3600/6416, lr 0.010000, loss 4.488471
+INFO 2021-01-22 12:57:52 train.py: 78] Epoch 11, iter 3800/6416, lr 0.010000, loss 4.448806
+INFO 2021-01-22 12:59:55 train.py: 78] Epoch 11, iter 4000/6416, lr 0.010000, loss 4.486610
+INFO 2021-01-22 13:01:57 train.py: 78] Epoch 11, iter 4200/6416, lr 0.010000, loss 4.447952
+INFO 2021-01-22 13:04:00 train.py: 78] Epoch 11, iter 4400/6416, lr 0.010000, loss 4.481845
+INFO 2021-01-22 13:06:03 train.py: 78] Epoch 11, iter 4600/6416, lr 0.010000, loss 4.480989
+INFO 2021-01-22 13:08:06 train.py: 78] Epoch 11, iter 4800/6416, lr 0.010000, loss 4.489532
+INFO 2021-01-22 13:10:09 train.py: 78] Epoch 11, iter 5000/6416, lr 0.010000, loss 4.483109
+INFO 2021-01-22 13:12:12 train.py: 78] Epoch 11, iter 5200/6416, lr 0.010000, loss 4.470513
+INFO 2021-01-22 13:14:15 train.py: 78] Epoch 11, iter 5400/6416, lr 0.010000, loss 4.464391
+INFO 2021-01-22 13:16:18 train.py: 78] Epoch 11, iter 5600/6416, lr 0.010000, loss 4.468705
+INFO 2021-01-22 13:18:20 train.py: 78] Epoch 11, iter 5800/6416, lr 0.010000, loss 4.492000
+INFO 2021-01-22 13:20:23 train.py: 91] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-01-22 13:20:24 train.py: 78] Epoch 11, iter 6000/6416, lr 0.010000, loss 4.489486
+INFO 2021-01-22 13:22:26 train.py: 78] Epoch 11, iter 6200/6416, lr 0.010000, loss 4.481620
+INFO 2021-01-22 13:24:29 train.py: 78] Epoch 11, iter 6400/6416, lr 0.010000, loss 4.488010
+INFO 2021-01-22 13:24:39 train.py: 96] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-01-22 13:24:40 train.py: 78] Epoch 12, iter 0/6416, lr 0.010000, loss 4.577611
+INFO 2021-01-22 13:26:43 train.py: 78] Epoch 12, iter 200/6416, lr 0.010000, loss 4.146712
+INFO 2021-01-22 13:28:46 train.py: 78] Epoch 12, iter 400/6416, lr 0.010000, loss 4.163031
+INFO 2021-01-22 13:30:49 train.py: 78] Epoch 12, iter 600/6416, lr 0.010000, loss 4.160265
+INFO 2021-01-22 13:32:52 train.py: 78] Epoch 12, iter 800/6416, lr 0.010000, loss 4.194707
+INFO 2021-01-22 13:34:54 train.py: 78] Epoch 12, iter 1000/6416, lr 0.010000, loss 4.186946
+INFO 2021-01-22 13:36:57 train.py: 78] Epoch 12, iter 1200/6416, lr 0.010000, loss 4.218983
+INFO 2021-01-22 13:38:59 train.py: 78] Epoch 12, iter 1400/6416, lr 0.010000, loss 4.198204
+INFO 2021-01-22 13:41:02 train.py: 78] Epoch 12, iter 1600/6416, lr 0.010000, loss 4.218284
+INFO 2021-01-22 13:43:05 train.py: 78] Epoch 12, iter 1800/6416, lr 0.010000, loss 4.222625
+INFO 2021-01-22 13:45:08 train.py: 78] Epoch 12, iter 2000/6416, lr 0.010000, loss 4.225643
+INFO 2021-01-22 13:47:10 train.py: 78] Epoch 12, iter 2200/6416, lr 0.010000, loss 4.251512
+INFO 2021-01-22 13:49:13 train.py: 78] Epoch 12, iter 2400/6416, lr 0.010000, loss 4.264884
+INFO 2021-01-22 13:51:16 train.py: 78] Epoch 12, iter 2600/6416, lr 0.010000, loss 4.240960
+INFO 2021-01-22 13:53:18 train.py: 78] Epoch 12, iter 2800/6416, lr 0.010000, loss 4.277109
+INFO 2021-01-22 13:55:21 train.py: 91] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-01-22 13:55:21 train.py: 78] Epoch 12, iter 3000/6416, lr 0.010000, loss 4.265947
+INFO 2021-01-22 13:57:24 train.py: 78] Epoch 12, iter 3200/6416, lr 0.010000, loss 4.277185
+INFO 2021-01-22 13:59:27 train.py: 78] Epoch 12, iter 3400/6416, lr 0.010000, loss 4.286415
+INFO 2021-01-22 14:01:29 train.py: 78] Epoch 12, iter 3600/6416, lr 0.010000, loss 4.283371
+INFO 2021-01-22 14:03:32 train.py: 78] Epoch 12, iter 3800/6416, lr 0.010000, loss 4.297545
+INFO 2021-01-22 14:05:35 train.py: 78] Epoch 12, iter 4000/6416, lr 0.010000, loss 4.311358
+INFO 2021-01-22 14:07:37 train.py: 78] Epoch 12, iter 4200/6416, lr 0.010000, loss 4.314255
+INFO 2021-01-22 14:09:40 train.py: 78] Epoch 12, iter 4400/6416, lr 0.010000, loss 4.308063
+INFO 2021-01-22 14:11:43 train.py: 78] Epoch 12, iter 4600/6416, lr 0.010000, loss 4.321885
+INFO 2021-01-22 14:13:45 train.py: 78] Epoch 12, iter 4800/6416, lr 0.010000, loss 4.330787
+INFO 2021-01-22 14:15:48 train.py: 78] Epoch 12, iter 5000/6416, lr 0.010000, loss 4.343471
+INFO 2021-01-22 14:17:51 train.py: 78] Epoch 12, iter 5200/6416, lr 0.010000, loss 4.347062
+INFO 2021-01-22 14:19:54 train.py: 78] Epoch 12, iter 5400/6416, lr 0.010000, loss 4.344270
+INFO 2021-01-22 14:21:56 train.py: 78] Epoch 12, iter 5600/6416, lr 0.010000, loss 4.342152
+INFO 2021-01-22 14:23:59 train.py: 78] Epoch 12, iter 5800/6416, lr 0.010000, loss 4.342138
+INFO 2021-01-22 14:26:01 train.py: 91] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-01-22 14:26:02 train.py: 78] Epoch 12, iter 6000/6416, lr 0.010000, loss 4.369567
+INFO 2021-01-22 14:28:05 train.py: 78] Epoch 12, iter 6200/6416, lr 0.010000, loss 4.361533
+INFO 2021-01-22 14:30:07 train.py: 78] Epoch 12, iter 6400/6416, lr 0.010000, loss 4.379941
+INFO 2021-01-22 14:30:17 train.py: 96] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-01-22 14:30:18 train.py: 78] Epoch 13, iter 0/6416, lr 0.001000, loss 4.262749
+INFO 2021-01-22 14:32:21 train.py: 78] Epoch 13, iter 200/6416, lr 0.001000, loss 3.927313
+INFO 2021-01-22 14:34:24 train.py: 78] Epoch 13, iter 400/6416, lr 0.001000, loss 3.897762
+INFO 2021-01-22 14:36:27 train.py: 78] Epoch 13, iter 600/6416, lr 0.001000, loss 3.922929
+INFO 2021-01-22 14:38:29 train.py: 78] Epoch 13, iter 800/6416, lr 0.001000, loss 3.892005
+INFO 2021-01-22 14:40:32 train.py: 78] Epoch 13, iter 1000/6416, lr 0.001000, loss 3.868123
+INFO 2021-01-22 14:42:34 train.py: 78] Epoch 13, iter 1200/6416, lr 0.001000, loss 3.876276
+INFO 2021-01-22 14:44:37 train.py: 78] Epoch 13, iter 1400/6416, lr 0.001000, loss 3.889019
+INFO 2021-01-22 14:46:40 train.py: 78] Epoch 13, iter 1600/6416, lr 0.001000, loss 3.871777
+INFO 2021-01-22 14:48:42 train.py: 78] Epoch 13, iter 1800/6416, lr 0.001000, loss 3.871109
+INFO 2021-01-22 14:50:45 train.py: 78] Epoch 13, iter 2000/6416, lr 0.001000, loss 3.888027
+INFO 2021-01-22 14:52:47 train.py: 78] Epoch 13, iter 2200/6416, lr 0.001000, loss 3.864406
+INFO 2021-01-22 14:54:50 train.py: 78] Epoch 13, iter 2400/6416, lr 0.001000, loss 3.873461
+INFO 2021-01-22 14:56:52 train.py: 78] Epoch 13, iter 2600/6416, lr 0.001000, loss 3.881583
+INFO 2021-01-22 14:58:55 train.py: 78] Epoch 13, iter 2800/6416, lr 0.001000, loss 3.881827
+INFO 2021-01-22 15:00:57 train.py: 91] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-01-22 15:00:58 train.py: 78] Epoch 13, iter 3000/6416, lr 0.001000, loss 3.882074
+INFO 2021-01-22 15:03:00 train.py: 78] Epoch 13, iter 3200/6416, lr 0.001000, loss 3.860324
+INFO 2021-01-22 15:05:03 train.py: 78] Epoch 13, iter 3400/6416, lr 0.001000, loss 3.876697
+INFO 2021-01-22 15:07:05 train.py: 78] Epoch 13, iter 3600/6416, lr 0.001000, loss 3.882277
+INFO 2021-01-22 15:09:08 train.py: 78] Epoch 13, iter 3800/6416, lr 0.001000, loss 3.881810
+INFO 2021-01-22 15:11:11 train.py: 78] Epoch 13, iter 4000/6416, lr 0.001000, loss 3.876003
+INFO 2021-01-22 15:13:14 train.py: 78] Epoch 13, iter 4200/6416, lr 0.001000, loss 3.891757
+INFO 2021-01-22 15:15:16 train.py: 78] Epoch 13, iter 4400/6416, lr 0.001000, loss 3.877914
+INFO 2021-01-22 15:17:19 train.py: 78] Epoch 13, iter 4600/6416, lr 0.001000, loss 3.865260
+INFO 2021-01-22 15:19:21 train.py: 78] Epoch 13, iter 4800/6416, lr 0.001000, loss 3.893300
+INFO 2021-01-22 15:21:24 train.py: 78] Epoch 13, iter 5000/6416, lr 0.001000, loss 3.864890
+INFO 2021-01-22 15:23:27 train.py: 78] Epoch 13, iter 5200/6416, lr 0.001000, loss 3.879290
+INFO 2021-01-22 15:25:30 train.py: 78] Epoch 13, iter 5400/6416, lr 0.001000, loss 3.870605
+INFO 2021-01-22 15:27:32 train.py: 78] Epoch 13, iter 5600/6416, lr 0.001000, loss 3.877739
+INFO 2021-01-22 15:29:35 train.py: 78] Epoch 13, iter 5800/6416, lr 0.001000, loss 3.865206
+INFO 2021-01-22 15:31:37 train.py: 91] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-01-22 15:31:38 train.py: 78] Epoch 13, iter 6000/6416, lr 0.001000, loss 3.879849
+INFO 2021-01-22 15:33:40 train.py: 78] Epoch 13, iter 6200/6416, lr 0.001000, loss 3.907357
+INFO 2021-01-22 15:35:43 train.py: 78] Epoch 13, iter 6400/6416, lr 0.001000, loss 3.893919
+INFO 2021-01-22 15:35:52 train.py: 96] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-01-22 15:35:54 train.py: 78] Epoch 14, iter 0/6416, lr 0.001000, loss 3.904979
+INFO 2021-01-22 15:37:57 train.py: 78] Epoch 14, iter 200/6416, lr 0.001000, loss 3.828881
+INFO 2021-01-22 15:40:00 train.py: 78] Epoch 14, iter 400/6416, lr 0.001000, loss 3.829152
+INFO 2021-01-22 15:42:03 train.py: 78] Epoch 14, iter 600/6416, lr 0.001000, loss 3.822321
+INFO 2021-01-22 15:44:05 train.py: 78] Epoch 14, iter 800/6416, lr 0.001000, loss 3.809073
+INFO 2021-01-22 15:46:08 train.py: 78] Epoch 14, iter 1000/6416, lr 0.001000, loss 3.828521
+INFO 2021-01-22 15:48:10 train.py: 78] Epoch 14, iter 1200/6416, lr 0.001000, loss 3.837055
+INFO 2021-01-22 15:50:13 train.py: 78] Epoch 14, iter 1400/6416, lr 0.001000, loss 3.864429
+INFO 2021-01-22 15:52:15 train.py: 78] Epoch 14, iter 1600/6416, lr 0.001000, loss 3.846405
+INFO 2021-01-22 15:54:18 train.py: 78] Epoch 14, iter 1800/6416, lr 0.001000, loss 3.825020
+INFO 2021-01-22 15:56:21 train.py: 78] Epoch 14, iter 2000/6416, lr 0.001000, loss 3.847449
+INFO 2021-01-22 15:58:23 train.py: 78] Epoch 14, iter 2200/6416, lr 0.001000, loss 3.819373
+INFO 2021-01-22 16:00:26 train.py: 78] Epoch 14, iter 2400/6416, lr 0.001000, loss 3.853569
+INFO 2021-01-22 16:02:29 train.py: 78] Epoch 14, iter 2600/6416, lr 0.001000, loss 3.869389
+INFO 2021-01-22 16:04:33 train.py: 78] Epoch 14, iter 2800/6416, lr 0.001000, loss 3.855815
+INFO 2021-01-22 16:06:36 train.py: 91] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-01-22 16:06:37 train.py: 78] Epoch 14, iter 3000/6416, lr 0.001000, loss 3.813049
+INFO 2021-01-22 16:08:40 train.py: 78] Epoch 14, iter 3200/6416, lr 0.001000, loss 3.871111
+INFO 2021-01-22 16:10:44 train.py: 78] Epoch 14, iter 3400/6416, lr 0.001000, loss 3.818367
+INFO 2021-01-22 16:12:47 train.py: 78] Epoch 14, iter 3600/6416, lr 0.001000, loss 3.839541
+INFO 2021-01-22 16:14:51 train.py: 78] Epoch 14, iter 3800/6416, lr 0.001000, loss 3.823516
+INFO 2021-01-22 16:16:54 train.py: 78] Epoch 14, iter 4000/6416, lr 0.001000, loss 3.835068
+INFO 2021-01-22 16:18:58 train.py: 78] Epoch 14, iter 4200/6416, lr 0.001000, loss 3.841915
+INFO 2021-01-22 16:21:01 train.py: 78] Epoch 14, iter 4400/6416, lr 0.001000, loss 3.854630
+INFO 2021-01-22 16:23:05 train.py: 78] Epoch 14, iter 4600/6416, lr 0.001000, loss 3.850053
+INFO 2021-01-22 16:25:08 train.py: 78] Epoch 14, iter 4800/6416, lr 0.001000, loss 3.854303
+INFO 2021-01-22 16:27:12 train.py: 78] Epoch 14, iter 5000/6416, lr 0.001000, loss 3.825909
+INFO 2021-01-22 16:29:16 train.py: 78] Epoch 14, iter 5200/6416, lr 0.001000, loss 3.846189
+INFO 2021-01-22 16:31:19 train.py: 78] Epoch 14, iter 5400/6416, lr 0.001000, loss 3.859406
+INFO 2021-01-22 16:33:23 train.py: 78] Epoch 14, iter 5600/6416, lr 0.001000, loss 3.847844
+INFO 2021-01-22 16:35:27 train.py: 78] Epoch 14, iter 5800/6416, lr 0.001000, loss 3.865903
+INFO 2021-01-22 16:37:30 train.py: 91] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-01-22 16:37:31 train.py: 78] Epoch 14, iter 6000/6416, lr 0.001000, loss 3.869699
+INFO 2021-01-22 16:39:34 train.py: 78] Epoch 14, iter 6200/6416, lr 0.001000, loss 3.861083
+INFO 2021-01-22 16:41:38 train.py: 78] Epoch 14, iter 6400/6416, lr 0.001000, loss 3.880113
+INFO 2021-01-22 16:41:47 train.py: 96] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-01-22 16:41:49 train.py: 78] Epoch 15, iter 0/6416, lr 0.001000, loss 3.850110
+INFO 2021-01-22 16:43:52 train.py: 78] Epoch 15, iter 200/6416, lr 0.001000, loss 3.821365
+INFO 2021-01-22 16:45:55 train.py: 78] Epoch 15, iter 400/6416, lr 0.001000, loss 3.790749
+INFO 2021-01-22 16:47:58 train.py: 78] Epoch 15, iter 600/6416, lr 0.001000, loss 3.810499
+INFO 2021-01-22 16:50:01 train.py: 78] Epoch 15, iter 800/6416, lr 0.001000, loss 3.809179
+INFO 2021-01-22 16:52:04 train.py: 78] Epoch 15, iter 1000/6416, lr 0.001000, loss 3.807130
+INFO 2021-01-22 16:54:07 train.py: 78] Epoch 15, iter 1200/6416, lr 0.001000, loss 3.820730
+INFO 2021-01-22 16:56:09 train.py: 78] Epoch 15, iter 1400/6416, lr 0.001000, loss 3.782620
+INFO 2021-01-22 16:58:12 train.py: 78] Epoch 15, iter 1600/6416, lr 0.001000, loss 3.804616
+INFO 2021-01-22 17:00:15 train.py: 78] Epoch 15, iter 1800/6416, lr 0.001000, loss 3.829590
+INFO 2021-01-22 17:02:18 train.py: 78] Epoch 15, iter 2000/6416, lr 0.001000, loss 3.796985
+INFO 2021-01-22 17:04:20 train.py: 78] Epoch 15, iter 2200/6416, lr 0.001000, loss 3.834459
+INFO 2021-01-22 17:06:23 train.py: 78] Epoch 15, iter 2400/6416, lr 0.001000, loss 3.817646
+INFO 2021-01-22 17:08:26 train.py: 78] Epoch 15, iter 2600/6416, lr 0.001000, loss 3.825184
+INFO 2021-01-22 17:10:29 train.py: 78] Epoch 15, iter 2800/6416, lr 0.001000, loss 3.810892
+INFO 2021-01-22 17:12:31 train.py: 91] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-01-22 17:12:32 train.py: 78] Epoch 15, iter 3000/6416, lr 0.001000, loss 3.807870
+INFO 2021-01-22 17:14:34 train.py: 78] Epoch 15, iter 3200/6416, lr 0.001000, loss 3.803119
+INFO 2021-01-22 17:16:37 train.py: 78] Epoch 15, iter 3400/6416, lr 0.001000, loss 3.809163
+INFO 2021-01-22 17:18:40 train.py: 78] Epoch 15, iter 3600/6416, lr 0.001000, loss 3.815229
+INFO 2021-01-22 17:20:43 train.py: 78] Epoch 15, iter 3800/6416, lr 0.001000, loss 3.833905
+INFO 2021-01-22 17:22:46 train.py: 78] Epoch 15, iter 4000/6416, lr 0.001000, loss 3.840918
+INFO 2021-01-22 17:24:49 train.py: 78] Epoch 15, iter 4200/6416, lr 0.001000, loss 3.819103
+INFO 2021-01-22 17:26:52 train.py: 78] Epoch 15, iter 4400/6416, lr 0.001000, loss 3.820613
+INFO 2021-01-22 17:28:54 train.py: 78] Epoch 15, iter 4600/6416, lr 0.001000, loss 3.821484
+INFO 2021-01-22 17:30:57 train.py: 78] Epoch 15, iter 4800/6416, lr 0.001000, loss 3.830079
+INFO 2021-01-22 17:33:00 train.py: 78] Epoch 15, iter 5000/6416, lr 0.001000, loss 3.833368
+INFO 2021-01-22 17:35:03 train.py: 78] Epoch 15, iter 5200/6416, lr 0.001000, loss 3.835595
+INFO 2021-01-22 17:37:06 train.py: 78] Epoch 15, iter 5400/6416, lr 0.001000, loss 3.833158
+INFO 2021-01-22 17:39:09 train.py: 78] Epoch 15, iter 5600/6416, lr 0.001000, loss 3.833666
+INFO 2021-01-22 17:41:12 train.py: 78] Epoch 15, iter 5800/6416, lr 0.001000, loss 3.848253
+INFO 2021-01-22 17:43:15 train.py: 91] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-01-22 17:43:15 train.py: 78] Epoch 15, iter 6000/6416, lr 0.001000, loss 3.829368
+INFO 2021-01-22 17:45:18 train.py: 78] Epoch 15, iter 6200/6416, lr 0.001000, loss 3.819348
+INFO 2021-01-22 17:47:21 train.py: 78] Epoch 15, iter 6400/6416, lr 0.001000, loss 3.830857
+INFO 2021-01-22 17:47:30 train.py: 96] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-01-22 17:47:32 train.py: 78] Epoch 16, iter 0/6416, lr 0.000100, loss 3.846121
+INFO 2021-01-22 17:49:35 train.py: 78] Epoch 16, iter 200/6416, lr 0.000100, loss 3.775344
+INFO 2021-01-22 17:51:38 train.py: 78] Epoch 16, iter 400/6416, lr 0.000100, loss 3.765746
+INFO 2021-01-22 17:53:41 train.py: 78] Epoch 16, iter 600/6416, lr 0.000100, loss 3.783891
+INFO 2021-01-22 17:55:44 train.py: 78] Epoch 16, iter 800/6416, lr 0.000100, loss 3.773140
+INFO 2021-01-22 17:57:47 train.py: 78] Epoch 16, iter 1000/6416, lr 0.000100, loss 3.769097
+INFO 2021-01-22 17:59:50 train.py: 78] Epoch 16, iter 1200/6416, lr 0.000100, loss 3.778376
+INFO 2021-01-22 18:01:52 train.py: 78] Epoch 16, iter 1400/6416, lr 0.000100, loss 3.781148
+INFO 2021-01-22 18:03:55 train.py: 78] Epoch 16, iter 1600/6416, lr 0.000100, loss 3.774220
+INFO 2021-01-22 18:05:58 train.py: 78] Epoch 16, iter 1800/6416, lr 0.000100, loss 3.774145
+INFO 2021-01-22 18:08:01 train.py: 78] Epoch 16, iter 2000/6416, lr 0.000100, loss 3.789803
+INFO 2021-01-22 18:10:04 train.py: 78] Epoch 16, iter 2200/6416, lr 0.000100, loss 3.804328
+INFO 2021-01-22 18:12:07 train.py: 78] Epoch 16, iter 2400/6416, lr 0.000100, loss 3.792517
+INFO 2021-01-22 18:14:10 train.py: 78] Epoch 16, iter 2600/6416, lr 0.000100, loss 3.750824
+INFO 2021-01-22 18:16:13 train.py: 78] Epoch 16, iter 2800/6416, lr 0.000100, loss 3.783069
+INFO 2021-01-22 18:18:15 train.py: 91] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-01-22 18:18:16 train.py: 78] Epoch 16, iter 3000/6416, lr 0.000100, loss 3.765859
+INFO 2021-01-22 18:20:19 train.py: 78] Epoch 16, iter 3200/6416, lr 0.000100, loss 3.753619
+INFO 2021-01-22 18:22:22 train.py: 78] Epoch 16, iter 3400/6416, lr 0.000100, loss 3.790236
+INFO 2021-01-22 18:24:25 train.py: 78] Epoch 16, iter 3600/6416, lr 0.000100, loss 3.750794
+INFO 2021-01-22 18:26:27 train.py: 78] Epoch 16, iter 3800/6416, lr 0.000100, loss 3.761451
+INFO 2021-01-22 18:28:30 train.py: 78] Epoch 16, iter 4000/6416, lr 0.000100, loss 3.791543
+INFO 2021-01-22 18:30:33 train.py: 78] Epoch 16, iter 4200/6416, lr 0.000100, loss 3.763095
+INFO 2021-01-22 18:32:36 train.py: 78] Epoch 16, iter 4400/6416, lr 0.000100, loss 3.783858
+INFO 2021-01-22 18:34:39 train.py: 78] Epoch 16, iter 4600/6416, lr 0.000100, loss 3.760115
+INFO 2021-01-22 18:36:42 train.py: 78] Epoch 16, iter 4800/6416, lr 0.000100, loss 3.761558
+INFO 2021-01-22 18:38:45 train.py: 78] Epoch 16, iter 5000/6416, lr 0.000100, loss 3.760175
+INFO 2021-01-22 18:40:47 train.py: 78] Epoch 16, iter 5200/6416, lr 0.000100, loss 3.765993
+INFO 2021-01-22 18:42:50 train.py: 78] Epoch 16, iter 5400/6416, lr 0.000100, loss 3.781831
+INFO 2021-01-22 18:44:54 train.py: 78] Epoch 16, iter 5600/6416, lr 0.000100, loss 3.756862
+INFO 2021-01-22 18:46:57 train.py: 78] Epoch 16, iter 5800/6416, lr 0.000100, loss 3.779427
+INFO 2021-01-22 18:48:59 train.py: 91] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-01-22 18:49:00 train.py: 78] Epoch 16, iter 6000/6416, lr 0.000100, loss 3.783866
+INFO 2021-01-22 18:51:03 train.py: 78] Epoch 16, iter 6200/6416, lr 0.000100, loss 3.755966
+INFO 2021-01-22 18:53:06 train.py: 78] Epoch 16, iter 6400/6416, lr 0.000100, loss 3.745360
+INFO 2021-01-22 18:53:15 train.py: 96] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-01-22 18:53:17 train.py: 78] Epoch 17, iter 0/6416, lr 0.000100, loss 3.712460
+INFO 2021-01-22 18:55:20 train.py: 78] Epoch 17, iter 200/6416, lr 0.000100, loss 3.780867
+INFO 2021-01-22 18:57:23 train.py: 78] Epoch 17, iter 400/6416, lr 0.000100, loss 3.741587
+INFO 2021-01-22 18:59:26 train.py: 78] Epoch 17, iter 600/6416, lr 0.000100, loss 3.764056
+INFO 2021-01-22 19:01:29 train.py: 78] Epoch 17, iter 800/6416, lr 0.000100, loss 3.764818
+INFO 2021-01-22 19:03:32 train.py: 78] Epoch 17, iter 1000/6416, lr 0.000100, loss 3.777415
+INFO 2021-01-22 19:05:35 train.py: 78] Epoch 17, iter 1200/6416, lr 0.000100, loss 3.769012
+INFO 2021-01-22 19:07:38 train.py: 78] Epoch 17, iter 1400/6416, lr 0.000100, loss 3.770929
+INFO 2021-01-22 19:09:41 train.py: 78] Epoch 17, iter 1600/6416, lr 0.000100, loss 3.742262
+INFO 2021-01-22 19:11:43 train.py: 78] Epoch 17, iter 1800/6416, lr 0.000100, loss 3.755050
+INFO 2021-01-22 19:13:46 train.py: 78] Epoch 17, iter 2000/6416, lr 0.000100, loss 3.749648
+INFO 2021-01-22 19:15:49 train.py: 78] Epoch 17, iter 2200/6416, lr 0.000100, loss 3.780187
+INFO 2021-01-22 19:17:52 train.py: 78] Epoch 17, iter 2400/6416, lr 0.000100, loss 3.785705
+INFO 2021-01-22 19:19:55 train.py: 78] Epoch 17, iter 2600/6416, lr 0.000100, loss 3.786001
+INFO 2021-01-22 19:21:58 train.py: 78] Epoch 17, iter 2800/6416, lr 0.000100, loss 3.760028
+INFO 2021-01-22 19:24:01 train.py: 91] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-01-22 19:24:01 train.py: 78] Epoch 17, iter 3000/6416, lr 0.000100, loss 3.782912
+INFO 2021-01-22 19:26:04 train.py: 78] Epoch 17, iter 3200/6416, lr 0.000100, loss 3.764997
+INFO 2021-01-22 19:28:07 train.py: 78] Epoch 17, iter 3400/6416, lr 0.000100, loss 3.776418
+INFO 2021-01-22 19:30:10 train.py: 78] Epoch 17, iter 3600/6416, lr 0.000100, loss 3.756580
+INFO 2021-01-22 19:32:13 train.py: 78] Epoch 17, iter 3800/6416, lr 0.000100, loss 3.774761
+INFO 2021-01-22 19:34:16 train.py: 78] Epoch 17, iter 4000/6416, lr 0.000100, loss 3.760095
+INFO 2021-01-22 19:36:19 train.py: 78] Epoch 17, iter 4200/6416, lr 0.000100, loss 3.758935
+INFO 2021-01-22 19:38:22 train.py: 78] Epoch 17, iter 4400/6416, lr 0.000100, loss 3.774396
+INFO 2021-01-22 19:40:25 train.py: 78] Epoch 17, iter 4600/6416, lr 0.000100, loss 3.775727
+INFO 2021-01-22 19:42:28 train.py: 78] Epoch 17, iter 4800/6416, lr 0.000100, loss 3.755226
+INFO 2021-01-22 19:44:31 train.py: 78] Epoch 17, iter 5000/6416, lr 0.000100, loss 3.780132
+INFO 2021-01-22 19:46:34 train.py: 78] Epoch 17, iter 5200/6416, lr 0.000100, loss 3.766347
+INFO 2021-01-22 19:48:37 train.py: 78] Epoch 17, iter 5400/6416, lr 0.000100, loss 3.761327
+INFO 2021-01-22 19:50:40 train.py: 78] Epoch 17, iter 5600/6416, lr 0.000100, loss 3.781232
+INFO 2021-01-22 19:52:43 train.py: 78] Epoch 17, iter 5800/6416, lr 0.000100, loss 3.765382
+INFO 2021-01-22 19:54:46 train.py: 91] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-01-22 19:54:47 train.py: 78] Epoch 17, iter 6000/6416, lr 0.000100, loss 3.781601
+INFO 2021-01-22 19:56:50 train.py: 78] Epoch 17, iter 6200/6416, lr 0.000100, loss 3.761133
+INFO 2021-01-22 19:58:53 train.py: 78] Epoch 17, iter 6400/6416, lr 0.000100, loss 3.759499
+INFO 2021-01-22 19:59:02 train.py: 96] Save checkpoint Epoch_17.pt to disk...
+INFO 2021-01-22 19:59:03 train.py: 179] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/.gitkeep b/bob/bio/facexzoo/models/backbones/HRNet/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..9fce2b3d4d3c08825ff8ba5ff1f7e763c3d7989f
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9781666666666666 | 0.0021723571995574545 |
+| Epoch_12.pt | 0.9781666666666666 | 0.00232339151848072 |
+| Epoch_17_batch_5999.pt | 0.9781666666666666 | 0.002349809553617393 |
+| Epoch_14.pt | 0.978 | 0.0022054925823643554 |
+| Epoch_13_batch_2999.pt | 0.9778333333333334 | 0.002618193703988483 |
+| Epoch_16_batch_5999.pt | 0.9778333333333332 | 0.002304718723632388 |
+| Epoch_15_batch_2999.pt | 0.9776666666666666 | 0.002372684056006956 |
+| Epoch_17_batch_2999.pt | 0.9776666666666666 | 0.0021401511426953506 |
+| Epoch_13.pt | 0.9773333333333334 | 0.0020667861375264717 |
+| Epoch_13_batch_5999.pt | 0.9773333333333332 | 0.002111111111111109 |
+| Epoch_12_batch_2999.pt | 0.9773333333333332 | 0.0021401511426953563 |
+| Epoch_16.pt | 0.9773333333333332 | 0.0024620577562400312 |
+| Epoch_15_batch_5999.pt | 0.9771666666666666 | 0.002497529643665448 |
+| Epoch_11_batch_5999.pt | 0.977 | 0.0024190601174530232 |
+| Epoch_16_batch_2999.pt | 0.977 | 0.002444444444444442 |
+| Epoch_17.pt | 0.9766666666666668 | 0.0022222222222222227 |
+| Epoch_10_batch_5999.pt | 0.9766666666666668 | 0.0020934937423796453 |
+| Epoch_14_batch_2999.pt | 0.9763333333333334 | 0.00223330569358242 |
+| Epoch_11.pt | 0.9763333333333334 | 0.002328036315528544 |
+| Epoch_14_batch_5999.pt | 0.9763333333333332 | 0.0021052550357218178 |
+| Epoch_12_batch_5999.pt | 0.9761666666666666 | 0.00223675801546637 |
+| Epoch_11_batch_2999.pt | 0.9743333333333333 | 0.0022388268532899792 |
+| Epoch_10_batch_2999.pt | 0.9743333333333333 | 0.0023333333333333383 |
+| Epoch_10.pt | 0.9741666666666667 | 0.0022939803135625025 |
+| Epoch_9_batch_5999.pt | 0.9718333333333333 | 0.002362907813126301 |
+| Epoch_8_batch_2999.pt | 0.9706666666666666 | 0.0024241582476968245 |
+| Epoch_8_batch_5999.pt | 0.9696666666666666 | 0.0023674750836291713 |
+| Epoch_9_batch_2999.pt | 0.9695 | 0.0027335365778094547 |
+| Epoch_7_batch_5999.pt | 0.9693333333333334 | 0.002845832994414608 |
+| Epoch_5_batch_2999.pt | 0.9688333333333334 | 0.0026416652062152745 |
+| Epoch_9.pt | 0.9688333333333332 | 0.0031725092300598353 |
+| Epoch_7_batch_2999.pt | 0.9684999999999999 | 0.0022831913986705054 |
+| Epoch_4_batch_5999.pt | 0.9683333333333334 | 0.0020336672464136792 |
+| Epoch_6_batch_2999.pt | 0.968 | 0.00281968389787767 |
+| Epoch_5_batch_5999.pt | 0.9678333333333334 | 0.002434956333423467 |
+| Epoch_4_batch_2999.pt | 0.9675 | 0.002499999999999999 |
+| Epoch_6_batch_5999.pt | 0.9673333333333334 | 0.002572408200620051 |
+| Epoch_6.pt | 0.9668333333333333 | 0.0023366378716458995 |
+| Epoch_3_batch_5999.pt | 0.9663333333333333 | 0.0023544022333796743 |
+| Epoch_7.pt | 0.9658333333333333 | 0.002900085141364049 |
+| Epoch_5.pt | 0.9656666666666668 | 0.0029418227321941584 |
+| Epoch_8.pt | 0.9656666666666667 | 0.0031051530249960307 |
+| Epoch_3_batch_2999.pt | 0.9644999999999999 | 0.0026181937039884804 |
+| Epoch_3.pt | 0.9630000000000001 | 0.003111111111111119 |
+| Epoch_4.pt | 0.9626666666666667 | 0.0026550673656330066 |
+| Epoch_2_batch_5999.pt | 0.9625 | 0.002573008039313206 |
+| Epoch_2.pt | 0.9606666666666666 | 0.0025724082006200466 |
+| Epoch_2_batch_2999.pt | 0.9566666666666667 | 0.00268741924943285 |
+| Epoch_1_batch_5999.pt | 0.9503333333333334 | 0.0032375116187407633 |
+| Epoch_1.pt | 0.9476666666666667 | 0.0038666028091789597 |
+| Epoch_1_batch_2999.pt | 0.9348333333333333 | 0.0035611934467860754 |
+| Epoch_0_batch_5999.pt | 0.8800000000000001 | 0.0056546713999997625 |
+| Epoch_0.pt | 0.8766666666666667 | 0.00615138246029618 |
+| Epoch_0_batch_2999.pt | 0.6321666666666667 | 0.0053115876122834475 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..8fb782e3daffa7e9b7d4bafabb62e0a1ef430388
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.9548333333333334 | 0.003963179294891515 |
+| Epoch_16_batch_2999.pt | 0.9545 | 0.003998842425095136 |
+| Epoch_17.pt | 0.9541666666666666 | 0.004159252663165509 |
+| Epoch_14_batch_5999.pt | 0.9541666666666666 | 0.004015786748555824 |
+| Epoch_10_batch_5999.pt | 0.9541666666666666 | 0.004015786748555818 |
+| Epoch_16_batch_5999.pt | 0.9541666666666666 | 0.003961621441334908 |
+| Epoch_17_batch_2999.pt | 0.9540000000000001 | 0.004038395965641386 |
+| Epoch_14.pt | 0.9538333333333334 | 0.0037022682403435835 |
+| Epoch_15_batch_5999.pt | 0.9538333333333332 | 0.003873381780789634 |
+| Epoch_11_batch_2999.pt | 0.9536666666666667 | 0.0036072629068328705 |
+| Epoch_15_batch_2999.pt | 0.9536666666666667 | 0.004004626953542246 |
+| Epoch_11_batch_5999.pt | 0.9535 | 0.003970959395224096 |
+| Epoch_10_batch_2999.pt | 0.9533333333333334 | 0.00373505251421592 |
+| Epoch_17_batch_5999.pt | 0.9533333333333334 | 0.0037184890068181157 |
+| Epoch_11.pt | 0.953 | 0.0038070759331137343 |
+| Epoch_13_batch_5999.pt | 0.9528333333333332 | 0.00377655208874607 |
+| Epoch_15.pt | 0.9526666666666666 | 0.0037777777777777714 |
+| Epoch_12.pt | 0.9525 | 0.0038107223643650007 |
+| Epoch_10.pt | 0.9523333333333334 | 0.004076430295076466 |
+| Epoch_13.pt | 0.952 | 0.0036072629068328705 |
+| Epoch_13_batch_2999.pt | 0.952 | 0.0037908271353848783 |
+| Epoch_12_batch_2999.pt | 0.9518333333333334 | 0.003908284963769235 |
+| Epoch_14_batch_2999.pt | 0.9515 | 0.003655285366576879 |
+| Epoch_12_batch_5999.pt | 0.9515 | 0.0035088072610306937 |
+| Epoch_6_batch_5999.pt | 0.9486666666666667 | 0.002884612219054927 |
+| Epoch_7.pt | 0.9486666666666667 | 0.0037498971179302635 |
+| Epoch_8_batch_5999.pt | 0.9483333333333335 | 0.003583225665910459 |
+| Epoch_9_batch_5999.pt | 0.9476666666666667 | 0.004261513091281139 |
+| Epoch_4_batch_2999.pt | 0.9476666666666667 | 0.003818408950542129 |
+| Epoch_8_batch_2999.pt | 0.9475 | 0.004232808366400096 |
+| Epoch_6_batch_2999.pt | 0.9471666666666666 | 0.004006553273800622 |
+| Epoch_5_batch_5999.pt | 0.9471666666666666 | 0.00412796844183554 |
+| Epoch_5_batch_2999.pt | 0.9466666666666667 | 0.003487189961438936 |
+| Epoch_7_batch_5999.pt | 0.9453333333333334 | 0.004498284995554929 |
+| Epoch_7_batch_2999.pt | 0.9450000000000001 | 0.00394405318873308 |
+| Epoch_4_batch_5999.pt | 0.9448333333333334 | 0.004032659876435444 |
+| Epoch_9_batch_2999.pt | 0.9445 | 0.00426042657111952 |
+| Epoch_6.pt | 0.9443333333333334 | 0.003593547028621383 |
+| Epoch_3_batch_2999.pt | 0.9441666666666666 | 0.003533350803590633 |
+| Epoch_8.pt | 0.944 | 0.0034533933928711236 |
+| Epoch_9.pt | 0.9436666666666668 | 0.002937623125867173 |
+| Epoch_3.pt | 0.9433333333333334 | 0.004975247372719581 |
+| Epoch_5.pt | 0.9433333333333331 | 0.003093202423794456 |
+| Epoch_3_batch_5999.pt | 0.9424999999999999 | 0.0036025539637496566 |
+| Epoch_4.pt | 0.9410000000000001 | 0.00409154503100488 |
+| Epoch_2.pt | 0.9393333333333332 | 0.004232443767720201 |
+| Epoch_2_batch_5999.pt | 0.9391666666666667 | 0.0037618126704955963 |
+| Epoch_2_batch_2999.pt | 0.9358333333333333 | 0.0032984283575105315 |
+| Epoch_1.pt | 0.9338333333333333 | 0.003333796264150551 |
+| Epoch_1_batch_5999.pt | 0.9326666666666666 | 0.0037367048268444917 |
+| Epoch_1_batch_2999.pt | 0.9183333333333333 | 0.0037433067839828904 |
+| Epoch_0.pt | 0.8793333333333333 | 0.004642796092394702 |
+| Epoch_0_batch_5999.pt | 0.8744999999999999 | 0.005483139138742954 |
+| Epoch_0_batch_2999.pt | 0.677 | 0.005707910906240586 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f229d1ab85b26d70aa430f0e67af9530618c9b63
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.8888333333333336 | 0.004817983287949738 |
+| Epoch_14_batch_2999.pt | 0.8886666666666668 | 0.004633479880442894 |
+| Epoch_17_batch_5999.pt | 0.8880000000000001 | 0.004905275567698694 |
+| Epoch_17.pt | 0.8875 | 0.004945692726323565 |
+| Epoch_13_batch_5999.pt | 0.8871666666666667 | 0.004981756842176048 |
+| Epoch_14.pt | 0.8868333333333334 | 0.004921921234840643 |
+| Epoch_15_batch_2999.pt | 0.8866666666666667 | 0.004687782913273109 |
+| Epoch_13_batch_2999.pt | 0.8863333333333333 | 0.005120763831912408 |
+| Epoch_11_batch_2999.pt | 0.8856666666666666 | 0.004818944098266984 |
+| Epoch_11.pt | 0.8855000000000001 | 0.005386597450967117 |
+| Epoch_13.pt | 0.8848333333333332 | 0.005172636932890828 |
+| Epoch_16.pt | 0.8848333333333332 | 0.004978038187632841 |
+| Epoch_12.pt | 0.8845000000000001 | 0.005409468151215269 |
+| Epoch_17_batch_2999.pt | 0.8843333333333334 | 0.004945380685392872 |
+| Epoch_15_batch_5999.pt | 0.8843333333333332 | 0.005062571444397565 |
+| Epoch_16_batch_2999.pt | 0.8841666666666665 | 0.004995368225036448 |
+| Epoch_14_batch_5999.pt | 0.8838333333333332 | 0.00496313570733088 |
+| Epoch_16_batch_5999.pt | 0.8836666666666666 | 0.0050723165380059275 |
+| Epoch_11_batch_5999.pt | 0.8831666666666667 | 0.0058502400499445524 |
+| Epoch_12_batch_2999.pt | 0.8821666666666668 | 0.004486261607382764 |
+| Epoch_12_batch_5999.pt | 0.8801666666666668 | 0.005823801736552048 |
+| Epoch_10.pt | 0.8795 | 0.005194073070503383 |
+| Epoch_10_batch_5999.pt | 0.8785000000000001 | 0.0050701863992812755 |
+| Epoch_10_batch_2999.pt | 0.8748333333333334 | 0.005574689273298304 |
+| Epoch_9_batch_2999.pt | 0.8548333333333332 | 0.005552499159259752 |
+| Epoch_6_batch_2999.pt | 0.8536666666666667 | 0.006269660435407056 |
+| Epoch_8_batch_5999.pt | 0.8518333333333334 | 0.005377421934967227 |
+| Epoch_7_batch_2999.pt | 0.85 | 0.006440611887195305 |
+| Epoch_8_batch_2999.pt | 0.849 | 0.005864730320176283 |
+| Epoch_7_batch_5999.pt | 0.8488333333333333 | 0.006392752454977215 |
+| Epoch_5_batch_5999.pt | 0.8480000000000001 | 0.005587684871413404 |
+| Epoch_5_batch_2999.pt | 0.8478333333333333 | 0.005763068884107053 |
+| Epoch_9_batch_5999.pt | 0.8471666666666666 | 0.007495060101555282 |
+| Epoch_6_batch_5999.pt | 0.8463333333333335 | 0.005756370598661609 |
+| Epoch_3_batch_5999.pt | 0.8461666666666666 | 0.007277777777777777 |
+| Epoch_9.pt | 0.8448333333333332 | 0.006774361616317952 |
+| Epoch_4_batch_5999.pt | 0.8446666666666667 | 0.006230153675056087 |
+| Epoch_7.pt | 0.8416666666666668 | 0.007131914015947205 |
+| Epoch_8.pt | 0.841 | 0.005779914134951097 |
+| Epoch_4_batch_2999.pt | 0.8406666666666667 | 0.00649881280706211 |
+| Epoch_5.pt | 0.8401666666666667 | 0.007099165313482155 |
+| Epoch_6.pt | 0.8386666666666667 | 0.006821353545853706 |
+| Epoch_3_batch_2999.pt | 0.8386666666666667 | 0.007323225357015151 |
+| Epoch_2.pt | 0.8355 | 0.0071709720056748775 |
+| Epoch_2_batch_5999.pt | 0.8341666666666667 | 0.006446599256608765 |
+| Epoch_2_batch_2999.pt | 0.833 | 0.006807766080906838 |
+| Epoch_3.pt | 0.8316666666666667 | 0.0064501890105333306 |
+| Epoch_4.pt | 0.8313333333333335 | 0.0066471937827315706 |
+| Epoch_1.pt | 0.8181666666666667 | 0.006375347969200977 |
+| Epoch_1_batch_5999.pt | 0.818 | 0.007765068982069397 |
+| Epoch_1_batch_2999.pt | 0.8055 | 0.0073411153994908706 |
+| Epoch_0.pt | 0.7518333333333332 | 0.007513157183514632 |
+| Epoch_0_batch_5999.pt | 0.7456666666666666 | 0.007722821719335615 |
+| Epoch_0_batch_2999.pt | 0.6133333333333333 | 0.010567244989431571 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..73a9ca6ec03872e035bc369263be5ae00cd90529
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_10.pt | 0.9979999999999999 | 0.0005983516452371659 |
+| Epoch_8_batch_5999.pt | 0.9978333333333333 | 0.0005583264233956013 |
+| Epoch_11_batch_5999.pt | 0.9978333333333333 | 0.0005583264233956013 |
+| Epoch_14_batch_2999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_15_batch_2999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_14_batch_5999.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_15.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_15_batch_5999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_13_batch_5999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_12_batch_2999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_17.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_13.pt | 0.9974999999999999 | 0.0007954345035153557 |
+| Epoch_11.pt | 0.9974999999999999 | 0.0006689774765995705 |
+| Epoch_17_batch_2999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_16.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_17_batch_5999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_14.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_16_batch_2999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_11_batch_2999.pt | 0.9973333333333333 | 0.0006666666666666665 |
+| Epoch_8.pt | 0.9973333333333333 | 0.0006666666666666665 |
+| Epoch_12_batch_5999.pt | 0.9973333333333333 | 0.0007114582486036506 |
+| Epoch_16_batch_5999.pt | 0.9973333333333333 | 0.0006666666666666665 |
+| Epoch_4_batch_5999.pt | 0.9971666666666668 | 0.001025899184034411 |
+| Epoch_4_batch_2999.pt | 0.9971666666666665 | 0.0007474235581707637 |
+| Epoch_10_batch_2999.pt | 0.9971666666666665 | 0.0007049209744694181 |
+| Epoch_10_batch_5999.pt | 0.9971666666666665 | 0.0007049209744694181 |
+| Epoch_9_batch_5999.pt | 0.9970000000000001 | 0.0006478835438716991 |
+| Epoch_12.pt | 0.9969999999999999 | 0.0007370277311900913 |
+| Epoch_3_batch_2999.pt | 0.9969999999999999 | 0.0006938886664887115 |
+| Epoch_13_batch_2999.pt | 0.9969999999999999 | 0.0007370277311900913 |
+| Epoch_6_batch_2999.pt | 0.9966666666666667 | 0.0009296222517045277 |
+| Epoch_7_batch_2999.pt | 0.9966666666666667 | 0.0008240220541217368 |
+| Epoch_6_batch_5999.pt | 0.9966666666666665 | 0.0007453559924999332 |
+| Epoch_5.pt | 0.9966666666666665 | 0.0010829771494232166 |
+| Epoch_9_batch_2999.pt | 0.9964999999999999 | 0.0005241100628920327 |
+| Epoch_8_batch_2999.pt | 0.9963333333333333 | 0.001105541596785132 |
+| Epoch_3_batch_5999.pt | 0.9963333333333333 | 0.0008534606386520669 |
+| Epoch_7_batch_5999.pt | 0.9961666666666666 | 0.0008975274678557517 |
+| Epoch_5_batch_2999.pt | 0.9961666666666666 | 0.0008258927081843649 |
+| Epoch_9.pt | 0.9959999999999999 | 0.0007114582486036552 |
+| Epoch_3.pt | 0.9959999999999999 | 0.0009026709338484399 |
+| Epoch_1_batch_5999.pt | 0.9956666666666665 | 0.0010886621079036305 |
+| Epoch_7.pt | 0.9955 | 0.001025899184034412 |
+| Epoch_2.pt | 0.9955 | 0.0011666666666666696 |
+| Epoch_5_batch_5999.pt | 0.9955 | 0.0008258927081843637 |
+| Epoch_2_batch_5999.pt | 0.9953333333333333 | 0.0010183501544346308 |
+| Epoch_2_batch_2999.pt | 0.9953333333333332 | 0.0011863420280034786 |
+| Epoch_6.pt | 0.9951666666666666 | 0.0007222222222222252 |
+| Epoch_4.pt | 0.9949999999999999 | 0.0010829771494232224 |
+| Epoch_1_batch_2999.pt | 0.9943333333333332 | 0.0010599324460188284 |
+| Epoch_1.pt | 0.9941666666666666 | 0.0008333333333333327 |
+| Epoch_0.pt | 0.9824999999999999 | 0.002037457868988028 |
+| Epoch_0_batch_5999.pt | 0.9795 | 0.0017924739783224 |
+| Epoch_0_batch_2999.pt | 0.9171666666666667 | 0.005170249653689995 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..8e43dc3c36fa783b6a2d1b9f33ae6cdca4790b89
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.9586666666666666 | 0.0025555555555555644 |
+| Epoch_15_batch_2999.pt | 0.9584999999999999 | 0.002715410979966653 |
+| Epoch_15_batch_5999.pt | 0.958 | 0.0028306087117460008 |
+| Epoch_17_batch_2999.pt | 0.9578333333333333 | 0.00281146242864 |
+| Epoch_17.pt | 0.9576666666666667 | 0.00281310864470493 |
+| Epoch_16_batch_5999.pt | 0.9568333333333332 | 0.0028158501994387116 |
+| Epoch_13_batch_2999.pt | 0.9566666666666667 | 0.0028327886186626625 |
+| Epoch_13.pt | 0.9563333333333335 | 0.002916534388534827 |
+| Epoch_13_batch_5999.pt | 0.9563333333333333 | 0.0030307070437746338 |
+| Epoch_14_batch_5999.pt | 0.9563333333333333 | 0.0025915341754868 |
+| Epoch_16_batch_2999.pt | 0.9560000000000001 | 0.002631715396072674 |
+| Epoch_16.pt | 0.9558333333333333 | 0.0028136571693556903 |
+| Epoch_15.pt | 0.9553333333333333 | 0.003101174608211751 |
+| Epoch_14_batch_2999.pt | 0.9550000000000001 | 0.00329608821648696 |
+| Epoch_12.pt | 0.9548333333333334 | 0.0024017740356908116 |
+| Epoch_14.pt | 0.9546666666666667 | 0.003431876713662335 |
+| Epoch_12_batch_2999.pt | 0.9541666666666668 | 0.0026087459737497536 |
+| Epoch_11_batch_5999.pt | 0.9528333333333334 | 0.0021666666666666657 |
+| Epoch_11_batch_2999.pt | 0.9516666666666665 | 0.0028760398012321786 |
+| Epoch_12_batch_5999.pt | 0.9515 | 0.0025270145367875954 |
+| Epoch_10_batch_2999.pt | 0.9513333333333334 | 0.002673602092336885 |
+| Epoch_11.pt | 0.9508333333333334 | 0.003164716748583581 |
+| Epoch_10.pt | 0.9501666666666665 | 0.0035611934467860715 |
+| Epoch_10_batch_5999.pt | 0.9479999999999998 | 0.002884612219054929 |
+| Epoch_9_batch_5999.pt | 0.9276666666666665 | 0.004548775442324165 |
+| Epoch_8_batch_5999.pt | 0.9235 | 0.003491170520747769 |
+| Epoch_7_batch_5999.pt | 0.9233333333333335 | 0.004090036070956877 |
+| Epoch_7_batch_2999.pt | 0.9228333333333334 | 0.0037105954398881226 |
+| Epoch_9_batch_2999.pt | 0.9225 | 0.0036281648593901586 |
+| Epoch_8_batch_2999.pt | 0.9188333333333333 | 0.0041279684418355405 |
+| Epoch_7.pt | 0.9181666666666667 | 0.00463114789919276 |
+| Epoch_6_batch_2999.pt | 0.9161666666666667 | 0.005264896561421613 |
+| Epoch_6_batch_5999.pt | 0.9158333333333333 | 0.0031056499687497096 |
+| Epoch_5_batch_2999.pt | 0.9145 | 0.003379769149834472 |
+| Epoch_3_batch_5999.pt | 0.9125 | 0.004397319610067758 |
+| Epoch_4_batch_2999.pt | 0.9113333333333333 | 0.004222222222222225 |
+| Epoch_5_batch_5999.pt | 0.9106666666666665 | 0.005455770532084851 |
+| Epoch_4_batch_5999.pt | 0.9101666666666667 | 0.0046378075918756395 |
+| Epoch_6.pt | 0.9091666666666665 | 0.0034359213546813834 |
+| Epoch_9.pt | 0.9075 | 0.00458964212273842 |
+| Epoch_8.pt | 0.9038333333333334 | 0.002919178812315477 |
+| Epoch_3_batch_2999.pt | 0.9021666666666667 | 0.004766459461955927 |
+| Epoch_5.pt | 0.9006666666666666 | 0.0038506052113696536 |
+| Epoch_2_batch_5999.pt | 0.8985 | 0.005970348543292513 |
+| Epoch_3.pt | 0.8944999999999999 | 0.0038654052859553216 |
+| Epoch_2.pt | 0.8885 | 0.006103277807866855 |
+| Epoch_2_batch_2999.pt | 0.8876666666666665 | 0.005224585753559441 |
+| Epoch_4.pt | 0.8871666666666667 | 0.005024937810560444 |
+| Epoch_1.pt | 0.8753333333333334 | 0.004058218303944371 |
+| Epoch_1_batch_5999.pt | 0.8733333333333333 | 0.005134006689843522 |
+| Epoch_1_batch_2999.pt | 0.8525 | 0.005195261373669352 |
+| Epoch_0.pt | 0.7766666666666667 | 0.007765863891114027 |
+| Epoch_0_batch_5999.pt | 0.765 | 0.007361897232656637 |
+| Epoch_0_batch_2999.pt | 0.6043333333333334 | 0.00654141601312667 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d49e541d68ace63f912ec7f7096dc48420fd2bde
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9476666666666669 | 0.002993820796734991 |
+| Epoch_16_batch_5999.pt | 0.9470000000000001 | 0.003010270485365348 |
+| Epoch_17.pt | 0.9468333333333334 | 0.002880864924992028 |
+| Epoch_17_batch_2999.pt | 0.9466666666666667 | 0.0029917582261858307 |
+| Epoch_16.pt | 0.9466666666666667 | 0.0028867513459481246 |
+| Epoch_17_batch_5999.pt | 0.9466666666666667 | 0.002687419249432845 |
+| Epoch_16_batch_2999.pt | 0.9466666666666667 | 0.0026759099063982834 |
+| Epoch_13.pt | 0.9465 | 0.0023498095536174015 |
+| Epoch_15_batch_2999.pt | 0.9463333333333332 | 0.0030102704853653523 |
+| Epoch_14_batch_2999.pt | 0.9461666666666668 | 0.0034431000960769226 |
+| Epoch_15_batch_5999.pt | 0.9461666666666668 | 0.0036094013046179493 |
+| Epoch_11_batch_2999.pt | 0.9456666666666667 | 0.0029627314724385264 |
+| Epoch_13_batch_5999.pt | 0.9453333333333334 | 0.0026034165586355526 |
+| Epoch_14_batch_5999.pt | 0.9448333333333334 | 0.002782773285859601 |
+| Epoch_13_batch_2999.pt | 0.9445 | 0.003277777777777774 |
+| Epoch_14.pt | 0.9443333333333334 | 0.003297960462145741 |
+| Epoch_12_batch_5999.pt | 0.9441666666666666 | 0.0024626844732557203 |
+| Epoch_10_batch_2999.pt | 0.9438333333333334 | 0.0023837153281027683 |
+| Epoch_12.pt | 0.9431666666666667 | 0.003262677080005785 |
+| Epoch_10_batch_5999.pt | 0.943 | 0.002808716591058782 |
+| Epoch_11_batch_5999.pt | 0.9418333333333333 | 0.003455627008907546 |
+| Epoch_10.pt | 0.9416666666666668 | 0.002557969874049187 |
+| Epoch_12_batch_2999.pt | 0.9408333333333333 | 0.0029000851413640383 |
+| Epoch_11.pt | 0.9393333333333332 | 0.002779999111821506 |
+| Epoch_9_batch_2999.pt | 0.9195 | 0.0027448042948968097 |
+| Epoch_7_batch_5999.pt | 0.9169999999999998 | 0.002895292049065638 |
+| Epoch_6_batch_2999.pt | 0.9168333333333333 | 0.003931904950937892 |
+| Epoch_6_batch_5999.pt | 0.916 | 0.002372684056006958 |
+| Epoch_9_batch_5999.pt | 0.9153333333333332 | 0.003958114029012641 |
+| Epoch_5_batch_2999.pt | 0.9153333333333332 | 0.002948110924760355 |
+| Epoch_5_batch_5999.pt | 0.9131666666666666 | 0.004226971013694952 |
+| Epoch_7_batch_2999.pt | 0.9128333333333334 | 0.002034425935955622 |
+| Epoch_8_batch_5999.pt | 0.9115 | 0.0034466838598324646 |
+| Epoch_9.pt | 0.9091666666666667 | 0.0030756912301480865 |
+| Epoch_4_batch_5999.pt | 0.9088333333333335 | 0.0030535346852580717 |
+| Epoch_4_batch_2999.pt | 0.9085000000000001 | 0.003955383891406124 |
+| Epoch_3_batch_5999.pt | 0.907 | 0.0037085153929508094 |
+| Epoch_7.pt | 0.9066666666666666 | 0.002810913475705224 |
+| Epoch_3_batch_2999.pt | 0.9055 | 0.005471869700997218 |
+| Epoch_6.pt | 0.9041666666666666 | 0.003890872509976246 |
+| Epoch_8_batch_2999.pt | 0.9038333333333334 | 0.003693922268862062 |
+| Epoch_2_batch_5999.pt | 0.9003333333333334 | 0.004214905941864592 |
+| Epoch_5.pt | 0.8985 | 0.002837687286471586 |
+| Epoch_8.pt | 0.8961666666666666 | 0.0032111880036929806 |
+| Epoch_3.pt | 0.8928333333333333 | 0.0035403319925232943 |
+| Epoch_2_batch_2999.pt | 0.891 | 0.0036362373715452395 |
+| Epoch_4.pt | 0.8896666666666666 | 0.0036413265795942067 |
+| Epoch_1.pt | 0.8798333333333334 | 0.0036552853665768833 |
+| Epoch_2.pt | 0.8775000000000001 | 0.0035767596895155337 |
+| Epoch_1_batch_5999.pt | 0.8741666666666668 | 0.004895513188902264 |
+| Epoch_1_batch_2999.pt | 0.8598333333333334 | 0.0038845213569807364 |
+| Epoch_0.pt | 0.7858333333333334 | 0.004735276074968898 |
+| Epoch_0_batch_5999.pt | 0.7783333333333333 | 0.005006169033808361 |
+| Epoch_0_batch_2999.pt | 0.6736666666666666 | 0.006785059840343121 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f341991acfb383c6378b37f95ce75b951ea085a6
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9908333333333333 | 0.0014958791130929184 |
+| Epoch_13_batch_5999.pt | 0.9896666666666667 | 0.00130998068028351 |
+| Epoch_12_batch_2999.pt | 0.9896666666666667 | 0.0015674151088517618 |
+| Epoch_17_batch_5999.pt | 0.9896666666666667 | 0.0013562839573037402 |
+| Epoch_17_batch_2999.pt | 0.9894999999999999 | 0.0016675923355337415 |
+| Epoch_14_batch_2999.pt | 0.9893333333333333 | 0.0017069212773041323 |
+| Epoch_13_batch_2999.pt | 0.9893333333333333 | 0.0011967032904743398 |
+| Epoch_17.pt | 0.9889999999999999 | 0.0013877773329774223 |
+| Epoch_14.pt | 0.9889999999999999 | 0.0013193713430042172 |
+| Epoch_16_batch_2999.pt | 0.9888333333333333 | 0.001166666666666665 |
+| Epoch_11_batch_2999.pt | 0.9886666666666667 | 0.0014865653511399646 |
+| Epoch_16_batch_5999.pt | 0.9884999999999999 | 0.0014792807728549206 |
+| Epoch_15_batch_5999.pt | 0.9883333333333333 | 0.0014698618394803254 |
+| Epoch_11_batch_5999.pt | 0.9883333333333333 | 0.0009938079899999067 |
+| Epoch_15.pt | 0.9881666666666666 | 0.0012777777777777837 |
+| Epoch_13.pt | 0.9881666666666666 | 0.0015801625170364317 |
+| Epoch_11.pt | 0.9881666666666666 | 0.001277777777777784 |
+| Epoch_10_batch_5999.pt | 0.9879999999999999 | 0.0011331154474650668 |
+| Epoch_16.pt | 0.9879999999999999 | 0.0014229164972073 |
+| Epoch_14_batch_5999.pt | 0.9876666666666665 | 0.0011166528467912076 |
+| Epoch_12_batch_5999.pt | 0.9874999999999998 | 0.0017786456215091253 |
+| Epoch_12.pt | 0.9873333333333333 | 0.0014098419489388361 |
+| Epoch_10.pt | 0.9871666666666666 | 0.0015918387535438156 |
+| Epoch_10_batch_2999.pt | 0.9866666666666666 | 0.001511274500970596 |
+| Epoch_8_batch_5999.pt | 0.9748333333333333 | 0.001711435755638821 |
+| Epoch_9_batch_5999.pt | 0.974 | 0.001651785416368722 |
+| Epoch_5_batch_2999.pt | 0.9739999999999999 | 0.002962731472438528 |
+| Epoch_6_batch_5999.pt | 0.9738333333333333 | 0.002331348362039695 |
+| Epoch_8_batch_2999.pt | 0.9726666666666667 | 0.0019277057303219403 |
+| Epoch_6_batch_2999.pt | 0.9718333333333333 | 0.002527014536787595 |
+| Epoch_7_batch_2999.pt | 0.9716666666666669 | 0.0022498285257018446 |
+| Epoch_7_batch_5999.pt | 0.9711666666666666 | 0.0017751717009633866 |
+| Epoch_9_batch_2999.pt | 0.9701666666666666 | 0.0017993483045224145 |
+| Epoch_5_batch_5999.pt | 0.9698333333333332 | 0.0026346457114176454 |
+| Epoch_4_batch_2999.pt | 0.9691666666666666 | 0.0022939803135625064 |
+| Epoch_4_batch_5999.pt | 0.9675 | 0.002197782271650279 |
+| Epoch_9.pt | 0.9668333333333333 | 0.0032150302880511765 |
+| Epoch_6.pt | 0.9663333333333334 | 0.002219442706159796 |
+| Epoch_5.pt | 0.9653333333333333 | 0.002431785403137702 |
+| Epoch_8.pt | 0.9651666666666665 | 0.0011235415786753711 |
+| Epoch_3_batch_5999.pt | 0.9644999999999999 | 0.0020344259359556176 |
+| Epoch_7.pt | 0.9638333333333332 | 0.0023837153281027735 |
+| Epoch_3_batch_2999.pt | 0.9628333333333334 | 0.0022912878474779276 |
+| Epoch_2_batch_5999.pt | 0.9628333333333334 | 0.002687993422978439 |
+| Epoch_3.pt | 0.9628333333333334 | 0.0018765939726727238 |
+| Epoch_4.pt | 0.9606666666666666 | 0.0025361582690029572 |
+| Epoch_2_batch_2999.pt | 0.9556666666666667 | 0.0023465235646603203 |
+| Epoch_2.pt | 0.9490000000000001 | 0.003362832433427013 |
+| Epoch_1_batch_5999.pt | 0.9469999999999998 | 0.002354402233379671 |
+| Epoch_1.pt | 0.9465 | 0.002726753598179559 |
+| Epoch_1_batch_2999.pt | 0.9275 | 0.002022252747022367 |
+| Epoch_0.pt | 0.8716666666666667 | 0.004097575314352394 |
+| Epoch_0_batch_5999.pt | 0.8703333333333333 | 0.003101174608211747 |
+| Epoch_0_batch_2999.pt | 0.7348333333333333 | 0.004335113594422084 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..485365dd3f3698015b155906da5bb76b07edee05
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9593333333333336 | 0.0024113927126900793 |
+| Epoch_12.pt | 0.9591666666666667 | 0.002678791878053596 |
+| Epoch_14_batch_5999.pt | 0.9586666666666668 | 0.002119864892037654 |
+| Epoch_14_batch_2999.pt | 0.958 | 0.0025067809272618885 |
+| Epoch_13.pt | 0.958 | 0.002301368353023112 |
+| Epoch_17.pt | 0.9576666666666668 | 0.002346523564660319 |
+| Epoch_15.pt | 0.9576666666666667 | 0.0021829869671542773 |
+| Epoch_15_batch_2999.pt | 0.9573333333333333 | 0.002372684056006963 |
+| Epoch_15_batch_5999.pt | 0.9571666666666667 | 0.00234454976066769 |
+| Epoch_13_batch_5999.pt | 0.9570000000000002 | 0.00241906011745303 |
+| Epoch_12_batch_2999.pt | 0.9568333333333333 | 0.002681095224891925 |
+| Epoch_13_batch_2999.pt | 0.9568333333333333 | 0.002414590423566748 |
+| Epoch_17_batch_2999.pt | 0.9564999999999999 | 0.0019476164603286817 |
+| Epoch_11_batch_2999.pt | 0.9563333333333333 | 0.0031011746082117465 |
+| Epoch_16_batch_5999.pt | 0.9561666666666667 | 0.0028658267503397705 |
+| Epoch_10_batch_5999.pt | 0.9560000000000001 | 0.0029313124351717655 |
+| Epoch_17_batch_5999.pt | 0.9560000000000001 | 0.002372684056006956 |
+| Epoch_16.pt | 0.9558333333333333 | 0.0022395160411940486 |
+| Epoch_12_batch_5999.pt | 0.9555000000000001 | 0.002396628290013671 |
+| Epoch_14.pt | 0.9543333333333333 | 0.0024241582476968292 |
+| Epoch_11_batch_5999.pt | 0.9541666666666666 | 0.0028571979712497288 |
+| Epoch_10.pt | 0.9536666666666666 | 0.0025190631219454748 |
+| Epoch_11.pt | 0.9533333333333334 | 0.0025215123817578225 |
+| Epoch_10_batch_2999.pt | 0.9523333333333334 | 0.002802115602870783 |
+| Epoch_8_batch_2999.pt | 0.9390000000000001 | 0.003711842908553353 |
+| Epoch_9_batch_5999.pt | 0.9376666666666665 | 0.002712567914607482 |
+| Epoch_7_batch_5999.pt | 0.9366666666666668 | 0.004059739090321136 |
+| Epoch_7_batch_2999.pt | 0.9348333333333334 | 0.004070747800382742 |
+| Epoch_5_batch_2999.pt | 0.9338333333333333 | 0.004513696577099906 |
+| Epoch_8_batch_5999.pt | 0.9335000000000001 | 0.003224615969095877 |
+| Epoch_6_batch_5999.pt | 0.9334999999999999 | 0.003924047419202286 |
+| Epoch_7.pt | 0.9333333333333333 | 0.0029080560729560874 |
+| Epoch_6_batch_2999.pt | 0.9328333333333333 | 0.0025825865109265498 |
+| Epoch_5_batch_5999.pt | 0.932 | 0.003051006715054657 |
+| Epoch_4_batch_2999.pt | 0.9309999999999998 | 0.0039220805766050585 |
+| Epoch_9.pt | 0.93 | 0.0031622776601683746 |
+| Epoch_9_batch_2999.pt | 0.9298333333333332 | 0.0034016154477425876 |
+| Epoch_4_batch_5999.pt | 0.9265000000000001 | 0.0024651897480370876 |
+| Epoch_3.pt | 0.9265000000000001 | 0.0033925299201715605 |
+| Epoch_3_batch_5999.pt | 0.9259999999999999 | 0.0028458329944146014 |
+| Epoch_5.pt | 0.9241666666666667 | 0.003761812670495604 |
+| Epoch_3_batch_2999.pt | 0.9238333333333333 | 0.0024094720491334926 |
+| Epoch_2_batch_5999.pt | 0.9225000000000001 | 0.0033263816399828248 |
+| Epoch_6.pt | 0.9221666666666668 | 0.0038494027114044497 |
+| Epoch_4.pt | 0.9203333333333333 | 0.002393406580948671 |
+| Epoch_8.pt | 0.9196666666666665 | 0.0028087165910587806 |
+| Epoch_2_batch_2999.pt | 0.9163333333333332 | 0.0033314809667922105 |
+| Epoch_1_batch_5999.pt | 0.9121666666666666 | 0.00395226142485165 |
+| Epoch_2.pt | 0.9104999999999999 | 0.003460981806038338 |
+| Epoch_1.pt | 0.9078333333333333 | 0.0031234872881934677 |
+| Epoch_1_batch_2999.pt | 0.8908333333333331 | 0.00355077804028329 |
+| Epoch_0.pt | 0.8371666666666666 | 0.004963135707330873 |
+| Epoch_0_batch_5999.pt | 0.8319999999999999 | 0.005697086108130785 |
+| Epoch_0_batch_2999.pt | 0.6973333333333332 | 0.005917123091419712 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/HRNet/log.log b/bob/bio/facexzoo/models/backbones/HRNet/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..9ace30a9682a474c433a1d2154a39f836214a137
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/HRNet/log.log
@@ -0,0 +1,657 @@
+INFO 2020-12-04 16:45:02 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2020-12-04 16:45:02 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2020-12-04 16:45:02 train.py: 177] Start optimization.
+INFO 2020-12-04 16:45:02 train.py: 178] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='HRNet', batch_size=512, data_root='/home/wangjun492/wj_data/facex-zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10,13,16', tensorboardx_logdir='mv-hrnet', train_file='/home/wangjun492/wj_data/facex-zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f0b2852fef0>)
+backbone param:
+{'NAME': 'cls_hrnet', 'out_h': 7, 'out_w': 7, 'feat_dim': 512, 'IMAGE_SIZE': [112, 112], 'EXTRA': {'STAGE1': {'NUM_MODULES': 1, 'NUM_RANCHES': 1, 'BLOCK': 'BOTTLENECK', 'NUM_BLOCKS': [4], 'NUM_CHANNELS': [64], 'FUSE_METHOD': 'SUM'}, 'STAGE2': {'NUM_MODULES': 1, 'NUM_BRANCHES': 2, 'BLOCK': 'BASIC', 'NUM_BLOCKS': [4, 4], 'NUM_CHANNELS': [18, 36], 'FUSE_METHOD': 'SUM'}, 'STAGE3': {'NUM_MODULES': 4, 'NUM_BRANCHES': 3, 'BLOCK': 'BASIC', 'NUM_BLOCKS': [4, 4, 4], 'NUM_CHANNELS': [18, 36, 72], 'FUSE_METHOD': 'SUM'}, 'STAGE4': {'NUM_MODULES': 3, 'NUM_BRANCHES': 4, 'BLOCK': 'BASIC', 'NUM_BLOCKS': [4, 4, 4, 4], 'NUM_CHANNELS': [18, 36, 72, 144], 'FUSE_METHOD': 'SUM'}}}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-12-04 16:45:31 train.py: 79] Epoch 0, iter 0/6416, lr 0.100000, loss 16.185728
+INFO 2020-12-04 16:51:55 train.py: 79] Epoch 0, iter 200/6416, lr 0.100000, loss 15.620639
+INFO 2020-12-04 16:58:20 train.py: 79] Epoch 0, iter 400/6416, lr 0.100000, loss 15.381243
+INFO 2020-12-04 17:04:45 train.py: 79] Epoch 0, iter 600/6416, lr 0.100000, loss 15.335378
+INFO 2020-12-04 17:11:11 train.py: 79] Epoch 0, iter 800/6416, lr 0.100000, loss 15.296765
+INFO 2020-12-04 17:17:36 train.py: 79] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.239479
+INFO 2020-12-04 17:24:02 train.py: 79] Epoch 0, iter 1200/6416, lr 0.100000, loss 15.121688
+INFO 2020-12-04 17:30:28 train.py: 79] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.964488
+INFO 2020-12-04 17:36:53 train.py: 79] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.738796
+INFO 2020-12-04 17:43:18 train.py: 79] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.461993
+INFO 2020-12-04 17:49:44 train.py: 79] Epoch 0, iter 2000/6416, lr 0.100000, loss 14.175071
+INFO 2020-12-04 17:56:09 train.py: 79] Epoch 0, iter 2200/6416, lr 0.100000, loss 13.885569
+INFO 2020-12-04 18:02:35 train.py: 79] Epoch 0, iter 2400/6416, lr 0.100000, loss 13.617845
+INFO 2020-12-04 18:08:59 train.py: 79] Epoch 0, iter 2600/6416, lr 0.100000, loss 13.278031
+INFO 2020-12-04 18:15:25 train.py: 79] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.839781
+INFO 2020-12-04 18:21:49 train.py: 92] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-12-04 18:21:51 train.py: 79] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.470365
+INFO 2020-12-04 18:28:16 train.py: 79] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.110523
+INFO 2020-12-04 18:34:42 train.py: 79] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.842208
+INFO 2020-12-04 18:41:08 train.py: 79] Epoch 0, iter 3600/6416, lr 0.100000, loss 11.675521
+INFO 2020-12-04 18:47:33 train.py: 79] Epoch 0, iter 3800/6416, lr 0.100000, loss 11.650650
+INFO 2020-12-04 18:53:57 train.py: 79] Epoch 0, iter 4000/6416, lr 0.100000, loss 11.749744
+INFO 2020-12-04 19:00:21 train.py: 79] Epoch 0, iter 4200/6416, lr 0.100000, loss 11.984227
+INFO 2020-12-04 19:06:44 train.py: 79] Epoch 0, iter 4400/6416, lr 0.100000, loss 12.304579
+INFO 2020-12-04 19:13:07 train.py: 79] Epoch 0, iter 4600/6416, lr 0.100000, loss 12.625556
+INFO 2020-12-04 19:19:30 train.py: 79] Epoch 0, iter 4800/6416, lr 0.100000, loss 12.969390
+INFO 2020-12-04 19:25:53 train.py: 79] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.193763
+INFO 2020-12-04 19:32:14 train.py: 79] Epoch 0, iter 5200/6416, lr 0.100000, loss 13.388941
+INFO 2020-12-04 19:38:36 train.py: 79] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.473217
+INFO 2020-12-04 19:44:57 train.py: 79] Epoch 0, iter 5600/6416, lr 0.100000, loss 13.433913
+INFO 2020-12-04 19:51:19 train.py: 79] Epoch 0, iter 5800/6416, lr 0.100000, loss 13.340193
+INFO 2020-12-04 19:57:39 train.py: 92] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-12-04 19:57:41 train.py: 79] Epoch 0, iter 6000/6416, lr 0.100000, loss 13.174352
+INFO 2020-12-04 20:04:03 train.py: 79] Epoch 0, iter 6200/6416, lr 0.100000, loss 12.932402
+INFO 2020-12-04 20:10:23 train.py: 79] Epoch 0, iter 6400/6416, lr 0.100000, loss 12.625778
+INFO 2020-12-04 20:10:52 train.py: 97] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-12-04 20:10:55 train.py: 79] Epoch 1, iter 0/6416, lr 0.100000, loss 12.427638
+INFO 2020-12-04 20:17:15 train.py: 79] Epoch 1, iter 200/6416, lr 0.100000, loss 12.079420
+INFO 2020-12-04 20:23:35 train.py: 79] Epoch 1, iter 400/6416, lr 0.100000, loss 11.768555
+INFO 2020-12-04 20:29:55 train.py: 79] Epoch 1, iter 600/6416, lr 0.100000, loss 11.424818
+INFO 2020-12-04 20:36:16 train.py: 79] Epoch 1, iter 800/6416, lr 0.100000, loss 11.129945
+INFO 2020-12-04 20:42:36 train.py: 79] Epoch 1, iter 1000/6416, lr 0.100000, loss 10.825961
+INFO 2020-12-04 20:48:57 train.py: 79] Epoch 1, iter 1200/6416, lr 0.100000, loss 10.503853
+INFO 2020-12-04 20:55:16 train.py: 79] Epoch 1, iter 1400/6416, lr 0.100000, loss 10.213863
+INFO 2020-12-04 21:01:36 train.py: 79] Epoch 1, iter 1600/6416, lr 0.100000, loss 9.947780
+INFO 2020-12-04 21:07:56 train.py: 79] Epoch 1, iter 1800/6416, lr 0.100000, loss 9.703285
+INFO 2020-12-04 21:14:16 train.py: 79] Epoch 1, iter 2000/6416, lr 0.100000, loss 9.421163
+INFO 2020-12-04 21:20:34 train.py: 79] Epoch 1, iter 2200/6416, lr 0.100000, loss 9.202775
+INFO 2020-12-04 21:26:55 train.py: 79] Epoch 1, iter 2400/6416, lr 0.100000, loss 8.958940
+INFO 2020-12-04 21:33:14 train.py: 79] Epoch 1, iter 2600/6416, lr 0.100000, loss 8.743585
+INFO 2020-12-04 21:39:34 train.py: 79] Epoch 1, iter 2800/6416, lr 0.100000, loss 8.536712
+INFO 2020-12-04 21:45:53 train.py: 92] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-12-04 21:45:54 train.py: 79] Epoch 1, iter 3000/6416, lr 0.100000, loss 8.372091
+INFO 2020-12-04 21:52:14 train.py: 79] Epoch 1, iter 3200/6416, lr 0.100000, loss 8.190922
+INFO 2020-12-04 21:58:35 train.py: 79] Epoch 1, iter 3400/6416, lr 0.100000, loss 8.022299
+INFO 2020-12-04 22:04:54 train.py: 79] Epoch 1, iter 3600/6416, lr 0.100000, loss 7.868789
+INFO 2020-12-04 22:11:14 train.py: 79] Epoch 1, iter 3800/6416, lr 0.100000, loss 7.726336
+INFO 2020-12-04 22:17:34 train.py: 79] Epoch 1, iter 4000/6416, lr 0.100000, loss 7.586878
+INFO 2020-12-04 22:23:54 train.py: 79] Epoch 1, iter 4200/6416, lr 0.100000, loss 7.453715
+INFO 2020-12-04 22:30:14 train.py: 79] Epoch 1, iter 4400/6416, lr 0.100000, loss 7.349989
+INFO 2020-12-04 22:36:36 train.py: 79] Epoch 1, iter 4600/6416, lr 0.100000, loss 7.243860
+INFO 2020-12-04 22:42:55 train.py: 79] Epoch 1, iter 4800/6416, lr 0.100000, loss 7.157616
+INFO 2020-12-04 22:49:16 train.py: 79] Epoch 1, iter 5000/6416, lr 0.100000, loss 7.052401
+INFO 2020-12-04 22:55:36 train.py: 79] Epoch 1, iter 5200/6416, lr 0.100000, loss 6.954512
+INFO 2020-12-04 23:01:57 train.py: 79] Epoch 1, iter 5400/6416, lr 0.100000, loss 6.891003
+INFO 2020-12-04 23:08:16 train.py: 79] Epoch 1, iter 5600/6416, lr 0.100000, loss 6.785696
+INFO 2020-12-04 23:14:36 train.py: 79] Epoch 1, iter 5800/6416, lr 0.100000, loss 6.731474
+INFO 2020-12-04 23:20:55 train.py: 92] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-12-04 23:20:57 train.py: 79] Epoch 1, iter 6000/6416, lr 0.100000, loss 6.633651
+INFO 2020-12-04 23:27:17 train.py: 79] Epoch 1, iter 6200/6416, lr 0.100000, loss 6.575952
+INFO 2020-12-04 23:33:38 train.py: 79] Epoch 1, iter 6400/6416, lr 0.100000, loss 6.521167
+INFO 2020-12-04 23:34:07 train.py: 97] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-12-04 23:34:10 train.py: 79] Epoch 2, iter 0/6416, lr 0.100000, loss 6.475440
+INFO 2020-12-04 23:40:30 train.py: 79] Epoch 2, iter 200/6416, lr 0.100000, loss 5.856195
+INFO 2020-12-04 23:46:49 train.py: 79] Epoch 2, iter 400/6416, lr 0.100000, loss 5.842575
+INFO 2020-12-04 23:53:07 train.py: 79] Epoch 2, iter 600/6416, lr 0.100000, loss 5.849176
+INFO 2020-12-04 23:59:26 train.py: 79] Epoch 2, iter 800/6416, lr 0.100000, loss 5.904193
+INFO 2020-12-05 00:05:45 train.py: 79] Epoch 2, iter 1000/6416, lr 0.100000, loss 5.938872
+INFO 2020-12-05 00:12:04 train.py: 79] Epoch 2, iter 1200/6416, lr 0.100000, loss 5.929337
+INFO 2020-12-05 00:18:24 train.py: 79] Epoch 2, iter 1400/6416, lr 0.100000, loss 5.914740
+INFO 2020-12-05 00:24:44 train.py: 79] Epoch 2, iter 1600/6416, lr 0.100000, loss 5.925265
+INFO 2020-12-05 00:31:04 train.py: 79] Epoch 2, iter 1800/6416, lr 0.100000, loss 5.898670
+INFO 2020-12-05 00:37:23 train.py: 79] Epoch 2, iter 2000/6416, lr 0.100000, loss 5.856751
+INFO 2020-12-05 00:43:45 train.py: 79] Epoch 2, iter 2200/6416, lr 0.100000, loss 5.851210
+INFO 2020-12-05 00:50:03 train.py: 79] Epoch 2, iter 2400/6416, lr 0.100000, loss 5.837268
+INFO 2020-12-05 00:56:23 train.py: 79] Epoch 2, iter 2600/6416, lr 0.100000, loss 5.805586
+INFO 2020-12-05 01:02:42 train.py: 79] Epoch 2, iter 2800/6416, lr 0.100000, loss 5.777944
+INFO 2020-12-05 01:09:01 train.py: 92] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-12-05 01:09:03 train.py: 79] Epoch 2, iter 3000/6416, lr 0.100000, loss 5.754814
+INFO 2020-12-05 01:15:23 train.py: 79] Epoch 2, iter 3200/6416, lr 0.100000, loss 5.704849
+INFO 2020-12-05 01:21:44 train.py: 79] Epoch 2, iter 3400/6416, lr 0.100000, loss 5.659240
+INFO 2020-12-05 01:28:05 train.py: 79] Epoch 2, iter 3600/6416, lr 0.100000, loss 5.678851
+INFO 2020-12-05 01:34:25 train.py: 79] Epoch 2, iter 3800/6416, lr 0.100000, loss 5.638702
+INFO 2020-12-05 01:40:44 train.py: 79] Epoch 2, iter 4000/6416, lr 0.100000, loss 5.607063
+INFO 2020-12-05 01:47:04 train.py: 79] Epoch 2, iter 4200/6416, lr 0.100000, loss 5.572677
+INFO 2020-12-05 01:53:24 train.py: 79] Epoch 2, iter 4400/6416, lr 0.100000, loss 5.526475
+INFO 2020-12-05 01:59:44 train.py: 79] Epoch 2, iter 4600/6416, lr 0.100000, loss 5.516565
+INFO 2020-12-05 02:06:05 train.py: 79] Epoch 2, iter 4800/6416, lr 0.100000, loss 5.509953
+INFO 2020-12-05 02:12:24 train.py: 79] Epoch 2, iter 5000/6416, lr 0.100000, loss 5.461623
+INFO 2020-12-05 02:18:44 train.py: 79] Epoch 2, iter 5200/6416, lr 0.100000, loss 5.435464
+INFO 2020-12-05 02:25:05 train.py: 79] Epoch 2, iter 5400/6416, lr 0.100000, loss 5.446301
+INFO 2020-12-05 02:31:26 train.py: 79] Epoch 2, iter 5600/6416, lr 0.100000, loss 5.395694
+INFO 2020-12-05 02:37:46 train.py: 79] Epoch 2, iter 5800/6416, lr 0.100000, loss 5.363104
+INFO 2020-12-05 02:44:05 train.py: 92] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-12-05 02:44:07 train.py: 79] Epoch 2, iter 6000/6416, lr 0.100000, loss 5.338967
+INFO 2020-12-05 02:50:27 train.py: 79] Epoch 2, iter 6200/6416, lr 0.100000, loss 5.322866
+INFO 2020-12-05 02:56:49 train.py: 79] Epoch 2, iter 6400/6416, lr 0.100000, loss 5.283296
+INFO 2020-12-05 02:57:17 train.py: 97] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-12-05 02:57:20 train.py: 79] Epoch 3, iter 0/6416, lr 0.100000, loss 5.215018
+INFO 2020-12-05 03:03:40 train.py: 79] Epoch 3, iter 200/6416, lr 0.100000, loss 4.757631
+INFO 2020-12-05 03:09:59 train.py: 79] Epoch 3, iter 400/6416, lr 0.100000, loss 4.673967
+INFO 2020-12-05 03:16:19 train.py: 79] Epoch 3, iter 600/6416, lr 0.100000, loss 4.763493
+INFO 2020-12-05 03:22:37 train.py: 79] Epoch 3, iter 800/6416, lr 0.100000, loss 4.818247
+INFO 2020-12-05 03:28:57 train.py: 79] Epoch 3, iter 1000/6416, lr 0.100000, loss 4.869078
+INFO 2020-12-05 03:35:16 train.py: 79] Epoch 3, iter 1200/6416, lr 0.100000, loss 4.898946
+INFO 2020-12-05 03:41:36 train.py: 79] Epoch 3, iter 1400/6416, lr 0.100000, loss 4.918872
+INFO 2020-12-05 03:47:56 train.py: 79] Epoch 3, iter 1600/6416, lr 0.100000, loss 4.921781
+INFO 2020-12-05 03:54:16 train.py: 79] Epoch 3, iter 1800/6416, lr 0.100000, loss 4.938617
+INFO 2020-12-05 04:00:36 train.py: 79] Epoch 3, iter 2000/6416, lr 0.100000, loss 4.953668
+INFO 2020-12-05 04:06:56 train.py: 79] Epoch 3, iter 2200/6416, lr 0.100000, loss 4.923477
+INFO 2020-12-05 04:13:17 train.py: 79] Epoch 3, iter 2400/6416, lr 0.100000, loss 4.927328
+INFO 2020-12-05 04:19:38 train.py: 79] Epoch 3, iter 2600/6416, lr 0.100000, loss 4.919026
+INFO 2020-12-05 04:25:58 train.py: 79] Epoch 3, iter 2800/6416, lr 0.100000, loss 4.920281
+INFO 2020-12-05 04:32:17 train.py: 92] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-12-05 04:32:19 train.py: 79] Epoch 3, iter 3000/6416, lr 0.100000, loss 4.913823
+INFO 2020-12-05 04:38:39 train.py: 79] Epoch 3, iter 3200/6416, lr 0.100000, loss 4.895135
+INFO 2020-12-05 04:44:59 train.py: 79] Epoch 3, iter 3400/6416, lr 0.100000, loss 4.895300
+INFO 2020-12-05 04:51:21 train.py: 79] Epoch 3, iter 3600/6416, lr 0.100000, loss 4.880908
+INFO 2020-12-05 04:57:40 train.py: 79] Epoch 3, iter 3800/6416, lr 0.100000, loss 4.865934
+INFO 2020-12-05 05:04:03 train.py: 79] Epoch 3, iter 4000/6416, lr 0.100000, loss 4.888324
+INFO 2020-12-05 05:10:23 train.py: 79] Epoch 3, iter 4200/6416, lr 0.100000, loss 4.886136
+INFO 2020-12-05 05:16:45 train.py: 79] Epoch 3, iter 4400/6416, lr 0.100000, loss 4.870280
+INFO 2020-12-05 05:23:04 train.py: 79] Epoch 3, iter 4600/6416, lr 0.100000, loss 4.828404
+INFO 2020-12-05 05:29:26 train.py: 79] Epoch 3, iter 4800/6416, lr 0.100000, loss 4.779614
+INFO 2020-12-05 05:35:45 train.py: 79] Epoch 3, iter 5000/6416, lr 0.100000, loss 4.809465
+INFO 2020-12-05 05:42:06 train.py: 79] Epoch 3, iter 5200/6416, lr 0.100000, loss 4.775767
+INFO 2020-12-05 05:48:26 train.py: 79] Epoch 3, iter 5400/6416, lr 0.100000, loss 4.803928
+INFO 2020-12-05 05:54:48 train.py: 79] Epoch 3, iter 5600/6416, lr 0.100000, loss 4.781010
+INFO 2020-12-05 06:01:08 train.py: 79] Epoch 3, iter 5800/6416, lr 0.100000, loss 4.747395
+INFO 2020-12-05 06:07:29 train.py: 92] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-12-05 06:07:30 train.py: 79] Epoch 3, iter 6000/6416, lr 0.100000, loss 4.735175
+INFO 2020-12-05 06:13:51 train.py: 79] Epoch 3, iter 6200/6416, lr 0.100000, loss 4.755426
+INFO 2020-12-05 06:20:12 train.py: 79] Epoch 3, iter 6400/6416, lr 0.100000, loss 4.720754
+INFO 2020-12-05 06:20:41 train.py: 97] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-12-05 06:20:44 train.py: 79] Epoch 4, iter 0/6416, lr 0.100000, loss 4.737357
+INFO 2020-12-05 06:27:03 train.py: 79] Epoch 4, iter 200/6416, lr 0.100000, loss 4.189590
+INFO 2020-12-05 06:33:22 train.py: 79] Epoch 4, iter 400/6416, lr 0.100000, loss 4.142774
+INFO 2020-12-05 06:39:42 train.py: 79] Epoch 4, iter 600/6416, lr 0.100000, loss 4.217251
+INFO 2020-12-05 06:46:01 train.py: 79] Epoch 4, iter 800/6416, lr 0.100000, loss 4.274831
+INFO 2020-12-05 06:52:20 train.py: 79] Epoch 4, iter 1000/6416, lr 0.100000, loss 4.344759
+INFO 2020-12-05 06:58:40 train.py: 79] Epoch 4, iter 1200/6416, lr 0.100000, loss 4.357843
+INFO 2020-12-05 07:04:59 train.py: 79] Epoch 4, iter 1400/6416, lr 0.100000, loss 4.373999
+INFO 2020-12-05 07:11:18 train.py: 79] Epoch 4, iter 1600/6416, lr 0.100000, loss 4.431645
+INFO 2020-12-05 07:17:38 train.py: 79] Epoch 4, iter 1800/6416, lr 0.100000, loss 4.459793
+INFO 2020-12-05 07:23:58 train.py: 79] Epoch 4, iter 2000/6416, lr 0.100000, loss 4.469297
+INFO 2020-12-05 07:30:17 train.py: 79] Epoch 4, iter 2200/6416, lr 0.100000, loss 4.484062
+INFO 2020-12-05 07:36:37 train.py: 79] Epoch 4, iter 2400/6416, lr 0.100000, loss 4.457344
+INFO 2020-12-05 07:42:56 train.py: 79] Epoch 4, iter 2600/6416, lr 0.100000, loss 4.484542
+INFO 2020-12-05 07:49:16 train.py: 79] Epoch 4, iter 2800/6416, lr 0.100000, loss 4.492071
+INFO 2020-12-05 07:55:37 train.py: 92] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-12-05 07:55:39 train.py: 79] Epoch 4, iter 3000/6416, lr 0.100000, loss 4.465983
+INFO 2020-12-05 08:01:59 train.py: 79] Epoch 4, iter 3200/6416, lr 0.100000, loss 4.478256
+INFO 2020-12-05 08:08:18 train.py: 79] Epoch 4, iter 3400/6416, lr 0.100000, loss 4.476951
+INFO 2020-12-05 08:14:38 train.py: 79] Epoch 4, iter 3600/6416, lr 0.100000, loss 4.488107
+INFO 2020-12-05 08:20:58 train.py: 79] Epoch 4, iter 3800/6416, lr 0.100000, loss 4.439960
+INFO 2020-12-05 08:27:18 train.py: 79] Epoch 4, iter 4000/6416, lr 0.100000, loss 4.466088
+INFO 2020-12-05 08:33:39 train.py: 79] Epoch 4, iter 4200/6416, lr 0.100000, loss 4.460900
+INFO 2020-12-05 08:39:58 train.py: 79] Epoch 4, iter 4400/6416, lr 0.100000, loss 4.428038
+INFO 2020-12-05 08:46:19 train.py: 79] Epoch 4, iter 4600/6416, lr 0.100000, loss 4.463020
+INFO 2020-12-05 08:52:39 train.py: 79] Epoch 4, iter 4800/6416, lr 0.100000, loss 4.422682
+INFO 2020-12-05 08:59:01 train.py: 79] Epoch 4, iter 5000/6416, lr 0.100000, loss 4.437150
+INFO 2020-12-05 09:05:22 train.py: 79] Epoch 4, iter 5200/6416, lr 0.100000, loss 4.443850
+INFO 2020-12-05 09:11:43 train.py: 79] Epoch 4, iter 5400/6416, lr 0.100000, loss 4.396583
+INFO 2020-12-05 09:18:02 train.py: 79] Epoch 4, iter 5600/6416, lr 0.100000, loss 4.386658
+INFO 2020-12-05 09:24:23 train.py: 79] Epoch 4, iter 5800/6416, lr 0.100000, loss 4.389865
+INFO 2020-12-05 09:30:42 train.py: 92] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-12-05 09:30:44 train.py: 79] Epoch 4, iter 6000/6416, lr 0.100000, loss 4.412408
+INFO 2020-12-05 09:37:05 train.py: 79] Epoch 4, iter 6200/6416, lr 0.100000, loss 4.365090
+INFO 2020-12-05 09:43:25 train.py: 79] Epoch 4, iter 6400/6416, lr 0.100000, loss 4.405208
+INFO 2020-12-05 09:43:53 train.py: 97] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-12-05 09:43:56 train.py: 79] Epoch 5, iter 0/6416, lr 0.100000, loss 4.373674
+INFO 2020-12-05 09:50:15 train.py: 79] Epoch 5, iter 200/6416, lr 0.100000, loss 3.871244
+INFO 2020-12-05 09:56:35 train.py: 79] Epoch 5, iter 400/6416, lr 0.100000, loss 3.822520
+INFO 2020-12-05 10:02:53 train.py: 79] Epoch 5, iter 600/6416, lr 0.100000, loss 3.850365
+INFO 2020-12-05 10:09:13 train.py: 79] Epoch 5, iter 800/6416, lr 0.100000, loss 3.959184
+INFO 2020-12-05 10:15:32 train.py: 79] Epoch 5, iter 1000/6416, lr 0.100000, loss 4.012598
+INFO 2020-12-05 10:21:52 train.py: 79] Epoch 5, iter 1200/6416, lr 0.100000, loss 4.065327
+INFO 2020-12-05 10:28:10 train.py: 79] Epoch 5, iter 1400/6416, lr 0.100000, loss 4.085495
+INFO 2020-12-05 10:34:32 train.py: 79] Epoch 5, iter 1600/6416, lr 0.100000, loss 4.128043
+INFO 2020-12-05 10:40:50 train.py: 79] Epoch 5, iter 1800/6416, lr 0.100000, loss 4.149458
+INFO 2020-12-05 10:47:10 train.py: 79] Epoch 5, iter 2000/6416, lr 0.100000, loss 4.148260
+INFO 2020-12-05 10:53:29 train.py: 79] Epoch 5, iter 2200/6416, lr 0.100000, loss 4.157678
+INFO 2020-12-05 10:59:50 train.py: 79] Epoch 5, iter 2400/6416, lr 0.100000, loss 4.198733
+INFO 2020-12-05 11:06:09 train.py: 79] Epoch 5, iter 2600/6416, lr 0.100000, loss 4.190339
+INFO 2020-12-05 11:12:29 train.py: 79] Epoch 5, iter 2800/6416, lr 0.100000, loss 4.218966
+INFO 2020-12-05 11:18:46 train.py: 92] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-12-05 11:18:48 train.py: 79] Epoch 5, iter 3000/6416, lr 0.100000, loss 4.185984
+INFO 2020-12-05 11:25:08 train.py: 79] Epoch 5, iter 3200/6416, lr 0.100000, loss 4.204819
+INFO 2020-12-05 11:31:28 train.py: 79] Epoch 5, iter 3400/6416, lr 0.100000, loss 4.218199
+INFO 2020-12-05 11:37:47 train.py: 79] Epoch 5, iter 3600/6416, lr 0.100000, loss 4.224664
+INFO 2020-12-05 11:44:08 train.py: 79] Epoch 5, iter 3800/6416, lr 0.100000, loss 4.192363
+INFO 2020-12-05 11:50:27 train.py: 79] Epoch 5, iter 4000/6416, lr 0.100000, loss 4.190297
+INFO 2020-12-05 11:56:47 train.py: 79] Epoch 5, iter 4200/6416, lr 0.100000, loss 4.215930
+INFO 2020-12-05 12:03:06 train.py: 79] Epoch 5, iter 4400/6416, lr 0.100000, loss 4.185400
+INFO 2020-12-05 12:09:27 train.py: 79] Epoch 5, iter 4600/6416, lr 0.100000, loss 4.187785
+INFO 2020-12-05 12:15:47 train.py: 79] Epoch 5, iter 4800/6416, lr 0.100000, loss 4.178934
+INFO 2020-12-05 12:22:07 train.py: 79] Epoch 5, iter 5000/6416, lr 0.100000, loss 4.166705
+INFO 2020-12-05 12:28:28 train.py: 79] Epoch 5, iter 5200/6416, lr 0.100000, loss 4.147225
+INFO 2020-12-05 12:34:49 train.py: 79] Epoch 5, iter 5400/6416, lr 0.100000, loss 4.188363
+INFO 2020-12-05 12:41:09 train.py: 79] Epoch 5, iter 5600/6416, lr 0.100000, loss 4.192761
+INFO 2020-12-05 12:47:30 train.py: 79] Epoch 5, iter 5800/6416, lr 0.100000, loss 4.176323
+INFO 2020-12-05 12:53:48 train.py: 92] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-12-05 12:53:50 train.py: 79] Epoch 5, iter 6000/6416, lr 0.100000, loss 4.150935
+INFO 2020-12-05 13:00:10 train.py: 79] Epoch 5, iter 6200/6416, lr 0.100000, loss 4.153237
+INFO 2020-12-05 13:06:31 train.py: 79] Epoch 5, iter 6400/6416, lr 0.100000, loss 4.163235
+INFO 2020-12-05 13:07:00 train.py: 97] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-12-05 13:07:03 train.py: 79] Epoch 6, iter 0/6416, lr 0.100000, loss 4.141977
+INFO 2020-12-05 13:13:22 train.py: 79] Epoch 6, iter 200/6416, lr 0.100000, loss 3.673644
+INFO 2020-12-05 13:19:41 train.py: 79] Epoch 6, iter 400/6416, lr 0.100000, loss 3.594239
+INFO 2020-12-05 13:26:00 train.py: 79] Epoch 6, iter 600/6416, lr 0.100000, loss 3.647684
+INFO 2020-12-05 13:32:19 train.py: 79] Epoch 6, iter 800/6416, lr 0.100000, loss 3.756734
+INFO 2020-12-05 13:38:38 train.py: 79] Epoch 6, iter 1000/6416, lr 0.100000, loss 3.808701
+INFO 2020-12-05 13:44:57 train.py: 79] Epoch 6, iter 1200/6416, lr 0.100000, loss 3.845456
+INFO 2020-12-05 13:51:16 train.py: 79] Epoch 6, iter 1400/6416, lr 0.100000, loss 3.872381
+INFO 2020-12-05 13:57:36 train.py: 79] Epoch 6, iter 1600/6416, lr 0.100000, loss 3.909472
+INFO 2020-12-05 14:03:56 train.py: 79] Epoch 6, iter 1800/6416, lr 0.100000, loss 3.922629
+INFO 2020-12-05 14:10:14 train.py: 79] Epoch 6, iter 2000/6416, lr 0.100000, loss 3.950967
+INFO 2020-12-05 14:16:34 train.py: 79] Epoch 6, iter 2200/6416, lr 0.100000, loss 3.966451
+INFO 2020-12-05 14:22:53 train.py: 79] Epoch 6, iter 2400/6416, lr 0.100000, loss 4.004148
+INFO 2020-12-05 14:29:13 train.py: 79] Epoch 6, iter 2600/6416, lr 0.100000, loss 3.990190
+INFO 2020-12-05 14:35:32 train.py: 79] Epoch 6, iter 2800/6416, lr 0.100000, loss 3.976946
+INFO 2020-12-05 14:41:51 train.py: 92] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-12-05 14:41:53 train.py: 79] Epoch 6, iter 3000/6416, lr 0.100000, loss 4.006734
+INFO 2020-12-05 14:48:12 train.py: 79] Epoch 6, iter 3200/6416, lr 0.100000, loss 4.028544
+INFO 2020-12-05 14:54:32 train.py: 79] Epoch 6, iter 3400/6416, lr 0.100000, loss 4.024532
+INFO 2020-12-05 15:00:51 train.py: 79] Epoch 6, iter 3600/6416, lr 0.100000, loss 3.988816
+INFO 2020-12-05 15:07:11 train.py: 79] Epoch 6, iter 3800/6416, lr 0.100000, loss 4.010228
+INFO 2020-12-05 15:13:31 train.py: 79] Epoch 6, iter 4000/6416, lr 0.100000, loss 4.008840
+INFO 2020-12-05 15:19:52 train.py: 79] Epoch 6, iter 4200/6416, lr 0.100000, loss 3.994895
+INFO 2020-12-05 15:26:12 train.py: 79] Epoch 6, iter 4400/6416, lr 0.100000, loss 4.003911
+INFO 2020-12-05 15:32:33 train.py: 79] Epoch 6, iter 4600/6416, lr 0.100000, loss 3.998039
+INFO 2020-12-05 15:38:52 train.py: 79] Epoch 6, iter 4800/6416, lr 0.100000, loss 3.976100
+INFO 2020-12-05 15:45:13 train.py: 79] Epoch 6, iter 5000/6416, lr 0.100000, loss 4.007036
+INFO 2020-12-05 15:51:33 train.py: 79] Epoch 6, iter 5200/6416, lr 0.100000, loss 4.014025
+INFO 2020-12-05 15:57:54 train.py: 79] Epoch 6, iter 5400/6416, lr 0.100000, loss 4.004497
+INFO 2020-12-05 16:04:14 train.py: 79] Epoch 6, iter 5600/6416, lr 0.100000, loss 4.014919
+INFO 2020-12-05 16:10:36 train.py: 79] Epoch 6, iter 5800/6416, lr 0.100000, loss 4.015862
+INFO 2020-12-05 16:16:56 train.py: 92] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-12-05 16:16:58 train.py: 79] Epoch 6, iter 6000/6416, lr 0.100000, loss 4.011972
+INFO 2020-12-05 16:23:18 train.py: 79] Epoch 6, iter 6200/6416, lr 0.100000, loss 3.993713
+INFO 2020-12-05 16:29:39 train.py: 79] Epoch 6, iter 6400/6416, lr 0.100000, loss 3.981413
+INFO 2020-12-05 16:30:07 train.py: 97] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-12-05 16:30:10 train.py: 79] Epoch 7, iter 0/6416, lr 0.100000, loss 4.081297
+INFO 2020-12-05 16:36:30 train.py: 79] Epoch 7, iter 200/6416, lr 0.100000, loss 3.492259
+INFO 2020-12-05 16:42:48 train.py: 79] Epoch 7, iter 400/6416, lr 0.100000, loss 3.437357
+INFO 2020-12-05 16:49:08 train.py: 79] Epoch 7, iter 600/6416, lr 0.100000, loss 3.509000
+INFO 2020-12-05 16:55:27 train.py: 79] Epoch 7, iter 800/6416, lr 0.100000, loss 3.573317
+INFO 2020-12-05 17:01:47 train.py: 79] Epoch 7, iter 1000/6416, lr 0.100000, loss 3.628325
+INFO 2020-12-05 17:08:06 train.py: 79] Epoch 7, iter 1200/6416, lr 0.100000, loss 3.669659
+INFO 2020-12-05 17:14:26 train.py: 79] Epoch 7, iter 1400/6416, lr 0.100000, loss 3.703262
+INFO 2020-12-05 17:20:44 train.py: 79] Epoch 7, iter 1600/6416, lr 0.100000, loss 3.754398
+INFO 2020-12-05 17:27:04 train.py: 79] Epoch 7, iter 1800/6416, lr 0.100000, loss 3.783757
+INFO 2020-12-05 17:33:23 train.py: 79] Epoch 7, iter 2000/6416, lr 0.100000, loss 3.811007
+INFO 2020-12-05 17:39:45 train.py: 79] Epoch 7, iter 2200/6416, lr 0.100000, loss 3.805011
+INFO 2020-12-05 17:46:03 train.py: 79] Epoch 7, iter 2400/6416, lr 0.100000, loss 3.814418
+INFO 2020-12-05 17:52:23 train.py: 79] Epoch 7, iter 2600/6416, lr 0.100000, loss 3.857767
+INFO 2020-12-05 17:58:43 train.py: 79] Epoch 7, iter 2800/6416, lr 0.100000, loss 3.829525
+INFO 2020-12-05 18:05:03 train.py: 92] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-12-05 18:05:05 train.py: 79] Epoch 7, iter 3000/6416, lr 0.100000, loss 3.835798
+INFO 2020-12-05 18:11:26 train.py: 79] Epoch 7, iter 3200/6416, lr 0.100000, loss 3.864374
+INFO 2020-12-05 18:17:46 train.py: 79] Epoch 7, iter 3400/6416, lr 0.100000, loss 3.860680
+INFO 2020-12-05 18:24:07 train.py: 79] Epoch 7, iter 3600/6416, lr 0.100000, loss 3.877339
+INFO 2020-12-05 18:30:27 train.py: 79] Epoch 7, iter 3800/6416, lr 0.100000, loss 3.860563
+INFO 2020-12-05 18:36:48 train.py: 79] Epoch 7, iter 4000/6416, lr 0.100000, loss 3.869084
+INFO 2020-12-05 18:43:08 train.py: 79] Epoch 7, iter 4200/6416, lr 0.100000, loss 3.892466
+INFO 2020-12-05 18:49:30 train.py: 79] Epoch 7, iter 4400/6416, lr 0.100000, loss 3.884217
+INFO 2020-12-05 18:55:49 train.py: 79] Epoch 7, iter 4600/6416, lr 0.100000, loss 3.884133
+INFO 2020-12-05 19:02:11 train.py: 79] Epoch 7, iter 4800/6416, lr 0.100000, loss 3.868462
+INFO 2020-12-05 19:08:30 train.py: 79] Epoch 7, iter 5000/6416, lr 0.100000, loss 3.902635
+INFO 2020-12-05 19:14:51 train.py: 79] Epoch 7, iter 5200/6416, lr 0.100000, loss 3.890947
+INFO 2020-12-05 19:21:11 train.py: 79] Epoch 7, iter 5400/6416, lr 0.100000, loss 3.874897
+INFO 2020-12-05 19:27:31 train.py: 79] Epoch 7, iter 5600/6416, lr 0.100000, loss 3.873735
+INFO 2020-12-05 19:33:51 train.py: 79] Epoch 7, iter 5800/6416, lr 0.100000, loss 3.853881
+INFO 2020-12-05 19:40:10 train.py: 92] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-12-05 19:40:12 train.py: 79] Epoch 7, iter 6000/6416, lr 0.100000, loss 3.878101
+INFO 2020-12-05 19:46:33 train.py: 79] Epoch 7, iter 6200/6416, lr 0.100000, loss 3.876496
+INFO 2020-12-05 19:52:54 train.py: 79] Epoch 7, iter 6400/6416, lr 0.100000, loss 3.890721
+INFO 2020-12-05 19:53:22 train.py: 97] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-12-05 19:53:25 train.py: 79] Epoch 8, iter 0/6416, lr 0.100000, loss 3.938865
+INFO 2020-12-05 19:59:46 train.py: 79] Epoch 8, iter 200/6416, lr 0.100000, loss 3.389026
+INFO 2020-12-05 20:06:05 train.py: 79] Epoch 8, iter 400/6416, lr 0.100000, loss 3.325515
+INFO 2020-12-05 20:12:26 train.py: 79] Epoch 8, iter 600/6416, lr 0.100000, loss 3.380000
+INFO 2020-12-05 20:18:45 train.py: 79] Epoch 8, iter 800/6416, lr 0.100000, loss 3.468307
+INFO 2020-12-05 20:25:04 train.py: 79] Epoch 8, iter 1000/6416, lr 0.100000, loss 3.493926
+INFO 2020-12-05 20:31:23 train.py: 79] Epoch 8, iter 1200/6416, lr 0.100000, loss 3.543463
+INFO 2020-12-05 20:37:42 train.py: 79] Epoch 8, iter 1400/6416, lr 0.100000, loss 3.604663
+INFO 2020-12-05 20:44:02 train.py: 79] Epoch 8, iter 1600/6416, lr 0.100000, loss 3.634113
+INFO 2020-12-05 20:50:22 train.py: 79] Epoch 8, iter 1800/6416, lr 0.100000, loss 3.654702
+INFO 2020-12-05 20:56:42 train.py: 79] Epoch 8, iter 2000/6416, lr 0.100000, loss 3.684589
+INFO 2020-12-05 21:03:02 train.py: 79] Epoch 8, iter 2200/6416, lr 0.100000, loss 3.692347
+INFO 2020-12-05 21:09:22 train.py: 79] Epoch 8, iter 2400/6416, lr 0.100000, loss 3.726210
+INFO 2020-12-05 21:15:41 train.py: 79] Epoch 8, iter 2600/6416, lr 0.100000, loss 3.717885
+INFO 2020-12-05 21:22:01 train.py: 79] Epoch 8, iter 2800/6416, lr 0.100000, loss 3.734899
+INFO 2020-12-05 21:28:19 train.py: 92] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-12-05 21:28:21 train.py: 79] Epoch 8, iter 3000/6416, lr 0.100000, loss 3.747616
+INFO 2020-12-05 21:34:42 train.py: 79] Epoch 8, iter 3200/6416, lr 0.100000, loss 3.740754
+INFO 2020-12-05 21:41:02 train.py: 79] Epoch 8, iter 3400/6416, lr 0.100000, loss 3.754269
+INFO 2020-12-05 21:47:22 train.py: 79] Epoch 8, iter 3600/6416, lr 0.100000, loss 3.764816
+INFO 2020-12-05 21:53:42 train.py: 79] Epoch 8, iter 3800/6416, lr 0.100000, loss 3.789393
+INFO 2020-12-05 22:00:02 train.py: 79] Epoch 8, iter 4000/6416, lr 0.100000, loss 3.750610
+INFO 2020-12-05 22:06:23 train.py: 79] Epoch 8, iter 4200/6416, lr 0.100000, loss 3.767294
+INFO 2020-12-05 22:12:43 train.py: 79] Epoch 8, iter 4400/6416, lr 0.100000, loss 3.754622
+INFO 2020-12-05 22:19:02 train.py: 79] Epoch 8, iter 4600/6416, lr 0.100000, loss 3.767717
+INFO 2020-12-05 22:25:22 train.py: 79] Epoch 8, iter 4800/6416, lr 0.100000, loss 3.763514
+INFO 2020-12-05 22:31:43 train.py: 79] Epoch 8, iter 5000/6416, lr 0.100000, loss 3.777894
+INFO 2020-12-05 22:38:03 train.py: 79] Epoch 8, iter 5200/6416, lr 0.100000, loss 3.768596
+INFO 2020-12-05 22:44:22 train.py: 79] Epoch 8, iter 5400/6416, lr 0.100000, loss 3.775835
+INFO 2020-12-05 22:50:42 train.py: 79] Epoch 8, iter 5600/6416, lr 0.100000, loss 3.772843
+INFO 2020-12-05 22:57:03 train.py: 79] Epoch 8, iter 5800/6416, lr 0.100000, loss 3.773340
+INFO 2020-12-05 23:03:22 train.py: 92] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-12-05 23:03:24 train.py: 79] Epoch 8, iter 6000/6416, lr 0.100000, loss 3.763247
+INFO 2020-12-05 23:09:46 train.py: 79] Epoch 8, iter 6200/6416, lr 0.100000, loss 3.756311
+INFO 2020-12-05 23:16:06 train.py: 79] Epoch 8, iter 6400/6416, lr 0.100000, loss 3.776171
+INFO 2020-12-05 23:16:35 train.py: 97] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-12-05 23:16:38 train.py: 79] Epoch 9, iter 0/6416, lr 0.100000, loss 3.747921
+INFO 2020-12-05 23:22:57 train.py: 79] Epoch 9, iter 200/6416, lr 0.100000, loss 3.303898
+INFO 2020-12-05 23:29:18 train.py: 79] Epoch 9, iter 400/6416, lr 0.100000, loss 3.208295
+INFO 2020-12-05 23:35:37 train.py: 79] Epoch 9, iter 600/6416, lr 0.100000, loss 3.306917
+INFO 2020-12-05 23:41:57 train.py: 79] Epoch 9, iter 800/6416, lr 0.100000, loss 3.366264
+INFO 2020-12-05 23:48:16 train.py: 79] Epoch 9, iter 1000/6416, lr 0.100000, loss 3.412550
+INFO 2020-12-05 23:54:36 train.py: 79] Epoch 9, iter 1200/6416, lr 0.100000, loss 3.461481
+INFO 2020-12-06 00:00:54 train.py: 79] Epoch 9, iter 1400/6416, lr 0.100000, loss 3.508088
+INFO 2020-12-06 00:07:15 train.py: 79] Epoch 9, iter 1600/6416, lr 0.100000, loss 3.538460
+INFO 2020-12-06 00:13:34 train.py: 79] Epoch 9, iter 1800/6416, lr 0.100000, loss 3.571782
+INFO 2020-12-06 00:19:54 train.py: 79] Epoch 9, iter 2000/6416, lr 0.100000, loss 3.602450
+INFO 2020-12-06 00:26:14 train.py: 79] Epoch 9, iter 2200/6416, lr 0.100000, loss 3.612209
+INFO 2020-12-06 00:32:34 train.py: 79] Epoch 9, iter 2400/6416, lr 0.100000, loss 3.592770
+INFO 2020-12-06 00:38:53 train.py: 79] Epoch 9, iter 2600/6416, lr 0.100000, loss 3.640832
+INFO 2020-12-06 00:45:13 train.py: 79] Epoch 9, iter 2800/6416, lr 0.100000, loss 3.644980
+INFO 2020-12-06 00:51:31 train.py: 92] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-12-06 00:51:33 train.py: 79] Epoch 9, iter 3000/6416, lr 0.100000, loss 3.668094
+INFO 2020-12-06 00:57:54 train.py: 79] Epoch 9, iter 3200/6416, lr 0.100000, loss 3.678232
+INFO 2020-12-06 01:04:14 train.py: 79] Epoch 9, iter 3400/6416, lr 0.100000, loss 3.666057
+INFO 2020-12-06 01:10:34 train.py: 79] Epoch 9, iter 3600/6416, lr 0.100000, loss 3.675592
+INFO 2020-12-06 01:16:54 train.py: 79] Epoch 9, iter 3800/6416, lr 0.100000, loss 3.674710
+INFO 2020-12-06 01:23:15 train.py: 79] Epoch 9, iter 4000/6416, lr 0.100000, loss 3.666389
+INFO 2020-12-06 01:29:36 train.py: 79] Epoch 9, iter 4200/6416, lr 0.100000, loss 3.679206
+INFO 2020-12-06 01:35:56 train.py: 79] Epoch 9, iter 4400/6416, lr 0.100000, loss 3.708740
+INFO 2020-12-06 01:42:16 train.py: 79] Epoch 9, iter 4600/6416, lr 0.100000, loss 3.664768
+INFO 2020-12-06 01:48:37 train.py: 79] Epoch 9, iter 4800/6416, lr 0.100000, loss 3.676176
+INFO 2020-12-06 01:54:57 train.py: 79] Epoch 9, iter 5000/6416, lr 0.100000, loss 3.713797
+INFO 2020-12-06 02:01:17 train.py: 79] Epoch 9, iter 5200/6416, lr 0.100000, loss 3.704025
+INFO 2020-12-06 02:07:37 train.py: 79] Epoch 9, iter 5400/6416, lr 0.100000, loss 3.667043
+INFO 2020-12-06 02:13:57 train.py: 79] Epoch 9, iter 5600/6416, lr 0.100000, loss 3.698179
+INFO 2020-12-06 02:20:19 train.py: 79] Epoch 9, iter 5800/6416, lr 0.100000, loss 3.678683
+INFO 2020-12-06 02:26:40 train.py: 92] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-12-06 02:26:42 train.py: 79] Epoch 9, iter 6000/6416, lr 0.100000, loss 3.679635
+INFO 2020-12-06 02:33:02 train.py: 79] Epoch 9, iter 6200/6416, lr 0.100000, loss 3.697636
+INFO 2020-12-06 02:39:22 train.py: 79] Epoch 9, iter 6400/6416, lr 0.100000, loss 3.663839
+INFO 2020-12-06 02:39:50 train.py: 97] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-12-06 02:39:53 train.py: 79] Epoch 10, iter 0/6416, lr 0.010000, loss 3.736168
+INFO 2020-12-06 02:46:13 train.py: 79] Epoch 10, iter 200/6416, lr 0.010000, loss 2.609726
+INFO 2020-12-06 02:52:33 train.py: 79] Epoch 10, iter 400/6416, lr 0.010000, loss 2.384023
+INFO 2020-12-06 02:58:52 train.py: 79] Epoch 10, iter 600/6416, lr 0.010000, loss 2.298823
+INFO 2020-12-06 03:05:12 train.py: 79] Epoch 10, iter 800/6416, lr 0.010000, loss 2.262948
+INFO 2020-12-06 03:11:31 train.py: 79] Epoch 10, iter 1000/6416, lr 0.010000, loss 2.177132
+INFO 2020-12-06 03:17:51 train.py: 79] Epoch 10, iter 1200/6416, lr 0.010000, loss 2.148990
+INFO 2020-12-06 03:24:10 train.py: 79] Epoch 10, iter 1400/6416, lr 0.010000, loss 2.120849
+INFO 2020-12-06 03:30:29 train.py: 79] Epoch 10, iter 1600/6416, lr 0.010000, loss 2.086983
+INFO 2020-12-06 03:36:50 train.py: 79] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.047091
+INFO 2020-12-06 03:43:10 train.py: 79] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.040680
+INFO 2020-12-06 03:49:30 train.py: 79] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.001162
+INFO 2020-12-06 03:55:50 train.py: 79] Epoch 10, iter 2400/6416, lr 0.010000, loss 1.978806
+INFO 2020-12-06 04:02:09 train.py: 79] Epoch 10, iter 2600/6416, lr 0.010000, loss 1.972213
+INFO 2020-12-06 04:08:29 train.py: 79] Epoch 10, iter 2800/6416, lr 0.010000, loss 1.933587
+INFO 2020-12-06 04:14:46 train.py: 92] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-12-06 04:14:48 train.py: 79] Epoch 10, iter 3000/6416, lr 0.010000, loss 1.916844
+INFO 2020-12-06 04:21:09 train.py: 79] Epoch 10, iter 3200/6416, lr 0.010000, loss 1.903127
+INFO 2020-12-06 04:27:29 train.py: 79] Epoch 10, iter 3400/6416, lr 0.010000, loss 1.881840
+INFO 2020-12-06 04:33:48 train.py: 79] Epoch 10, iter 3600/6416, lr 0.010000, loss 1.878274
+INFO 2020-12-06 04:40:11 train.py: 79] Epoch 10, iter 3800/6416, lr 0.010000, loss 1.875843
+INFO 2020-12-06 04:46:30 train.py: 79] Epoch 10, iter 4000/6416, lr 0.010000, loss 1.851889
+INFO 2020-12-06 04:52:51 train.py: 79] Epoch 10, iter 4200/6416, lr 0.010000, loss 1.833614
+INFO 2020-12-06 04:59:11 train.py: 79] Epoch 10, iter 4400/6416, lr 0.010000, loss 1.810368
+INFO 2020-12-06 05:05:33 train.py: 79] Epoch 10, iter 4600/6416, lr 0.010000, loss 1.816090
+INFO 2020-12-06 05:11:54 train.py: 79] Epoch 10, iter 4800/6416, lr 0.010000, loss 1.794370
+INFO 2020-12-06 05:18:14 train.py: 79] Epoch 10, iter 5000/6416, lr 0.010000, loss 1.794215
+INFO 2020-12-06 05:24:33 train.py: 79] Epoch 10, iter 5200/6416, lr 0.010000, loss 1.756364
+INFO 2020-12-06 05:30:54 train.py: 79] Epoch 10, iter 5400/6416, lr 0.010000, loss 1.757093
+INFO 2020-12-06 05:37:13 train.py: 79] Epoch 10, iter 5600/6416, lr 0.010000, loss 1.757041
+INFO 2020-12-06 05:43:36 train.py: 79] Epoch 10, iter 5800/6416, lr 0.010000, loss 1.748112
+INFO 2020-12-06 05:49:54 train.py: 92] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-12-06 05:49:56 train.py: 79] Epoch 10, iter 6000/6416, lr 0.010000, loss 1.749999
+INFO 2020-12-06 05:56:15 train.py: 79] Epoch 10, iter 6200/6416, lr 0.010000, loss 1.738492
+INFO 2020-12-06 06:02:37 train.py: 79] Epoch 10, iter 6400/6416, lr 0.010000, loss 1.734276
+INFO 2020-12-06 06:03:05 train.py: 97] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-12-06 06:03:08 train.py: 79] Epoch 11, iter 0/6416, lr 0.010000, loss 1.749832
+INFO 2020-12-06 06:09:28 train.py: 79] Epoch 11, iter 200/6416, lr 0.010000, loss 1.456448
+INFO 2020-12-06 06:15:48 train.py: 79] Epoch 11, iter 400/6416, lr 0.010000, loss 1.416619
+INFO 2020-12-06 06:22:08 train.py: 79] Epoch 11, iter 600/6416, lr 0.010000, loss 1.442494
+INFO 2020-12-06 06:28:28 train.py: 79] Epoch 11, iter 800/6416, lr 0.010000, loss 1.437556
+INFO 2020-12-06 06:34:47 train.py: 79] Epoch 11, iter 1000/6416, lr 0.010000, loss 1.435519
+INFO 2020-12-06 06:41:05 train.py: 79] Epoch 11, iter 1200/6416, lr 0.010000, loss 1.446615
+INFO 2020-12-06 06:47:24 train.py: 79] Epoch 11, iter 1400/6416, lr 0.010000, loss 1.445880
+INFO 2020-12-06 06:53:43 train.py: 79] Epoch 11, iter 1600/6416, lr 0.010000, loss 1.451455
+INFO 2020-12-06 07:00:02 train.py: 79] Epoch 11, iter 1800/6416, lr 0.010000, loss 1.436330
+INFO 2020-12-06 07:06:21 train.py: 79] Epoch 11, iter 2000/6416, lr 0.010000, loss 1.446569
+INFO 2020-12-06 07:12:40 train.py: 79] Epoch 11, iter 2200/6416, lr 0.010000, loss 1.442806
+INFO 2020-12-06 07:18:58 train.py: 79] Epoch 11, iter 2400/6416, lr 0.010000, loss 1.447099
+INFO 2020-12-06 07:25:17 train.py: 79] Epoch 11, iter 2600/6416, lr 0.010000, loss 1.443769
+INFO 2020-12-06 07:31:37 train.py: 79] Epoch 11, iter 2800/6416, lr 0.010000, loss 1.451199
+INFO 2020-12-06 07:37:54 train.py: 92] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-12-06 07:37:56 train.py: 79] Epoch 11, iter 3000/6416, lr 0.010000, loss 1.446435
+INFO 2020-12-06 07:44:16 train.py: 79] Epoch 11, iter 3200/6416, lr 0.010000, loss 1.449132
+INFO 2020-12-06 07:50:35 train.py: 79] Epoch 11, iter 3400/6416, lr 0.010000, loss 1.439864
+INFO 2020-12-06 07:56:55 train.py: 79] Epoch 11, iter 3600/6416, lr 0.010000, loss 1.441093
+INFO 2020-12-06 08:03:14 train.py: 79] Epoch 11, iter 3800/6416, lr 0.010000, loss 1.436291
+INFO 2020-12-06 08:09:34 train.py: 79] Epoch 11, iter 4000/6416, lr 0.010000, loss 1.437486
+INFO 2020-12-06 08:15:54 train.py: 79] Epoch 11, iter 4200/6416, lr 0.010000, loss 1.441966
+INFO 2020-12-06 08:22:14 train.py: 79] Epoch 11, iter 4400/6416, lr 0.010000, loss 1.416469
+INFO 2020-12-06 08:28:35 train.py: 79] Epoch 11, iter 4600/6416, lr 0.010000, loss 1.443927
+INFO 2020-12-06 08:34:54 train.py: 79] Epoch 11, iter 4800/6416, lr 0.010000, loss 1.432850
+INFO 2020-12-06 08:41:14 train.py: 79] Epoch 11, iter 5000/6416, lr 0.010000, loss 1.436703
+INFO 2020-12-06 08:47:34 train.py: 79] Epoch 11, iter 5200/6416, lr 0.010000, loss 1.430472
+INFO 2020-12-06 08:53:53 train.py: 79] Epoch 11, iter 5400/6416, lr 0.010000, loss 1.421570
+INFO 2020-12-06 09:00:13 train.py: 79] Epoch 11, iter 5600/6416, lr 0.010000, loss 1.439003
+INFO 2020-12-06 09:06:33 train.py: 79] Epoch 11, iter 5800/6416, lr 0.010000, loss 1.431989
+INFO 2020-12-06 09:12:52 train.py: 92] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-12-06 09:12:54 train.py: 79] Epoch 11, iter 6000/6416, lr 0.010000, loss 1.420511
+INFO 2020-12-06 09:19:14 train.py: 79] Epoch 11, iter 6200/6416, lr 0.010000, loss 1.423586
+INFO 2020-12-06 09:25:34 train.py: 79] Epoch 11, iter 6400/6416, lr 0.010000, loss 1.429725
+INFO 2020-12-06 09:26:02 train.py: 97] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-12-06 09:26:05 train.py: 79] Epoch 12, iter 0/6416, lr 0.010000, loss 1.396104
+INFO 2020-12-06 09:32:26 train.py: 79] Epoch 12, iter 200/6416, lr 0.010000, loss 1.188492
+INFO 2020-12-06 09:38:45 train.py: 79] Epoch 12, iter 400/6416, lr 0.010000, loss 1.187144
+INFO 2020-12-06 09:45:06 train.py: 79] Epoch 12, iter 600/6416, lr 0.010000, loss 1.176856
+INFO 2020-12-06 09:51:24 train.py: 79] Epoch 12, iter 800/6416, lr 0.010000, loss 1.190689
+INFO 2020-12-06 09:57:44 train.py: 79] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.209835
+INFO 2020-12-06 10:04:03 train.py: 79] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.210686
+INFO 2020-12-06 10:10:23 train.py: 79] Epoch 12, iter 1400/6416, lr 0.010000, loss 1.194799
+INFO 2020-12-06 10:16:43 train.py: 79] Epoch 12, iter 1600/6416, lr 0.010000, loss 1.217039
+INFO 2020-12-06 10:23:03 train.py: 79] Epoch 12, iter 1800/6416, lr 0.010000, loss 1.216754
+INFO 2020-12-06 10:29:22 train.py: 79] Epoch 12, iter 2000/6416, lr 0.010000, loss 1.201908
+INFO 2020-12-06 10:35:42 train.py: 79] Epoch 12, iter 2200/6416, lr 0.010000, loss 1.226034
+INFO 2020-12-06 10:42:00 train.py: 79] Epoch 12, iter 2400/6416, lr 0.010000, loss 1.220114
+INFO 2020-12-06 10:48:20 train.py: 79] Epoch 12, iter 2600/6416, lr 0.010000, loss 1.219304
+INFO 2020-12-06 10:54:39 train.py: 79] Epoch 12, iter 2800/6416, lr 0.010000, loss 1.239668
+INFO 2020-12-06 11:00:58 train.py: 92] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-12-06 11:00:59 train.py: 79] Epoch 12, iter 3000/6416, lr 0.010000, loss 1.222346
+INFO 2020-12-06 11:07:19 train.py: 79] Epoch 12, iter 3200/6416, lr 0.010000, loss 1.237528
+INFO 2020-12-06 11:13:38 train.py: 79] Epoch 12, iter 3400/6416, lr 0.010000, loss 1.243870
+INFO 2020-12-06 11:19:58 train.py: 79] Epoch 12, iter 3600/6416, lr 0.010000, loss 1.262808
+INFO 2020-12-06 11:26:18 train.py: 79] Epoch 12, iter 3800/6416, lr 0.010000, loss 1.249614
+INFO 2020-12-06 11:32:40 train.py: 79] Epoch 12, iter 4000/6416, lr 0.010000, loss 1.252200
+INFO 2020-12-06 11:38:59 train.py: 79] Epoch 12, iter 4200/6416, lr 0.010000, loss 1.247521
+INFO 2020-12-06 11:45:20 train.py: 79] Epoch 12, iter 4400/6416, lr 0.010000, loss 1.262831
+INFO 2020-12-06 11:51:40 train.py: 79] Epoch 12, iter 4600/6416, lr 0.010000, loss 1.266734
+INFO 2020-12-06 11:58:01 train.py: 79] Epoch 12, iter 4800/6416, lr 0.010000, loss 1.255959
+INFO 2020-12-06 12:04:22 train.py: 79] Epoch 12, iter 5000/6416, lr 0.010000, loss 1.275324
+INFO 2020-12-06 12:10:43 train.py: 79] Epoch 12, iter 5200/6416, lr 0.010000, loss 1.275397
+INFO 2020-12-06 12:17:02 train.py: 79] Epoch 12, iter 5400/6416, lr 0.010000, loss 1.282276
+INFO 2020-12-06 12:23:23 train.py: 79] Epoch 12, iter 5600/6416, lr 0.010000, loss 1.265771
+INFO 2020-12-06 12:29:42 train.py: 79] Epoch 12, iter 5800/6416, lr 0.010000, loss 1.260887
+INFO 2020-12-06 12:36:02 train.py: 92] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-12-06 12:36:04 train.py: 79] Epoch 12, iter 6000/6416, lr 0.010000, loss 1.284875
+INFO 2020-12-06 12:42:25 train.py: 79] Epoch 12, iter 6200/6416, lr 0.010000, loss 1.275401
+INFO 2020-12-06 12:48:46 train.py: 79] Epoch 12, iter 6400/6416, lr 0.010000, loss 1.269966
+INFO 2020-12-06 12:49:14 train.py: 97] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-12-06 12:49:17 train.py: 79] Epoch 13, iter 0/6416, lr 0.001000, loss 1.307253
+INFO 2020-12-06 12:55:37 train.py: 79] Epoch 13, iter 200/6416, lr 0.001000, loss 1.020010
+INFO 2020-12-06 13:01:58 train.py: 79] Epoch 13, iter 400/6416, lr 0.001000, loss 0.976545
+INFO 2020-12-06 13:08:17 train.py: 79] Epoch 13, iter 600/6416, lr 0.001000, loss 0.995507
+INFO 2020-12-06 13:14:36 train.py: 79] Epoch 13, iter 800/6416, lr 0.001000, loss 0.994149
+INFO 2020-12-06 13:20:56 train.py: 79] Epoch 13, iter 1000/6416, lr 0.001000, loss 0.986863
+INFO 2020-12-06 13:27:15 train.py: 79] Epoch 13, iter 1200/6416, lr 0.001000, loss 0.986045
+INFO 2020-12-06 13:33:34 train.py: 79] Epoch 13, iter 1400/6416, lr 0.001000, loss 0.978853
+INFO 2020-12-06 13:39:53 train.py: 79] Epoch 13, iter 1600/6416, lr 0.001000, loss 0.983476
+INFO 2020-12-06 13:46:12 train.py: 79] Epoch 13, iter 1800/6416, lr 0.001000, loss 0.978783
+INFO 2020-12-06 13:52:31 train.py: 79] Epoch 13, iter 2000/6416, lr 0.001000, loss 0.983774
+INFO 2020-12-06 13:58:51 train.py: 79] Epoch 13, iter 2200/6416, lr 0.001000, loss 0.971636
+INFO 2020-12-06 14:05:11 train.py: 79] Epoch 13, iter 2400/6416, lr 0.001000, loss 0.969167
+INFO 2020-12-06 14:11:30 train.py: 79] Epoch 13, iter 2600/6416, lr 0.001000, loss 0.983853
+INFO 2020-12-06 14:17:49 train.py: 79] Epoch 13, iter 2800/6416, lr 0.001000, loss 0.978536
+INFO 2020-12-06 14:24:07 train.py: 92] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-12-06 14:24:09 train.py: 79] Epoch 13, iter 3000/6416, lr 0.001000, loss 0.978534
+INFO 2020-12-06 14:30:29 train.py: 79] Epoch 13, iter 3200/6416, lr 0.001000, loss 0.980241
+INFO 2020-12-06 14:36:48 train.py: 79] Epoch 13, iter 3400/6416, lr 0.001000, loss 0.971680
+INFO 2020-12-06 14:43:08 train.py: 79] Epoch 13, iter 3600/6416, lr 0.001000, loss 0.982299
+INFO 2020-12-06 14:49:27 train.py: 79] Epoch 13, iter 3800/6416, lr 0.001000, loss 0.981664
+INFO 2020-12-06 14:55:47 train.py: 79] Epoch 13, iter 4000/6416, lr 0.001000, loss 0.983848
+INFO 2020-12-06 15:02:08 train.py: 79] Epoch 13, iter 4200/6416, lr 0.001000, loss 0.966763
+INFO 2020-12-06 15:08:28 train.py: 79] Epoch 13, iter 4400/6416, lr 0.001000, loss 0.974734
+INFO 2020-12-06 15:14:48 train.py: 79] Epoch 13, iter 4600/6416, lr 0.001000, loss 0.967398
+INFO 2020-12-06 15:21:08 train.py: 79] Epoch 13, iter 4800/6416, lr 0.001000, loss 0.969599
+INFO 2020-12-06 15:27:28 train.py: 79] Epoch 13, iter 5000/6416, lr 0.001000, loss 0.968946
+INFO 2020-12-06 15:33:48 train.py: 79] Epoch 13, iter 5200/6416, lr 0.001000, loss 0.978118
+INFO 2020-12-06 15:40:08 train.py: 79] Epoch 13, iter 5400/6416, lr 0.001000, loss 0.973513
+INFO 2020-12-06 15:46:28 train.py: 79] Epoch 13, iter 5600/6416, lr 0.001000, loss 0.980251
+INFO 2020-12-06 15:52:48 train.py: 79] Epoch 13, iter 5800/6416, lr 0.001000, loss 0.968728
+INFO 2020-12-06 15:59:07 train.py: 92] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-12-06 15:59:09 train.py: 79] Epoch 13, iter 6000/6416, lr 0.001000, loss 0.972907
+INFO 2020-12-06 16:05:30 train.py: 79] Epoch 13, iter 6200/6416, lr 0.001000, loss 0.978645
+INFO 2020-12-06 16:11:50 train.py: 79] Epoch 13, iter 6400/6416, lr 0.001000, loss 0.984186
+INFO 2020-12-06 16:12:18 train.py: 97] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-12-06 16:12:21 train.py: 79] Epoch 14, iter 0/6416, lr 0.001000, loss 0.965483
+INFO 2020-12-06 16:18:40 train.py: 79] Epoch 14, iter 200/6416, lr 0.001000, loss 0.950053
+INFO 2020-12-06 16:25:01 train.py: 79] Epoch 14, iter 400/6416, lr 0.001000, loss 0.942763
+INFO 2020-12-06 16:31:19 train.py: 79] Epoch 14, iter 600/6416, lr 0.001000, loss 0.949050
+INFO 2020-12-06 16:37:38 train.py: 79] Epoch 14, iter 800/6416, lr 0.001000, loss 0.946663
+INFO 2020-12-06 16:43:57 train.py: 79] Epoch 14, iter 1000/6416, lr 0.001000, loss 0.939141
+INFO 2020-12-06 16:50:17 train.py: 79] Epoch 14, iter 1200/6416, lr 0.001000, loss 0.948596
+INFO 2020-12-06 16:56:36 train.py: 79] Epoch 14, iter 1400/6416, lr 0.001000, loss 0.939278
+INFO 2020-12-06 17:02:55 train.py: 79] Epoch 14, iter 1600/6416, lr 0.001000, loss 0.942789
+INFO 2020-12-06 17:09:13 train.py: 79] Epoch 14, iter 1800/6416, lr 0.001000, loss 0.955463
+INFO 2020-12-06 17:15:34 train.py: 79] Epoch 14, iter 2000/6416, lr 0.001000, loss 0.939293
+INFO 2020-12-06 17:21:53 train.py: 79] Epoch 14, iter 2200/6416, lr 0.001000, loss 0.948789
+INFO 2020-12-06 17:28:14 train.py: 79] Epoch 14, iter 2400/6416, lr 0.001000, loss 0.945787
+INFO 2020-12-06 17:34:33 train.py: 79] Epoch 14, iter 2600/6416, lr 0.001000, loss 0.947393
+INFO 2020-12-06 17:40:53 train.py: 79] Epoch 14, iter 2800/6416, lr 0.001000, loss 0.950612
+INFO 2020-12-06 17:47:12 train.py: 92] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-12-06 17:47:14 train.py: 79] Epoch 14, iter 3000/6416, lr 0.001000, loss 0.942860
+INFO 2020-12-06 17:53:34 train.py: 79] Epoch 14, iter 3200/6416, lr 0.001000, loss 0.943538
+INFO 2020-12-06 17:59:54 train.py: 79] Epoch 14, iter 3400/6416, lr 0.001000, loss 0.952549
+INFO 2020-12-06 18:06:13 train.py: 79] Epoch 14, iter 3600/6416, lr 0.001000, loss 0.943889
+INFO 2020-12-06 18:12:33 train.py: 79] Epoch 14, iter 3800/6416, lr 0.001000, loss 0.944875
+INFO 2020-12-06 18:18:52 train.py: 79] Epoch 14, iter 4000/6416, lr 0.001000, loss 0.942350
+INFO 2020-12-06 18:25:12 train.py: 79] Epoch 14, iter 4200/6416, lr 0.001000, loss 0.951103
+INFO 2020-12-06 18:31:32 train.py: 79] Epoch 14, iter 4400/6416, lr 0.001000, loss 0.954832
+INFO 2020-12-06 18:37:52 train.py: 79] Epoch 14, iter 4600/6416, lr 0.001000, loss 0.955230
+INFO 2020-12-06 18:44:12 train.py: 79] Epoch 14, iter 4800/6416, lr 0.001000, loss 0.954810
+INFO 2020-12-06 18:50:32 train.py: 79] Epoch 14, iter 5000/6416, lr 0.001000, loss 0.948358
+INFO 2020-12-06 18:56:52 train.py: 79] Epoch 14, iter 5200/6416, lr 0.001000, loss 0.941443
+INFO 2020-12-06 19:03:12 train.py: 79] Epoch 14, iter 5400/6416, lr 0.001000, loss 0.944014
+INFO 2020-12-06 19:09:31 train.py: 79] Epoch 14, iter 5600/6416, lr 0.001000, loss 0.962197
+INFO 2020-12-06 19:15:52 train.py: 79] Epoch 14, iter 5800/6416, lr 0.001000, loss 0.951441
+INFO 2020-12-06 19:22:10 train.py: 92] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-12-06 19:22:12 train.py: 79] Epoch 14, iter 6000/6416, lr 0.001000, loss 0.961924
+INFO 2020-12-06 19:28:32 train.py: 79] Epoch 14, iter 6200/6416, lr 0.001000, loss 0.943430
+INFO 2020-12-06 19:34:52 train.py: 79] Epoch 14, iter 6400/6416, lr 0.001000, loss 0.953354
+INFO 2020-12-06 19:35:20 train.py: 97] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-12-06 19:35:23 train.py: 79] Epoch 15, iter 0/6416, lr 0.001000, loss 0.935825
+INFO 2020-12-06 19:41:43 train.py: 79] Epoch 15, iter 200/6416, lr 0.001000, loss 0.909370
+INFO 2020-12-06 19:48:03 train.py: 79] Epoch 15, iter 400/6416, lr 0.001000, loss 0.920209
+INFO 2020-12-06 19:54:23 train.py: 79] Epoch 15, iter 600/6416, lr 0.001000, loss 0.914311
+INFO 2020-12-06 20:00:43 train.py: 79] Epoch 15, iter 800/6416, lr 0.001000, loss 0.926013
+INFO 2020-12-06 20:07:02 train.py: 79] Epoch 15, iter 1000/6416, lr 0.001000, loss 0.916556
+INFO 2020-12-06 20:13:21 train.py: 79] Epoch 15, iter 1200/6416, lr 0.001000, loss 0.919629
+INFO 2020-12-06 20:19:40 train.py: 79] Epoch 15, iter 1400/6416, lr 0.001000, loss 0.919652
+INFO 2020-12-06 20:25:59 train.py: 79] Epoch 15, iter 1600/6416, lr 0.001000, loss 0.924657
+INFO 2020-12-06 20:32:19 train.py: 79] Epoch 15, iter 1800/6416, lr 0.001000, loss 0.933341
+INFO 2020-12-06 20:38:39 train.py: 79] Epoch 15, iter 2000/6416, lr 0.001000, loss 0.921353
+INFO 2020-12-06 20:44:59 train.py: 79] Epoch 15, iter 2200/6416, lr 0.001000, loss 0.922088
+INFO 2020-12-06 20:51:18 train.py: 79] Epoch 15, iter 2400/6416, lr 0.001000, loss 0.931816
+INFO 2020-12-06 20:57:38 train.py: 79] Epoch 15, iter 2600/6416, lr 0.001000, loss 0.934314
+INFO 2020-12-06 21:03:57 train.py: 79] Epoch 15, iter 2800/6416, lr 0.001000, loss 0.924028
+INFO 2020-12-06 21:10:16 train.py: 92] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-12-06 21:10:18 train.py: 79] Epoch 15, iter 3000/6416, lr 0.001000, loss 0.917808
+INFO 2020-12-06 21:16:37 train.py: 79] Epoch 15, iter 3200/6416, lr 0.001000, loss 0.930806
+INFO 2020-12-06 21:22:57 train.py: 79] Epoch 15, iter 3400/6416, lr 0.001000, loss 0.938937
+INFO 2020-12-06 21:29:16 train.py: 79] Epoch 15, iter 3600/6416, lr 0.001000, loss 0.932077
+INFO 2020-12-06 21:35:36 train.py: 79] Epoch 15, iter 3800/6416, lr 0.001000, loss 0.930898
+INFO 2020-12-06 21:41:57 train.py: 79] Epoch 15, iter 4000/6416, lr 0.001000, loss 0.922503
+INFO 2020-12-06 21:48:17 train.py: 79] Epoch 15, iter 4200/6416, lr 0.001000, loss 0.922279
+INFO 2020-12-06 21:54:36 train.py: 79] Epoch 15, iter 4400/6416, lr 0.001000, loss 0.931669
+INFO 2020-12-06 22:00:56 train.py: 79] Epoch 15, iter 4600/6416, lr 0.001000, loss 0.934641
+INFO 2020-12-06 22:07:15 train.py: 79] Epoch 15, iter 4800/6416, lr 0.001000, loss 0.933393
+INFO 2020-12-06 22:13:36 train.py: 79] Epoch 15, iter 5000/6416, lr 0.001000, loss 0.926171
+INFO 2020-12-06 22:19:55 train.py: 79] Epoch 15, iter 5200/6416, lr 0.001000, loss 0.932479
+INFO 2020-12-06 22:26:17 train.py: 79] Epoch 15, iter 5400/6416, lr 0.001000, loss 0.934811
+INFO 2020-12-06 22:32:36 train.py: 79] Epoch 15, iter 5600/6416, lr 0.001000, loss 0.928534
+INFO 2020-12-06 22:38:57 train.py: 79] Epoch 15, iter 5800/6416, lr 0.001000, loss 0.941098
+INFO 2020-12-06 22:45:17 train.py: 92] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-12-06 22:45:19 train.py: 79] Epoch 15, iter 6000/6416, lr 0.001000, loss 0.929192
+INFO 2020-12-06 22:51:39 train.py: 79] Epoch 15, iter 6200/6416, lr 0.001000, loss 0.933308
+INFO 2020-12-06 22:58:00 train.py: 79] Epoch 15, iter 6400/6416, lr 0.001000, loss 0.937846
+INFO 2020-12-06 22:58:29 train.py: 97] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-12-06 22:58:32 train.py: 79] Epoch 16, iter 0/6416, lr 0.000100, loss 0.971559
+INFO 2020-12-06 23:04:51 train.py: 79] Epoch 16, iter 200/6416, lr 0.000100, loss 0.899312
+INFO 2020-12-06 23:11:10 train.py: 79] Epoch 16, iter 400/6416, lr 0.000100, loss 0.899546
+INFO 2020-12-06 23:17:30 train.py: 79] Epoch 16, iter 600/6416, lr 0.000100, loss 0.900199
+INFO 2020-12-06 23:23:49 train.py: 79] Epoch 16, iter 800/6416, lr 0.000100, loss 0.891348
+INFO 2020-12-06 23:30:08 train.py: 79] Epoch 16, iter 1000/6416, lr 0.000100, loss 0.893778
+INFO 2020-12-06 23:36:27 train.py: 79] Epoch 16, iter 1200/6416, lr 0.000100, loss 0.890369
+INFO 2020-12-06 23:42:47 train.py: 79] Epoch 16, iter 1400/6416, lr 0.000100, loss 0.895213
+INFO 2020-12-06 23:49:06 train.py: 79] Epoch 16, iter 1600/6416, lr 0.000100, loss 0.905256
+INFO 2020-12-06 23:55:25 train.py: 79] Epoch 16, iter 1800/6416, lr 0.000100, loss 0.898873
+INFO 2020-12-07 00:01:45 train.py: 79] Epoch 16, iter 2000/6416, lr 0.000100, loss 0.893949
+INFO 2020-12-07 00:08:05 train.py: 79] Epoch 16, iter 2200/6416, lr 0.000100, loss 0.903722
+INFO 2020-12-07 00:14:24 train.py: 79] Epoch 16, iter 2400/6416, lr 0.000100, loss 0.897051
+INFO 2020-12-07 00:20:44 train.py: 79] Epoch 16, iter 2600/6416, lr 0.000100, loss 0.898794
+INFO 2020-12-07 00:27:03 train.py: 79] Epoch 16, iter 2800/6416, lr 0.000100, loss 0.903484
+INFO 2020-12-07 00:33:23 train.py: 92] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-12-07 00:33:25 train.py: 79] Epoch 16, iter 3000/6416, lr 0.000100, loss 0.902564
+INFO 2020-12-07 00:39:45 train.py: 79] Epoch 16, iter 3200/6416, lr 0.000100, loss 0.900985
+INFO 2020-12-07 00:46:06 train.py: 79] Epoch 16, iter 3400/6416, lr 0.000100, loss 0.914455
+INFO 2020-12-07 00:52:26 train.py: 79] Epoch 16, iter 3600/6416, lr 0.000100, loss 0.901060
+INFO 2020-12-07 00:58:46 train.py: 79] Epoch 16, iter 3800/6416, lr 0.000100, loss 0.910659
+INFO 2020-12-07 01:05:06 train.py: 79] Epoch 16, iter 4000/6416, lr 0.000100, loss 0.890063
+INFO 2020-12-07 01:11:27 train.py: 79] Epoch 16, iter 4200/6416, lr 0.000100, loss 0.892856
+INFO 2020-12-07 01:17:48 train.py: 79] Epoch 16, iter 4400/6416, lr 0.000100, loss 0.899108
+INFO 2020-12-07 01:24:08 train.py: 79] Epoch 16, iter 4600/6416, lr 0.000100, loss 0.891107
+INFO 2020-12-07 01:30:29 train.py: 79] Epoch 16, iter 4800/6416, lr 0.000100, loss 0.902412
+INFO 2020-12-07 01:36:49 train.py: 79] Epoch 16, iter 5000/6416, lr 0.000100, loss 0.893027
+INFO 2020-12-07 01:43:09 train.py: 79] Epoch 16, iter 5200/6416, lr 0.000100, loss 0.895569
+INFO 2020-12-07 01:49:30 train.py: 79] Epoch 16, iter 5400/6416, lr 0.000100, loss 0.887312
+INFO 2020-12-07 01:55:51 train.py: 79] Epoch 16, iter 5600/6416, lr 0.000100, loss 0.899679
+INFO 2020-12-07 02:02:11 train.py: 79] Epoch 16, iter 5800/6416, lr 0.000100, loss 0.894838
+INFO 2020-12-07 02:08:30 train.py: 92] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-12-07 02:08:32 train.py: 79] Epoch 16, iter 6000/6416, lr 0.000100, loss 0.891069
+INFO 2020-12-07 02:14:52 train.py: 79] Epoch 16, iter 6200/6416, lr 0.000100, loss 0.904488
+INFO 2020-12-07 02:21:14 train.py: 79] Epoch 16, iter 6400/6416, lr 0.000100, loss 0.885888
+INFO 2020-12-07 02:21:41 train.py: 97] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-12-07 02:21:45 train.py: 79] Epoch 17, iter 0/6416, lr 0.000100, loss 0.934706
+INFO 2020-12-07 02:28:05 train.py: 79] Epoch 17, iter 200/6416, lr 0.000100, loss 0.893623
+INFO 2020-12-07 02:34:24 train.py: 79] Epoch 17, iter 400/6416, lr 0.000100, loss 0.901666
+INFO 2020-12-07 02:40:45 train.py: 79] Epoch 17, iter 600/6416, lr 0.000100, loss 0.892125
+INFO 2020-12-07 02:47:04 train.py: 79] Epoch 17, iter 800/6416, lr 0.000100, loss 0.898280
+INFO 2020-12-07 02:53:24 train.py: 79] Epoch 17, iter 1000/6416, lr 0.000100, loss 0.891789
+INFO 2020-12-07 02:59:43 train.py: 79] Epoch 17, iter 1200/6416, lr 0.000100, loss 0.888356
+INFO 2020-12-07 03:06:03 train.py: 79] Epoch 17, iter 1400/6416, lr 0.000100, loss 0.897336
+INFO 2020-12-07 03:12:22 train.py: 79] Epoch 17, iter 1600/6416, lr 0.000100, loss 0.892398
+INFO 2020-12-07 03:18:43 train.py: 79] Epoch 17, iter 1800/6416, lr 0.000100, loss 0.905767
+INFO 2020-12-07 03:25:02 train.py: 79] Epoch 17, iter 2000/6416, lr 0.000100, loss 0.885561
+INFO 2020-12-07 03:31:23 train.py: 79] Epoch 17, iter 2200/6416, lr 0.000100, loss 0.887983
+INFO 2020-12-07 03:37:42 train.py: 79] Epoch 17, iter 2400/6416, lr 0.000100, loss 0.899450
+INFO 2020-12-07 03:44:01 train.py: 79] Epoch 17, iter 2600/6416, lr 0.000100, loss 0.892069
+INFO 2020-12-07 03:50:21 train.py: 79] Epoch 17, iter 2800/6416, lr 0.000100, loss 0.894267
+INFO 2020-12-07 03:56:39 train.py: 92] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-12-07 03:56:41 train.py: 79] Epoch 17, iter 3000/6416, lr 0.000100, loss 0.890403
+INFO 2020-12-07 04:03:01 train.py: 79] Epoch 17, iter 3200/6416, lr 0.000100, loss 0.879247
+INFO 2020-12-07 04:09:22 train.py: 79] Epoch 17, iter 3400/6416, lr 0.000100, loss 0.901325
+INFO 2020-12-07 04:15:43 train.py: 79] Epoch 17, iter 3600/6416, lr 0.000100, loss 0.903118
+INFO 2020-12-07 04:22:03 train.py: 79] Epoch 17, iter 3800/6416, lr 0.000100, loss 0.894275
+INFO 2020-12-07 04:28:24 train.py: 79] Epoch 17, iter 4000/6416, lr 0.000100, loss 0.897160
+INFO 2020-12-07 04:34:43 train.py: 79] Epoch 17, iter 4200/6416, lr 0.000100, loss 0.889665
+INFO 2020-12-07 04:41:04 train.py: 79] Epoch 17, iter 4400/6416, lr 0.000100, loss 0.900363
+INFO 2020-12-07 04:47:25 train.py: 79] Epoch 17, iter 4600/6416, lr 0.000100, loss 0.898280
+INFO 2020-12-07 04:53:46 train.py: 79] Epoch 17, iter 4800/6416, lr 0.000100, loss 0.890538
+INFO 2020-12-07 05:00:06 train.py: 79] Epoch 17, iter 5000/6416, lr 0.000100, loss 0.899692
+INFO 2020-12-07 05:06:27 train.py: 79] Epoch 17, iter 5200/6416, lr 0.000100, loss 0.900445
+INFO 2020-12-07 05:12:46 train.py: 79] Epoch 17, iter 5400/6416, lr 0.000100, loss 0.891245
+INFO 2020-12-07 05:19:07 train.py: 79] Epoch 17, iter 5600/6416, lr 0.000100, loss 0.901533
+INFO 2020-12-07 05:25:27 train.py: 79] Epoch 17, iter 5800/6416, lr 0.000100, loss 0.886352
+INFO 2020-12-07 05:31:46 train.py: 92] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-12-07 05:31:48 train.py: 79] Epoch 17, iter 6000/6416, lr 0.000100, loss 0.902356
+INFO 2020-12-07 05:38:10 train.py: 79] Epoch 17, iter 6200/6416, lr 0.000100, loss 0.891394
+INFO 2020-12-07 05:44:30 train.py: 79] Epoch 17, iter 6400/6416, lr 0.000100, loss 0.888566
+INFO 2020-12-07 05:44:58 train.py: 97] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-12-07 05:44:59 train.py: 180] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/.gitkeep b/bob/bio/facexzoo/models/backbones/LightCNN29/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_African.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6050b372952bdab6b7362e0ecbdc89882fbdea10
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.8831666666666667 | 0.004335113594422089 |
+| Epoch_17.pt | 0.8825 | 0.0037039351779518488 |
+| Epoch_16_batch_5999.pt | 0.882 | 0.003973678831610594 |
+| Epoch_13_batch_5999.pt | 0.8816666666666666 | 0.0037101795237919973 |
+| Epoch_14_batch_5999.pt | 0.8815000000000002 | 0.004017323597731317 |
+| Epoch_14_batch_2999.pt | 0.8815 | 0.003986474044646715 |
+| Epoch_16.pt | 0.8813333333333334 | 0.0037989602216175035 |
+| Epoch_17_batch_2999.pt | 0.8811666666666668 | 0.0036603480911175826 |
+| Epoch_16_batch_2999.pt | 0.8811666666666668 | 0.0039287638240527 |
+| Epoch_15.pt | 0.881 | 0.004354648431614538 |
+| Epoch_14.pt | 0.8806666666666668 | 0.0036362373715452404 |
+| Epoch_17_batch_5999.pt | 0.8806666666666667 | 0.0038825344910346986 |
+| Epoch_13_batch_2999.pt | 0.8798333333333332 | 0.0038204291659898176 |
+| Epoch_13.pt | 0.8793333333333333 | 0.0038425814368423503 |
+| Epoch_15_batch_2999.pt | 0.8785000000000001 | 0.003713921091857393 |
+| Epoch_12_batch_2999.pt | 0.8763333333333334 | 0.004429140317332167 |
+| Epoch_11_batch_2999.pt | 0.8743333333333332 | 0.004734950167258339 |
+| Epoch_10_batch_5999.pt | 0.8738333333333334 | 0.004052510272185977 |
+| Epoch_11_batch_5999.pt | 0.8738333333333334 | 0.004367740891333548 |
+| Epoch_12_batch_5999.pt | 0.8736666666666666 | 0.004559618784151823 |
+| Epoch_12.pt | 0.8711666666666668 | 0.004052510272185977 |
+| Epoch_11.pt | 0.8706666666666667 | 0.005013561854523771 |
+| Epoch_10.pt | 0.8683333333333334 | 0.004694362260950576 |
+| Epoch_10_batch_2999.pt | 0.8636666666666667 | 0.004725815626252604 |
+| Epoch_9_batch_2999.pt | 0.8348333333333333 | 0.0029128281619818486 |
+| Epoch_9_batch_5999.pt | 0.8343333333333334 | 0.003728435941236115 |
+| Epoch_8_batch_2999.pt | 0.834 | 0.00438290731629245 |
+| Epoch_9.pt | 0.8333333333333334 | 0.0032584173996922723 |
+| Epoch_8_batch_5999.pt | 0.8323333333333334 | 0.0044527699798919745 |
+| Epoch_7_batch_2999.pt | 0.8320000000000001 | 0.004126098806139364 |
+| Epoch_7_batch_5999.pt | 0.8311666666666666 | 0.0036519063176766236 |
+| Epoch_6_batch_2999.pt | 0.8281666666666666 | 0.004781974900667853 |
+| Epoch_8.pt | 0.8268333333333334 | 0.0046643512774560865 |
+| Epoch_6_batch_5999.pt | 0.826 | 0.004275973645531971 |
+| Epoch_7.pt | 0.8258333333333333 | 0.003145151078826256 |
+| Epoch_5_batch_5999.pt | 0.8234999999999999 | 0.0031957726707265836 |
+| Epoch_6.pt | 0.8231666666666666 | 0.00403265987643544 |
+| Epoch_5_batch_2999.pt | 0.8183333333333334 | 0.004906533814626583 |
+| Epoch_5.pt | 0.8108333333333334 | 0.0037944891814509288 |
+| Epoch_4_batch_2999.pt | 0.808 | 0.00353815185068681 |
+| Epoch_4.pt | 0.8056666666666666 | 0.0028781852993308003 |
+| Epoch_4_batch_5999.pt | 0.8028333333333334 | 0.005988930117359638 |
+| Epoch_3_batch_5999.pt | 0.7991666666666666 | 0.005796177965896643 |
+| Epoch_3_batch_2999.pt | 0.7853333333333333 | 0.005004935835357872 |
+| Epoch_3.pt | 0.7761666666666668 | 0.0036771734985273107 |
+| Epoch_2_batch_5999.pt | 0.7735000000000001 | 0.005051891223915218 |
+| Epoch_2_batch_2999.pt | 0.7615 | 0.0048460878965650815 |
+| Epoch_2.pt | 0.7581666666666667 | 0.006238322424070975 |
+| Epoch_1_batch_5999.pt | 0.7323333333333333 | 0.004038395965641385 |
+| Epoch_1.pt | 0.7021666666666666 | 0.004981756842176052 |
+| Epoch_1_batch_2999.pt | 0.6933333333333334 | 0.0037018513886572595 |
+| Epoch_0.pt | 0.6323333333333333 | 0.004548775442324159 |
+| Epoch_0_batch_5999.pt | 0.6188333333333332 | 0.0038252733300226695 |
+| Epoch_0_batch_2999.pt | 0.5650000000000001 | 0.006039990188856396 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Asian.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d8e3ff24063ab6fc1c6cd762cd5d7d27118c2bda
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.8783333333333335 | 0.005055250296034368 |
+| Epoch_13_batch_2999.pt | 0.8773333333333333 | 0.0046956770242478725 |
+| Epoch_16.pt | 0.8771666666666667 | 0.004727448088778852 |
+| Epoch_13_batch_5999.pt | 0.8755000000000001 | 0.005466226269523583 |
+| Epoch_16_batch_5999.pt | 0.8755000000000001 | 0.005104161942552606 |
+| Epoch_13.pt | 0.8751666666666669 | 0.005850240049944562 |
+| Epoch_17_batch_2999.pt | 0.8751666666666666 | 0.005273097339852125 |
+| Epoch_17.pt | 0.8751666666666666 | 0.005220153254455276 |
+| Epoch_14_batch_5999.pt | 0.8751666666666666 | 0.005273097339852127 |
+| Epoch_17_batch_5999.pt | 0.875 | 0.004887626099538391 |
+| Epoch_14_batch_2999.pt | 0.875 | 0.005351819812796579 |
+| Epoch_11_batch_5999.pt | 0.8745 | 0.005352108157389482 |
+| Epoch_15.pt | 0.8739999999999999 | 0.004901498888913948 |
+| Epoch_11.pt | 0.8736666666666668 | 0.005537749241945386 |
+| Epoch_15_batch_5999.pt | 0.8735000000000002 | 0.0057437566749245885 |
+| Epoch_16_batch_2999.pt | 0.8731666666666668 | 0.004833333333333336 |
+| Epoch_12.pt | 0.8726666666666667 | 0.005763872155263528 |
+| Epoch_12_batch_5999.pt | 0.8723333333333333 | 0.005444444444444448 |
+| Epoch_12_batch_2999.pt | 0.8708333333333333 | 0.005623199300253397 |
+| Epoch_11_batch_2999.pt | 0.8705 | 0.006339422024517554 |
+| Epoch_15_batch_2999.pt | 0.8703333333333333 | 0.005344894875913647 |
+| Epoch_10.pt | 0.8683333333333334 | 0.004987639041658544 |
+| Epoch_10_batch_2999.pt | 0.868 | 0.00532754315321638 |
+| Epoch_10_batch_5999.pt | 0.8664999999999999 | 0.005474125444554232 |
+| Epoch_7_batch_5999.pt | 0.8493333333333333 | 0.004695677024247878 |
+| Epoch_9_batch_2999.pt | 0.8461666666666666 | 0.005746979883188888 |
+| Epoch_8_batch_5999.pt | 0.8458333333333332 | 0.005260204662027873 |
+| Epoch_9_batch_5999.pt | 0.8455 | 0.004875296762097517 |
+| Epoch_7_batch_2999.pt | 0.841 | 0.005890984951916249 |
+| Epoch_8_batch_2999.pt | 0.8401666666666667 | 0.00572222222222222 |
+| Epoch_6_batch_2999.pt | 0.8378333333333334 | 0.007009913614937693 |
+| Epoch_6_batch_5999.pt | 0.8353333333333334 | 0.004387130449422137 |
+| Epoch_9.pt | 0.8346666666666668 | 0.005017254179944733 |
+| Epoch_7.pt | 0.8343333333333334 | 0.005450110430413657 |
+| Epoch_5_batch_5999.pt | 0.833 | 0.006155395104206466 |
+| Epoch_6.pt | 0.8310000000000001 | 0.0037449554547450505 |
+| Epoch_8.pt | 0.8293333333333333 | 0.005963883066808729 |
+| Epoch_5_batch_2999.pt | 0.829 | 0.005068664323536467 |
+| Epoch_4_batch_5999.pt | 0.8248333333333333 | 0.006218500871304899 |
+| Epoch_4_batch_2999.pt | 0.8173333333333334 | 0.0040990814990437136 |
+| Epoch_3_batch_5999.pt | 0.8163333333333334 | 0.005836242660741735 |
+| Epoch_5.pt | 0.8133333333333332 | 0.005499719409228701 |
+| Epoch_4.pt | 0.813 | 0.004880042501332823 |
+| Epoch_3_batch_2999.pt | 0.8006666666666666 | 0.005578840093614753 |
+| Epoch_3.pt | 0.7973333333333332 | 0.004275973645531967 |
+| Epoch_2_batch_5999.pt | 0.7956666666666667 | 0.0036784323016104143 |
+| Epoch_2_batch_2999.pt | 0.7798333333333333 | 0.0034377174454233357 |
+| Epoch_2.pt | 0.7770000000000001 | 0.004841946348777979 |
+| Epoch_1_batch_5999.pt | 0.7476666666666667 | 0.004225145187073671 |
+| Epoch_1.pt | 0.7288333333333333 | 0.003936611941790417 |
+| Epoch_1_batch_2999.pt | 0.7208333333333332 | 0.006142595663193328 |
+| Epoch_0_batch_5999.pt | 0.6743333333333335 | 0.005001234415523062 |
+| Epoch_0.pt | 0.6696666666666666 | 0.00611515018036312 |
+| Epoch_0_batch_2999.pt | 0.6291666666666667 | 0.00735035902373331 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Caucasian.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..84a1efb43636a5902d0f62c02579722d283302c6
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_2999.pt | 0.9533333333333335 | 0.0025700074458267466 |
+| Epoch_17_batch_5999.pt | 0.9528333333333332 | 0.002434956333423463 |
+| Epoch_15.pt | 0.9521666666666666 | 0.0020645449084513356 |
+| Epoch_16_batch_2999.pt | 0.9521666666666666 | 0.0022089883724523353 |
+| Epoch_17.pt | 0.9518333333333333 | 0.0021865187393504924 |
+| Epoch_14_batch_2999.pt | 0.9518333333333333 | 0.0027827732858596048 |
+| Epoch_16_batch_5999.pt | 0.9515 | 0.002388888888888892 |
+| Epoch_15_batch_2999.pt | 0.9513333333333334 | 0.002519063121945468 |
+| Epoch_13_batch_2999.pt | 0.9508333333333333 | 0.002655648533937139 |
+| Epoch_14.pt | 0.9508333333333333 | 0.002573008039313208 |
+| Epoch_13_batch_5999.pt | 0.9506666666666665 | 0.0023596400646216935 |
+| Epoch_16.pt | 0.9503333333333334 | 0.0025795970591653114 |
+| Epoch_13.pt | 0.9498333333333333 | 0.0024400212506664183 |
+| Epoch_14_batch_5999.pt | 0.9498333333333331 | 0.0025391988626725414 |
+| Epoch_15_batch_5999.pt | 0.9490000000000001 | 0.0025724082006200496 |
+| Epoch_11.pt | 0.9486666666666668 | 0.0027755546659548463 |
+| Epoch_12_batch_5999.pt | 0.9475 | 0.0035939764421413062 |
+| Epoch_12_batch_2999.pt | 0.9469999999999998 | 0.0025795970591653136 |
+| Epoch_10.pt | 0.9466666666666667 | 0.0021081851067789197 |
+| Epoch_11_batch_5999.pt | 0.9466666666666667 | 0.0034960294939005055 |
+| Epoch_12.pt | 0.9450000000000001 | 0.003360996324945372 |
+| Epoch_11_batch_2999.pt | 0.9450000000000001 | 0.0030020569080236145 |
+| Epoch_10_batch_5999.pt | 0.9446666666666668 | 0.0027193862778934247 |
+| Epoch_10_batch_2999.pt | 0.9433333333333334 | 0.002151657414559672 |
+| Epoch_9_batch_2999.pt | 0.923 | 0.0028306087117460016 |
+| Epoch_9_batch_5999.pt | 0.9206666666666667 | 0.0036021255727660796 |
+| Epoch_9.pt | 0.9198333333333334 | 0.0034377174454233318 |
+| Epoch_8_batch_2999.pt | 0.9183333333333333 | 0.003496029493900503 |
+| Epoch_8_batch_5999.pt | 0.9175000000000001 | 0.002253255532297023 |
+| Epoch_7_batch_5999.pt | 0.9163333333333334 | 0.0029270977494043325 |
+| Epoch_7_batch_2999.pt | 0.9151666666666666 | 0.003057575090181562 |
+| Epoch_6_batch_2999.pt | 0.9136666666666666 | 0.0026851213274654653 |
+| Epoch_8.pt | 0.9110000000000001 | 0.002572408200620055 |
+| Epoch_6_batch_5999.pt | 0.9084999999999999 | 0.002388888888888887 |
+| Epoch_6.pt | 0.906 | 0.004452769979891972 |
+| Epoch_5_batch_5999.pt | 0.9059999999999999 | 0.003802208584948599 |
+| Epoch_5.pt | 0.9036666666666665 | 0.0048291808471334925 |
+| Epoch_5_batch_2999.pt | 0.9028333333333333 | 0.004374801582802231 |
+| Epoch_7.pt | 0.9026666666666667 | 0.00404603143467403 |
+| Epoch_4_batch_2999.pt | 0.8981666666666666 | 0.004226971013694946 |
+| Epoch_4_batch_5999.pt | 0.8973333333333333 | 0.0025843785221362214 |
+| Epoch_3_batch_5999.pt | 0.8965000000000002 | 0.003939746811442531 |
+| Epoch_4.pt | 0.889 | 0.003703518513888653 |
+| Epoch_3.pt | 0.8875 | 0.003055555555555558 |
+| Epoch_3_batch_2999.pt | 0.8861666666666667 | 0.0020191979828206915 |
+| Epoch_2_batch_5999.pt | 0.8719999999999999 | 0.0025676044462869677 |
+| Epoch_2.pt | 0.8666666666666668 | 0.0038167920082928207 |
+| Epoch_2_batch_2999.pt | 0.8643333333333334 | 0.004151453709393199 |
+| Epoch_1_batch_5999.pt | 0.8426666666666668 | 0.003969015799887057 |
+| Epoch_1.pt | 0.8198333333333332 | 0.00472352923088888 |
+| Epoch_1_batch_2999.pt | 0.8191666666666666 | 0.0050567764093660435 |
+| Epoch_0_batch_5999.pt | 0.7596666666666667 | 0.005598721194375167 |
+| Epoch_0.pt | 0.7383333333333334 | 0.005499719409228705 |
+| Epoch_0_batch_2999.pt | 0.6643333333333333 | 0.005584369721390692 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Indian.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d82c0539a420294164515bebfe605cf6eab4d7c4
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_RFW_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.9088333333333333 | 0.00327777777777777 |
+| Epoch_16_batch_5999.pt | 0.9073333333333332 | 0.0037035185138886567 |
+| Epoch_13_batch_5999.pt | 0.9071666666666667 | 0.0033614554460901854 |
+| Epoch_11_batch_5999.pt | 0.9068333333333334 | 0.0034377174454233296 |
+| Epoch_16_batch_2999.pt | 0.9068333333333334 | 0.003697262918216627 |
+| Epoch_12_batch_5999.pt | 0.9066666666666666 | 0.0033609963249453673 |
+| Epoch_15.pt | 0.9063333333333334 | 0.003520662115056634 |
+| Epoch_15_batch_2999.pt | 0.9061666666666666 | 0.0034251250315107413 |
+| Epoch_17_batch_5999.pt | 0.9058333333333334 | 0.0033998002846209437 |
+| Epoch_16.pt | 0.9056666666666668 | 0.003670032125536387 |
+| Epoch_17.pt | 0.9056666666666666 | 0.0031836775070876464 |
+| Epoch_10.pt | 0.9053333333333334 | 0.004178132372658483 |
+| Epoch_14_batch_2999.pt | 0.9053333333333334 | 0.0037085153929508037 |
+| Epoch_15_batch_5999.pt | 0.9051666666666666 | 0.003224615969095884 |
+| Epoch_11.pt | 0.9051666666666666 | 0.0033282368446111424 |
+| Epoch_17_batch_2999.pt | 0.9048333333333332 | 0.0033467323292953374 |
+| Epoch_13_batch_2999.pt | 0.9040000000000001 | 0.004151453709393202 |
+| Epoch_14.pt | 0.9039999999999999 | 0.003929942040850535 |
+| Epoch_12_batch_2999.pt | 0.9036666666666665 | 0.0032942149067496855 |
+| Epoch_12.pt | 0.9030000000000001 | 0.004020011670027902 |
+| Epoch_13.pt | 0.9023333333333333 | 0.0033811386788228774 |
+| Epoch_11_batch_2999.pt | 0.9021666666666668 | 0.0033614554460901806 |
+| Epoch_10_batch_5999.pt | 0.9021666666666667 | 0.0037022682403435835 |
+| Epoch_10_batch_2999.pt | 0.899 | 0.0032697642154582585 |
+| Epoch_9.pt | 0.8866666666666667 | 0.004381498700695629 |
+| Epoch_9_batch_5999.pt | 0.8826666666666666 | 0.004375859703892272 |
+| Epoch_9_batch_2999.pt | 0.8811666666666668 | 0.003952261424851644 |
+| Epoch_8_batch_5999.pt | 0.8785000000000001 | 0.004624478615923729 |
+| Epoch_8_batch_2999.pt | 0.8748333333333334 | 0.0038845213569807385 |
+| Epoch_7_batch_2999.pt | 0.874 | 0.003984538017120242 |
+| Epoch_7_batch_5999.pt | 0.873 | 0.0034765528549249003 |
+| Epoch_6.pt | 0.8698333333333335 | 0.0036972629182166353 |
+| Epoch_7.pt | 0.8684999999999998 | 0.0030575750901815552 |
+| Epoch_6_batch_2999.pt | 0.8673333333333332 | 0.005455770532084851 |
+| Epoch_8.pt | 0.867 | 0.0043871304494221336 |
+| Epoch_6_batch_5999.pt | 0.8664999999999999 | 0.004093430448841922 |
+| Epoch_5.pt | 0.8658333333333335 | 0.003938179688543837 |
+| Epoch_5_batch_5999.pt | 0.8623333333333333 | 0.003953432638812715 |
+| Epoch_5_batch_2999.pt | 0.8619999999999999 | 0.00294811092476035 |
+| Epoch_4_batch_2999.pt | 0.8576666666666666 | 0.004347555010336406 |
+| Epoch_4_batch_5999.pt | 0.8548333333333333 | 0.004292183535937485 |
+| Epoch_3_batch_5999.pt | 0.8473333333333333 | 0.0030245905752924905 |
+| Epoch_4.pt | 0.8460000000000001 | 0.0043829073162924516 |
+| Epoch_3_batch_2999.pt | 0.843 | 0.004065816547451559 |
+| Epoch_3.pt | 0.8418333333333333 | 0.0041533119314590364 |
+| Epoch_2_batch_5999.pt | 0.8328333333333333 | 0.0038892856940415865 |
+| Epoch_2_batch_2999.pt | 0.8221666666666666 | 0.004325134694255983 |
+| Epoch_2.pt | 0.7996666666666667 | 0.003529417816504129 |
+| Epoch_1.pt | 0.7958333333333335 | 0.006276794416591016 |
+| Epoch_1_batch_5999.pt | 0.7888333333333333 | 0.004075294430020735 |
+| Epoch_1_batch_2999.pt | 0.7613333333333333 | 0.005162782291328418 |
+| Epoch_0_batch_5999.pt | 0.7138333333333333 | 0.006275810901320709 |
+| Epoch_0.pt | 0.7083333333333333 | 0.0035311663515712665 |
+| Epoch_0_batch_2999.pt | 0.6281666666666667 | 0.0036637193541772285 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_agedb30.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_agedb30.txt
new file mode 100644
index 0000000000000000000000000000000000000000..cd9c3e166749e94af91b34d4ac934e5e4a7ec202
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_agedb30.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.9578333333333333 | 0.0027448042948968066 |
+| Epoch_13.pt | 0.9578333333333333 | 0.0027894200468073817 |
+| Epoch_13_batch_5999.pt | 0.9574999999999999 | 0.0022804861946415294 |
+| Epoch_16_batch_2999.pt | 0.9573333333333333 | 0.002643417167415629 |
+| Epoch_17_batch_2999.pt | 0.9568333333333333 | 0.003097689302100148 |
+| Epoch_14_batch_5999.pt | 0.9568333333333333 | 0.0024273391416615017 |
+| Epoch_14_batch_2999.pt | 0.9566666666666667 | 0.002421610524189266 |
+| Epoch_16_batch_5999.pt | 0.9564999999999999 | 0.0026579719364234885 |
+| Epoch_17.pt | 0.9563333333333333 | 0.0027193862778934286 |
+| Epoch_15_batch_2999.pt | 0.9553333333333333 | 0.0032087842395985937 |
+| Epoch_12_batch_5999.pt | 0.9553333333333333 | 0.0026152449546532892 |
+| Epoch_13_batch_2999.pt | 0.9551666666666667 | 0.002539198862672546 |
+| Epoch_17_batch_5999.pt | 0.9551666666666666 | 0.0030877096070200576 |
+| Epoch_15.pt | 0.9550000000000001 | 0.003152502435358025 |
+| Epoch_12_batch_2999.pt | 0.9548333333333334 | 0.0029860788111948245 |
+| Epoch_15_batch_5999.pt | 0.9546666666666667 | 0.0030711722135745054 |
+| Epoch_16.pt | 0.9543333333333333 | 0.0031249691356500537 |
+| Epoch_11_batch_5999.pt | 0.9533333333333334 | 0.002843663087126604 |
+| Epoch_12.pt | 0.9525 | 0.0027805541680538245 |
+| Epoch_11.pt | 0.952 | 0.002958561545709854 |
+| Epoch_11_batch_2999.pt | 0.952 | 0.003071172213574502 |
+| Epoch_10_batch_5999.pt | 0.9514999999999999 | 0.0026346457114176484 |
+| Epoch_10.pt | 0.9503333333333334 | 0.0035555555555555553 |
+| Epoch_10_batch_2999.pt | 0.9463333333333332 | 0.0033222036417169015 |
+| Epoch_9.pt | 0.9396666666666667 | 0.0037085153929508033 |
+| Epoch_8_batch_5999.pt | 0.9396666666666667 | 0.004330483393312235 |
+| Epoch_9_batch_5999.pt | 0.9393333333333335 | 0.004368800722519448 |
+| Epoch_7_batch_5999.pt | 0.9381666666666668 | 0.006418770698813799 |
+| Epoch_8.pt | 0.9369999999999999 | 0.0046201384191759546 |
+| Epoch_8_batch_2999.pt | 0.9360000000000002 | 0.004333333333333336 |
+| Epoch_9_batch_2999.pt | 0.9356666666666668 | 0.005080828162299415 |
+| Epoch_7_batch_2999.pt | 0.9323333333333335 | 0.00445969605341988 |
+| Epoch_7.pt | 0.9323333333333332 | 0.004203173404306162 |
+| Epoch_5_batch_5999.pt | 0.9279999999999999 | 0.005481731726019557 |
+| Epoch_6_batch_5999.pt | 0.9268333333333334 | 0.005786585167543516 |
+| Epoch_5_batch_2999.pt | 0.9258333333333335 | 0.006112373606964084 |
+| Epoch_6_batch_2999.pt | 0.9256666666666666 | 0.0048253446112463275 |
+| Epoch_6.pt | 0.9254999999999999 | 0.006161659628924344 |
+| Epoch_5.pt | 0.9231666666666666 | 0.005960000414284514 |
+| Epoch_4_batch_2999.pt | 0.9188333333333333 | 0.005049446858839779 |
+| Epoch_4_batch_5999.pt | 0.9173333333333333 | 0.006320650068145731 |
+| Epoch_4.pt | 0.9145000000000001 | 0.005805754914174107 |
+| Epoch_3_batch_5999.pt | 0.9141666666666668 | 0.005796177965896635 |
+| Epoch_3_batch_2999.pt | 0.9041666666666668 | 0.006325775217217684 |
+| Epoch_2_batch_5999.pt | 0.8955 | 0.00756882005579714 |
+| Epoch_3.pt | 0.8896666666666666 | 0.005054029073575178 |
+| Epoch_2_batch_2999.pt | 0.8813333333333333 | 0.006372200426195405 |
+| Epoch_2.pt | 0.8719999999999999 | 0.008655976312694364 |
+| Epoch_1_batch_5999.pt | 0.8571666666666667 | 0.007449625063065048 |
+| Epoch_1.pt | 0.8341666666666667 | 0.005710343657903437 |
+| Epoch_1_batch_2999.pt | 0.8200000000000001 | 0.008954618092648347 |
+| Epoch_0_batch_5999.pt | 0.7471666666666666 | 0.005905897868201722 |
+| Epoch_0.pt | 0.7146666666666666 | 0.0049366355314495675 |
+| Epoch_0_batch_2999.pt | 0.5974999999999999 | 0.009132934568207554 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..9252be912581525dc5210a8e2e1e682c18a79a68
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9386666666666666 | 0.0034765528549248834 |
+| Epoch_17_batch_2999.pt | 0.9383333333333332 | 0.003406602159279081 |
+| Epoch_15_batch_5999.pt | 0.9381666666666668 | 0.0035175924707281595 |
+| Epoch_16_batch_2999.pt | 0.9381666666666666 | 0.0034645470728152986 |
+| Epoch_14.pt | 0.9380000000000001 | 0.0032565224199450674 |
+| Epoch_15.pt | 0.9376666666666666 | 0.00365317382728302 |
+| Epoch_14_batch_2999.pt | 0.9375 | 0.0036872318902322434 |
+| Epoch_16.pt | 0.9373333333333334 | 0.003550343401926765 |
+| Epoch_11.pt | 0.9371666666666668 | 0.0031036617163252033 |
+| Epoch_12_batch_5999.pt | 0.9371666666666668 | 0.00367717349852731 |
+| Epoch_12_batch_2999.pt | 0.9371666666666668 | 0.0031036617163251994 |
+| Epoch_17_batch_5999.pt | 0.9366666666666668 | 0.0038086970002228033 |
+| Epoch_15_batch_2999.pt | 0.9366666666666668 | 0.004112612338515937 |
+| Epoch_14_batch_5999.pt | 0.9365 | 0.003796115623523671 |
+| Epoch_17.pt | 0.9364999999999999 | 0.0039475730941090055 |
+| Epoch_13.pt | 0.9359999999999999 | 0.0037777777777777736 |
+| Epoch_13_batch_5999.pt | 0.9359999999999999 | 0.003728435941236117 |
+| Epoch_13_batch_2999.pt | 0.9356666666666668 | 0.0037201486761740425 |
+| Epoch_11_batch_2999.pt | 0.9348333333333333 | 0.003473444229409015 |
+| Epoch_10_batch_5999.pt | 0.9345000000000001 | 0.003833333333333333 |
+| Epoch_10.pt | 0.9343333333333336 | 0.003976784481816283 |
+| Epoch_12.pt | 0.9331666666666667 | 0.0036721339710357125 |
+| Epoch_10_batch_2999.pt | 0.9331666666666667 | 0.003578485092263983 |
+| Epoch_11_batch_5999.pt | 0.9316666666666666 | 0.0038968173148333737 |
+| Epoch_9.pt | 0.9271666666666667 | 0.0033152286106301544 |
+| Epoch_9_batch_5999.pt | 0.9255000000000001 | 0.004157768276015034 |
+| Epoch_7_batch_2999.pt | 0.9238333333333333 | 0.003897213312732352 |
+| Epoch_9_batch_2999.pt | 0.9226666666666666 | 0.004396968652757638 |
+| Epoch_7_batch_5999.pt | 0.9218333333333334 | 0.004377622670687092 |
+| Epoch_8_batch_5999.pt | 0.9211666666666666 | 0.004486261607382761 |
+| Epoch_8_batch_2999.pt | 0.9201666666666665 | 0.0032341732395173087 |
+| Epoch_6_batch_2999.pt | 0.9191666666666668 | 0.0033724556023947117 |
+| Epoch_8.pt | 0.9178333333333335 | 0.0037928620419329538 |
+| Epoch_5_batch_2999.pt | 0.9166666666666666 | 0.004281744192888379 |
+| Epoch_6_batch_5999.pt | 0.9164999999999999 | 0.0038924586790229725 |
+| Epoch_6.pt | 0.9161666666666667 | 0.0037437190197403196 |
+| Epoch_5_batch_5999.pt | 0.9146666666666668 | 0.003996912388578883 |
+| Epoch_7.pt | 0.9111666666666667 | 0.003434124324561246 |
+| Epoch_4_batch_5999.pt | 0.9099999999999999 | 0.004194352464039306 |
+| Epoch_5.pt | 0.9096666666666666 | 0.004980207740225552 |
+| Epoch_4_batch_2999.pt | 0.9061666666666666 | 0.004655078204114274 |
+| Epoch_3_batch_5999.pt | 0.9026666666666667 | 0.0040307460327149 |
+| Epoch_4.pt | 0.898 | 0.003649792823479302 |
+| Epoch_3_batch_2999.pt | 0.8955 | 0.00545492189123787 |
+| Epoch_3.pt | 0.8946666666666665 | 0.003839367231815773 |
+| Epoch_2_batch_5999.pt | 0.8943333333333333 | 0.004951617767812707 |
+| Epoch_2.pt | 0.8821666666666668 | 0.004956913117487637 |
+| Epoch_2_batch_2999.pt | 0.8815000000000002 | 0.006228419532798287 |
+| Epoch_1_batch_5999.pt | 0.8521666666666666 | 0.005403759550900705 |
+| Epoch_1.pt | 0.8471666666666667 | 0.004574823750231149 |
+| Epoch_1_batch_2999.pt | 0.8285 | 0.006645568466799026 |
+| Epoch_0_batch_5999.pt | 0.7381666666666666 | 0.004651098370211358 |
+| Epoch_0.pt | 0.7313333333333334 | 0.005138813813265414 |
+| Epoch_0_batch_2999.pt | 0.5836666666666666 | 0.005815050714724233 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..04e90766503d8019ca66bd9f1a690ee620eb1cce
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.826 | 0.006621140851306477 |
+| Epoch_13.pt | 0.8256666666666665 | 0.006046119049072346 |
+| Epoch_16_batch_2999.pt | 0.8256666666666665 | 0.006484549538406609 |
+| Epoch_17.pt | 0.8254999999999999 | 0.006961312846238844 |
+| Epoch_15_batch_5999.pt | 0.8253333333333334 | 0.006572482853853442 |
+| Epoch_16.pt | 0.8251666666666667 | 0.0064952499262875335 |
+| Epoch_17_batch_5999.pt | 0.8248333333333333 | 0.006783467557139595 |
+| Epoch_17_batch_2999.pt | 0.8244999999999999 | 0.0069657451100561835 |
+| Epoch_13_batch_5999.pt | 0.8243333333333334 | 0.0062122936659974965 |
+| Epoch_14.pt | 0.8238333333333333 | 0.006502848378751958 |
+| Epoch_15.pt | 0.8238333333333333 | 0.006831526409699357 |
+| Epoch_13_batch_2999.pt | 0.8234999999999999 | 0.006566140230825425 |
+| Epoch_15_batch_2999.pt | 0.8228333333333333 | 0.007201037305250187 |
+| Epoch_12_batch_5999.pt | 0.8228333333333332 | 0.006426459578652105 |
+| Epoch_12.pt | 0.8218333333333334 | 0.006243267979330084 |
+| Epoch_11_batch_5999.pt | 0.8216666666666667 | 0.006141339380193547 |
+| Epoch_11.pt | 0.8213333333333332 | 0.007199965706365514 |
+| Epoch_14_batch_2999.pt | 0.8211666666666668 | 0.006652995241168621 |
+| Epoch_10.pt | 0.8185 | 0.006457123542603505 |
+| Epoch_14_batch_5999.pt | 0.8183333333333334 | 0.006867349840161087 |
+| Epoch_10_batch_2999.pt | 0.8183333333333334 | 0.006191391873668897 |
+| Epoch_12_batch_2999.pt | 0.818 | 0.0063673550141752884 |
+| Epoch_11_batch_2999.pt | 0.8164999999999999 | 0.00723695079367847 |
+| Epoch_10_batch_5999.pt | 0.8148333333333333 | 0.007683757118544943 |
+| Epoch_9_batch_2999.pt | 0.8005000000000001 | 0.007320063751999247 |
+| Epoch_8_batch_2999.pt | 0.7929999999999999 | 0.008205689083394117 |
+| Epoch_9_batch_5999.pt | 0.7928333333333334 | 0.006831526409699353 |
+| Epoch_9.pt | 0.792 | 0.007592231247047343 |
+| Epoch_7_batch_5999.pt | 0.7914999999999999 | 0.006909690815115667 |
+| Epoch_8_batch_5999.pt | 0.7898333333333334 | 0.007767652139037026 |
+| Epoch_8.pt | 0.7896666666666667 | 0.005745637099487956 |
+| Epoch_7_batch_2999.pt | 0.7891666666666667 | 0.009037814933125827 |
+| Epoch_6_batch_5999.pt | 0.7863333333333333 | 0.007473409653688997 |
+| Epoch_5_batch_5999.pt | 0.7843333333333333 | 0.008462699566368722 |
+| Epoch_6_batch_2999.pt | 0.7843333333333332 | 0.007970625081982645 |
+| Epoch_5_batch_2999.pt | 0.7815 | 0.009383627717242542 |
+| Epoch_5.pt | 0.7781666666666667 | 0.008025614856642478 |
+| Epoch_4_batch_5999.pt | 0.7776666666666667 | 0.008143011849144552 |
+| Epoch_6.pt | 0.7731666666666667 | 0.008371947573464582 |
+| Epoch_7.pt | 0.7728333333333335 | 0.009753283714210121 |
+| Epoch_4_batch_2999.pt | 0.77 | 0.007681949341086048 |
+| Epoch_3_batch_5999.pt | 0.765 | 0.009036961141150638 |
+| Epoch_4.pt | 0.7615 | 0.007905498945686281 |
+| Epoch_3.pt | 0.7586666666666667 | 0.008853401383049493 |
+| Epoch_3_batch_2999.pt | 0.756 | 0.006881716696534232 |
+| Epoch_2_batch_5999.pt | 0.7545 | 0.008628298260039318 |
+| Epoch_2_batch_2999.pt | 0.7441666666666666 | 0.007015195147888828 |
+| Epoch_2.pt | 0.7328333333333333 | 0.008251636575421838 |
+| Epoch_1_batch_5999.pt | 0.7155 | 0.007872635839522139 |
+| Epoch_1.pt | 0.6944999999999999 | 0.008965124471535437 |
+| Epoch_1_batch_2999.pt | 0.6793333333333333 | 0.007662639844251856 |
+| Epoch_0_batch_5999.pt | 0.6295 | 0.007519727142409915 |
+| Epoch_0.pt | 0.6101666666666665 | 0.006418770698813802 |
+| Epoch_0_batch_2999.pt | 0.5516666666666665 | 0.008292492514242356 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..903ac41cad0d718c995b9f7b2046827a694410e1
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11.pt | 0.9956666666666667 | 0.0008678055195451773 |
+| Epoch_15_batch_2999.pt | 0.9955 | 0.0008624541497922234 |
+| Epoch_10_batch_5999.pt | 0.9951666666666668 | 0.001150684176511553 |
+| Epoch_13.pt | 0.9951666666666668 | 0.0009444444444444424 |
+| Epoch_16.pt | 0.9951666666666666 | 0.0008766518798921936 |
+| Epoch_16_batch_2999.pt | 0.9950000000000001 | 0.0009938079899999026 |
+| Epoch_13_batch_5999.pt | 0.9948333333333335 | 0.0008766518798921935 |
+| Epoch_16_batch_5999.pt | 0.9948333333333335 | 0.0011772011166898341 |
+| Epoch_17_batch_2999.pt | 0.9946666666666667 | 0.001212079123848407 |
+| Epoch_15_batch_5999.pt | 0.9946666666666667 | 0.0011055415967851326 |
+| Epoch_12_batch_5999.pt | 0.9946666666666667 | 0.0011055415967851344 |
+| Epoch_14_batch_5999.pt | 0.9946666666666667 | 0.0010772621905369608 |
+| Epoch_17_batch_5999.pt | 0.9945 | 0.0011666666666666629 |
+| Epoch_14_batch_2999.pt | 0.9945 | 0.0011666666666666629 |
+| Epoch_14.pt | 0.9943333333333333 | 0.0011706281947614122 |
+| Epoch_15.pt | 0.9943333333333333 | 0.0011967032904743294 |
+| Epoch_11_batch_5999.pt | 0.9943333333333333 | 0.0010304020550550787 |
+| Epoch_17.pt | 0.9941666666666669 | 0.0011180339887498939 |
+| Epoch_10_batch_2999.pt | 0.9941666666666666 | 0.0011453071182271292 |
+| Epoch_13_batch_2999.pt | 0.9940000000000001 | 0.0009362388636862645 |
+| Epoch_11_batch_2999.pt | 0.9940000000000001 | 0.0011706281947614103 |
+| Epoch_12_batch_2999.pt | 0.9940000000000001 | 0.0011166528467912104 |
+| Epoch_10.pt | 0.9936666666666667 | 0.0012120791238484118 |
+| Epoch_9_batch_2999.pt | 0.9936666666666667 | 0.0011600340565456157 |
+| Epoch_12.pt | 0.9935 | 0.0012285191326386663 |
+| Epoch_8_batch_2999.pt | 0.9926666666666668 | 0.001271724793584401 |
+| Epoch_7_batch_5999.pt | 0.9926666666666668 | 0.0015752718754175358 |
+| Epoch_9_batch_5999.pt | 0.9919999999999998 | 0.0009875771574795098 |
+| Epoch_5.pt | 0.9918333333333335 | 0.0009765775461803875 |
+| Epoch_6_batch_2999.pt | 0.9918333333333333 | 0.0014792807728549258 |
+| Epoch_7.pt | 0.9918333333333333 | 0.0013933262448871614 |
+| Epoch_6.pt | 0.9916666666666666 | 0.0012422599874998802 |
+| Epoch_8.pt | 0.9914999999999999 | 0.001253390463630947 |
+| Epoch_6_batch_5999.pt | 0.9913333333333332 | 0.0014865653511399643 |
+| Epoch_5_batch_5999.pt | 0.9911666666666668 | 0.0011124991330278195 |
+| Epoch_4_batch_5999.pt | 0.9911666666666668 | 0.001192828364087999 |
+| Epoch_4.pt | 0.991 | 0.0009362388636862651 |
+| Epoch_7_batch_2999.pt | 0.9906666666666666 | 0.001271724793584403 |
+| Epoch_8_batch_5999.pt | 0.9906666666666666 | 0.0013425606637327372 |
+| Epoch_5_batch_2999.pt | 0.9906666666666666 | 0.001474055462380183 |
+| Epoch_9.pt | 0.9903333333333334 | 0.0014444444444444507 |
+| Epoch_4_batch_2999.pt | 0.9901666666666668 | 0.0013252067157640682 |
+| Epoch_3_batch_5999.pt | 0.9884999999999999 | 0.0015406027359846706 |
+| Epoch_3.pt | 0.9876666666666667 | 0.0018291197370171471 |
+| Epoch_3_batch_2999.pt | 0.9868333333333335 | 0.0013709958532503398 |
+| Epoch_2_batch_5999.pt | 0.9858333333333332 | 0.0015163715626618011 |
+| Epoch_2_batch_2999.pt | 0.9818333333333333 | 0.0016749792701868098 |
+| Epoch_2.pt | 0.9795 | 0.001757699113284037 |
+| Epoch_1_batch_5999.pt | 0.9743333333333333 | 0.001555555555555555 |
+| Epoch_1.pt | 0.9696666666666666 | 0.0015869840952317527 |
+| Epoch_1_batch_2999.pt | 0.9625 | 0.002942347261525639 |
+| Epoch_0_batch_5999.pt | 0.931 | 0.00397678448181629 |
+| Epoch_0.pt | 0.9164999999999999 | 0.003860611478464824 |
+| Epoch_0_batch_2999.pt | 0.8228333333333333 | 0.005527987154046995 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..c054db27d05380ce07f656ad56784b05b7d06123
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.8931616699428512 |
+| 2 | 0.9143744223283908 |
+| 3 | 0.9242225028314045 |
+| 4 | 0.9299243657003007 |
+| 5 | 0.9340901102620514 |
+| 6 | 0.9369931135035213 |
+| 7 | 0.9397008474686592 |
+| 8 | 0.9419594620982334 |
+| 9 | 0.9436127419711783 |
+| 10 | 0.9451423513024461 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/LightCNN29/log.log b/bob/bio/facexzoo/models/backbones/LightCNN29/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..f5d84c17f9ede65f4c0d38e4a235d25993cdaeef
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/LightCNN29/log.log
@@ -0,0 +1,654 @@
+INFO 2021-10-11 14:25:20 train.py: 180] Start optimization.
+INFO 2021-10-11 14:25:20 train.py: 181] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='LightCNN', batch_size=512, data_root='/export/home/wangjun492/wj_data/msra_crop', epoches=180000, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mag-mobile', train_file='/export/home/wangjun492/wj_data/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f3ac422af28>)
+backbone param:
+{'depth': 29, 'drop_ratio': 0.4, 'out_h': 7, 'out_w': 7, 'feat_dim': 512, 'dropout_ratio': 0.2}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2021-10-11 14:25:45 train.py: 82] Epoch 0, iter 0/6416, lr 0.100000, loss 16.369417
+INFO 2021-10-11 14:31:18 train.py: 82] Epoch 0, iter 200/6416, lr 0.100000, loss 15.651273
+INFO 2021-10-11 14:38:05 train.py: 82] Epoch 0, iter 400/6416, lr 0.100000, loss 15.345418
+INFO 2021-10-11 14:45:57 train.py: 82] Epoch 0, iter 600/6416, lr 0.100000, loss 15.261409
+INFO 2021-10-11 14:53:00 train.py: 82] Epoch 0, iter 800/6416, lr 0.100000, loss 15.173952
+INFO 2021-10-11 15:00:32 train.py: 82] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.093871
+INFO 2021-10-11 15:07:48 train.py: 82] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.985372
+INFO 2021-10-11 15:14:40 train.py: 82] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.856655
+INFO 2021-10-11 15:21:58 train.py: 82] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.715352
+INFO 2021-10-11 15:28:46 train.py: 82] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.533554
+INFO 2021-10-11 15:36:21 train.py: 82] Epoch 0, iter 2000/6416, lr 0.100000, loss 14.381878
+INFO 2021-10-11 15:43:41 train.py: 82] Epoch 0, iter 2200/6416, lr 0.100000, loss 14.243026
+INFO 2021-10-11 15:50:29 train.py: 82] Epoch 0, iter 2400/6416, lr 0.100000, loss 14.102276
+INFO 2021-10-11 15:58:30 train.py: 82] Epoch 0, iter 2600/6416, lr 0.100000, loss 13.979956
+INFO 2021-10-11 16:05:00 train.py: 82] Epoch 0, iter 2800/6416, lr 0.100000, loss 13.843389
+INFO 2021-10-11 16:12:15 train.py: 95] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-10-11 16:12:16 train.py: 82] Epoch 0, iter 3000/6416, lr 0.100000, loss 13.708714
+INFO 2021-10-11 16:19:17 train.py: 82] Epoch 0, iter 3200/6416, lr 0.100000, loss 13.543101
+INFO 2021-10-11 16:26:22 train.py: 82] Epoch 0, iter 3400/6416, lr 0.100000, loss 13.357659
+INFO 2021-10-11 16:34:02 train.py: 82] Epoch 0, iter 3600/6416, lr 0.100000, loss 13.181470
+INFO 2021-10-11 16:40:23 train.py: 82] Epoch 0, iter 3800/6416, lr 0.100000, loss 13.011384
+INFO 2021-10-11 16:47:37 train.py: 82] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.825429
+INFO 2021-10-11 16:54:31 train.py: 82] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.656490
+INFO 2021-10-11 17:02:05 train.py: 82] Epoch 0, iter 4400/6416, lr 0.100000, loss 12.513968
+INFO 2021-10-11 17:08:40 train.py: 82] Epoch 0, iter 4600/6416, lr 0.100000, loss 12.413372
+INFO 2021-10-11 17:16:32 train.py: 82] Epoch 0, iter 4800/6416, lr 0.100000, loss 12.334254
+INFO 2021-10-11 17:23:52 train.py: 82] Epoch 0, iter 5000/6416, lr 0.100000, loss 12.348694
+INFO 2021-10-11 17:31:08 train.py: 82] Epoch 0, iter 5200/6416, lr 0.100000, loss 12.375680
+INFO 2021-10-11 17:37:42 train.py: 82] Epoch 0, iter 5400/6416, lr 0.100000, loss 12.418322
+INFO 2021-10-11 17:45:19 train.py: 82] Epoch 0, iter 5600/6416, lr 0.100000, loss 12.548591
+INFO 2021-10-11 17:52:17 train.py: 82] Epoch 0, iter 5800/6416, lr 0.100000, loss 12.691089
+INFO 2021-10-11 17:59:50 train.py: 95] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-10-11 17:59:50 train.py: 82] Epoch 0, iter 6000/6416, lr 0.100000, loss 12.901197
+INFO 2021-10-11 18:06:10 train.py: 82] Epoch 0, iter 6200/6416, lr 0.100000, loss 13.147470
+INFO 2021-10-11 18:14:01 train.py: 82] Epoch 0, iter 6400/6416, lr 0.100000, loss 13.373587
+INFO 2021-10-11 18:14:31 train.py: 100] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-10-11 18:14:33 train.py: 82] Epoch 1, iter 0/6416, lr 0.100000, loss 13.489206
+INFO 2021-10-11 18:15:47 train.py: 82] Epoch 1, iter 200/6416, lr 0.100000, loss 13.788728
+INFO 2021-10-11 18:17:01 train.py: 82] Epoch 1, iter 400/6416, lr 0.100000, loss 14.003860
+INFO 2021-10-11 18:18:15 train.py: 82] Epoch 1, iter 600/6416, lr 0.100000, loss 14.274572
+INFO 2021-10-11 18:19:28 train.py: 82] Epoch 1, iter 800/6416, lr 0.100000, loss 14.500683
+INFO 2021-10-11 18:20:42 train.py: 82] Epoch 1, iter 1000/6416, lr 0.100000, loss 14.732706
+INFO 2021-10-11 18:21:55 train.py: 82] Epoch 1, iter 1200/6416, lr 0.100000, loss 14.938140
+INFO 2021-10-11 18:23:08 train.py: 82] Epoch 1, iter 1400/6416, lr 0.100000, loss 15.113757
+INFO 2021-10-11 18:24:21 train.py: 82] Epoch 1, iter 1600/6416, lr 0.100000, loss 15.284691
+INFO 2021-10-11 18:25:34 train.py: 82] Epoch 1, iter 1800/6416, lr 0.100000, loss 15.397329
+INFO 2021-10-11 18:26:48 train.py: 82] Epoch 1, iter 2000/6416, lr 0.100000, loss 15.498385
+INFO 2021-10-11 18:28:01 train.py: 82] Epoch 1, iter 2200/6416, lr 0.100000, loss 15.572148
+INFO 2021-10-11 18:29:14 train.py: 82] Epoch 1, iter 2400/6416, lr 0.100000, loss 15.580485
+INFO 2021-10-11 18:30:26 train.py: 82] Epoch 1, iter 2600/6416, lr 0.100000, loss 15.591917
+INFO 2021-10-11 18:31:39 train.py: 82] Epoch 1, iter 2800/6416, lr 0.100000, loss 15.571841
+INFO 2021-10-11 18:32:53 train.py: 95] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-10-11 18:32:53 train.py: 82] Epoch 1, iter 3000/6416, lr 0.100000, loss 15.535668
+INFO 2021-10-11 18:34:06 train.py: 82] Epoch 1, iter 3200/6416, lr 0.100000, loss 15.511756
+INFO 2021-10-11 18:35:19 train.py: 82] Epoch 1, iter 3400/6416, lr 0.100000, loss 15.403087
+INFO 2021-10-11 18:36:32 train.py: 82] Epoch 1, iter 3600/6416, lr 0.100000, loss 15.328490
+INFO 2021-10-11 18:37:44 train.py: 82] Epoch 1, iter 3800/6416, lr 0.100000, loss 15.226699
+INFO 2021-10-11 18:38:57 train.py: 82] Epoch 1, iter 4000/6416, lr 0.100000, loss 15.136332
+INFO 2021-10-11 18:40:09 train.py: 82] Epoch 1, iter 4200/6416, lr 0.100000, loss 15.005565
+INFO 2021-10-11 18:41:22 train.py: 82] Epoch 1, iter 4400/6416, lr 0.100000, loss 14.881872
+INFO 2021-10-11 18:42:35 train.py: 82] Epoch 1, iter 4600/6416, lr 0.100000, loss 14.741101
+INFO 2021-10-11 18:43:48 train.py: 82] Epoch 1, iter 4800/6416, lr 0.100000, loss 14.616261
+INFO 2021-10-11 18:45:00 train.py: 82] Epoch 1, iter 5000/6416, lr 0.100000, loss 14.476123
+INFO 2021-10-11 18:46:13 train.py: 82] Epoch 1, iter 5200/6416, lr 0.100000, loss 14.333634
+INFO 2021-10-11 18:47:26 train.py: 82] Epoch 1, iter 5400/6416, lr 0.100000, loss 14.233344
+INFO 2021-10-11 18:48:39 train.py: 82] Epoch 1, iter 5600/6416, lr 0.100000, loss 14.102279
+INFO 2021-10-11 18:49:52 train.py: 82] Epoch 1, iter 5800/6416, lr 0.100000, loss 13.976093
+INFO 2021-10-11 18:51:05 train.py: 95] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-10-11 18:51:06 train.py: 82] Epoch 1, iter 6000/6416, lr 0.100000, loss 13.862471
+INFO 2021-10-11 18:52:18 train.py: 82] Epoch 1, iter 6200/6416, lr 0.100000, loss 13.684381
+INFO 2021-10-11 18:53:31 train.py: 82] Epoch 1, iter 6400/6416, lr 0.100000, loss 13.578435
+INFO 2021-10-11 18:53:38 train.py: 100] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-10-11 18:53:39 train.py: 82] Epoch 2, iter 0/6416, lr 0.100000, loss 13.555868
+INFO 2021-10-11 18:54:52 train.py: 82] Epoch 2, iter 200/6416, lr 0.100000, loss 13.156655
+INFO 2021-10-11 18:56:05 train.py: 82] Epoch 2, iter 400/6416, lr 0.100000, loss 13.023767
+INFO 2021-10-11 18:57:17 train.py: 82] Epoch 2, iter 600/6416, lr 0.100000, loss 13.002100
+INFO 2021-10-11 18:58:30 train.py: 82] Epoch 2, iter 800/6416, lr 0.100000, loss 12.910942
+INFO 2021-10-11 18:59:43 train.py: 82] Epoch 2, iter 1000/6416, lr 0.100000, loss 12.831745
+INFO 2021-10-11 19:00:55 train.py: 82] Epoch 2, iter 1200/6416, lr 0.100000, loss 12.767485
+INFO 2021-10-11 19:02:08 train.py: 82] Epoch 2, iter 1400/6416, lr 0.100000, loss 12.699910
+INFO 2021-10-11 19:03:20 train.py: 82] Epoch 2, iter 1600/6416, lr 0.100000, loss 12.583787
+INFO 2021-10-11 19:04:33 train.py: 82] Epoch 2, iter 1800/6416, lr 0.100000, loss 12.496349
+INFO 2021-10-11 19:05:46 train.py: 82] Epoch 2, iter 2000/6416, lr 0.100000, loss 12.416905
+INFO 2021-10-11 19:06:59 train.py: 82] Epoch 2, iter 2200/6416, lr 0.100000, loss 12.325140
+INFO 2021-10-11 19:08:11 train.py: 82] Epoch 2, iter 2400/6416, lr 0.100000, loss 12.252379
+INFO 2021-10-11 19:09:24 train.py: 82] Epoch 2, iter 2600/6416, lr 0.100000, loss 12.144799
+INFO 2021-10-11 19:10:36 train.py: 82] Epoch 2, iter 2800/6416, lr 0.100000, loss 12.092999
+INFO 2021-10-11 19:11:50 train.py: 95] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-10-11 19:11:50 train.py: 82] Epoch 2, iter 3000/6416, lr 0.100000, loss 11.976919
+INFO 2021-10-11 19:13:03 train.py: 82] Epoch 2, iter 3200/6416, lr 0.100000, loss 11.905363
+INFO 2021-10-11 19:14:16 train.py: 82] Epoch 2, iter 3400/6416, lr 0.100000, loss 11.826677
+INFO 2021-10-11 19:15:28 train.py: 82] Epoch 2, iter 3600/6416, lr 0.100000, loss 11.736064
+INFO 2021-10-11 19:16:41 train.py: 82] Epoch 2, iter 3800/6416, lr 0.100000, loss 11.647443
+INFO 2021-10-11 19:17:54 train.py: 82] Epoch 2, iter 4000/6416, lr 0.100000, loss 11.564908
+INFO 2021-10-11 19:19:07 train.py: 82] Epoch 2, iter 4200/6416, lr 0.100000, loss 11.495774
+INFO 2021-10-11 19:20:19 train.py: 82] Epoch 2, iter 4400/6416, lr 0.100000, loss 11.453334
+INFO 2021-10-11 19:21:32 train.py: 82] Epoch 2, iter 4600/6416, lr 0.100000, loss 11.364244
+INFO 2021-10-11 19:22:45 train.py: 82] Epoch 2, iter 4800/6416, lr 0.100000, loss 11.289079
+INFO 2021-10-11 19:23:58 train.py: 82] Epoch 2, iter 5000/6416, lr 0.100000, loss 11.251229
+INFO 2021-10-11 19:25:10 train.py: 82] Epoch 2, iter 5200/6416, lr 0.100000, loss 11.161038
+INFO 2021-10-11 19:26:23 train.py: 82] Epoch 2, iter 5400/6416, lr 0.100000, loss 11.068827
+INFO 2021-10-11 19:27:36 train.py: 82] Epoch 2, iter 5600/6416, lr 0.100000, loss 11.022828
+INFO 2021-10-11 19:28:48 train.py: 82] Epoch 2, iter 5800/6416, lr 0.100000, loss 10.995260
+INFO 2021-10-11 19:30:02 train.py: 95] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-10-11 19:30:02 train.py: 82] Epoch 2, iter 6000/6416, lr 0.100000, loss 10.949003
+INFO 2021-10-11 19:31:15 train.py: 82] Epoch 2, iter 6200/6416, lr 0.100000, loss 10.853169
+INFO 2021-10-11 19:32:28 train.py: 82] Epoch 2, iter 6400/6416, lr 0.100000, loss 10.819425
+INFO 2021-10-11 19:32:35 train.py: 100] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-10-11 19:32:36 train.py: 82] Epoch 3, iter 0/6416, lr 0.100000, loss 10.901023
+INFO 2021-10-11 19:33:49 train.py: 82] Epoch 3, iter 200/6416, lr 0.100000, loss 10.308744
+INFO 2021-10-11 19:35:02 train.py: 82] Epoch 3, iter 400/6416, lr 0.100000, loss 10.295135
+INFO 2021-10-11 19:36:15 train.py: 82] Epoch 3, iter 600/6416, lr 0.100000, loss 10.310412
+INFO 2021-10-11 19:37:27 train.py: 82] Epoch 3, iter 800/6416, lr 0.100000, loss 10.352543
+INFO 2021-10-11 19:38:40 train.py: 82] Epoch 3, iter 1000/6416, lr 0.100000, loss 10.338906
+INFO 2021-10-11 19:39:53 train.py: 82] Epoch 3, iter 1200/6416, lr 0.100000, loss 10.303761
+INFO 2021-10-11 19:41:06 train.py: 82] Epoch 3, iter 1400/6416, lr 0.100000, loss 10.279404
+INFO 2021-10-11 19:42:18 train.py: 82] Epoch 3, iter 1600/6416, lr 0.100000, loss 10.291302
+INFO 2021-10-11 19:43:31 train.py: 82] Epoch 3, iter 1800/6416, lr 0.100000, loss 10.236655
+INFO 2021-10-11 19:44:44 train.py: 82] Epoch 3, iter 2000/6416, lr 0.100000, loss 10.200403
+INFO 2021-10-11 19:45:57 train.py: 82] Epoch 3, iter 2200/6416, lr 0.100000, loss 10.172110
+INFO 2021-10-11 19:47:10 train.py: 82] Epoch 3, iter 2400/6416, lr 0.100000, loss 10.128775
+INFO 2021-10-11 19:48:23 train.py: 82] Epoch 3, iter 2600/6416, lr 0.100000, loss 10.089414
+INFO 2021-10-11 19:49:35 train.py: 82] Epoch 3, iter 2800/6416, lr 0.100000, loss 10.076917
+INFO 2021-10-11 19:50:49 train.py: 95] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-10-11 19:50:49 train.py: 82] Epoch 3, iter 3000/6416, lr 0.100000, loss 10.048779
+INFO 2021-10-11 19:52:02 train.py: 82] Epoch 3, iter 3200/6416, lr 0.100000, loss 9.983377
+INFO 2021-10-11 19:53:14 train.py: 82] Epoch 3, iter 3400/6416, lr 0.100000, loss 9.966386
+INFO 2021-10-11 19:54:27 train.py: 82] Epoch 3, iter 3600/6416, lr 0.100000, loss 9.905903
+INFO 2021-10-11 19:55:40 train.py: 82] Epoch 3, iter 3800/6416, lr 0.100000, loss 9.883424
+INFO 2021-10-11 19:56:52 train.py: 82] Epoch 3, iter 4000/6416, lr 0.100000, loss 9.859765
+INFO 2021-10-11 19:58:05 train.py: 82] Epoch 3, iter 4200/6416, lr 0.100000, loss 9.824201
+INFO 2021-10-11 19:59:18 train.py: 82] Epoch 3, iter 4400/6416, lr 0.100000, loss 9.762341
+INFO 2021-10-11 20:00:30 train.py: 82] Epoch 3, iter 4600/6416, lr 0.100000, loss 9.717520
+INFO 2021-10-11 20:01:43 train.py: 82] Epoch 3, iter 4800/6416, lr 0.100000, loss 9.733386
+INFO 2021-10-11 20:02:56 train.py: 82] Epoch 3, iter 5000/6416, lr 0.100000, loss 9.691353
+INFO 2021-10-11 20:04:08 train.py: 82] Epoch 3, iter 5200/6416, lr 0.100000, loss 9.672284
+INFO 2021-10-11 20:05:21 train.py: 82] Epoch 3, iter 5400/6416, lr 0.100000, loss 9.624590
+INFO 2021-10-11 20:06:34 train.py: 82] Epoch 3, iter 5600/6416, lr 0.100000, loss 9.559540
+INFO 2021-10-11 20:07:47 train.py: 82] Epoch 3, iter 5800/6416, lr 0.100000, loss 9.554033
+INFO 2021-10-11 20:09:01 train.py: 95] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-10-11 20:09:01 train.py: 82] Epoch 3, iter 6000/6416, lr 0.100000, loss 9.531322
+INFO 2021-10-11 20:10:14 train.py: 82] Epoch 3, iter 6200/6416, lr 0.100000, loss 9.491122
+INFO 2021-10-11 20:11:26 train.py: 82] Epoch 3, iter 6400/6416, lr 0.100000, loss 9.447590
+INFO 2021-10-11 20:11:33 train.py: 100] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-10-11 20:11:35 train.py: 82] Epoch 4, iter 0/6416, lr 0.100000, loss 9.546006
+INFO 2021-10-11 20:12:47 train.py: 82] Epoch 4, iter 200/6416, lr 0.100000, loss 8.969079
+INFO 2021-10-11 20:14:00 train.py: 82] Epoch 4, iter 400/6416, lr 0.100000, loss 8.960844
+INFO 2021-10-11 20:15:13 train.py: 82] Epoch 4, iter 600/6416, lr 0.100000, loss 9.052503
+INFO 2021-10-11 20:16:25 train.py: 82] Epoch 4, iter 800/6416, lr 0.100000, loss 9.093162
+INFO 2021-10-11 20:17:38 train.py: 82] Epoch 4, iter 1000/6416, lr 0.100000, loss 9.110755
+INFO 2021-10-11 20:18:50 train.py: 82] Epoch 4, iter 1200/6416, lr 0.100000, loss 9.120912
+INFO 2021-10-11 20:20:03 train.py: 82] Epoch 4, iter 1400/6416, lr 0.100000, loss 9.106434
+INFO 2021-10-11 20:21:15 train.py: 82] Epoch 4, iter 1600/6416, lr 0.100000, loss 9.096911
+INFO 2021-10-11 20:22:28 train.py: 82] Epoch 4, iter 1800/6416, lr 0.100000, loss 9.085510
+INFO 2021-10-11 20:23:40 train.py: 82] Epoch 4, iter 2000/6416, lr 0.100000, loss 9.106281
+INFO 2021-10-11 20:24:53 train.py: 82] Epoch 4, iter 2200/6416, lr 0.100000, loss 9.088951
+INFO 2021-10-11 20:26:06 train.py: 82] Epoch 4, iter 2400/6416, lr 0.100000, loss 9.057558
+INFO 2021-10-11 20:27:18 train.py: 82] Epoch 4, iter 2600/6416, lr 0.100000, loss 9.048886
+INFO 2021-10-11 20:28:31 train.py: 82] Epoch 4, iter 2800/6416, lr 0.100000, loss 9.036140
+INFO 2021-10-11 20:29:43 train.py: 95] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-10-11 20:29:44 train.py: 82] Epoch 4, iter 3000/6416, lr 0.100000, loss 9.003127
+INFO 2021-10-11 20:30:56 train.py: 82] Epoch 4, iter 3200/6416, lr 0.100000, loss 8.995553
+INFO 2021-10-11 20:32:09 train.py: 82] Epoch 4, iter 3400/6416, lr 0.100000, loss 8.969836
+INFO 2021-10-11 20:33:21 train.py: 82] Epoch 4, iter 3600/6416, lr 0.100000, loss 8.909684
+INFO 2021-10-11 20:34:34 train.py: 82] Epoch 4, iter 3800/6416, lr 0.100000, loss 8.933936
+INFO 2021-10-11 20:35:46 train.py: 82] Epoch 4, iter 4000/6416, lr 0.100000, loss 8.955494
+INFO 2021-10-11 20:36:59 train.py: 82] Epoch 4, iter 4200/6416, lr 0.100000, loss 8.894087
+INFO 2021-10-11 20:38:11 train.py: 82] Epoch 4, iter 4400/6416, lr 0.100000, loss 8.874055
+INFO 2021-10-11 20:39:24 train.py: 82] Epoch 4, iter 4600/6416, lr 0.100000, loss 8.855272
+INFO 2021-10-11 20:40:36 train.py: 82] Epoch 4, iter 4800/6416, lr 0.100000, loss 8.830998
+INFO 2021-10-11 20:41:49 train.py: 82] Epoch 4, iter 5000/6416, lr 0.100000, loss 8.791327
+INFO 2021-10-11 20:43:02 train.py: 82] Epoch 4, iter 5200/6416, lr 0.100000, loss 8.810705
+INFO 2021-10-11 20:44:14 train.py: 82] Epoch 4, iter 5400/6416, lr 0.100000, loss 8.804400
+INFO 2021-10-11 20:45:27 train.py: 82] Epoch 4, iter 5600/6416, lr 0.100000, loss 8.762913
+INFO 2021-10-11 20:46:39 train.py: 82] Epoch 4, iter 5800/6416, lr 0.100000, loss 8.709885
+INFO 2021-10-11 20:47:52 train.py: 95] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-10-11 20:47:53 train.py: 82] Epoch 4, iter 6000/6416, lr 0.100000, loss 8.712645
+INFO 2021-10-11 20:49:05 train.py: 82] Epoch 4, iter 6200/6416, lr 0.100000, loss 8.681430
+INFO 2021-10-11 20:50:18 train.py: 82] Epoch 4, iter 6400/6416, lr 0.100000, loss 8.680779
+INFO 2021-10-11 20:50:24 train.py: 100] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-10-11 20:50:26 train.py: 82] Epoch 5, iter 0/6416, lr 0.100000, loss 8.733378
+INFO 2021-10-11 20:51:39 train.py: 82] Epoch 5, iter 200/6416, lr 0.100000, loss 8.232566
+INFO 2021-10-11 20:52:51 train.py: 82] Epoch 5, iter 400/6416, lr 0.100000, loss 8.210535
+INFO 2021-10-11 20:54:03 train.py: 82] Epoch 5, iter 600/6416, lr 0.100000, loss 8.279316
+INFO 2021-10-11 20:55:16 train.py: 82] Epoch 5, iter 800/6416, lr 0.100000, loss 8.305984
+INFO 2021-10-11 20:56:28 train.py: 82] Epoch 5, iter 1000/6416, lr 0.100000, loss 8.359419
+INFO 2021-10-11 20:57:40 train.py: 82] Epoch 5, iter 1200/6416, lr 0.100000, loss 8.416656
+INFO 2021-10-11 20:58:53 train.py: 82] Epoch 5, iter 1400/6416, lr 0.100000, loss 8.409195
+INFO 2021-10-11 21:00:05 train.py: 82] Epoch 5, iter 1600/6416, lr 0.100000, loss 8.425412
+INFO 2021-10-11 21:01:17 train.py: 82] Epoch 5, iter 1800/6416, lr 0.100000, loss 8.429231
+INFO 2021-10-11 21:02:30 train.py: 82] Epoch 5, iter 2000/6416, lr 0.100000, loss 8.407216
+INFO 2021-10-11 21:03:42 train.py: 82] Epoch 5, iter 2200/6416, lr 0.100000, loss 8.415463
+INFO 2021-10-11 21:04:55 train.py: 82] Epoch 5, iter 2400/6416, lr 0.100000, loss 8.440240
+INFO 2021-10-11 21:06:07 train.py: 82] Epoch 5, iter 2600/6416, lr 0.100000, loss 8.419984
+INFO 2021-10-11 21:07:20 train.py: 82] Epoch 5, iter 2800/6416, lr 0.100000, loss 8.446846
+INFO 2021-10-11 21:08:33 train.py: 95] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-10-11 21:08:33 train.py: 82] Epoch 5, iter 3000/6416, lr 0.100000, loss 8.382041
+INFO 2021-10-11 21:09:46 train.py: 82] Epoch 5, iter 3200/6416, lr 0.100000, loss 8.327706
+INFO 2021-10-11 21:10:59 train.py: 82] Epoch 5, iter 3400/6416, lr 0.100000, loss 8.367816
+INFO 2021-10-11 21:12:11 train.py: 82] Epoch 5, iter 3600/6416, lr 0.100000, loss 8.370680
+INFO 2021-10-11 21:13:24 train.py: 82] Epoch 5, iter 3800/6416, lr 0.100000, loss 8.301980
+INFO 2021-10-11 21:14:36 train.py: 82] Epoch 5, iter 4000/6416, lr 0.100000, loss 8.332644
+INFO 2021-10-11 21:15:49 train.py: 82] Epoch 5, iter 4200/6416, lr 0.100000, loss 8.299320
+INFO 2021-10-11 21:17:02 train.py: 82] Epoch 5, iter 4400/6416, lr 0.100000, loss 8.336908
+INFO 2021-10-11 21:18:15 train.py: 82] Epoch 5, iter 4600/6416, lr 0.100000, loss 8.296913
+INFO 2021-10-11 21:19:27 train.py: 82] Epoch 5, iter 4800/6416, lr 0.100000, loss 8.277619
+INFO 2021-10-11 21:20:40 train.py: 82] Epoch 5, iter 5000/6416, lr 0.100000, loss 8.251506
+INFO 2021-10-11 21:21:53 train.py: 82] Epoch 5, iter 5200/6416, lr 0.100000, loss 8.299996
+INFO 2021-10-11 21:23:05 train.py: 82] Epoch 5, iter 5400/6416, lr 0.100000, loss 8.213327
+INFO 2021-10-11 21:24:18 train.py: 82] Epoch 5, iter 5600/6416, lr 0.100000, loss 8.275252
+INFO 2021-10-11 21:25:30 train.py: 82] Epoch 5, iter 5800/6416, lr 0.100000, loss 8.217992
+INFO 2021-10-11 21:26:43 train.py: 95] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-10-11 21:26:44 train.py: 82] Epoch 5, iter 6000/6416, lr 0.100000, loss 8.180969
+INFO 2021-10-11 21:27:56 train.py: 82] Epoch 5, iter 6200/6416, lr 0.100000, loss 8.200027
+INFO 2021-10-11 21:29:09 train.py: 82] Epoch 5, iter 6400/6416, lr 0.100000, loss 8.183986
+INFO 2021-10-11 21:29:15 train.py: 100] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-10-11 21:29:17 train.py: 82] Epoch 6, iter 0/6416, lr 0.100000, loss 8.173692
+INFO 2021-10-11 21:30:30 train.py: 82] Epoch 6, iter 200/6416, lr 0.100000, loss 7.717168
+INFO 2021-10-11 21:31:42 train.py: 82] Epoch 6, iter 400/6416, lr 0.100000, loss 7.686439
+INFO 2021-10-11 21:32:55 train.py: 82] Epoch 6, iter 600/6416, lr 0.100000, loss 7.784537
+INFO 2021-10-11 21:34:07 train.py: 82] Epoch 6, iter 800/6416, lr 0.100000, loss 7.852187
+INFO 2021-10-11 21:35:19 train.py: 82] Epoch 6, iter 1000/6416, lr 0.100000, loss 7.849350
+INFO 2021-10-11 21:36:32 train.py: 82] Epoch 6, iter 1200/6416, lr 0.100000, loss 7.895397
+INFO 2021-10-11 21:37:44 train.py: 82] Epoch 6, iter 1400/6416, lr 0.100000, loss 7.968666
+INFO 2021-10-11 21:38:56 train.py: 82] Epoch 6, iter 1600/6416, lr 0.100000, loss 7.918271
+INFO 2021-10-11 21:40:09 train.py: 82] Epoch 6, iter 1800/6416, lr 0.100000, loss 7.965510
+INFO 2021-10-11 21:41:22 train.py: 82] Epoch 6, iter 2000/6416, lr 0.100000, loss 7.977845
+INFO 2021-10-11 21:42:34 train.py: 82] Epoch 6, iter 2200/6416, lr 0.100000, loss 8.002042
+INFO 2021-10-11 21:43:47 train.py: 82] Epoch 6, iter 2400/6416, lr 0.100000, loss 8.005938
+INFO 2021-10-11 21:44:59 train.py: 82] Epoch 6, iter 2600/6416, lr 0.100000, loss 7.950871
+INFO 2021-10-11 21:46:12 train.py: 82] Epoch 6, iter 2800/6416, lr 0.100000, loss 7.954270
+INFO 2021-10-11 21:47:25 train.py: 95] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-10-11 21:47:25 train.py: 82] Epoch 6, iter 3000/6416, lr 0.100000, loss 7.975647
+INFO 2021-10-11 21:48:38 train.py: 82] Epoch 6, iter 3200/6416, lr 0.100000, loss 7.940947
+INFO 2021-10-11 21:49:50 train.py: 82] Epoch 6, iter 3400/6416, lr 0.100000, loss 7.963245
+INFO 2021-10-11 21:51:03 train.py: 82] Epoch 6, iter 3600/6416, lr 0.100000, loss 7.953733
+INFO 2021-10-11 21:52:15 train.py: 82] Epoch 6, iter 3800/6416, lr 0.100000, loss 7.977181
+INFO 2021-10-11 21:53:28 train.py: 82] Epoch 6, iter 4000/6416, lr 0.100000, loss 7.939420
+INFO 2021-10-11 21:54:40 train.py: 82] Epoch 6, iter 4200/6416, lr 0.100000, loss 7.950298
+INFO 2021-10-11 21:55:52 train.py: 82] Epoch 6, iter 4400/6416, lr 0.100000, loss 7.890088
+INFO 2021-10-11 21:57:05 train.py: 82] Epoch 6, iter 4600/6416, lr 0.100000, loss 7.883477
+INFO 2021-10-11 21:58:17 train.py: 82] Epoch 6, iter 4800/6416, lr 0.100000, loss 7.905051
+INFO 2021-10-11 21:59:30 train.py: 82] Epoch 6, iter 5000/6416, lr 0.100000, loss 7.891187
+INFO 2021-10-11 22:00:42 train.py: 82] Epoch 6, iter 5200/6416, lr 0.100000, loss 7.893691
+INFO 2021-10-11 22:01:55 train.py: 82] Epoch 6, iter 5400/6416, lr 0.100000, loss 7.848986
+INFO 2021-10-11 22:03:07 train.py: 82] Epoch 6, iter 5600/6416, lr 0.100000, loss 7.846524
+INFO 2021-10-11 22:04:19 train.py: 82] Epoch 6, iter 5800/6416, lr 0.100000, loss 7.852815
+INFO 2021-10-11 22:05:32 train.py: 95] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-10-11 22:05:33 train.py: 82] Epoch 6, iter 6000/6416, lr 0.100000, loss 7.860579
+INFO 2021-10-11 22:06:45 train.py: 82] Epoch 6, iter 6200/6416, lr 0.100000, loss 7.820628
+INFO 2021-10-11 22:07:57 train.py: 82] Epoch 6, iter 6400/6416, lr 0.100000, loss 7.811384
+INFO 2021-10-11 22:08:04 train.py: 100] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-10-11 22:08:06 train.py: 82] Epoch 7, iter 0/6416, lr 0.100000, loss 7.858979
+INFO 2021-10-11 22:09:18 train.py: 82] Epoch 7, iter 200/6416, lr 0.100000, loss 7.347155
+INFO 2021-10-11 22:10:31 train.py: 82] Epoch 7, iter 400/6416, lr 0.100000, loss 7.338178
+INFO 2021-10-11 22:11:43 train.py: 82] Epoch 7, iter 600/6416, lr 0.100000, loss 7.432431
+INFO 2021-10-11 22:12:56 train.py: 82] Epoch 7, iter 800/6416, lr 0.100000, loss 7.462470
+INFO 2021-10-11 22:14:08 train.py: 82] Epoch 7, iter 1000/6416, lr 0.100000, loss 7.529243
+INFO 2021-10-11 22:15:21 train.py: 82] Epoch 7, iter 1200/6416, lr 0.100000, loss 7.553025
+INFO 2021-10-11 22:16:34 train.py: 82] Epoch 7, iter 1400/6416, lr 0.100000, loss 7.597005
+INFO 2021-10-11 22:17:46 train.py: 82] Epoch 7, iter 1600/6416, lr 0.100000, loss 7.621427
+INFO 2021-10-11 22:18:59 train.py: 82] Epoch 7, iter 1800/6416, lr 0.100000, loss 7.642597
+INFO 2021-10-11 22:20:11 train.py: 82] Epoch 7, iter 2000/6416, lr 0.100000, loss 7.652329
+INFO 2021-10-11 22:21:23 train.py: 82] Epoch 7, iter 2200/6416, lr 0.100000, loss 7.604213
+INFO 2021-10-11 22:22:36 train.py: 82] Epoch 7, iter 2400/6416, lr 0.100000, loss 7.682101
+INFO 2021-10-11 22:23:49 train.py: 82] Epoch 7, iter 2600/6416, lr 0.100000, loss 7.678351
+INFO 2021-10-11 22:25:01 train.py: 82] Epoch 7, iter 2800/6416, lr 0.100000, loss 7.654919
+INFO 2021-10-11 22:26:14 train.py: 95] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-10-11 22:26:15 train.py: 82] Epoch 7, iter 3000/6416, lr 0.100000, loss 7.633514
+INFO 2021-10-11 22:27:27 train.py: 82] Epoch 7, iter 3200/6416, lr 0.100000, loss 7.621909
+INFO 2021-10-11 22:28:40 train.py: 82] Epoch 7, iter 3400/6416, lr 0.100000, loss 7.646492
+INFO 2021-10-11 22:29:53 train.py: 82] Epoch 7, iter 3600/6416, lr 0.100000, loss 7.642772
+INFO 2021-10-11 22:31:05 train.py: 82] Epoch 7, iter 3800/6416, lr 0.100000, loss 7.635411
+INFO 2021-10-11 22:32:18 train.py: 82] Epoch 7, iter 4000/6416, lr 0.100000, loss 7.627056
+INFO 2021-10-11 22:33:30 train.py: 82] Epoch 7, iter 4200/6416, lr 0.100000, loss 7.626884
+INFO 2021-10-11 22:34:43 train.py: 82] Epoch 7, iter 4400/6416, lr 0.100000, loss 7.654327
+INFO 2021-10-11 22:35:56 train.py: 82] Epoch 7, iter 4600/6416, lr 0.100000, loss 7.635016
+INFO 2021-10-11 22:37:09 train.py: 82] Epoch 7, iter 4800/6416, lr 0.100000, loss 7.628870
+INFO 2021-10-11 22:38:21 train.py: 82] Epoch 7, iter 5000/6416, lr 0.100000, loss 7.582794
+INFO 2021-10-11 22:39:34 train.py: 82] Epoch 7, iter 5200/6416, lr 0.100000, loss 7.625343
+INFO 2021-10-11 22:40:46 train.py: 82] Epoch 7, iter 5400/6416, lr 0.100000, loss 7.570694
+INFO 2021-10-11 22:41:59 train.py: 82] Epoch 7, iter 5600/6416, lr 0.100000, loss 7.544982
+INFO 2021-10-11 22:43:11 train.py: 82] Epoch 7, iter 5800/6416, lr 0.100000, loss 7.556763
+INFO 2021-10-11 22:44:24 train.py: 95] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-10-11 22:44:25 train.py: 82] Epoch 7, iter 6000/6416, lr 0.100000, loss 7.569464
+INFO 2021-10-11 22:45:37 train.py: 82] Epoch 7, iter 6200/6416, lr 0.100000, loss 7.551716
+INFO 2021-10-11 22:46:50 train.py: 82] Epoch 7, iter 6400/6416, lr 0.100000, loss 7.574602
+INFO 2021-10-11 22:46:56 train.py: 100] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-10-11 22:46:58 train.py: 82] Epoch 8, iter 0/6416, lr 0.100000, loss 7.628828
+INFO 2021-10-11 22:48:11 train.py: 82] Epoch 8, iter 200/6416, lr 0.100000, loss 7.097268
+INFO 2021-10-11 22:49:23 train.py: 82] Epoch 8, iter 400/6416, lr 0.100000, loss 7.102408
+INFO 2021-10-11 22:50:36 train.py: 82] Epoch 8, iter 600/6416, lr 0.100000, loss 7.184297
+INFO 2021-10-11 22:51:49 train.py: 82] Epoch 8, iter 800/6416, lr 0.100000, loss 7.239732
+INFO 2021-10-11 22:53:01 train.py: 82] Epoch 8, iter 1000/6416, lr 0.100000, loss 7.260486
+INFO 2021-10-11 22:54:13 train.py: 82] Epoch 8, iter 1200/6416, lr 0.100000, loss 7.328599
+INFO 2021-10-11 22:55:26 train.py: 82] Epoch 8, iter 1400/6416, lr 0.100000, loss 7.362567
+INFO 2021-10-11 22:56:39 train.py: 82] Epoch 8, iter 1600/6416, lr 0.100000, loss 7.380589
+INFO 2021-10-11 22:57:51 train.py: 82] Epoch 8, iter 1800/6416, lr 0.100000, loss 7.389757
+INFO 2021-10-11 22:59:03 train.py: 82] Epoch 8, iter 2000/6416, lr 0.100000, loss 7.399941
+INFO 2021-10-11 23:00:16 train.py: 82] Epoch 8, iter 2200/6416, lr 0.100000, loss 7.447249
+INFO 2021-10-11 23:01:29 train.py: 82] Epoch 8, iter 2400/6416, lr 0.100000, loss 7.388890
+INFO 2021-10-11 23:02:41 train.py: 82] Epoch 8, iter 2600/6416, lr 0.100000, loss 7.379715
+INFO 2021-10-11 23:03:53 train.py: 82] Epoch 8, iter 2800/6416, lr 0.100000, loss 7.402969
+INFO 2021-10-11 23:05:06 train.py: 95] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-10-11 23:05:07 train.py: 82] Epoch 8, iter 3000/6416, lr 0.100000, loss 7.407772
+INFO 2021-10-11 23:06:19 train.py: 82] Epoch 8, iter 3200/6416, lr 0.100000, loss 7.372286
+INFO 2021-10-11 23:07:31 train.py: 82] Epoch 8, iter 3400/6416, lr 0.100000, loss 7.388764
+INFO 2021-10-11 23:08:43 train.py: 82] Epoch 8, iter 3600/6416, lr 0.100000, loss 7.412668
+INFO 2021-10-11 23:09:56 train.py: 82] Epoch 8, iter 3800/6416, lr 0.100000, loss 7.406136
+INFO 2021-10-11 23:11:08 train.py: 82] Epoch 8, iter 4000/6416, lr 0.100000, loss 7.412901
+INFO 2021-10-11 23:12:20 train.py: 82] Epoch 8, iter 4200/6416, lr 0.100000, loss 7.397339
+INFO 2021-10-11 23:13:33 train.py: 82] Epoch 8, iter 4400/6416, lr 0.100000, loss 7.385983
+INFO 2021-10-11 23:14:45 train.py: 82] Epoch 8, iter 4600/6416, lr 0.100000, loss 7.372269
+INFO 2021-10-11 23:15:57 train.py: 82] Epoch 8, iter 4800/6416, lr 0.100000, loss 7.367348
+INFO 2021-10-11 23:17:10 train.py: 82] Epoch 8, iter 5000/6416, lr 0.100000, loss 7.402361
+INFO 2021-10-11 23:18:22 train.py: 82] Epoch 8, iter 5200/6416, lr 0.100000, loss 7.385488
+INFO 2021-10-11 23:19:34 train.py: 82] Epoch 8, iter 5400/6416, lr 0.100000, loss 7.330408
+INFO 2021-10-11 23:20:46 train.py: 82] Epoch 8, iter 5600/6416, lr 0.100000, loss 7.318797
+INFO 2021-10-11 23:21:59 train.py: 82] Epoch 8, iter 5800/6416, lr 0.100000, loss 7.353994
+INFO 2021-10-11 23:23:12 train.py: 95] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-10-11 23:23:12 train.py: 82] Epoch 8, iter 6000/6416, lr 0.100000, loss 7.320303
+INFO 2021-10-11 23:24:25 train.py: 82] Epoch 8, iter 6200/6416, lr 0.100000, loss 7.347665
+INFO 2021-10-11 23:25:37 train.py: 82] Epoch 8, iter 6400/6416, lr 0.100000, loss 7.330069
+INFO 2021-10-11 23:25:43 train.py: 100] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-10-11 23:25:45 train.py: 82] Epoch 9, iter 0/6416, lr 0.100000, loss 7.362451
+INFO 2021-10-11 23:26:58 train.py: 82] Epoch 9, iter 200/6416, lr 0.100000, loss 6.870590
+INFO 2021-10-11 23:28:10 train.py: 82] Epoch 9, iter 400/6416, lr 0.100000, loss 6.889367
+INFO 2021-10-11 23:29:22 train.py: 82] Epoch 9, iter 600/6416, lr 0.100000, loss 6.982569
+INFO 2021-10-11 23:30:35 train.py: 82] Epoch 9, iter 800/6416, lr 0.100000, loss 7.039750
+INFO 2021-10-11 23:31:47 train.py: 82] Epoch 9, iter 1000/6416, lr 0.100000, loss 7.072858
+INFO 2021-10-11 23:33:00 train.py: 82] Epoch 9, iter 1200/6416, lr 0.100000, loss 7.088371
+INFO 2021-10-11 23:34:12 train.py: 82] Epoch 9, iter 1400/6416, lr 0.100000, loss 7.112397
+INFO 2021-10-11 23:35:25 train.py: 82] Epoch 9, iter 1600/6416, lr 0.100000, loss 7.142042
+INFO 2021-10-11 23:36:37 train.py: 82] Epoch 9, iter 1800/6416, lr 0.100000, loss 7.182000
+INFO 2021-10-11 23:37:50 train.py: 82] Epoch 9, iter 2000/6416, lr 0.100000, loss 7.214905
+INFO 2021-10-11 23:39:02 train.py: 82] Epoch 9, iter 2200/6416, lr 0.100000, loss 7.226594
+INFO 2021-10-11 23:40:15 train.py: 82] Epoch 9, iter 2400/6416, lr 0.100000, loss 7.230421
+INFO 2021-10-11 23:41:27 train.py: 82] Epoch 9, iter 2600/6416, lr 0.100000, loss 7.233995
+INFO 2021-10-11 23:42:40 train.py: 82] Epoch 9, iter 2800/6416, lr 0.100000, loss 7.210237
+INFO 2021-10-11 23:43:53 train.py: 95] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-10-11 23:43:53 train.py: 82] Epoch 9, iter 3000/6416, lr 0.100000, loss 7.191278
+INFO 2021-10-11 23:45:06 train.py: 82] Epoch 9, iter 3200/6416, lr 0.100000, loss 7.224431
+INFO 2021-10-11 23:46:18 train.py: 82] Epoch 9, iter 3400/6416, lr 0.100000, loss 7.214607
+INFO 2021-10-11 23:47:31 train.py: 82] Epoch 9, iter 3600/6416, lr 0.100000, loss 7.231178
+INFO 2021-10-11 23:48:43 train.py: 82] Epoch 9, iter 3800/6416, lr 0.100000, loss 7.198881
+INFO 2021-10-11 23:49:56 train.py: 82] Epoch 9, iter 4000/6416, lr 0.100000, loss 7.224024
+INFO 2021-10-11 23:51:08 train.py: 82] Epoch 9, iter 4200/6416, lr 0.100000, loss 7.201266
+INFO 2021-10-11 23:52:21 train.py: 82] Epoch 9, iter 4400/6416, lr 0.100000, loss 7.226752
+INFO 2021-10-11 23:53:33 train.py: 82] Epoch 9, iter 4600/6416, lr 0.100000, loss 7.200576
+INFO 2021-10-11 23:54:46 train.py: 82] Epoch 9, iter 4800/6416, lr 0.100000, loss 7.239975
+INFO 2021-10-11 23:55:59 train.py: 82] Epoch 9, iter 5000/6416, lr 0.100000, loss 7.169514
+INFO 2021-10-11 23:57:12 train.py: 82] Epoch 9, iter 5200/6416, lr 0.100000, loss 7.204526
+INFO 2021-10-11 23:58:25 train.py: 82] Epoch 9, iter 5400/6416, lr 0.100000, loss 7.183902
+INFO 2021-10-11 23:59:37 train.py: 82] Epoch 9, iter 5600/6416, lr 0.100000, loss 7.213397
+INFO 2021-10-12 00:00:50 train.py: 82] Epoch 9, iter 5800/6416, lr 0.100000, loss 7.157720
+INFO 2021-10-12 00:02:03 train.py: 95] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-10-12 00:02:04 train.py: 82] Epoch 9, iter 6000/6416, lr 0.100000, loss 7.179854
+INFO 2021-10-12 00:03:16 train.py: 82] Epoch 9, iter 6200/6416, lr 0.100000, loss 7.212065
+INFO 2021-10-12 00:04:29 train.py: 82] Epoch 9, iter 6400/6416, lr 0.100000, loss 7.167078
+INFO 2021-10-12 00:04:36 train.py: 100] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-10-12 00:04:37 train.py: 82] Epoch 10, iter 0/6416, lr 0.010000, loss 7.270155
+INFO 2021-10-12 00:05:50 train.py: 82] Epoch 10, iter 200/6416, lr 0.010000, loss 5.900540
+INFO 2021-10-12 00:07:03 train.py: 82] Epoch 10, iter 400/6416, lr 0.010000, loss 5.672721
+INFO 2021-10-12 00:08:16 train.py: 82] Epoch 10, iter 600/6416, lr 0.010000, loss 5.576178
+INFO 2021-10-12 00:09:29 train.py: 82] Epoch 10, iter 800/6416, lr 0.010000, loss 5.506370
+INFO 2021-10-12 00:10:41 train.py: 82] Epoch 10, iter 1000/6416, lr 0.010000, loss 5.439568
+INFO 2021-10-12 00:11:54 train.py: 82] Epoch 10, iter 1200/6416, lr 0.010000, loss 5.392367
+INFO 2021-10-12 00:13:07 train.py: 82] Epoch 10, iter 1400/6416, lr 0.010000, loss 5.365546
+INFO 2021-10-12 00:14:19 train.py: 82] Epoch 10, iter 1600/6416, lr 0.010000, loss 5.316992
+INFO 2021-10-12 00:15:32 train.py: 82] Epoch 10, iter 1800/6416, lr 0.010000, loss 5.242922
+INFO 2021-10-12 00:16:44 train.py: 82] Epoch 10, iter 2000/6416, lr 0.010000, loss 5.259774
+INFO 2021-10-12 00:17:56 train.py: 82] Epoch 10, iter 2200/6416, lr 0.010000, loss 5.228217
+INFO 2021-10-12 00:19:08 train.py: 82] Epoch 10, iter 2400/6416, lr 0.010000, loss 5.224589
+INFO 2021-10-12 00:20:21 train.py: 82] Epoch 10, iter 2600/6416, lr 0.010000, loss 5.205215
+INFO 2021-10-12 00:21:33 train.py: 82] Epoch 10, iter 2800/6416, lr 0.010000, loss 5.154026
+INFO 2021-10-12 00:22:46 train.py: 95] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-10-12 00:22:47 train.py: 82] Epoch 10, iter 3000/6416, lr 0.010000, loss 5.146273
+INFO 2021-10-12 00:23:59 train.py: 82] Epoch 10, iter 3200/6416, lr 0.010000, loss 5.140429
+INFO 2021-10-12 00:25:11 train.py: 82] Epoch 10, iter 3400/6416, lr 0.010000, loss 5.099019
+INFO 2021-10-12 00:26:23 train.py: 82] Epoch 10, iter 3600/6416, lr 0.010000, loss 5.096363
+INFO 2021-10-12 00:27:36 train.py: 82] Epoch 10, iter 3800/6416, lr 0.010000, loss 5.044957
+INFO 2021-10-12 00:28:48 train.py: 82] Epoch 10, iter 4000/6416, lr 0.010000, loss 5.017373
+INFO 2021-10-12 00:30:01 train.py: 82] Epoch 10, iter 4200/6416, lr 0.010000, loss 5.019155
+INFO 2021-10-12 00:31:13 train.py: 82] Epoch 10, iter 4400/6416, lr 0.010000, loss 5.017333
+INFO 2021-10-12 00:32:25 train.py: 82] Epoch 10, iter 4600/6416, lr 0.010000, loss 4.956304
+INFO 2021-10-12 00:33:38 train.py: 82] Epoch 10, iter 4800/6416, lr 0.010000, loss 4.972423
+INFO 2021-10-12 00:34:50 train.py: 82] Epoch 10, iter 5000/6416, lr 0.010000, loss 4.951248
+INFO 2021-10-12 00:36:03 train.py: 82] Epoch 10, iter 5200/6416, lr 0.010000, loss 4.946939
+INFO 2021-10-12 00:37:15 train.py: 82] Epoch 10, iter 5400/6416, lr 0.010000, loss 4.910794
+INFO 2021-10-12 00:38:28 train.py: 82] Epoch 10, iter 5600/6416, lr 0.010000, loss 4.937050
+INFO 2021-10-12 00:39:40 train.py: 82] Epoch 10, iter 5800/6416, lr 0.010000, loss 4.889030
+INFO 2021-10-12 00:40:53 train.py: 95] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-10-12 00:40:53 train.py: 82] Epoch 10, iter 6000/6416, lr 0.010000, loss 4.889353
+INFO 2021-10-12 00:42:05 train.py: 82] Epoch 10, iter 6200/6416, lr 0.010000, loss 4.874124
+INFO 2021-10-12 00:43:18 train.py: 82] Epoch 10, iter 6400/6416, lr 0.010000, loss 4.865207
+INFO 2021-10-12 00:43:24 train.py: 100] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-10-12 00:43:26 train.py: 82] Epoch 11, iter 0/6416, lr 0.010000, loss 4.836659
+INFO 2021-10-12 00:44:39 train.py: 82] Epoch 11, iter 200/6416, lr 0.010000, loss 4.540447
+INFO 2021-10-12 00:45:51 train.py: 82] Epoch 11, iter 400/6416, lr 0.010000, loss 4.533320
+INFO 2021-10-12 00:47:04 train.py: 82] Epoch 11, iter 600/6416, lr 0.010000, loss 4.538823
+INFO 2021-10-12 00:48:16 train.py: 82] Epoch 11, iter 800/6416, lr 0.010000, loss 4.511605
+INFO 2021-10-12 00:49:28 train.py: 82] Epoch 11, iter 1000/6416, lr 0.010000, loss 4.536511
+INFO 2021-10-12 00:50:40 train.py: 82] Epoch 11, iter 1200/6416, lr 0.010000, loss 4.540965
+INFO 2021-10-12 00:51:52 train.py: 82] Epoch 11, iter 1400/6416, lr 0.010000, loss 4.535392
+INFO 2021-10-12 00:53:04 train.py: 82] Epoch 11, iter 1600/6416, lr 0.010000, loss 4.522141
+INFO 2021-10-12 00:54:16 train.py: 82] Epoch 11, iter 1800/6416, lr 0.010000, loss 4.516510
+INFO 2021-10-12 00:55:28 train.py: 82] Epoch 11, iter 2000/6416, lr 0.010000, loss 4.544702
+INFO 2021-10-12 00:56:40 train.py: 82] Epoch 11, iter 2200/6416, lr 0.010000, loss 4.504144
+INFO 2021-10-12 00:57:52 train.py: 82] Epoch 11, iter 2400/6416, lr 0.010000, loss 4.526364
+INFO 2021-10-12 00:59:04 train.py: 82] Epoch 11, iter 2600/6416, lr 0.010000, loss 4.545413
+INFO 2021-10-12 01:00:16 train.py: 82] Epoch 11, iter 2800/6416, lr 0.010000, loss 4.545825
+INFO 2021-10-12 01:01:28 train.py: 95] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-10-12 01:01:29 train.py: 82] Epoch 11, iter 3000/6416, lr 0.010000, loss 4.521823
+INFO 2021-10-12 01:02:40 train.py: 82] Epoch 11, iter 3200/6416, lr 0.010000, loss 4.536296
+INFO 2021-10-12 01:03:52 train.py: 82] Epoch 11, iter 3400/6416, lr 0.010000, loss 4.540962
+INFO 2021-10-12 01:05:04 train.py: 82] Epoch 11, iter 3600/6416, lr 0.010000, loss 4.553539
+INFO 2021-10-12 01:06:16 train.py: 82] Epoch 11, iter 3800/6416, lr 0.010000, loss 4.544167
+INFO 2021-10-12 01:07:28 train.py: 82] Epoch 11, iter 4000/6416, lr 0.010000, loss 4.529852
+INFO 2021-10-12 01:08:40 train.py: 82] Epoch 11, iter 4200/6416, lr 0.010000, loss 4.529248
+INFO 2021-10-12 01:09:52 train.py: 82] Epoch 11, iter 4400/6416, lr 0.010000, loss 4.549093
+INFO 2021-10-12 01:11:04 train.py: 82] Epoch 11, iter 4600/6416, lr 0.010000, loss 4.558291
+INFO 2021-10-12 01:12:16 train.py: 82] Epoch 11, iter 4800/6416, lr 0.010000, loss 4.548318
+INFO 2021-10-12 01:13:28 train.py: 82] Epoch 11, iter 5000/6416, lr 0.010000, loss 4.538473
+INFO 2021-10-12 01:14:40 train.py: 82] Epoch 11, iter 5200/6416, lr 0.010000, loss 4.545289
+INFO 2021-10-12 01:15:52 train.py: 82] Epoch 11, iter 5400/6416, lr 0.010000, loss 4.522825
+INFO 2021-10-12 01:17:04 train.py: 82] Epoch 11, iter 5600/6416, lr 0.010000, loss 4.514425
+INFO 2021-10-12 01:18:16 train.py: 82] Epoch 11, iter 5800/6416, lr 0.010000, loss 4.535403
+INFO 2021-10-12 01:19:29 train.py: 95] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-10-12 01:19:29 train.py: 82] Epoch 11, iter 6000/6416, lr 0.010000, loss 4.505395
+INFO 2021-10-12 01:20:42 train.py: 82] Epoch 11, iter 6200/6416, lr 0.010000, loss 4.523374
+INFO 2021-10-12 01:21:54 train.py: 82] Epoch 11, iter 6400/6416, lr 0.010000, loss 4.517419
+INFO 2021-10-12 01:22:00 train.py: 100] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-10-12 01:22:02 train.py: 82] Epoch 12, iter 0/6416, lr 0.010000, loss 4.438810
+INFO 2021-10-12 01:23:15 train.py: 82] Epoch 12, iter 200/6416, lr 0.010000, loss 4.212799
+INFO 2021-10-12 01:24:27 train.py: 82] Epoch 12, iter 400/6416, lr 0.010000, loss 4.208974
+INFO 2021-10-12 01:25:40 train.py: 82] Epoch 12, iter 600/6416, lr 0.010000, loss 4.215587
+INFO 2021-10-12 01:26:52 train.py: 82] Epoch 12, iter 800/6416, lr 0.010000, loss 4.189658
+INFO 2021-10-12 01:28:05 train.py: 82] Epoch 12, iter 1000/6416, lr 0.010000, loss 4.238931
+INFO 2021-10-12 01:29:17 train.py: 82] Epoch 12, iter 1200/6416, lr 0.010000, loss 4.224834
+INFO 2021-10-12 01:30:29 train.py: 82] Epoch 12, iter 1400/6416, lr 0.010000, loss 4.263417
+INFO 2021-10-12 01:31:41 train.py: 82] Epoch 12, iter 1600/6416, lr 0.010000, loss 4.264576
+INFO 2021-10-12 01:32:53 train.py: 82] Epoch 12, iter 1800/6416, lr 0.010000, loss 4.274143
+INFO 2021-10-12 01:34:05 train.py: 82] Epoch 12, iter 2000/6416, lr 0.010000, loss 4.256176
+INFO 2021-10-12 01:35:16 train.py: 82] Epoch 12, iter 2200/6416, lr 0.010000, loss 4.274632
+INFO 2021-10-12 01:36:28 train.py: 82] Epoch 12, iter 2400/6416, lr 0.010000, loss 4.259171
+INFO 2021-10-12 01:37:40 train.py: 82] Epoch 12, iter 2600/6416, lr 0.010000, loss 4.299415
+INFO 2021-10-12 01:38:52 train.py: 82] Epoch 12, iter 2800/6416, lr 0.010000, loss 4.305869
+INFO 2021-10-12 01:40:05 train.py: 95] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-10-12 01:40:05 train.py: 82] Epoch 12, iter 3000/6416, lr 0.010000, loss 4.312959
+INFO 2021-10-12 01:41:17 train.py: 82] Epoch 12, iter 3200/6416, lr 0.010000, loss 4.305828
+INFO 2021-10-12 01:42:29 train.py: 82] Epoch 12, iter 3400/6416, lr 0.010000, loss 4.324463
+INFO 2021-10-12 01:43:41 train.py: 82] Epoch 12, iter 3600/6416, lr 0.010000, loss 4.344051
+INFO 2021-10-12 01:44:53 train.py: 82] Epoch 12, iter 3800/6416, lr 0.010000, loss 4.323265
+INFO 2021-10-12 01:46:05 train.py: 82] Epoch 12, iter 4000/6416, lr 0.010000, loss 4.349889
+INFO 2021-10-12 01:47:16 train.py: 82] Epoch 12, iter 4200/6416, lr 0.010000, loss 4.357720
+INFO 2021-10-12 01:48:28 train.py: 82] Epoch 12, iter 4400/6416, lr 0.010000, loss 4.352079
+INFO 2021-10-12 01:49:39 train.py: 82] Epoch 12, iter 4600/6416, lr 0.010000, loss 4.359271
+INFO 2021-10-12 01:50:51 train.py: 82] Epoch 12, iter 4800/6416, lr 0.010000, loss 4.381599
+INFO 2021-10-12 01:52:03 train.py: 82] Epoch 12, iter 5000/6416, lr 0.010000, loss 4.335753
+INFO 2021-10-12 01:53:14 train.py: 82] Epoch 12, iter 5200/6416, lr 0.010000, loss 4.353476
+INFO 2021-10-12 01:54:26 train.py: 82] Epoch 12, iter 5400/6416, lr 0.010000, loss 4.373205
+INFO 2021-10-12 01:55:37 train.py: 82] Epoch 12, iter 5600/6416, lr 0.010000, loss 4.366847
+INFO 2021-10-12 01:56:49 train.py: 82] Epoch 12, iter 5800/6416, lr 0.010000, loss 4.364130
+INFO 2021-10-12 01:58:01 train.py: 95] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-10-12 01:58:02 train.py: 82] Epoch 12, iter 6000/6416, lr 0.010000, loss 4.368445
+INFO 2021-10-12 01:59:13 train.py: 82] Epoch 12, iter 6200/6416, lr 0.010000, loss 4.399129
+INFO 2021-10-12 02:00:25 train.py: 82] Epoch 12, iter 6400/6416, lr 0.010000, loss 4.361923
+INFO 2021-10-12 02:00:31 train.py: 100] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-10-12 02:00:33 train.py: 82] Epoch 13, iter 0/6416, lr 0.001000, loss 4.441228
+INFO 2021-10-12 02:01:45 train.py: 82] Epoch 13, iter 200/6416, lr 0.001000, loss 3.953566
+INFO 2021-10-12 02:02:57 train.py: 82] Epoch 13, iter 400/6416, lr 0.001000, loss 3.924126
+INFO 2021-10-12 02:04:10 train.py: 82] Epoch 13, iter 600/6416, lr 0.001000, loss 3.894496
+INFO 2021-10-12 02:05:22 train.py: 82] Epoch 13, iter 800/6416, lr 0.001000, loss 3.897298
+INFO 2021-10-12 02:06:33 train.py: 82] Epoch 13, iter 1000/6416, lr 0.001000, loss 3.880164
+INFO 2021-10-12 02:07:44 train.py: 82] Epoch 13, iter 1200/6416, lr 0.001000, loss 3.864967
+INFO 2021-10-12 02:08:56 train.py: 82] Epoch 13, iter 1400/6416, lr 0.001000, loss 3.889341
+INFO 2021-10-12 02:10:08 train.py: 82] Epoch 13, iter 1600/6416, lr 0.001000, loss 3.868593
+INFO 2021-10-12 02:11:19 train.py: 82] Epoch 13, iter 1800/6416, lr 0.001000, loss 3.896852
+INFO 2021-10-12 02:12:31 train.py: 82] Epoch 13, iter 2000/6416, lr 0.001000, loss 3.883626
+INFO 2021-10-12 02:13:43 train.py: 82] Epoch 13, iter 2200/6416, lr 0.001000, loss 3.867222
+INFO 2021-10-12 02:14:55 train.py: 82] Epoch 13, iter 2400/6416, lr 0.001000, loss 3.887697
+INFO 2021-10-12 02:16:06 train.py: 82] Epoch 13, iter 2600/6416, lr 0.001000, loss 3.846236
+INFO 2021-10-12 02:17:18 train.py: 82] Epoch 13, iter 2800/6416, lr 0.001000, loss 3.853321
+INFO 2021-10-12 02:18:30 train.py: 95] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-10-12 02:18:31 train.py: 82] Epoch 13, iter 3000/6416, lr 0.001000, loss 3.845072
+INFO 2021-10-12 02:19:42 train.py: 82] Epoch 13, iter 3200/6416, lr 0.001000, loss 3.897856
+INFO 2021-10-12 02:20:54 train.py: 82] Epoch 13, iter 3400/6416, lr 0.001000, loss 3.884608
+INFO 2021-10-12 02:22:05 train.py: 82] Epoch 13, iter 3600/6416, lr 0.001000, loss 3.877761
+INFO 2021-10-12 02:23:17 train.py: 82] Epoch 13, iter 3800/6416, lr 0.001000, loss 3.876001
+INFO 2021-10-12 02:24:28 train.py: 82] Epoch 13, iter 4000/6416, lr 0.001000, loss 3.894900
+INFO 2021-10-12 02:25:40 train.py: 82] Epoch 13, iter 4200/6416, lr 0.001000, loss 3.891905
+INFO 2021-10-12 02:26:52 train.py: 82] Epoch 13, iter 4400/6416, lr 0.001000, loss 3.904178
+INFO 2021-10-12 02:28:03 train.py: 82] Epoch 13, iter 4600/6416, lr 0.001000, loss 3.890727
+INFO 2021-10-12 02:29:15 train.py: 82] Epoch 13, iter 4800/6416, lr 0.001000, loss 3.879863
+INFO 2021-10-12 02:30:27 train.py: 82] Epoch 13, iter 5000/6416, lr 0.001000, loss 3.891477
+INFO 2021-10-12 02:31:38 train.py: 82] Epoch 13, iter 5200/6416, lr 0.001000, loss 3.882521
+INFO 2021-10-12 02:32:50 train.py: 82] Epoch 13, iter 5400/6416, lr 0.001000, loss 3.868747
+INFO 2021-10-12 02:34:02 train.py: 82] Epoch 13, iter 5600/6416, lr 0.001000, loss 3.883568
+INFO 2021-10-12 02:35:13 train.py: 82] Epoch 13, iter 5800/6416, lr 0.001000, loss 3.894116
+INFO 2021-10-12 02:36:26 train.py: 95] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-10-12 02:36:26 train.py: 82] Epoch 13, iter 6000/6416, lr 0.001000, loss 3.906889
+INFO 2021-10-12 02:37:38 train.py: 82] Epoch 13, iter 6200/6416, lr 0.001000, loss 3.897078
+INFO 2021-10-12 02:38:49 train.py: 82] Epoch 13, iter 6400/6416, lr 0.001000, loss 3.871173
+INFO 2021-10-12 02:38:56 train.py: 100] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-10-12 02:38:58 train.py: 82] Epoch 14, iter 0/6416, lr 0.001000, loss 3.928265
+INFO 2021-10-12 02:40:10 train.py: 82] Epoch 14, iter 200/6416, lr 0.001000, loss 3.825548
+INFO 2021-10-12 02:41:23 train.py: 82] Epoch 14, iter 400/6416, lr 0.001000, loss 3.822929
+INFO 2021-10-12 02:42:35 train.py: 82] Epoch 14, iter 600/6416, lr 0.001000, loss 3.796972
+INFO 2021-10-12 02:43:46 train.py: 82] Epoch 14, iter 800/6416, lr 0.001000, loss 3.832615
+INFO 2021-10-12 02:44:58 train.py: 82] Epoch 14, iter 1000/6416, lr 0.001000, loss 3.844211
+INFO 2021-10-12 02:46:10 train.py: 82] Epoch 14, iter 1200/6416, lr 0.001000, loss 3.819662
+INFO 2021-10-12 02:47:22 train.py: 82] Epoch 14, iter 1400/6416, lr 0.001000, loss 3.810265
+INFO 2021-10-12 02:48:34 train.py: 82] Epoch 14, iter 1600/6416, lr 0.001000, loss 3.827449
+INFO 2021-10-12 02:49:45 train.py: 82] Epoch 14, iter 1800/6416, lr 0.001000, loss 3.810592
+INFO 2021-10-12 02:50:57 train.py: 82] Epoch 14, iter 2000/6416, lr 0.001000, loss 3.823550
+INFO 2021-10-12 02:52:08 train.py: 82] Epoch 14, iter 2200/6416, lr 0.001000, loss 3.811846
+INFO 2021-10-12 02:53:20 train.py: 82] Epoch 14, iter 2400/6416, lr 0.001000, loss 3.851203
+INFO 2021-10-12 02:54:31 train.py: 82] Epoch 14, iter 2600/6416, lr 0.001000, loss 3.828231
+INFO 2021-10-12 02:55:43 train.py: 82] Epoch 14, iter 2800/6416, lr 0.001000, loss 3.838185
+INFO 2021-10-12 02:56:55 train.py: 95] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-10-12 02:56:56 train.py: 82] Epoch 14, iter 3000/6416, lr 0.001000, loss 3.848820
+INFO 2021-10-12 02:58:08 train.py: 82] Epoch 14, iter 3200/6416, lr 0.001000, loss 3.863523
+INFO 2021-10-12 02:59:19 train.py: 82] Epoch 14, iter 3400/6416, lr 0.001000, loss 3.851798
+INFO 2021-10-12 03:00:31 train.py: 82] Epoch 14, iter 3600/6416, lr 0.001000, loss 3.837083
+INFO 2021-10-12 03:01:43 train.py: 82] Epoch 14, iter 3800/6416, lr 0.001000, loss 3.845924
+INFO 2021-10-12 03:02:54 train.py: 82] Epoch 14, iter 4000/6416, lr 0.001000, loss 3.839927
+INFO 2021-10-12 03:04:06 train.py: 82] Epoch 14, iter 4200/6416, lr 0.001000, loss 3.866790
+INFO 2021-10-12 03:05:18 train.py: 82] Epoch 14, iter 4400/6416, lr 0.001000, loss 3.830319
+INFO 2021-10-12 03:06:30 train.py: 82] Epoch 14, iter 4600/6416, lr 0.001000, loss 3.849122
+INFO 2021-10-12 03:07:42 train.py: 82] Epoch 14, iter 4800/6416, lr 0.001000, loss 3.867346
+INFO 2021-10-12 03:08:53 train.py: 82] Epoch 14, iter 5000/6416, lr 0.001000, loss 3.857668
+INFO 2021-10-12 03:10:05 train.py: 82] Epoch 14, iter 5200/6416, lr 0.001000, loss 3.830453
+INFO 2021-10-12 03:11:17 train.py: 82] Epoch 14, iter 5400/6416, lr 0.001000, loss 3.841674
+INFO 2021-10-12 03:12:28 train.py: 82] Epoch 14, iter 5600/6416, lr 0.001000, loss 3.848972
+INFO 2021-10-12 03:13:40 train.py: 82] Epoch 14, iter 5800/6416, lr 0.001000, loss 3.859697
+INFO 2021-10-12 03:14:53 train.py: 95] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-10-12 03:14:53 train.py: 82] Epoch 14, iter 6000/6416, lr 0.001000, loss 3.847249
+INFO 2021-10-12 03:16:05 train.py: 82] Epoch 14, iter 6200/6416, lr 0.001000, loss 3.848293
+INFO 2021-10-12 03:17:17 train.py: 82] Epoch 14, iter 6400/6416, lr 0.001000, loss 3.844851
+INFO 2021-10-12 03:17:23 train.py: 100] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-10-12 03:17:25 train.py: 82] Epoch 15, iter 0/6416, lr 0.001000, loss 3.839562
+INFO 2021-10-12 03:18:38 train.py: 82] Epoch 15, iter 200/6416, lr 0.001000, loss 3.784405
+INFO 2021-10-12 03:19:50 train.py: 82] Epoch 15, iter 400/6416, lr 0.001000, loss 3.797444
+INFO 2021-10-12 03:21:01 train.py: 82] Epoch 15, iter 600/6416, lr 0.001000, loss 3.786072
+INFO 2021-10-12 03:22:13 train.py: 82] Epoch 15, iter 800/6416, lr 0.001000, loss 3.799371
+INFO 2021-10-12 03:23:24 train.py: 82] Epoch 15, iter 1000/6416, lr 0.001000, loss 3.802272
+INFO 2021-10-12 03:24:36 train.py: 82] Epoch 15, iter 1200/6416, lr 0.001000, loss 3.777513
+INFO 2021-10-12 03:25:47 train.py: 82] Epoch 15, iter 1400/6416, lr 0.001000, loss 3.773127
+INFO 2021-10-12 03:26:59 train.py: 82] Epoch 15, iter 1600/6416, lr 0.001000, loss 3.805599
+INFO 2021-10-12 03:28:11 train.py: 82] Epoch 15, iter 1800/6416, lr 0.001000, loss 3.811088
+INFO 2021-10-12 03:29:22 train.py: 82] Epoch 15, iter 2000/6416, lr 0.001000, loss 3.792045
+INFO 2021-10-12 03:30:34 train.py: 82] Epoch 15, iter 2200/6416, lr 0.001000, loss 3.809186
+INFO 2021-10-12 03:31:45 train.py: 82] Epoch 15, iter 2400/6416, lr 0.001000, loss 3.778055
+INFO 2021-10-12 03:32:57 train.py: 82] Epoch 15, iter 2600/6416, lr 0.001000, loss 3.805090
+INFO 2021-10-12 03:34:08 train.py: 82] Epoch 15, iter 2800/6416, lr 0.001000, loss 3.802675
+INFO 2021-10-12 03:35:20 train.py: 95] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-10-12 03:35:21 train.py: 82] Epoch 15, iter 3000/6416, lr 0.001000, loss 3.824646
+INFO 2021-10-12 03:36:32 train.py: 82] Epoch 15, iter 3200/6416, lr 0.001000, loss 3.831821
+INFO 2021-10-12 03:37:44 train.py: 82] Epoch 15, iter 3400/6416, lr 0.001000, loss 3.814118
+INFO 2021-10-12 03:38:55 train.py: 82] Epoch 15, iter 3600/6416, lr 0.001000, loss 3.812598
+INFO 2021-10-12 03:40:07 train.py: 82] Epoch 15, iter 3800/6416, lr 0.001000, loss 3.811165
+INFO 2021-10-12 03:41:18 train.py: 82] Epoch 15, iter 4000/6416, lr 0.001000, loss 3.823861
+INFO 2021-10-12 03:42:30 train.py: 82] Epoch 15, iter 4200/6416, lr 0.001000, loss 3.814580
+INFO 2021-10-12 03:43:42 train.py: 82] Epoch 15, iter 4400/6416, lr 0.001000, loss 3.810455
+INFO 2021-10-12 03:44:53 train.py: 82] Epoch 15, iter 4600/6416, lr 0.001000, loss 3.810051
+INFO 2021-10-12 03:46:05 train.py: 82] Epoch 15, iter 4800/6416, lr 0.001000, loss 3.793287
+INFO 2021-10-12 03:47:17 train.py: 82] Epoch 15, iter 5000/6416, lr 0.001000, loss 3.823380
+INFO 2021-10-12 03:48:28 train.py: 82] Epoch 15, iter 5200/6416, lr 0.001000, loss 3.806899
+INFO 2021-10-12 03:49:40 train.py: 82] Epoch 15, iter 5400/6416, lr 0.001000, loss 3.830942
+INFO 2021-10-12 03:50:51 train.py: 82] Epoch 15, iter 5600/6416, lr 0.001000, loss 3.815722
+INFO 2021-10-12 03:52:03 train.py: 82] Epoch 15, iter 5800/6416, lr 0.001000, loss 3.804584
+INFO 2021-10-12 03:53:15 train.py: 95] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-10-12 03:53:16 train.py: 82] Epoch 15, iter 6000/6416, lr 0.001000, loss 3.816869
+INFO 2021-10-12 03:54:28 train.py: 82] Epoch 15, iter 6200/6416, lr 0.001000, loss 3.823887
+INFO 2021-10-12 03:55:39 train.py: 82] Epoch 15, iter 6400/6416, lr 0.001000, loss 3.854727
+INFO 2021-10-12 03:55:46 train.py: 100] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-10-12 03:55:48 train.py: 82] Epoch 16, iter 0/6416, lr 0.000100, loss 3.879567
+INFO 2021-10-12 03:57:00 train.py: 82] Epoch 16, iter 200/6416, lr 0.000100, loss 3.748064
+INFO 2021-10-12 03:58:13 train.py: 82] Epoch 16, iter 400/6416, lr 0.000100, loss 3.753644
+INFO 2021-10-12 03:59:25 train.py: 82] Epoch 16, iter 600/6416, lr 0.000100, loss 3.729019
+INFO 2021-10-12 04:00:37 train.py: 82] Epoch 16, iter 800/6416, lr 0.000100, loss 3.750805
+INFO 2021-10-12 04:01:48 train.py: 82] Epoch 16, iter 1000/6416, lr 0.000100, loss 3.740468
+INFO 2021-10-12 04:03:00 train.py: 82] Epoch 16, iter 1200/6416, lr 0.000100, loss 3.747286
+INFO 2021-10-12 04:04:11 train.py: 82] Epoch 16, iter 1400/6416, lr 0.000100, loss 3.750590
+INFO 2021-10-12 04:05:22 train.py: 82] Epoch 16, iter 1600/6416, lr 0.000100, loss 3.763460
+INFO 2021-10-12 04:06:34 train.py: 82] Epoch 16, iter 1800/6416, lr 0.000100, loss 3.760462
+INFO 2021-10-12 04:07:45 train.py: 82] Epoch 16, iter 2000/6416, lr 0.000100, loss 3.786539
+INFO 2021-10-12 04:08:57 train.py: 82] Epoch 16, iter 2200/6416, lr 0.000100, loss 3.753119
+INFO 2021-10-12 04:10:08 train.py: 82] Epoch 16, iter 2400/6416, lr 0.000100, loss 3.775878
+INFO 2021-10-12 04:11:19 train.py: 82] Epoch 16, iter 2600/6416, lr 0.000100, loss 3.749750
+INFO 2021-10-12 04:12:31 train.py: 82] Epoch 16, iter 2800/6416, lr 0.000100, loss 3.775911
+INFO 2021-10-12 04:13:42 train.py: 95] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-10-12 04:13:43 train.py: 82] Epoch 16, iter 3000/6416, lr 0.000100, loss 3.738060
+INFO 2021-10-12 04:14:54 train.py: 82] Epoch 16, iter 3200/6416, lr 0.000100, loss 3.763050
+INFO 2021-10-12 04:16:06 train.py: 82] Epoch 16, iter 3400/6416, lr 0.000100, loss 3.754281
+INFO 2021-10-12 04:17:17 train.py: 82] Epoch 16, iter 3600/6416, lr 0.000100, loss 3.744961
+INFO 2021-10-12 04:18:29 train.py: 82] Epoch 16, iter 3800/6416, lr 0.000100, loss 3.745488
+INFO 2021-10-12 04:19:40 train.py: 82] Epoch 16, iter 4000/6416, lr 0.000100, loss 3.750797
+INFO 2021-10-12 04:20:52 train.py: 82] Epoch 16, iter 4200/6416, lr 0.000100, loss 3.732517
+INFO 2021-10-12 04:22:03 train.py: 82] Epoch 16, iter 4400/6416, lr 0.000100, loss 3.783783
+INFO 2021-10-12 04:23:15 train.py: 82] Epoch 16, iter 4600/6416, lr 0.000100, loss 3.755100
+INFO 2021-10-12 04:24:27 train.py: 82] Epoch 16, iter 4800/6416, lr 0.000100, loss 3.760945
+INFO 2021-10-12 04:25:39 train.py: 82] Epoch 16, iter 5000/6416, lr 0.000100, loss 3.758861
+INFO 2021-10-12 04:26:51 train.py: 82] Epoch 16, iter 5200/6416, lr 0.000100, loss 3.745392
+INFO 2021-10-12 04:28:03 train.py: 82] Epoch 16, iter 5400/6416, lr 0.000100, loss 3.771717
+INFO 2021-10-12 04:29:15 train.py: 82] Epoch 16, iter 5600/6416, lr 0.000100, loss 3.758553
+INFO 2021-10-12 04:30:27 train.py: 82] Epoch 16, iter 5800/6416, lr 0.000100, loss 3.742543
+INFO 2021-10-12 04:31:39 train.py: 95] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-10-12 04:31:40 train.py: 82] Epoch 16, iter 6000/6416, lr 0.000100, loss 3.763480
+INFO 2021-10-12 04:32:51 train.py: 82] Epoch 16, iter 6200/6416, lr 0.000100, loss 3.752254
+INFO 2021-10-12 04:34:03 train.py: 82] Epoch 16, iter 6400/6416, lr 0.000100, loss 3.772104
+INFO 2021-10-12 04:34:10 train.py: 100] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-10-12 04:34:12 train.py: 82] Epoch 17, iter 0/6416, lr 0.000100, loss 3.821568
+INFO 2021-10-12 04:35:24 train.py: 82] Epoch 17, iter 200/6416, lr 0.000100, loss 3.745310
+INFO 2021-10-12 04:36:37 train.py: 82] Epoch 17, iter 400/6416, lr 0.000100, loss 3.759674
+INFO 2021-10-12 04:37:48 train.py: 82] Epoch 17, iter 600/6416, lr 0.000100, loss 3.741271
+INFO 2021-10-12 04:39:00 train.py: 82] Epoch 17, iter 800/6416, lr 0.000100, loss 3.771942
+INFO 2021-10-12 04:40:12 train.py: 82] Epoch 17, iter 1000/6416, lr 0.000100, loss 3.761821
+INFO 2021-10-12 04:41:23 train.py: 82] Epoch 17, iter 1200/6416, lr 0.000100, loss 3.751132
+INFO 2021-10-12 04:42:34 train.py: 82] Epoch 17, iter 1400/6416, lr 0.000100, loss 3.757315
+INFO 2021-10-12 04:43:45 train.py: 82] Epoch 17, iter 1600/6416, lr 0.000100, loss 3.759437
+INFO 2021-10-12 04:44:57 train.py: 82] Epoch 17, iter 1800/6416, lr 0.000100, loss 3.729860
+INFO 2021-10-12 04:46:08 train.py: 82] Epoch 17, iter 2000/6416, lr 0.000100, loss 3.740405
+INFO 2021-10-12 04:47:20 train.py: 82] Epoch 17, iter 2200/6416, lr 0.000100, loss 3.740128
+INFO 2021-10-12 04:48:31 train.py: 82] Epoch 17, iter 2400/6416, lr 0.000100, loss 3.754893
+INFO 2021-10-12 04:49:43 train.py: 82] Epoch 17, iter 2600/6416, lr 0.000100, loss 3.745712
+INFO 2021-10-12 04:50:54 train.py: 82] Epoch 17, iter 2800/6416, lr 0.000100, loss 3.748604
+INFO 2021-10-12 04:52:06 train.py: 95] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-10-12 04:52:07 train.py: 82] Epoch 17, iter 3000/6416, lr 0.000100, loss 3.761845
+INFO 2021-10-12 04:53:18 train.py: 82] Epoch 17, iter 3200/6416, lr 0.000100, loss 3.727124
+INFO 2021-10-12 04:54:29 train.py: 82] Epoch 17, iter 3400/6416, lr 0.000100, loss 3.750767
+INFO 2021-10-12 04:55:41 train.py: 82] Epoch 17, iter 3600/6416, lr 0.000100, loss 3.750870
+INFO 2021-10-12 04:56:53 train.py: 82] Epoch 17, iter 3800/6416, lr 0.000100, loss 3.753284
+INFO 2021-10-12 04:58:04 train.py: 82] Epoch 17, iter 4000/6416, lr 0.000100, loss 3.757116
+INFO 2021-10-12 04:59:16 train.py: 82] Epoch 17, iter 4200/6416, lr 0.000100, loss 3.739168
+INFO 2021-10-12 05:00:28 train.py: 82] Epoch 17, iter 4400/6416, lr 0.000100, loss 3.747804
+INFO 2021-10-12 05:01:39 train.py: 82] Epoch 17, iter 4600/6416, lr 0.000100, loss 3.749971
+INFO 2021-10-12 05:02:51 train.py: 82] Epoch 17, iter 4800/6416, lr 0.000100, loss 3.734868
+INFO 2021-10-12 05:04:03 train.py: 82] Epoch 17, iter 5000/6416, lr 0.000100, loss 3.743888
+INFO 2021-10-12 05:05:14 train.py: 82] Epoch 17, iter 5200/6416, lr 0.000100, loss 3.754829
+INFO 2021-10-12 05:06:26 train.py: 82] Epoch 17, iter 5400/6416, lr 0.000100, loss 3.775448
+INFO 2021-10-12 05:07:38 train.py: 82] Epoch 17, iter 5600/6416, lr 0.000100, loss 3.737936
+INFO 2021-10-12 05:08:49 train.py: 82] Epoch 17, iter 5800/6416, lr 0.000100, loss 3.755900
+INFO 2021-10-12 05:10:02 train.py: 95] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-10-12 05:10:02 train.py: 82] Epoch 17, iter 6000/6416, lr 0.000100, loss 3.740855
+INFO 2021-10-12 05:11:14 train.py: 82] Epoch 17, iter 6200/6416, lr 0.000100, loss 3.718625
+INFO 2021-10-12 05:12:26 train.py: 82] Epoch 17, iter 6400/6416, lr 0.000100, loss 3.770707
+INFO 2021-10-12 05:12:32 train.py: 100] Save checkpoint Epoch_17.pt to disk...
\ No newline at end of file
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/.gitkeep b/bob/bio/facexzoo/models/backbones/MobileFaceNet/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0db4976e32282361fe102cc7e26e28dd760be91c
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9596666666666666 | 0.002808716591058783 |
+| Epoch_16_batch_5999.pt | 0.9595 | 0.0024726903426964316 |
+| Epoch_13.pt | 0.9588333333333333 | 0.002236758015466374 |
+| Epoch_14.pt | 0.9586666666666666 | 0.002830608711745998 |
+| Epoch_17.pt | 0.9584999999999999 | 0.0026810952248919255 |
+| Epoch_15.pt | 0.9583333333333334 | 0.002277100170213248 |
+| Epoch_14_batch_5999.pt | 0.9580000000000002 | 0.0025067809272618815 |
+| Epoch_16.pt | 0.958 | 0.0029376231258671785 |
+| Epoch_17_batch_2999.pt | 0.9578333333333333 | 0.0027448042948968053 |
+| Epoch_14_batch_2999.pt | 0.9576666666666667 | 0.002678215731820883 |
+| Epoch_17_batch_5999.pt | 0.9576666666666667 | 0.0027011657291429437 |
+| Epoch_13_batch_5999.pt | 0.9570000000000001 | 0.0027307123838765626 |
+| Epoch_13_batch_2999.pt | 0.9570000000000001 | 0.0026620330112690966 |
+| Epoch_15_batch_5999.pt | 0.9565000000000001 | 0.0026229048075806453 |
+| Epoch_15_batch_2999.pt | 0.9565000000000001 | 0.0027154109799666527 |
+| Epoch_12_batch_5999.pt | 0.9563333333333333 | 0.002662033011269096 |
+| Epoch_11_batch_5999.pt | 0.9560000000000001 | 0.003353641838397019 |
+| Epoch_10_batch_2999.pt | 0.9558333333333333 | 0.0029000851413640456 |
+| Epoch_12.pt | 0.9556666666666667 | 0.003288588574877501 |
+| Epoch_12_batch_2999.pt | 0.9553333333333331 | 0.003237511618740774 |
+| Epoch_11.pt | 0.9550000000000001 | 0.0028544961285922577 |
+| Epoch_10_batch_5999.pt | 0.9536666666666666 | 0.0031407320055783453 |
+| Epoch_11_batch_2999.pt | 0.9531666666666666 | 0.0029234049148717527 |
+| Epoch_10.pt | 0.9531666666666666 | 0.0030169275516412093 |
+| Epoch_8_batch_2999.pt | 0.9455 | 0.003505287012077192 |
+| Epoch_8_batch_5999.pt | 0.9445 | 0.0031822229981377076 |
+| Epoch_7_batch_2999.pt | 0.9418333333333331 | 0.00419766248885858 |
+| Epoch_9.pt | 0.9413333333333332 | 0.003749897117930266 |
+| Epoch_9_batch_5999.pt | 0.9411666666666665 | 0.004245913067618996 |
+| Epoch_6_batch_5999.pt | 0.9405000000000001 | 0.00441692871340021 |
+| Epoch_7_batch_5999.pt | 0.9403333333333332 | 0.003839367231815777 |
+| Epoch_5_batch_5999.pt | 0.9401666666666667 | 0.004461425891758629 |
+| Epoch_6_batch_2999.pt | 0.9396666666666667 | 0.003798960221617497 |
+| Epoch_9_batch_2999.pt | 0.9383333333333332 | 0.003583225665910461 |
+| Epoch_5_batch_2999.pt | 0.9356666666666668 | 0.0043899436144752 |
+| Epoch_8.pt | 0.9335000000000001 | 0.0038845213569807416 |
+| Epoch_6.pt | 0.9328333333333333 | 0.005500000000000003 |
+| Epoch_4_batch_2999.pt | 0.9305 | 0.004486261607382757 |
+| Epoch_4_batch_5999.pt | 0.9303333333333332 | 0.004373037478700983 |
+| Epoch_4.pt | 0.9301666666666666 | 0.0031666666666666623 |
+| Epoch_3_batch_2999.pt | 0.9298333333333334 | 0.004047938051543746 |
+| Epoch_7.pt | 0.9286666666666668 | 0.0035030850600965497 |
+| Epoch_3_batch_5999.pt | 0.9275 | 0.004603071974772747 |
+| Epoch_2_batch_5999.pt | 0.9241666666666667 | 0.006087073938749425 |
+| Epoch_3.pt | 0.9235 | 0.00513671125031725 |
+| Epoch_5.pt | 0.9218333333333332 | 0.004419722911821962 |
+| Epoch_2_batch_2999.pt | 0.921 | 0.0051830683509736 |
+| Epoch_2.pt | 0.9126666666666667 | 0.00476095228569523 |
+| Epoch_1_batch_5999.pt | 0.9028333333333333 | 0.007777976187945486 |
+| Epoch_1.pt | 0.9019999999999999 | 0.005664487598458173 |
+| Epoch_1_batch_2999.pt | 0.893 | 0.006185406959874126 |
+| Epoch_0_batch_5999.pt | 0.8415000000000001 | 0.007224145043182491 |
+| Epoch_0.pt | 0.8406666666666667 | 0.006602468674109144 |
+| Epoch_0_batch_2999.pt | 0.7585 | 0.005871304984602846 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..77bce39df7a606a31ef76f7d74719a0a4d8ef50f
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9381666666666666 | 0.003713921091857389 |
+| Epoch_14_batch_2999.pt | 0.9378333333333334 | 0.0034609818060383395 |
+| Epoch_14.pt | 0.9376666666666666 | 0.0033166247903553946 |
+| Epoch_15.pt | 0.9375 | 0.0037944891814509227 |
+| Epoch_13.pt | 0.9375 | 0.003593976442141298 |
+| Epoch_16_batch_2999.pt | 0.9375 | 0.004031128874149273 |
+| Epoch_12_batch_2999.pt | 0.9373333333333334 | 0.004015402444353916 |
+| Epoch_10_batch_5999.pt | 0.9373333333333334 | 0.0036868133384526766 |
+| Epoch_11_batch_2999.pt | 0.9371666666666668 | 0.0035140810224152056 |
+| Epoch_17_batch_5999.pt | 0.9369999999999999 | 0.0040046269535422415 |
+| Epoch_16.pt | 0.9368333333333334 | 0.0036552853665768807 |
+| Epoch_15_batch_5999.pt | 0.9366666666666668 | 0.0039518709430619355 |
+| Epoch_12.pt | 0.9366666666666665 | 0.003944053188733076 |
+| Epoch_17_batch_2999.pt | 0.9365 | 0.0035870996571906477 |
+| Epoch_12_batch_5999.pt | 0.9361666666666666 | 0.0035316033500695124 |
+| Epoch_17.pt | 0.9359999999999999 | 0.0037035185138886554 |
+| Epoch_15_batch_2999.pt | 0.9359999999999999 | 0.0036700321255363913 |
+| Epoch_13_batch_2999.pt | 0.9359999999999999 | 0.0037284359412361094 |
+| Epoch_14_batch_5999.pt | 0.9356666666666668 | 0.0034623192113072995 |
+| Epoch_11_batch_5999.pt | 0.9356666666666668 | 0.004099081499043714 |
+| Epoch_10_batch_2999.pt | 0.9354999999999999 | 0.0037271940261598894 |
+| Epoch_13_batch_5999.pt | 0.9353333333333333 | 0.003503085060096543 |
+| Epoch_11.pt | 0.9351666666666667 | 0.003156905032261193 |
+| Epoch_10.pt | 0.9334999999999999 | 0.00389245867902296 |
+| Epoch_9_batch_5999.pt | 0.9271666666666667 | 0.004082860894816248 |
+| Epoch_7_batch_5999.pt | 0.9256666666666667 | 0.003644715437079277 |
+| Epoch_8_batch_2999.pt | 0.9255000000000001 | 0.00458156531555982 |
+| Epoch_9_batch_2999.pt | 0.9243333333333332 | 0.0038666028091789584 |
+| Epoch_7_batch_2999.pt | 0.9238333333333333 | 0.003324525400083818 |
+| Epoch_6_batch_5999.pt | 0.9236666666666666 | 0.0038393672318157825 |
+| Epoch_5_batch_5999.pt | 0.9225 | 0.0030251007533471062 |
+| Epoch_9.pt | 0.9221666666666666 | 0.0033152286106301523 |
+| Epoch_8.pt | 0.9214999999999998 | 0.003482318654269966 |
+| Epoch_6_batch_2999.pt | 0.9213333333333334 | 0.0034318767136623358 |
+| Epoch_8_batch_5999.pt | 0.9211666666666666 | 0.0034161020310318276 |
+| Epoch_4_batch_5999.pt | 0.9185000000000001 | 0.004040306185683719 |
+| Epoch_6.pt | 0.9185000000000001 | 0.0036972629182166314 |
+| Epoch_4.pt | 0.9178333333333333 | 0.004289306254879471 |
+| Epoch_5_batch_2999.pt | 0.9173333333333332 | 0.0035241670060341605 |
+| Epoch_3_batch_5999.pt | 0.915 | 0.004201704533597563 |
+| Epoch_3_batch_2999.pt | 0.9148333333333334 | 0.003796115623523669 |
+| Epoch_4_batch_2999.pt | 0.9145 | 0.003677173498527313 |
+| Epoch_3.pt | 0.9115 | 0.004769048873779113 |
+| Epoch_7.pt | 0.9108333333333334 | 0.0037122586382862485 |
+| Epoch_2_batch_2999.pt | 0.9106666666666667 | 0.005364492313143695 |
+| Epoch_2_batch_5999.pt | 0.908 | 0.005310134731266008 |
+| Epoch_2.pt | 0.9076666666666666 | 0.004268749491621903 |
+| Epoch_5.pt | 0.9049999999999999 | 0.002897423291201178 |
+| Epoch_1.pt | 0.9023333333333333 | 0.004876246279442598 |
+| Epoch_1_batch_5999.pt | 0.8965 | 0.004915646472938589 |
+| Epoch_1_batch_2999.pt | 0.8821666666666668 | 0.005832275036275763 |
+| Epoch_0_batch_5999.pt | 0.8460000000000001 | 0.005410323921751042 |
+| Epoch_0.pt | 0.8458333333333332 | 0.006513281776967183 |
+| Epoch_0_batch_2999.pt | 0.7321666666666666 | 0.008913334441397197 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..795f9f8bb3803af502a7f3832dca0e3791ee72c1
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.8333333333333333 | 0.007260581890090914 |
+| Epoch_16_batch_5999.pt | 0.8308333333333333 | 0.006541651922606642 |
+| Epoch_10_batch_5999.pt | 0.8305 | 0.006898962169221229 |
+| Epoch_15.pt | 0.8303333333333333 | 0.006544246365612105 |
+| Epoch_14.pt | 0.8301666666666667 | 0.006394683362656983 |
+| Epoch_12_batch_2999.pt | 0.8298333333333334 | 0.0068693719949452685 |
+| Epoch_12_batch_5999.pt | 0.8298333333333334 | 0.006806179109755634 |
+| Epoch_17_batch_2999.pt | 0.8296666666666667 | 0.006343072433863149 |
+| Epoch_17_batch_5999.pt | 0.8295 | 0.006912370374736211 |
+| Epoch_15_batch_2999.pt | 0.8285 | 0.006385022984963992 |
+| Epoch_16_batch_2999.pt | 0.8281666666666666 | 0.0066223061143082865 |
+| Epoch_16.pt | 0.828 | 0.006220237778790407 |
+| Epoch_13_batch_2999.pt | 0.8278333333333334 | 0.006781647347123024 |
+| Epoch_13.pt | 0.8275 | 0.006723140433157013 |
+| Epoch_14_batch_2999.pt | 0.8271666666666666 | 0.006804364975223588 |
+| Epoch_12.pt | 0.8266666666666665 | 0.007657804862272348 |
+| Epoch_15_batch_5999.pt | 0.8265 | 0.006622306114308287 |
+| Epoch_17.pt | 0.8265 | 0.005765210691812374 |
+| Epoch_11_batch_5999.pt | 0.8263333333333334 | 0.007473409653688988 |
+| Epoch_14_batch_5999.pt | 0.8256666666666665 | 0.007058835633008262 |
+| Epoch_11_batch_2999.pt | 0.825 | 0.007395360574352775 |
+| Epoch_10_batch_2999.pt | 0.8243333333333333 | 0.007400367024598542 |
+| Epoch_11.pt | 0.8234999999999999 | 0.006336500182900753 |
+| Epoch_10.pt | 0.8203333333333334 | 0.006059376979749321 |
+| Epoch_9_batch_5999.pt | 0.8091666666666667 | 0.005785518319236597 |
+| Epoch_8.pt | 0.8038333333333332 | 0.008326107978766194 |
+| Epoch_7_batch_5999.pt | 0.8023333333333333 | 0.006291283407507587 |
+| Epoch_9_batch_2999.pt | 0.8021666666666667 | 0.007833333333333335 |
+| Epoch_8_batch_2999.pt | 0.8015000000000001 | 0.006570839057983835 |
+| Epoch_8_batch_5999.pt | 0.8013333333333333 | 0.00795744856435284 |
+| Epoch_6_batch_5999.pt | 0.8001666666666667 | 0.007897686799951202 |
+| Epoch_9.pt | 0.7971666666666667 | 0.008927174493894722 |
+| Epoch_5_batch_5999.pt | 0.7953333333333333 | 0.008042326301865584 |
+| Epoch_5_batch_2999.pt | 0.7938333333333334 | 0.007856938429279016 |
+| Epoch_6_batch_2999.pt | 0.7936666666666666 | 0.0081187178942807 |
+| Epoch_7_batch_2999.pt | 0.7928333333333334 | 0.008516686598512587 |
+| Epoch_6.pt | 0.7921666666666667 | 0.0050738375181408595 |
+| Epoch_4_batch_2999.pt | 0.7898333333333334 | 0.005496631965391444 |
+| Epoch_4_batch_5999.pt | 0.7896666666666666 | 0.006269660435407063 |
+| Epoch_3.pt | 0.7871666666666667 | 0.00937902177109724 |
+| Epoch_3_batch_5999.pt | 0.7851666666666667 | 0.0068423607939488125 |
+| Epoch_2.pt | 0.7841666666666666 | 0.009048054152723514 |
+| Epoch_4.pt | 0.7835 | 0.006806179109755635 |
+| Epoch_7.pt | 0.7829999999999999 | 0.006235105709599193 |
+| Epoch_3_batch_2999.pt | 0.7828333333333333 | 0.00753612698118491 |
+| Epoch_5.pt | 0.7808333333333334 | 0.007646713164636313 |
+| Epoch_2_batch_5999.pt | 0.7789999999999999 | 0.008632410933135572 |
+| Epoch_2_batch_2999.pt | 0.7759999999999999 | 0.008274607894182611 |
+| Epoch_1_batch_5999.pt | 0.7626666666666667 | 0.007106769507976613 |
+| Epoch_1_batch_2999.pt | 0.7535000000000001 | 0.008799726426724021 |
+| Epoch_1.pt | 0.7526666666666667 | 0.006877230277975012 |
+| Epoch_0.pt | 0.7073333333333334 | 0.009102299261826758 |
+| Epoch_0_batch_5999.pt | 0.7026666666666668 | 0.007790465841314707 |
+| Epoch_0_batch_2999.pt | 0.6346666666666667 | 0.00881146839684748 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..265655b2c338ca798d30df93c9eef55630fb5125
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9956666666666665 | 0.0011706281947614146 |
+| Epoch_12_batch_2999.pt | 0.9956666666666665 | 0.0010886621079036374 |
+| Epoch_12.pt | 0.9956666666666665 | 0.0010599324460188284 |
+| Epoch_16_batch_2999.pt | 0.9956666666666665 | 0.0011706281947614146 |
+| Epoch_15_batch_5999.pt | 0.9955 | 0.0011399046960379555 |
+| Epoch_15_batch_2999.pt | 0.9955 | 0.0011399046960379555 |
+| Epoch_11_batch_5999.pt | 0.9955 | 0.0011666666666666642 |
+| Epoch_16.pt | 0.9953333333333333 | 0.001105541596785135 |
+| Epoch_13.pt | 0.9953333333333332 | 0.001133115447465063 |
+| Epoch_14_batch_2999.pt | 0.9951666666666666 | 0.0010957268290731133 |
+| Epoch_10.pt | 0.9951666666666666 | 0.0011772011166898393 |
+| Epoch_17_batch_2999.pt | 0.9951666666666666 | 0.0010378634273483006 |
+| Epoch_13_batch_2999.pt | 0.9949999999999999 | 0.0011111111111111128 |
+| Epoch_17_batch_5999.pt | 0.9949999999999999 | 0.0011111111111111128 |
+| Epoch_13_batch_5999.pt | 0.9948333333333332 | 0.001228519132638667 |
+| Epoch_17.pt | 0.9948333333333332 | 0.0011235415786753737 |
+| Epoch_14_batch_5999.pt | 0.9948333333333332 | 0.0010671873729054782 |
+| Epoch_12_batch_5999.pt | 0.9946666666666667 | 0.001018350154434636 |
+| Epoch_14.pt | 0.9945 | 0.0011399046960379542 |
+| Epoch_16_batch_5999.pt | 0.9944999999999998 | 0.0011666666666666696 |
+| Epoch_11_batch_2999.pt | 0.9943333333333333 | 0.0012222222222222198 |
+| Epoch_10_batch_5999.pt | 0.9941666666666666 | 0.0010613873985857113 |
+| Epoch_11.pt | 0.9940000000000001 | 0.0013193713430042124 |
+| Epoch_10_batch_2999.pt | 0.994 | 0.0012957670877434026 |
+| Epoch_7_batch_5999.pt | 0.9935 | 0.0013017082793177735 |
+| Epoch_8.pt | 0.9931666666666666 | 0.001301708279317778 |
+| Epoch_9_batch_2999.pt | 0.9931666666666666 | 0.0012533904636309486 |
+| Epoch_9_batch_5999.pt | 0.993 | 0.0011600340565456162 |
+| Epoch_5_batch_5999.pt | 0.993 | 0.0017177360926378114 |
+| Epoch_8_batch_5999.pt | 0.9926666666666666 | 0.0013653561919382778 |
+| Epoch_6_batch_5999.pt | 0.9924999999999999 | 0.0017078251276599358 |
+| Epoch_9.pt | 0.9923333333333332 | 0.0007934920476158749 |
+| Epoch_4_batch_5999.pt | 0.9921666666666666 | 0.0016859989894992859 |
+| Epoch_3.pt | 0.992 | 0.0017533037597843894 |
+| Epoch_3_batch_2999.pt | 0.9918333333333335 | 0.0012031337682059848 |
+| Epoch_8_batch_2999.pt | 0.9918333333333333 | 0.0016187558093703858 |
+| Epoch_6_batch_2999.pt | 0.9916666666666668 | 0.0015908690070307063 |
+| Epoch_5_batch_2999.pt | 0.9915 | 0.001479280772854929 |
+| Epoch_2_batch_2999.pt | 0.9915 | 0.0018333333333333322 |
+| Epoch_2_batch_5999.pt | 0.9913333333333334 | 0.0017177360926378127 |
+| Epoch_6.pt | 0.9911666666666665 | 0.0012435016269777464 |
+| Epoch_4_batch_2999.pt | 0.991 | 0.001742709682373126 |
+| Epoch_3_batch_5999.pt | 0.9909999999999999 | 0.0020964402515681346 |
+| Epoch_5.pt | 0.9908333333333333 | 0.001614937983749851 |
+| Epoch_4.pt | 0.9908333333333333 | 0.0015762512176790149 |
+| Epoch_7.pt | 0.9906666666666668 | 0.0017603310575283158 |
+| Epoch_7_batch_2999.pt | 0.9905000000000002 | 0.0017751717009633857 |
+| Epoch_2.pt | 0.9884999999999999 | 0.0016749792701868215 |
+| Epoch_1_batch_5999.pt | 0.9864999999999998 | 0.0015204369092671074 |
+| Epoch_1.pt | 0.9841666666666666 | 0.002620550314460163 |
+| Epoch_1_batch_2999.pt | 0.9823333333333334 | 0.0018954135676924433 |
+| Epoch_0.pt | 0.9703333333333333 | 0.0022879178091082257 |
+| Epoch_0_batch_5999.pt | 0.9655000000000001 | 0.0030434102055116375 |
+| Epoch_0_batch_2999.pt | 0.9373333333333334 | 0.003390254733864971 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..fdd509ff22c09c8d778b6039b9ea42abb155b89c
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8873333333333333 | 0.005117146195408648 |
+| Epoch_17_batch_2999.pt | 0.8868333333333334 | 0.004292183535937484 |
+| Epoch_16_batch_2999.pt | 0.8868333333333333 | 0.004657729537494039 |
+| Epoch_16_batch_5999.pt | 0.8866666666666667 | 0.00450651106459406 |
+| Epoch_14_batch_2999.pt | 0.8859999999999999 | 0.0049140765305546565 |
+| Epoch_15_batch_5999.pt | 0.8859999999999999 | 0.004622809791714378 |
+| Epoch_15_batch_2999.pt | 0.8844999999999998 | 0.0044655747698019105 |
+| Epoch_14_batch_5999.pt | 0.8836666666666666 | 0.0055154105093292724 |
+| Epoch_13.pt | 0.8835 | 0.004965622559901967 |
+| Epoch_13_batch_2999.pt | 0.8835 | 0.004703885969055426 |
+| Epoch_17_batch_5999.pt | 0.8831666666666667 | 0.004377622670687092 |
+| Epoch_12_batch_2999.pt | 0.8825 | 0.005617707893376856 |
+| Epoch_13_batch_5999.pt | 0.8818333333333334 | 0.005722222222222222 |
+| Epoch_16.pt | 0.8818333333333334 | 0.004677566635509283 |
+| Epoch_14.pt | 0.8815 | 0.0056023033182018836 |
+| Epoch_15.pt | 0.8813333333333334 | 0.004955356249106172 |
+| Epoch_12_batch_5999.pt | 0.8805 | 0.005443594037892865 |
+| Epoch_11_batch_5999.pt | 0.8796666666666667 | 0.005315943872846971 |
+| Epoch_12.pt | 0.8796666666666665 | 0.0045119867787215395 |
+| Epoch_11.pt | 0.8786666666666667 | 0.004936635531449571 |
+| Epoch_10.pt | 0.8785000000000001 | 0.00465109837021135 |
+| Epoch_11_batch_2999.pt | 0.8781666666666668 | 0.004577521567803227 |
+| Epoch_10_batch_2999.pt | 0.876 | 0.004375859703892269 |
+| Epoch_10_batch_5999.pt | 0.8758333333333332 | 0.005573581865803604 |
+| Epoch_9_batch_5999.pt | 0.8578333333333333 | 0.0053115876122834475 |
+| Epoch_9_batch_2999.pt | 0.8526666666666667 | 0.006522515609528517 |
+| Epoch_8_batch_2999.pt | 0.8496666666666666 | 0.004841946348777979 |
+| Epoch_7_batch_5999.pt | 0.8451666666666666 | 0.004530077807088045 |
+| Epoch_6_batch_5999.pt | 0.8445 | 0.007248030809625737 |
+| Epoch_4_batch_5999.pt | 0.8413333333333334 | 0.007070194786636901 |
+| Epoch_7_batch_2999.pt | 0.8393333333333333 | 0.0061824123303304696 |
+| Epoch_8_batch_5999.pt | 0.8391666666666667 | 0.0062274283775512725 |
+| Epoch_9.pt | 0.8381666666666666 | 0.003986474044646718 |
+| Epoch_5_batch_2999.pt | 0.8370000000000001 | 0.006807766080906841 |
+| Epoch_6_batch_2999.pt | 0.8348333333333333 | 0.0043493294502332976 |
+| Epoch_4_batch_2999.pt | 0.8328333333333333 | 0.006206577554358889 |
+| Epoch_5.pt | 0.8301666666666666 | 0.006302311941063307 |
+| Epoch_4.pt | 0.8298333333333334 | 0.006336500182900745 |
+| Epoch_6.pt | 0.8283333333333334 | 0.005746711351549225 |
+| Epoch_5_batch_5999.pt | 0.8283333333333334 | 0.00569275042553311 |
+| Epoch_3.pt | 0.8258333333333334 | 0.0055235187389253235 |
+| Epoch_8.pt | 0.8256666666666665 | 0.0067868791351446455 |
+| Epoch_7.pt | 0.8245000000000001 | 0.006554849302822761 |
+| Epoch_3_batch_2999.pt | 0.8243333333333334 | 0.00799768485019027 |
+| Epoch_3_batch_5999.pt | 0.8193333333333334 | 0.005726266100766724 |
+| Epoch_2_batch_5999.pt | 0.8168333333333335 | 0.005738380637922078 |
+| Epoch_2_batch_2999.pt | 0.8146666666666667 | 0.0063089198073681685 |
+| Epoch_2.pt | 0.7996666666666666 | 0.005788451678947156 |
+| Epoch_1_batch_5999.pt | 0.7956666666666667 | 0.0048317366455908695 |
+| Epoch_1.pt | 0.7776666666666666 | 0.00639347661369591 |
+| Epoch_1_batch_2999.pt | 0.7581666666666667 | 0.006471447259279004 |
+| Epoch_0.pt | 0.7028333333333332 | 0.005905897868201724 |
+| Epoch_0_batch_5999.pt | 0.7021666666666667 | 0.00636856671279407 |
+| Epoch_0_batch_2999.pt | 0.6210000000000001 | 0.008047697316547477 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..661f912477becb9158263e521e0795d01aca6912
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,57 @@
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.8801666666666665 | 0.004248819734444195 |
+| Epoch_16_batch_2999.pt | 0.8779999999999999 | 0.004653420353638202 |
+| Epoch_14_batch_5999.pt | 0.8775000000000001 | 0.0038908725099762597 |
+| Epoch_17_batch_5999.pt | 0.8771666666666667 | 0.004635144863025806 |
+| Epoch_15.pt | 0.8768333333333335 | 0.004953175811263309 |
+| Epoch_17_batch_2999.pt | 0.8765000000000001 | 0.004160736520745062 |
+| Epoch_15_batch_2999.pt | 0.8761666666666666 | 0.0035836563161933607 |
+| Epoch_12_batch_2999.pt | 0.8756666666666666 | 0.005056471223547888 |
+| Epoch_11_batch_5999.pt | 0.8755 | 0.0038574122970755527 |
+| Epoch_13_batch_2999.pt | 0.8753333333333334 | 0.00437303747870098 |
+| Epoch_16_batch_5999.pt | 0.8753333333333334 | 0.004552844721899529 |
+| Epoch_15_batch_5999.pt | 0.8751666666666666 | 0.004447568346583431 |
+| Epoch_13_batch_5999.pt | 0.8745 | 0.0035228530805528203 |
+| Epoch_13.pt | 0.8741666666666668 | 0.004596361953749566 |
+| Epoch_10_batch_5999.pt | 0.8741666666666668 | 0.0031549490810001577 |
+| Epoch_12_batch_5999.pt | 0.8736666666666668 | 0.003440858348267117 |
+| Epoch_17.pt | 0.8728333333333333 | 0.0047012606662311185 |
+| Epoch_11.pt | 0.8718333333333333 | 0.003955383891406126 |
+| Epoch_14.pt | 0.8718333333333333 | 0.004996603784843862 |
+| Epoch_14_batch_2999.pt | 0.8709999999999999 | 0.004417976744904675 |
+| Epoch_10.pt | 0.8706666666666667 | 0.004361730316975672 |
+| Epoch_11_batch_2999.pt | 0.8698333333333335 | 0.0038204291659898184 |
+| Epoch_10_batch_2999.pt | 0.8676666666666666 | 0.004121608220220315 |
+| Epoch_12.pt | 0.8655000000000002 | 0.003991116679035955 |
+| Epoch_9_batch_5999.pt | 0.849 | 0.004863570806275395 |
+| Epoch_8_batch_2999.pt | 0.8474999999999999 | 0.004844813951249549 |
+| Epoch_9_batch_2999.pt | 0.8471666666666667 | 0.0032303537244681383 |
+| Epoch_6_batch_5999.pt | 0.8446666666666667 | 0.00385541146064388 |
+| Epoch_4_batch_5999.pt | 0.8445 | 0.004759979769642659 |
+| Epoch_8.pt | 0.8425 | 0.005742681870148159 |
+| Epoch_5_batch_5999.pt | 0.8401666666666667 | 0.004509591971906815 |
+| Epoch_8_batch_5999.pt | 0.8398333333333332 | 0.004801298692624985 |
+| Epoch_4_batch_2999.pt | 0.8386666666666667 | 0.003218388523991166 |
+| Epoch_7_batch_2999.pt | 0.8383333333333333 | 0.005773502691896265 |
+| Epoch_6_batch_2999.pt | 0.8380000000000001 | 0.0038554114606438802 |
+| Epoch_6.pt | 0.8380000000000001 | 0.006175419214040115 |
+| Epoch_5_batch_2999.pt | 0.8346666666666668 | 0.0043871304494221405 |
+| Epoch_7_batch_5999.pt | 0.8341666666666667 | 0.004181455237263045 |
+| Epoch_3_batch_2999.pt | 0.8300000000000001 | 0.0031720227608044898 |
+| Epoch_3.pt | 0.8288333333333332 | 0.0050126383482031215 |
+| Epoch_4.pt | 0.8286666666666667 | 0.004930379496074179 |
+| Epoch_3_batch_5999.pt | 0.825 | 0.005085685551709239 |
+| Epoch_2_batch_2999.pt | 0.8246666666666667 | 0.0039031484600556194 |
+| Epoch_2_batch_5999.pt | 0.8236666666666667 | 0.0039346513799168375 |
+| Epoch_9.pt | 0.8233333333333335 | 0.004303314829119351 |
+| Epoch_7.pt | 0.8216666666666667 | 0.004164814403109187 |
+| Epoch_2.pt | 0.8099999999999999 | 0.005097808776424204 |
+| Epoch_5.pt | 0.8035 | 0.00479486608162738 |
+| Epoch_1.pt | 0.7941666666666667 | 0.006257094738606828 |
+| Epoch_1_batch_5999.pt | 0.7921666666666666 | 0.0047729303574985185 |
+| Epoch_1_batch_2999.pt | 0.7901666666666666 | 0.005342873411966292 |
+| Epoch_0.pt | 0.7403333333333333 | 0.005362190449783839 |
+| Epoch_0_batch_5999.pt | 0.7344999999999999 | 0.0041725883845834194 |
+| Epoch_0_batch_2999.pt | 0.6815 | 0.0070072713556599085 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..49ffb997aec3b5089eed91009926c5e6aa036ea4
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9570000000000001 | 0.00356422554052122 |
+| Epoch_16_batch_2999.pt | 0.9558333333333333 | 0.0036111111111111153 |
+| Epoch_14_batch_5999.pt | 0.9558333333333332 | 0.003326381639982827 |
+| Epoch_17_batch_2999.pt | 0.9551666666666667 | 0.0035000000000000005 |
+| Epoch_17.pt | 0.9551666666666666 | 0.00387656778320434 |
+| Epoch_14.pt | 0.9548333333333334 | 0.0031861002130977017 |
+| Epoch_17_batch_5999.pt | 0.9546666666666667 | 0.003546864377669419 |
+| Epoch_15_batch_5999.pt | 0.9545000000000001 | 0.0041577682760150425 |
+| Epoch_13_batch_2999.pt | 0.9545 | 0.003833333333333336 |
+| Epoch_15_batch_2999.pt | 0.9543333333333333 | 0.0035676876351116303 |
+| Epoch_13.pt | 0.9538333333333332 | 0.003735465660892057 |
+| Epoch_13_batch_5999.pt | 0.9533333333333334 | 0.0036767538017276236 |
+| Epoch_14_batch_2999.pt | 0.9531666666666666 | 0.0034645470728153086 |
+| Epoch_15.pt | 0.9523333333333334 | 0.0037531879453454498 |
+| Epoch_12_batch_2999.pt | 0.9523333333333331 | 0.0033993463423951887 |
+| Epoch_16.pt | 0.9521666666666666 | 0.003702268240343592 |
+| Epoch_11_batch_5999.pt | 0.9521666666666666 | 0.004127968441835539 |
+| Epoch_12_batch_5999.pt | 0.9516666666666665 | 0.003591828861165471 |
+| Epoch_12.pt | 0.9515 | 0.0034911705207477744 |
+| Epoch_11_batch_2999.pt | 0.9513333333333331 | 0.00288461221905493 |
+| Epoch_11.pt | 0.95 | 0.0034066021592790915 |
+| Epoch_10_batch_5999.pt | 0.9496666666666667 | 0.003879353388910243 |
+| Epoch_10.pt | 0.9488333333333335 | 0.0034964708839021266 |
+| Epoch_10_batch_2999.pt | 0.9481666666666667 | 0.003621352997273837 |
+| Epoch_8_batch_5999.pt | 0.9299999999999999 | 0.004201704533597566 |
+| Epoch_9_batch_5999.pt | 0.9295 | 0.004423910900359653 |
+| Epoch_9.pt | 0.9289999999999999 | 0.003154459903684086 |
+| Epoch_7_batch_5999.pt | 0.9263333333333333 | 0.0028087165910587845 |
+| Epoch_8_batch_2999.pt | 0.9256666666666666 | 0.005672110674711205 |
+| Epoch_9_batch_2999.pt | 0.9248333333333333 | 0.0038526085439079573 |
+| Epoch_6_batch_5999.pt | 0.9244999999999999 | 0.004245913067618998 |
+| Epoch_8.pt | 0.9225 | 0.002952818281315182 |
+| Epoch_5_batch_5999.pt | 0.9206666666666667 | 0.0039999999999999975 |
+| Epoch_6_batch_2999.pt | 0.9203333333333333 | 0.003208784239598597 |
+| Epoch_7_batch_2999.pt | 0.9196666666666667 | 0.004096068575814838 |
+| Epoch_5_batch_2999.pt | 0.9141666666666668 | 0.0031155721984022153 |
+| Epoch_4.pt | 0.9126666666666667 | 0.0015947444549341489 |
+| Epoch_6.pt | 0.9120000000000001 | 0.0028414915227876533 |
+| Epoch_4_batch_2999.pt | 0.9114999999999999 | 0.003787976429717954 |
+| Epoch_4_batch_5999.pt | 0.9106666666666667 | 0.004812535072823938 |
+| Epoch_7.pt | 0.9095000000000001 | 0.0029085866917129815 |
+| Epoch_3_batch_2999.pt | 0.9081666666666666 | 0.0029022128636908826 |
+| Epoch_3_batch_5999.pt | 0.9065000000000001 | 0.003057575090181559 |
+| Epoch_2_batch_2999.pt | 0.9025000000000001 | 0.0031353224576716543 |
+| Epoch_2_batch_5999.pt | 0.9010000000000001 | 0.003014368881389011 |
+| Epoch_3.pt | 0.8998333333333333 | 0.0016933056282364609 |
+| Epoch_5.pt | 0.8995000000000001 | 0.003668770440244237 |
+| Epoch_2.pt | 0.8978333333333332 | 0.003960062974932603 |
+| Epoch_1_batch_5999.pt | 0.8835000000000001 | 0.003621352997273838 |
+| Epoch_1.pt | 0.8811666666666668 | 0.0033059056770692987 |
+| Epoch_1_batch_2999.pt | 0.8671666666666666 | 0.004339383241150778 |
+| Epoch_0.pt | 0.8361666666666666 | 0.004289306254879478 |
+| Epoch_0_batch_5999.pt | 0.8266666666666668 | 0.006211299937499414 |
+| Epoch_0_batch_2999.pt | 0.7578333333333334 | 0.005931970297037765 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b2ac0c099ba21fef0489cd0b0d150457751a9e66
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.9085000000000001 | 0.003852608543907959 |
+| Epoch_16.pt | 0.9075 | 0.003930334701294229 |
+| Epoch_17.pt | 0.9066666666666666 | 0.003397529966982369 |
+| Epoch_17_batch_2999.pt | 0.9066666666666666 | 0.003659926464889028 |
+| Epoch_13_batch_5999.pt | 0.9065 | 0.003446683859832462 |
+| Epoch_12_batch_5999.pt | 0.9061666666666668 | 0.00407529443002074 |
+| Epoch_13.pt | 0.906 | 0.004046031434674032 |
+| Epoch_16_batch_5999.pt | 0.9059999999999999 | 0.003761402417735721 |
+| Epoch_12_batch_2999.pt | 0.9058333333333334 | 0.0037781862524265083 |
+| Epoch_17_batch_5999.pt | 0.9056666666666666 | 0.0036784323016104117 |
+| Epoch_15_batch_2999.pt | 0.9053333333333334 | 0.0034676636742949378 |
+| Epoch_16_batch_2999.pt | 0.9053333333333333 | 0.003839367231815781 |
+| Epoch_10.pt | 0.9045 | 0.003549039167485427 |
+| Epoch_14_batch_5999.pt | 0.9043333333333334 | 0.00375318794534545 |
+| Epoch_13_batch_2999.pt | 0.9041666666666666 | 0.0033724556023947074 |
+| Epoch_15.pt | 0.9033333333333333 | 0.003583225665910463 |
+| Epoch_14_batch_2999.pt | 0.9030000000000001 | 0.00391104797928845 |
+| Epoch_10_batch_2999.pt | 0.9024999999999999 | 0.003193840522623292 |
+| Epoch_14.pt | 0.9024999999999999 | 0.003859012219291616 |
+| Epoch_12.pt | 0.9021666666666667 | 0.004668319813010156 |
+| Epoch_10_batch_5999.pt | 0.9019999999999999 | 0.004192880503136267 |
+| Epoch_11_batch_5999.pt | 0.8996666666666668 | 0.004004626953542241 |
+| Epoch_11_batch_2999.pt | 0.899 | 0.0038984010650775177 |
+| Epoch_11.pt | 0.8986666666666666 | 0.004344714399114921 |
+| Epoch_9.pt | 0.8821666666666668 | 0.0052531589555567075 |
+| Epoch_9_batch_5999.pt | 0.8811666666666668 | 0.004981756842176049 |
+| Epoch_8_batch_2999.pt | 0.8796666666666665 | 0.004103596736137634 |
+| Epoch_7_batch_5999.pt | 0.8793333333333333 | 0.0050442486501405155 |
+| Epoch_7_batch_2999.pt | 0.8791666666666668 | 0.005710343657903436 |
+| Epoch_9_batch_2999.pt | 0.8789999999999999 | 0.003610683735393763 |
+| Epoch_6_batch_2999.pt | 0.8786666666666667 | 0.004148478822798768 |
+| Epoch_5_batch_5999.pt | 0.8785000000000001 | 0.004890466917040397 |
+| Epoch_6_batch_5999.pt | 0.8768333333333335 | 0.005196449405098938 |
+| Epoch_8_batch_5999.pt | 0.8756666666666666 | 0.003636237371545237 |
+| Epoch_8.pt | 0.8733333333333333 | 0.004694362260950578 |
+| Epoch_6.pt | 0.8720000000000001 | 0.005162782291328417 |
+| Epoch_4_batch_2999.pt | 0.8716666666666667 | 0.00477260702109212 |
+| Epoch_5_batch_2999.pt | 0.8705 | 0.00606574067007016 |
+| Epoch_4_batch_5999.pt | 0.8680000000000001 | 0.004816381511487389 |
+| Epoch_4.pt | 0.867 | 0.0035294178165041355 |
+| Epoch_5.pt | 0.8666666666666666 | 0.005594309277855158 |
+| Epoch_3_batch_5999.pt | 0.8621666666666666 | 0.005079916884709384 |
+| Epoch_2_batch_5999.pt | 0.8618333333333335 | 0.0044890126495590555 |
+| Epoch_3_batch_2999.pt | 0.8616666666666667 | 0.004296136650929155 |
+| Epoch_3.pt | 0.8536666666666666 | 0.004294699575575041 |
+| Epoch_2_batch_2999.pt | 0.8508333333333333 | 0.0049705925324307915 |
+| Epoch_2.pt | 0.8498333333333333 | 0.006238322424070969 |
+| Epoch_7.pt | 0.8488333333333333 | 0.00534633831078181 |
+| Epoch_1_batch_5999.pt | 0.8373333333333333 | 0.004528374193847736 |
+| Epoch_1.pt | 0.8318333333333333 | 0.005377421934967231 |
+| Epoch_1_batch_2999.pt | 0.8140000000000001 | 0.005495228008390329 |
+| Epoch_0_batch_5999.pt | 0.7881666666666666 | 0.004370566536714872 |
+| Epoch_0.pt | 0.7849999999999999 | 0.00481125224324688 |
+| Epoch_0_batch_2999.pt | 0.7068333333333332 | 0.005278947238855226 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/MobileFaceNet/log.log b/bob/bio/facexzoo/models/backbones/MobileFaceNet/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..d38e0a8f66c6c904fb3541adc812d096a3bddfdb
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/MobileFaceNet/log.log
@@ -0,0 +1,657 @@
+INFO 2020-11-24 17:02:43 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2020-11-24 17:02:43 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2020-11-24 17:02:43 train.py: 172] Start optimization.
+INFO 2020-11-24 17:02:43 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/home/wangjun492/wj_data/facex-zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='mv-softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-mobile', train_file='/home/wangjun492/wj_data/facex-zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7fa027e56b00>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-11-24 17:03:10 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 16.358618
+INFO 2020-11-24 17:04:32 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 15.827444
+INFO 2020-11-24 17:05:55 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 15.327547
+INFO 2020-11-24 17:07:17 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 15.027858
+INFO 2020-11-24 17:08:39 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 14.677421
+INFO 2020-11-24 17:10:01 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 14.355350
+INFO 2020-11-24 17:11:23 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.027037
+INFO 2020-11-24 17:12:45 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 13.690001
+INFO 2020-11-24 17:14:07 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 13.346130
+INFO 2020-11-24 17:15:29 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 13.024926
+INFO 2020-11-24 17:16:51 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 12.701884
+INFO 2020-11-24 17:18:13 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 12.395277
+INFO 2020-11-24 17:19:35 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 12.130202
+INFO 2020-11-24 17:20:57 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 11.987308
+INFO 2020-11-24 17:22:18 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 11.955337
+INFO 2020-11-24 17:23:39 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-24 17:23:39 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.067932
+INFO 2020-11-24 17:25:00 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.314615
+INFO 2020-11-24 17:26:21 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 12.647381
+INFO 2020-11-24 17:27:42 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 13.030897
+INFO 2020-11-24 17:29:02 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 13.441236
+INFO 2020-11-24 17:30:22 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 13.796649
+INFO 2020-11-24 17:31:41 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 14.154292
+INFO 2020-11-24 17:33:00 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 14.411892
+INFO 2020-11-24 17:34:19 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 14.619047
+INFO 2020-11-24 17:35:38 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 14.832919
+INFO 2020-11-24 17:36:56 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 14.915812
+INFO 2020-11-24 17:38:14 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 14.976967
+INFO 2020-11-24 17:39:32 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 14.973729
+INFO 2020-11-24 17:40:50 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 14.912079
+INFO 2020-11-24 17:42:08 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 14.853437
+INFO 2020-11-24 17:43:25 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-24 17:43:26 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 14.786114
+INFO 2020-11-24 17:44:43 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 14.640276
+INFO 2020-11-24 17:46:00 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 14.499856
+INFO 2020-11-24 17:46:07 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-24 17:46:08 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 14.468562
+INFO 2020-11-24 17:47:26 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 14.190097
+INFO 2020-11-24 17:48:43 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 14.034737
+INFO 2020-11-24 17:50:00 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 13.867991
+INFO 2020-11-24 17:51:17 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 13.711440
+INFO 2020-11-24 17:52:35 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 13.581579
+INFO 2020-11-24 17:53:51 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 13.396562
+INFO 2020-11-24 17:55:08 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 13.189663
+INFO 2020-11-24 17:56:25 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 13.039445
+INFO 2020-11-24 17:57:42 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 12.903281
+INFO 2020-11-24 17:58:59 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 12.729310
+INFO 2020-11-24 18:00:15 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 12.598813
+INFO 2020-11-24 18:01:32 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 12.446454
+INFO 2020-11-24 18:02:49 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 12.294402
+INFO 2020-11-24 18:04:06 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 12.151741
+INFO 2020-11-24 18:05:22 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-24 18:05:23 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 12.006855
+INFO 2020-11-24 18:06:39 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 11.907829
+INFO 2020-11-24 18:07:55 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 11.759257
+INFO 2020-11-24 18:09:11 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 11.656780
+INFO 2020-11-24 18:10:27 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 11.537830
+INFO 2020-11-24 18:11:43 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 11.460143
+INFO 2020-11-24 18:13:00 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 11.327951
+INFO 2020-11-24 18:14:16 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 11.268308
+INFO 2020-11-24 18:15:32 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 11.173050
+INFO 2020-11-24 18:16:48 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 11.074287
+INFO 2020-11-24 18:18:04 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 10.989137
+INFO 2020-11-24 18:19:21 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 10.935530
+INFO 2020-11-24 18:20:37 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 10.858660
+INFO 2020-11-24 18:21:53 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 10.765018
+INFO 2020-11-24 18:23:10 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 10.696050
+INFO 2020-11-24 18:24:26 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-24 18:24:26 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 10.650471
+INFO 2020-11-24 18:25:43 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 10.606789
+INFO 2020-11-24 18:27:00 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 10.507497
+INFO 2020-11-24 18:27:06 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-24 18:27:08 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 10.401783
+INFO 2020-11-24 18:28:25 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 9.958792
+INFO 2020-11-24 18:29:42 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 9.970688
+INFO 2020-11-24 18:30:58 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 10.018418
+INFO 2020-11-24 18:32:15 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 10.043185
+INFO 2020-11-24 18:33:32 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 10.026329
+INFO 2020-11-24 18:34:48 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 10.008143
+INFO 2020-11-24 18:36:05 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 10.019662
+INFO 2020-11-24 18:37:22 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 9.995654
+INFO 2020-11-24 18:38:38 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 9.985731
+INFO 2020-11-24 18:39:55 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 9.926832
+INFO 2020-11-24 18:41:11 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 9.920456
+INFO 2020-11-24 18:42:28 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 9.864497
+INFO 2020-11-24 18:43:45 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 9.836358
+INFO 2020-11-24 18:45:01 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 9.801993
+INFO 2020-11-24 18:46:18 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-24 18:46:18 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 9.784279
+INFO 2020-11-24 18:47:35 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 9.777114
+INFO 2020-11-24 18:48:51 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 9.716017
+INFO 2020-11-24 18:50:08 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 9.685481
+INFO 2020-11-24 18:51:25 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 9.638493
+INFO 2020-11-24 18:52:42 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 9.615591
+INFO 2020-11-24 18:53:58 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 9.573484
+INFO 2020-11-24 18:55:15 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 9.557758
+INFO 2020-11-24 18:56:32 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 9.530840
+INFO 2020-11-24 18:57:49 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 9.470006
+INFO 2020-11-24 18:59:06 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 9.479366
+INFO 2020-11-24 19:00:23 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 9.434593
+INFO 2020-11-24 19:01:40 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 9.401267
+INFO 2020-11-24 19:02:56 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 9.366447
+INFO 2020-11-24 19:04:13 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 9.358310
+INFO 2020-11-24 19:05:30 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-24 19:05:31 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 9.291458
+INFO 2020-11-24 19:06:47 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 9.317040
+INFO 2020-11-24 19:08:03 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 9.263851
+INFO 2020-11-24 19:08:09 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-24 19:08:11 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 9.158338
+INFO 2020-11-24 19:09:27 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 8.738770
+INFO 2020-11-24 19:10:43 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 8.747253
+INFO 2020-11-24 19:12:00 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 8.834588
+INFO 2020-11-24 19:13:16 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 8.904287
+INFO 2020-11-24 19:14:33 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 8.929899
+INFO 2020-11-24 19:15:50 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 8.972742
+INFO 2020-11-24 19:17:06 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 8.967419
+INFO 2020-11-24 19:18:23 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 8.950048
+INFO 2020-11-24 19:19:39 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 9.018943
+INFO 2020-11-24 19:20:56 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 8.952868
+INFO 2020-11-24 19:22:12 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 8.942784
+INFO 2020-11-24 19:23:29 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 8.956252
+INFO 2020-11-24 19:24:45 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 8.937395
+INFO 2020-11-24 19:26:02 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 8.942570
+INFO 2020-11-24 19:27:18 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-24 19:27:19 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 8.916036
+INFO 2020-11-24 19:28:35 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 8.912676
+INFO 2020-11-24 19:29:52 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 8.899460
+INFO 2020-11-24 19:31:09 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 8.846656
+INFO 2020-11-24 19:32:25 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 8.852470
+INFO 2020-11-24 19:33:42 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 8.813011
+INFO 2020-11-24 19:34:59 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 8.826520
+INFO 2020-11-24 19:36:15 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 8.782351
+INFO 2020-11-24 19:37:32 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 8.789090
+INFO 2020-11-24 19:38:49 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 8.784844
+INFO 2020-11-24 19:40:06 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 8.752145
+INFO 2020-11-24 19:41:22 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 8.755990
+INFO 2020-11-24 19:42:39 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 8.748737
+INFO 2020-11-24 19:43:56 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 8.742908
+INFO 2020-11-24 19:45:13 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 8.702594
+INFO 2020-11-24 19:46:30 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-24 19:46:30 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 8.683531
+INFO 2020-11-24 19:47:47 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 8.677554
+INFO 2020-11-24 19:49:04 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 8.687340
+INFO 2020-11-24 19:49:10 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-24 19:49:12 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 8.645050
+INFO 2020-11-24 19:50:29 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 8.138176
+INFO 2020-11-24 19:51:45 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 8.133244
+INFO 2020-11-24 19:53:02 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 8.266590
+INFO 2020-11-24 19:54:18 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 8.264567
+INFO 2020-11-24 19:55:35 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 8.324405
+INFO 2020-11-24 19:56:52 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 8.387397
+INFO 2020-11-24 19:58:08 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 8.381465
+INFO 2020-11-24 19:59:25 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 8.419101
+INFO 2020-11-24 20:00:41 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 8.418929
+INFO 2020-11-24 20:01:58 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 8.439405
+INFO 2020-11-24 20:03:14 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 8.440718
+INFO 2020-11-24 20:04:31 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 8.468213
+INFO 2020-11-24 20:05:47 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 8.412492
+INFO 2020-11-24 20:07:04 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 8.401329
+INFO 2020-11-24 20:08:20 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-24 20:08:20 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 8.462322
+INFO 2020-11-24 20:09:36 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 8.428143
+INFO 2020-11-24 20:10:52 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 8.421280
+INFO 2020-11-24 20:12:09 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 8.391416
+INFO 2020-11-24 20:13:25 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 8.353408
+INFO 2020-11-24 20:14:42 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 8.400384
+INFO 2020-11-24 20:15:59 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 8.367021
+INFO 2020-11-24 20:17:15 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 8.395899
+INFO 2020-11-24 20:18:32 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 8.344757
+INFO 2020-11-24 20:19:49 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 8.359471
+INFO 2020-11-24 20:21:06 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 8.358082
+INFO 2020-11-24 20:22:22 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 8.314536
+INFO 2020-11-24 20:23:39 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 8.286721
+INFO 2020-11-24 20:24:56 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 8.331915
+INFO 2020-11-24 20:26:13 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 8.285538
+INFO 2020-11-24 20:27:29 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-24 20:27:30 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 8.278215
+INFO 2020-11-24 20:28:47 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 8.267316
+INFO 2020-11-24 20:30:04 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 8.247428
+INFO 2020-11-24 20:30:10 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-24 20:30:12 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 8.371536
+INFO 2020-11-24 20:31:28 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 7.767822
+INFO 2020-11-24 20:32:45 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 7.743762
+INFO 2020-11-24 20:34:02 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 7.862978
+INFO 2020-11-24 20:35:18 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 7.930466
+INFO 2020-11-24 20:36:35 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 7.993809
+INFO 2020-11-24 20:37:51 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 8.014129
+INFO 2020-11-24 20:39:08 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 8.067885
+INFO 2020-11-24 20:40:24 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 8.064815
+INFO 2020-11-24 20:41:41 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 8.070696
+INFO 2020-11-24 20:42:57 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 8.081757
+INFO 2020-11-24 20:44:14 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 8.069307
+INFO 2020-11-24 20:45:30 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 8.108222
+INFO 2020-11-24 20:46:47 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 8.123253
+INFO 2020-11-24 20:48:03 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 8.108594
+INFO 2020-11-24 20:49:20 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-24 20:49:20 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 8.108960
+INFO 2020-11-24 20:50:36 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 8.088804
+INFO 2020-11-24 20:51:52 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 8.077876
+INFO 2020-11-24 20:53:08 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 8.086027
+INFO 2020-11-24 20:54:24 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 8.072146
+INFO 2020-11-24 20:55:40 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 8.084796
+INFO 2020-11-24 20:56:56 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 8.068487
+INFO 2020-11-24 20:58:12 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 8.052808
+INFO 2020-11-24 20:59:28 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 8.054618
+INFO 2020-11-24 21:00:44 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 8.083125
+INFO 2020-11-24 21:02:00 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 8.037382
+INFO 2020-11-24 21:03:16 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 8.025006
+INFO 2020-11-24 21:04:32 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 8.020393
+INFO 2020-11-24 21:05:48 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 7.998549
+INFO 2020-11-24 21:07:04 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 7.991002
+INFO 2020-11-24 21:08:20 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-24 21:08:21 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 8.000411
+INFO 2020-11-24 21:09:38 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 7.964763
+INFO 2020-11-24 21:10:55 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 7.980003
+INFO 2020-11-24 21:11:01 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-24 21:11:03 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 7.949017
+INFO 2020-11-24 21:12:19 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 7.518950
+INFO 2020-11-24 21:13:36 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 7.480380
+INFO 2020-11-24 21:14:53 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 7.594087
+INFO 2020-11-24 21:16:09 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 7.652602
+INFO 2020-11-24 21:17:26 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 7.707925
+INFO 2020-11-24 21:18:42 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 7.781615
+INFO 2020-11-24 21:19:59 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 7.807427
+INFO 2020-11-24 21:21:15 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 7.810504
+INFO 2020-11-24 21:22:32 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 7.841157
+INFO 2020-11-24 21:23:48 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 7.837509
+INFO 2020-11-24 21:25:05 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 7.858138
+INFO 2020-11-24 21:26:21 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 7.860692
+INFO 2020-11-24 21:27:38 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 7.850410
+INFO 2020-11-24 21:28:54 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 7.864597
+INFO 2020-11-24 21:30:10 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-24 21:30:11 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 7.829093
+INFO 2020-11-24 21:31:28 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 7.873271
+INFO 2020-11-24 21:32:44 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 7.859303
+INFO 2020-11-24 21:34:01 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 7.833756
+INFO 2020-11-24 21:35:17 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 7.868141
+INFO 2020-11-24 21:36:34 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 7.832145
+INFO 2020-11-24 21:37:51 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 7.827492
+INFO 2020-11-24 21:39:07 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 7.822211
+INFO 2020-11-24 21:40:24 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 7.803100
+INFO 2020-11-24 21:41:41 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 7.839389
+INFO 2020-11-24 21:42:57 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 7.828123
+INFO 2020-11-24 21:44:14 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 7.807350
+INFO 2020-11-24 21:45:31 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 7.840542
+INFO 2020-11-24 21:46:48 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 7.790327
+INFO 2020-11-24 21:48:05 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 7.800599
+INFO 2020-11-24 21:49:21 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-24 21:49:22 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 7.794503
+INFO 2020-11-24 21:50:38 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 7.811972
+INFO 2020-11-24 21:51:55 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 7.816250
+INFO 2020-11-24 21:52:01 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-24 21:52:03 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 7.721017
+INFO 2020-11-24 21:53:20 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 7.263689
+INFO 2020-11-24 21:54:37 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 7.302127
+INFO 2020-11-24 21:55:53 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 7.388004
+INFO 2020-11-24 21:57:10 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 7.476136
+INFO 2020-11-24 21:58:26 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 7.525854
+INFO 2020-11-24 21:59:43 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 7.569437
+INFO 2020-11-24 22:00:59 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 7.606572
+INFO 2020-11-24 22:02:16 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 7.646732
+INFO 2020-11-24 22:03:32 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 7.663091
+INFO 2020-11-24 22:04:49 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 7.641256
+INFO 2020-11-24 22:06:05 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 7.667531
+INFO 2020-11-24 22:07:21 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 7.688095
+INFO 2020-11-24 22:08:38 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 7.702703
+INFO 2020-11-24 22:09:54 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 7.673153
+INFO 2020-11-24 22:11:11 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-24 22:11:11 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 7.691864
+INFO 2020-11-24 22:12:28 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 7.671905
+INFO 2020-11-24 22:13:44 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 7.709454
+INFO 2020-11-24 22:15:01 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 7.666515
+INFO 2020-11-24 22:16:18 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 7.645696
+INFO 2020-11-24 22:17:34 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 7.697766
+INFO 2020-11-24 22:18:51 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 7.661309
+INFO 2020-11-24 22:20:08 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 7.704096
+INFO 2020-11-24 22:21:24 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 7.703111
+INFO 2020-11-24 22:22:41 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 7.686914
+INFO 2020-11-24 22:23:58 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 7.652473
+INFO 2020-11-24 22:25:15 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 7.657636
+INFO 2020-11-24 22:26:31 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 7.659919
+INFO 2020-11-24 22:27:48 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 7.633489
+INFO 2020-11-24 22:29:05 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 7.652294
+INFO 2020-11-24 22:30:22 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-24 22:30:22 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 7.629084
+INFO 2020-11-24 22:31:39 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 7.587239
+INFO 2020-11-24 22:32:56 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 7.620722
+INFO 2020-11-24 22:33:02 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-24 22:33:04 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 7.599243
+INFO 2020-11-24 22:34:21 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 7.169748
+INFO 2020-11-24 22:35:37 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 7.161929
+INFO 2020-11-24 22:36:54 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 7.237029
+INFO 2020-11-24 22:38:10 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 7.324340
+INFO 2020-11-24 22:39:27 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 7.391021
+INFO 2020-11-24 22:40:43 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 7.419794
+INFO 2020-11-24 22:42:00 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 7.462434
+INFO 2020-11-24 22:43:16 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 7.477895
+INFO 2020-11-24 22:44:33 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 7.508406
+INFO 2020-11-24 22:45:49 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 7.522996
+INFO 2020-11-24 22:47:06 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 7.525577
+INFO 2020-11-24 22:48:22 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 7.525456
+INFO 2020-11-24 22:49:38 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 7.531138
+INFO 2020-11-24 22:50:55 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 7.542735
+INFO 2020-11-24 22:52:11 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-24 22:52:12 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 7.531534
+INFO 2020-11-24 22:53:28 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 7.529461
+INFO 2020-11-24 22:54:45 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 7.560661
+INFO 2020-11-24 22:56:01 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 7.550087
+INFO 2020-11-24 22:57:18 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 7.508024
+INFO 2020-11-24 22:58:35 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 7.549892
+INFO 2020-11-24 22:59:51 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 7.559926
+INFO 2020-11-24 23:01:08 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 7.509007
+INFO 2020-11-24 23:02:25 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 7.568892
+INFO 2020-11-24 23:03:41 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 7.538015
+INFO 2020-11-24 23:04:58 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 7.519274
+INFO 2020-11-24 23:06:15 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 7.505941
+INFO 2020-11-24 23:07:32 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 7.555411
+INFO 2020-11-24 23:08:49 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 7.507525
+INFO 2020-11-24 23:10:05 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 7.528209
+INFO 2020-11-24 23:11:22 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-24 23:11:22 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 7.515858
+INFO 2020-11-24 23:12:39 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 7.534646
+INFO 2020-11-24 23:13:56 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 7.508126
+INFO 2020-11-24 23:14:02 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-24 23:14:04 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 7.443227
+INFO 2020-11-24 23:15:20 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 7.009048
+INFO 2020-11-24 23:16:36 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 7.035446
+INFO 2020-11-24 23:17:52 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 7.069120
+INFO 2020-11-24 23:19:08 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 7.197210
+INFO 2020-11-24 23:20:24 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 7.304525
+INFO 2020-11-24 23:21:40 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 7.299744
+INFO 2020-11-24 23:22:56 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 7.324403
+INFO 2020-11-24 23:24:11 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 7.391143
+INFO 2020-11-24 23:25:27 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 7.417595
+INFO 2020-11-24 23:26:43 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 7.439957
+INFO 2020-11-24 23:27:59 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 7.406044
+INFO 2020-11-24 23:29:15 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 7.424606
+INFO 2020-11-24 23:30:31 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 7.398502
+INFO 2020-11-24 23:31:47 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 7.423678
+INFO 2020-11-24 23:33:02 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-24 23:33:03 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 7.436243
+INFO 2020-11-24 23:34:19 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 7.473204
+INFO 2020-11-24 23:35:36 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 7.443741
+INFO 2020-11-24 23:36:53 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 7.441603
+INFO 2020-11-24 23:38:09 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 7.417943
+INFO 2020-11-24 23:39:26 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 7.423790
+INFO 2020-11-24 23:40:42 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 7.446940
+INFO 2020-11-24 23:41:59 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 7.406042
+INFO 2020-11-24 23:43:16 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 7.427912
+INFO 2020-11-24 23:44:33 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 7.437229
+INFO 2020-11-24 23:45:49 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 7.434042
+INFO 2020-11-24 23:47:06 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 7.401678
+INFO 2020-11-24 23:48:23 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 7.439736
+INFO 2020-11-24 23:49:40 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 7.417320
+INFO 2020-11-24 23:50:56 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 7.431844
+INFO 2020-11-24 23:52:13 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-24 23:52:14 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 7.436903
+INFO 2020-11-24 23:53:30 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 7.398268
+INFO 2020-11-24 23:54:47 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 7.397418
+INFO 2020-11-24 23:54:53 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-24 23:54:55 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 7.353138
+INFO 2020-11-24 23:56:12 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 6.257766
+INFO 2020-11-24 23:57:29 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 6.017529
+INFO 2020-11-24 23:58:45 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 5.915038
+INFO 2020-11-25 00:00:02 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 5.836611
+INFO 2020-11-25 00:01:18 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 5.783414
+INFO 2020-11-25 00:02:34 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 5.734762
+INFO 2020-11-25 00:03:51 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 5.738091
+INFO 2020-11-25 00:05:07 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 5.682976
+INFO 2020-11-25 00:06:23 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 5.647589
+INFO 2020-11-25 00:07:40 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 5.622531
+INFO 2020-11-25 00:08:56 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 5.601900
+INFO 2020-11-25 00:10:12 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 5.571118
+INFO 2020-11-25 00:11:29 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 5.534947
+INFO 2020-11-25 00:12:45 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 5.509576
+INFO 2020-11-25 00:14:01 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-25 00:14:02 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 5.511195
+INFO 2020-11-25 00:15:17 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 5.482364
+INFO 2020-11-25 00:16:33 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 5.489383
+INFO 2020-11-25 00:17:49 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 5.448496
+INFO 2020-11-25 00:19:05 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 5.419645
+INFO 2020-11-25 00:20:21 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 5.423076
+INFO 2020-11-25 00:21:37 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 5.412503
+INFO 2020-11-25 00:22:53 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 5.375420
+INFO 2020-11-25 00:24:08 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 5.357363
+INFO 2020-11-25 00:25:24 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 5.346682
+INFO 2020-11-25 00:26:40 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 5.338702
+INFO 2020-11-25 00:27:56 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 5.338754
+INFO 2020-11-25 00:29:12 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 5.340596
+INFO 2020-11-25 00:30:28 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 5.277478
+INFO 2020-11-25 00:31:44 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 5.315326
+INFO 2020-11-25 00:33:00 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-25 00:33:01 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 5.293450
+INFO 2020-11-25 00:34:17 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 5.272774
+INFO 2020-11-25 00:35:34 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 5.283323
+INFO 2020-11-25 00:35:40 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-25 00:35:42 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 5.223241
+INFO 2020-11-25 00:36:58 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 4.948623
+INFO 2020-11-25 00:38:15 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 4.906858
+INFO 2020-11-25 00:39:31 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 4.907461
+INFO 2020-11-25 00:40:48 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 4.904690
+INFO 2020-11-25 00:42:04 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 4.928326
+INFO 2020-11-25 00:43:21 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 4.924131
+INFO 2020-11-25 00:44:37 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 4.936636
+INFO 2020-11-25 00:45:53 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 4.918629
+INFO 2020-11-25 00:47:09 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 4.919805
+INFO 2020-11-25 00:48:26 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 4.960497
+INFO 2020-11-25 00:49:42 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 4.957431
+INFO 2020-11-25 00:50:58 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 4.956089
+INFO 2020-11-25 00:52:15 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 4.945728
+INFO 2020-11-25 00:53:31 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 4.957325
+INFO 2020-11-25 00:54:47 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-25 00:54:47 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 4.980482
+INFO 2020-11-25 00:56:04 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 4.983182
+INFO 2020-11-25 00:57:20 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 4.966438
+INFO 2020-11-25 00:58:37 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 4.987589
+INFO 2020-11-25 00:59:53 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 4.968704
+INFO 2020-11-25 01:01:10 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 4.980256
+INFO 2020-11-25 01:02:26 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 4.970413
+INFO 2020-11-25 01:03:43 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 5.002899
+INFO 2020-11-25 01:04:59 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 5.007192
+INFO 2020-11-25 01:06:16 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 4.998616
+INFO 2020-11-25 01:07:33 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 5.009392
+INFO 2020-11-25 01:08:49 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 4.986240
+INFO 2020-11-25 01:10:06 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 4.972138
+INFO 2020-11-25 01:11:23 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 4.976395
+INFO 2020-11-25 01:12:39 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 4.971344
+INFO 2020-11-25 01:13:56 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-25 01:13:56 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 4.987184
+INFO 2020-11-25 01:15:12 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 4.976141
+INFO 2020-11-25 01:16:28 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 5.002322
+INFO 2020-11-25 01:16:35 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-25 01:16:36 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 4.983368
+INFO 2020-11-25 01:17:53 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 4.611268
+INFO 2020-11-25 01:19:09 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 4.640412
+INFO 2020-11-25 01:20:26 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 4.673774
+INFO 2020-11-25 01:21:42 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 4.692424
+INFO 2020-11-25 01:22:59 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 4.691139
+INFO 2020-11-25 01:24:15 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 4.690612
+INFO 2020-11-25 01:25:31 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 4.695927
+INFO 2020-11-25 01:26:47 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 4.736963
+INFO 2020-11-25 01:28:04 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 4.741829
+INFO 2020-11-25 01:29:20 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 4.736865
+INFO 2020-11-25 01:30:36 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 4.762031
+INFO 2020-11-25 01:31:53 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 4.767115
+INFO 2020-11-25 01:33:09 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 4.782443
+INFO 2020-11-25 01:34:25 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 4.760673
+INFO 2020-11-25 01:35:41 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-25 01:35:42 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 4.802496
+INFO 2020-11-25 01:36:58 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 4.796419
+INFO 2020-11-25 01:38:15 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 4.792343
+INFO 2020-11-25 01:39:31 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 4.832982
+INFO 2020-11-25 01:40:48 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 4.821779
+INFO 2020-11-25 01:42:04 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 4.881370
+INFO 2020-11-25 01:43:20 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 4.884917
+INFO 2020-11-25 01:44:37 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 4.857210
+INFO 2020-11-25 01:45:53 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 4.858853
+INFO 2020-11-25 01:47:10 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 4.859920
+INFO 2020-11-25 01:48:27 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 4.852360
+INFO 2020-11-25 01:49:43 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 4.872837
+INFO 2020-11-25 01:51:00 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 4.871405
+INFO 2020-11-25 01:52:17 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 4.908140
+INFO 2020-11-25 01:53:33 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 4.893963
+INFO 2020-11-25 01:54:50 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-25 01:54:50 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 4.906605
+INFO 2020-11-25 01:56:07 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 4.916764
+INFO 2020-11-25 01:57:23 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 4.909206
+INFO 2020-11-25 01:57:30 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-25 01:57:31 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 4.876686
+INFO 2020-11-25 01:58:47 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 4.442451
+INFO 2020-11-25 02:00:03 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 4.408205
+INFO 2020-11-25 02:01:19 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 4.419447
+INFO 2020-11-25 02:02:35 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 4.387983
+INFO 2020-11-25 02:03:51 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 4.412583
+INFO 2020-11-25 02:05:06 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 4.399296
+INFO 2020-11-25 02:06:22 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 4.394198
+INFO 2020-11-25 02:07:38 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 4.380451
+INFO 2020-11-25 02:08:53 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 4.419471
+INFO 2020-11-25 02:10:09 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 4.381618
+INFO 2020-11-25 02:11:24 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 4.409688
+INFO 2020-11-25 02:12:40 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 4.408339
+INFO 2020-11-25 02:13:56 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 4.411750
+INFO 2020-11-25 02:15:11 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 4.413968
+INFO 2020-11-25 02:16:27 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-25 02:16:27 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 4.390786
+INFO 2020-11-25 02:17:44 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 4.429628
+INFO 2020-11-25 02:19:00 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 4.420643
+INFO 2020-11-25 02:20:16 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 4.388773
+INFO 2020-11-25 02:21:33 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 4.397255
+INFO 2020-11-25 02:22:49 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 4.405686
+INFO 2020-11-25 02:24:06 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 4.395911
+INFO 2020-11-25 02:25:22 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 4.432924
+INFO 2020-11-25 02:26:39 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 4.400759
+INFO 2020-11-25 02:27:55 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 4.384009
+INFO 2020-11-25 02:29:12 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 4.438662
+INFO 2020-11-25 02:30:29 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 4.423308
+INFO 2020-11-25 02:31:45 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 4.411874
+INFO 2020-11-25 02:33:02 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 4.426057
+INFO 2020-11-25 02:34:19 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 4.420076
+INFO 2020-11-25 02:35:35 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-25 02:35:35 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 4.426545
+INFO 2020-11-25 02:36:52 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 4.424432
+INFO 2020-11-25 02:38:09 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 4.414508
+INFO 2020-11-25 02:38:15 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-25 02:38:17 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 4.443413
+INFO 2020-11-25 02:39:33 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 4.337744
+INFO 2020-11-25 02:40:50 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 4.345555
+INFO 2020-11-25 02:42:06 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 4.335841
+INFO 2020-11-25 02:43:23 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 4.349915
+INFO 2020-11-25 02:44:39 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 4.377623
+INFO 2020-11-25 02:45:55 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 4.354851
+INFO 2020-11-25 02:47:12 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 4.369637
+INFO 2020-11-25 02:48:28 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 4.347162
+INFO 2020-11-25 02:49:44 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 4.375702
+INFO 2020-11-25 02:51:00 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 4.355511
+INFO 2020-11-25 02:52:17 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 4.351502
+INFO 2020-11-25 02:53:33 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 4.346628
+INFO 2020-11-25 02:54:49 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 4.385360
+INFO 2020-11-25 02:56:06 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 4.382407
+INFO 2020-11-25 02:57:22 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-25 02:57:22 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 4.368762
+INFO 2020-11-25 02:58:38 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 4.392034
+INFO 2020-11-25 02:59:54 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 4.361917
+INFO 2020-11-25 03:01:09 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 4.373660
+INFO 2020-11-25 03:02:25 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 4.387465
+INFO 2020-11-25 03:03:41 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 4.354591
+INFO 2020-11-25 03:04:57 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 4.383426
+INFO 2020-11-25 03:06:13 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 4.364166
+INFO 2020-11-25 03:07:29 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 4.376958
+INFO 2020-11-25 03:08:44 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 4.372547
+INFO 2020-11-25 03:10:00 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 4.359061
+INFO 2020-11-25 03:11:16 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 4.379103
+INFO 2020-11-25 03:12:32 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 4.410229
+INFO 2020-11-25 03:13:48 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 4.386184
+INFO 2020-11-25 03:15:04 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 4.415015
+INFO 2020-11-25 03:16:20 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-25 03:16:20 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 4.391684
+INFO 2020-11-25 03:17:37 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 4.421325
+INFO 2020-11-25 03:18:53 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 4.391710
+INFO 2020-11-25 03:19:00 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-25 03:19:01 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 4.447200
+INFO 2020-11-25 03:20:18 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 4.308515
+INFO 2020-11-25 03:21:34 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 4.339306
+INFO 2020-11-25 03:22:51 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 4.358757
+INFO 2020-11-25 03:24:07 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 4.343400
+INFO 2020-11-25 03:25:23 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 4.327294
+INFO 2020-11-25 03:26:40 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 4.339487
+INFO 2020-11-25 03:27:56 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 4.330323
+INFO 2020-11-25 03:29:12 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 4.349277
+INFO 2020-11-25 03:30:28 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 4.335158
+INFO 2020-11-25 03:31:45 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 4.341329
+INFO 2020-11-25 03:33:01 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 4.328359
+INFO 2020-11-25 03:34:17 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 4.342991
+INFO 2020-11-25 03:35:34 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 4.365566
+INFO 2020-11-25 03:36:50 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 4.348621
+INFO 2020-11-25 03:38:06 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-25 03:38:06 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 4.341549
+INFO 2020-11-25 03:39:23 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 4.353372
+INFO 2020-11-25 03:40:39 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 4.341762
+INFO 2020-11-25 03:41:56 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 4.357957
+INFO 2020-11-25 03:43:12 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 4.356401
+INFO 2020-11-25 03:44:28 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 4.337076
+INFO 2020-11-25 03:45:45 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 4.354352
+INFO 2020-11-25 03:47:01 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 4.380453
+INFO 2020-11-25 03:48:18 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 4.356988
+INFO 2020-11-25 03:49:34 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 4.371785
+INFO 2020-11-25 03:50:51 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 4.359285
+INFO 2020-11-25 03:52:07 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 4.362331
+INFO 2020-11-25 03:53:24 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 4.364020
+INFO 2020-11-25 03:54:41 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 4.396912
+INFO 2020-11-25 03:55:57 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 4.379699
+INFO 2020-11-25 03:57:14 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-25 03:57:14 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 4.372220
+INFO 2020-11-25 03:58:31 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 4.375687
+INFO 2020-11-25 03:59:47 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 4.387501
+INFO 2020-11-25 03:59:54 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-25 03:59:55 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 4.437425
+INFO 2020-11-25 04:01:12 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 4.312835
+INFO 2020-11-25 04:02:28 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 4.326099
+INFO 2020-11-25 04:03:45 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 4.280351
+INFO 2020-11-25 04:05:01 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 4.285148
+INFO 2020-11-25 04:06:18 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 4.303608
+INFO 2020-11-25 04:07:34 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 4.304034
+INFO 2020-11-25 04:08:50 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 4.293425
+INFO 2020-11-25 04:10:07 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 4.328923
+INFO 2020-11-25 04:11:23 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 4.299053
+INFO 2020-11-25 04:12:39 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 4.304958
+INFO 2020-11-25 04:13:55 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 4.297158
+INFO 2020-11-25 04:15:12 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 4.286107
+INFO 2020-11-25 04:16:28 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 4.298167
+INFO 2020-11-25 04:17:45 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 4.299586
+INFO 2020-11-25 04:19:01 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-25 04:19:01 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 4.286830
+INFO 2020-11-25 04:20:17 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 4.326733
+INFO 2020-11-25 04:21:33 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 4.311498
+INFO 2020-11-25 04:22:48 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 4.313252
+INFO 2020-11-25 04:24:04 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 4.304375
+INFO 2020-11-25 04:25:20 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 4.305652
+INFO 2020-11-25 04:26:36 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 4.332660
+INFO 2020-11-25 04:27:52 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 4.300166
+INFO 2020-11-25 04:29:08 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 4.277409
+INFO 2020-11-25 04:30:23 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 4.320323
+INFO 2020-11-25 04:31:39 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 4.266749
+INFO 2020-11-25 04:32:55 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 4.300358
+INFO 2020-11-25 04:34:11 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 4.313041
+INFO 2020-11-25 04:35:27 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 4.324255
+INFO 2020-11-25 04:36:43 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 4.296991
+INFO 2020-11-25 04:37:59 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-25 04:37:59 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 4.308512
+INFO 2020-11-25 04:39:16 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 4.318126
+INFO 2020-11-25 04:40:33 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 4.305930
+INFO 2020-11-25 04:40:39 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-25 04:40:41 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 4.295313
+INFO 2020-11-25 04:41:57 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 4.302628
+INFO 2020-11-25 04:43:14 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 4.303336
+INFO 2020-11-25 04:44:30 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 4.289142
+INFO 2020-11-25 04:45:46 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 4.317333
+INFO 2020-11-25 04:47:03 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 4.305242
+INFO 2020-11-25 04:48:19 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 4.287574
+INFO 2020-11-25 04:49:36 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 4.306370
+INFO 2020-11-25 04:50:52 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 4.315201
+INFO 2020-11-25 04:52:08 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 4.301431
+INFO 2020-11-25 04:53:24 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 4.299300
+INFO 2020-11-25 04:54:41 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 4.293645
+INFO 2020-11-25 04:55:57 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 4.291119
+INFO 2020-11-25 04:57:13 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 4.308664
+INFO 2020-11-25 04:58:29 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 4.319460
+INFO 2020-11-25 04:59:46 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-25 04:59:46 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 4.309168
+INFO 2020-11-25 05:01:02 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 4.292737
+INFO 2020-11-25 05:02:19 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 4.293627
+INFO 2020-11-25 05:03:35 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 4.306646
+INFO 2020-11-25 05:04:52 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 4.282552
+INFO 2020-11-25 05:06:08 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 4.314200
+INFO 2020-11-25 05:07:25 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 4.296638
+INFO 2020-11-25 05:08:41 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 4.324167
+INFO 2020-11-25 05:09:58 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 4.305802
+INFO 2020-11-25 05:11:14 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 4.278409
+INFO 2020-11-25 05:12:31 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 4.306388
+INFO 2020-11-25 05:13:47 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 4.309003
+INFO 2020-11-25 05:15:04 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 4.266980
+INFO 2020-11-25 05:16:21 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 4.294368
+INFO 2020-11-25 05:17:37 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 4.304988
+INFO 2020-11-25 05:18:54 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-25 05:18:54 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 4.306873
+INFO 2020-11-25 05:20:11 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 4.316005
+INFO 2020-11-25 05:21:28 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 4.300295
+INFO 2020-11-25 05:21:34 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-25 05:21:34 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/ReXNet_1/.gitkeep b/bob/bio/facexzoo/models/backbones/ReXNet_1/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_agedb30.txt b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_agedb30.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e41abdf5f5737ed2f07f1cceaf3dda2592920964
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_agedb30.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.967 | 0.002775554665954846 |
+| Epoch_13.pt | 0.967 | 0.0030510067150546602 |
+| Epoch_15_batch_5999.pt | 0.967 | 0.0026270200927859737 |
+| Epoch_14_batch_2999.pt | 0.9663333333333334 | 0.0030205060486818217 |
+| Epoch_14_batch_5999.pt | 0.9663333333333334 | 0.0031700761372638023 |
+| Epoch_14.pt | 0.9663333333333334 | 0.003265986323710909 |
+| Epoch_16_batch_5999.pt | 0.9663333333333333 | 0.0028523328117763193 |
+| Epoch_17_batch_2999.pt | 0.9661666666666665 | 0.0029715731378480784 |
+| Epoch_13_batch_5999.pt | 0.9658333333333333 | 0.0031451510788262655 |
+| Epoch_17_batch_5999.pt | 0.9656666666666667 | 0.002631715396072672 |
+| Epoch_16.pt | 0.9654999999999999 | 0.0026533230969465723 |
+| Epoch_15_batch_2999.pt | 0.9653333333333333 | 0.002591534175486803 |
+| Epoch_16_batch_2999.pt | 0.9653333333333333 | 0.002808716591058784 |
+| Epoch_17.pt | 0.9651666666666667 | 0.0026579719364234833 |
+| Epoch_11_batch_5999.pt | 0.9646666666666667 | 0.0026620330112690944 |
+| Epoch_15.pt | 0.9646666666666667 | 0.002884612219054929 |
+| Epoch_10_batch_5999.pt | 0.9641666666666667 | 0.0026323017201071394 |
+| Epoch_10.pt | 0.9641666666666667 | 0.0032417987690878175 |
+| Epoch_12_batch_5999.pt | 0.9640000000000001 | 0.003124969135650058 |
+| Epoch_12.pt | 0.9636666666666667 | 0.0035381518506868138 |
+| Epoch_11.pt | 0.9628333333333334 | 0.002778333277788882 |
+| Epoch_12_batch_2999.pt | 0.9628333333333334 | 0.003053534685258065 |
+| Epoch_11_batch_2999.pt | 0.9621666666666668 | 0.002653323096946567 |
+| Epoch_10_batch_2999.pt | 0.9616666666666667 | 0.003220305943597652 |
+| Epoch_9_batch_5999.pt | 0.9553333333333333 | 0.0037333994703136527 |
+| Epoch_9_batch_2999.pt | 0.9538333333333332 | 0.0019092044752710652 |
+| Epoch_7_batch_5999.pt | 0.9526666666666668 | 0.0035849479566448183 |
+| Epoch_8_batch_5999.pt | 0.9523333333333334 | 0.00326976421545826 |
+| Epoch_7.pt | 0.9514999999999999 | 0.0030271406057389614 |
+| Epoch_7_batch_2999.pt | 0.9511666666666667 | 0.0036519063176766236 |
+| Epoch_6_batch_2999.pt | 0.9504999999999999 | 0.002897955856832214 |
+| Epoch_6_batch_5999.pt | 0.95 | 0.002581988897471617 |
+| Epoch_8_batch_2999.pt | 0.9493333333333334 | 0.0029418227321941653 |
+| Epoch_5_batch_5999.pt | 0.9491666666666665 | 0.00341790853672712 |
+| Epoch_5_batch_2999.pt | 0.9486666666666667 | 0.003766322501484061 |
+| Epoch_8.pt | 0.9481666666666666 | 0.0026579719364234855 |
+| Epoch_4_batch_5999.pt | 0.9476666666666669 | 0.002867441755680877 |
+| Epoch_6.pt | 0.9475000000000001 | 0.0037205634778340485 |
+| Epoch_5.pt | 0.9470000000000001 | 0.0034765528549248912 |
+| Epoch_9.pt | 0.9458333333333332 | 0.003085709794888621 |
+| Epoch_4.pt | 0.9446666666666668 | 0.0031894889098682973 |
+| Epoch_4_batch_2999.pt | 0.9441666666666666 | 0.002747052292239137 |
+| Epoch_3_batch_5999.pt | 0.9413333333333334 | 0.003020506048681819 |
+| Epoch_3_batch_2999.pt | 0.9375 | 0.003550778040283289 |
+| Epoch_3.pt | 0.9346666666666668 | 0.003179797338056478 |
+| Epoch_2_batch_5999.pt | 0.9344999999999999 | 0.004289306254879465 |
+| Epoch_2.pt | 0.9323333333333332 | 0.002474561939035566 |
+| Epoch_2_batch_2999.pt | 0.9298333333333334 | 0.004040306185683716 |
+| Epoch_1_batch_5999.pt | 0.917 | 0.0036666666666666705 |
+| Epoch_1.pt | 0.9158333333333333 | 0.006727729608048495 |
+| Epoch_1_batch_2999.pt | 0.8963333333333333 | 0.003140732005578346 |
+| Epoch_0_batch_5999.pt | 0.8621666666666666 | 0.006573656742100472 |
+| Epoch_0.pt | 0.8461666666666666 | 0.0069657451100561835 |
+| Epoch_0_batch_2999.pt | 0.774 | 0.004061259307221572 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..095d7f28862a58f98e46c5278947404b79511ba0
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.9458333333333332 | 0.003363291303925624 |
+| Epoch_15.pt | 0.9453333333333334 | 0.002852332811776319 |
+| Epoch_17_batch_5999.pt | 0.9453333333333334 | 0.0032470309324894317 |
+| Epoch_15_batch_2999.pt | 0.9448333333333334 | 0.0028158501994387073 |
+| Epoch_11_batch_5999.pt | 0.9446666666666668 | 0.003275422882885256 |
+| Epoch_11.pt | 0.9446666666666665 | 0.003766322501484066 |
+| Epoch_17_batch_2999.pt | 0.9446666666666665 | 0.0032183885239911625 |
+| Epoch_15_batch_5999.pt | 0.9445 | 0.002992274002110054 |
+| Epoch_13.pt | 0.9440000000000002 | 0.0034084137000395436 |
+| Epoch_16_batch_2999.pt | 0.9440000000000002 | 0.002834966849371792 |
+| Epoch_13_batch_5999.pt | 0.9434999999999999 | 0.0031372906452212163 |
+| Epoch_16_batch_5999.pt | 0.9433333333333334 | 0.003002056908023617 |
+| Epoch_17.pt | 0.9431666666666667 | 0.0030676528205063375 |
+| Epoch_12_batch_2999.pt | 0.9430000000000002 | 0.003733399470313655 |
+| Epoch_14_batch_2999.pt | 0.943 | 0.00333148096679221 |
+| Epoch_14.pt | 0.943 | 0.0031797973380564893 |
+| Epoch_13_batch_2999.pt | 0.9428333333333334 | 0.0034609818060383317 |
+| Epoch_14_batch_5999.pt | 0.9428333333333334 | 0.0033797691498344677 |
+| Epoch_11_batch_2999.pt | 0.9425000000000001 | 0.003542075139843354 |
+| Epoch_10_batch_5999.pt | 0.9421666666666667 | 0.003277777777777781 |
+| Epoch_10.pt | 0.9411666666666667 | 0.003370624736026108 |
+| Epoch_10_batch_2999.pt | 0.9400000000000001 | 0.003539896071571994 |
+| Epoch_12.pt | 0.9398333333333333 | 0.0035175924707281573 |
+| Epoch_12_batch_5999.pt | 0.9395000000000001 | 0.0035316033500695163 |
+| Epoch_9_batch_5999.pt | 0.9366666666666668 | 0.003522414996477193 |
+| Epoch_8_batch_2999.pt | 0.9351666666666667 | 0.0034911705207477687 |
+| Epoch_8_batch_5999.pt | 0.9328333333333333 | 0.00423134978318436 |
+| Epoch_6_batch_2999.pt | 0.9328333333333332 | 0.003609401304617953 |
+| Epoch_9.pt | 0.9326666666666668 | 0.003866602809178961 |
+| Epoch_7_batch_5999.pt | 0.9321666666666666 | 0.0039051248379533216 |
+| Epoch_7_batch_2999.pt | 0.9318333333333333 | 0.004405734198423736 |
+| Epoch_5_batch_5999.pt | 0.9315 | 0.0035263557939018532 |
+| Epoch_6_batch_5999.pt | 0.9313333333333332 | 0.003950308629918041 |
+| Epoch_9_batch_2999.pt | 0.9313333333333332 | 0.003863408589047496 |
+| Epoch_8.pt | 0.9306666666666666 | 0.004487293445846374 |
+| Epoch_7.pt | 0.9303333333333335 | 0.0038713891978187373 |
+| Epoch_5_batch_2999.pt | 0.9296666666666666 | 0.005029542354558343 |
+| Epoch_4_batch_5999.pt | 0.9253333333333333 | 0.00350308506009654 |
+| Epoch_6.pt | 0.925 | 0.0034516054593353475 |
+| Epoch_3_batch_5999.pt | 0.9243333333333335 | 0.00340841370003955 |
+| Epoch_4.pt | 0.9240000000000002 | 0.004217833976911629 |
+| Epoch_4_batch_2999.pt | 0.9231666666666667 | 0.003763453234319994 |
+| Epoch_5.pt | 0.9228333333333332 | 0.004409935471659139 |
+| Epoch_3_batch_2999.pt | 0.9208333333333332 | 0.0028136571693556894 |
+| Epoch_3.pt | 0.9195 | 0.003143187813292365 |
+| Epoch_2_batch_5999.pt | 0.9173333333333333 | 0.004786813161897336 |
+| Epoch_2.pt | 0.9148333333333334 | 0.0037387691907536046 |
+| Epoch_2_batch_2999.pt | 0.9143333333333332 | 0.004166296279833929 |
+| Epoch_1_batch_5999.pt | 0.9073333333333332 | 0.003728435941236111 |
+| Epoch_1.pt | 0.905 | 0.0031426968052735466 |
+| Epoch_1_batch_2999.pt | 0.8939999999999999 | 0.006137317546507323 |
+| Epoch_0.pt | 0.8593333333333334 | 0.004254264382017035 |
+| Epoch_0_batch_5999.pt | 0.8548333333333333 | 0.0051786003161634194 |
+| Epoch_0_batch_2999.pt | 0.7398333333333332 | 0.006218500871304895 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..9b1270c5fd3b5061e4595734541f005de6ee05f3
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8468333333333332 | 0.0069007514351967746 |
+| Epoch_16.pt | 0.8456666666666667 | 0.006699917080747261 |
+| Epoch_11.pt | 0.845 | 0.006907680463275361 |
+| Epoch_13_batch_5999.pt | 0.8446666666666666 | 0.006748570950693432 |
+| Epoch_13.pt | 0.8441666666666666 | 0.007487644143174467 |
+| Epoch_12.pt | 0.843 | 0.007118052168020875 |
+| Epoch_17_batch_5999.pt | 0.8423333333333334 | 0.00727417213481463 |
+| Epoch_15.pt | 0.8421666666666667 | 0.007420567014373961 |
+| Epoch_17_batch_2999.pt | 0.842 | 0.006775956039328586 |
+| Epoch_12_batch_5999.pt | 0.8418333333333333 | 0.006042289239953619 |
+| Epoch_10_batch_5999.pt | 0.8414999999999999 | 0.0074429932250505365 |
+| Epoch_15_batch_5999.pt | 0.8413333333333334 | 0.006789607163312089 |
+| Epoch_16_batch_5999.pt | 0.8413333333333333 | 0.007242705994547676 |
+| Epoch_16_batch_2999.pt | 0.8411666666666667 | 0.007433034373659252 |
+| Epoch_15_batch_2999.pt | 0.841 | 0.006662961933584419 |
+| Epoch_14.pt | 0.8408333333333333 | 0.007015195147888832 |
+| Epoch_11_batch_5999.pt | 0.8406666666666668 | 0.007516237566794567 |
+| Epoch_10.pt | 0.8403333333333333 | 0.007169895913289633 |
+| Epoch_13_batch_2999.pt | 0.8400000000000001 | 0.007453559924999296 |
+| Epoch_14_batch_2999.pt | 0.8398333333333333 | 0.007279473932966638 |
+| Epoch_11_batch_2999.pt | 0.8390000000000001 | 0.006266706067646033 |
+| Epoch_12_batch_2999.pt | 0.8388333333333332 | 0.00796229540698837 |
+| Epoch_14_batch_5999.pt | 0.8383333333333333 | 0.007005289007176937 |
+| Epoch_10_batch_2999.pt | 0.8365 | 0.008200609733428357 |
+| Epoch_9_batch_5999.pt | 0.8188333333333333 | 0.006559556218051472 |
+| Epoch_9.pt | 0.8173333333333334 | 0.007869302759833675 |
+| Epoch_8_batch_5999.pt | 0.8151666666666667 | 0.007936670633253004 |
+| Epoch_9_batch_2999.pt | 0.8141666666666667 | 0.007610300037887246 |
+| Epoch_7_batch_5999.pt | 0.8131666666666666 | 0.008715142207887579 |
+| Epoch_7_batch_2999.pt | 0.8128333333333334 | 0.007453766965455026 |
+| Epoch_4_batch_2999.pt | 0.8128333333333332 | 0.006196623931476569 |
+| Epoch_7.pt | 0.8103333333333333 | 0.008981462390204987 |
+| Epoch_8_batch_2999.pt | 0.8098333333333334 | 0.007889866919029745 |
+| Epoch_4_batch_5999.pt | 0.8091666666666667 | 0.00833796167766868 |
+| Epoch_6_batch_5999.pt | 0.8089999999999999 | 0.008043861243439225 |
+| Epoch_8.pt | 0.8075000000000001 | 0.008492007862152758 |
+| Epoch_5_batch_2999.pt | 0.8068333333333333 | 0.007959969290839706 |
+| Epoch_3_batch_2999.pt | 0.8038333333333334 | 0.007536126981184914 |
+| Epoch_6_batch_2999.pt | 0.8035 | 0.006006426599387289 |
+| Epoch_5_batch_5999.pt | 0.8033333333333333 | 0.007286042804780001 |
+| Epoch_4.pt | 0.8024999999999999 | 0.00613756898832264 |
+| Epoch_5.pt | 0.8019999999999999 | 0.007302122133621225 |
+| Epoch_6.pt | 0.8015000000000001 | 0.008271809929282278 |
+| Epoch_3_batch_5999.pt | 0.7991666666666667 | 0.005922597444857044 |
+| Epoch_2.pt | 0.7988333333333333 | 0.007218161251444289 |
+| Epoch_3.pt | 0.7986666666666666 | 0.006299127903581504 |
+| Epoch_2_batch_5999.pt | 0.7965 | 0.007834909218946686 |
+| Epoch_1_batch_5999.pt | 0.7875 | 0.006881940940687528 |
+| Epoch_2_batch_2999.pt | 0.7836666666666667 | 0.007824463062870338 |
+| Epoch_1.pt | 0.7651666666666668 | 0.00775970121608835 |
+| Epoch_1_batch_2999.pt | 0.7588333333333332 | 0.005895436611439915 |
+| Epoch_0.pt | 0.7324999999999999 | 0.010491766319063114 |
+| Epoch_0_batch_5999.pt | 0.726 | 0.00955232503531192 |
+| Epoch_0_batch_2999.pt | 0.6325000000000001 | 0.009972957261675687 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..75f22e0200e6973523c01943e74e4bc663bda56d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9964999999999999 | 0.0008766518798921926 |
+| Epoch_15_batch_5999.pt | 0.9964999999999999 | 0.0009444444444444436 |
+| Epoch_17.pt | 0.9964999999999999 | 0.0009444444444444436 |
+| Epoch_11_batch_5999.pt | 0.9963333333333333 | 0.000922958206990897 |
+| Epoch_13_batch_5999.pt | 0.9963333333333333 | 0.0009229582069908971 |
+| Epoch_17_batch_2999.pt | 0.9963333333333333 | 0.000922958206990897 |
+| Epoch_10_batch_5999.pt | 0.9959999999999999 | 0.0009362388636862643 |
+| Epoch_11.pt | 0.9959999999999999 | 0.0010599324460188312 |
+| Epoch_12_batch_2999.pt | 0.9959999999999999 | 0.0010304020550550763 |
+| Epoch_12_batch_5999.pt | 0.9959999999999999 | 0.0010000000000000022 |
+| Epoch_14_batch_5999.pt | 0.9959999999999999 | 0.0009026709338484409 |
+| Epoch_17_batch_5999.pt | 0.9959999999999999 | 0.0009026709338484409 |
+| Epoch_16.pt | 0.9958333333333333 | 0.0010318986456114817 |
+| Epoch_14.pt | 0.9956666666666667 | 0.0010304020550550804 |
+| Epoch_15_batch_2999.pt | 0.9956666666666667 | 0.0010599324460188332 |
+| Epoch_16_batch_2999.pt | 0.9956666666666667 | 0.001143958904554112 |
+| Epoch_16_batch_5999.pt | 0.9956666666666667 | 0.0010599324460188332 |
+| Epoch_10.pt | 0.9956666666666665 | 0.001143958904554112 |
+| Epoch_11_batch_2999.pt | 0.9956666666666665 | 0.0010886621079036361 |
+| Epoch_14_batch_2999.pt | 0.9955 | 0.0009312808119022318 |
+| Epoch_15.pt | 0.9955 | 0.0010258991840344167 |
+| Epoch_12.pt | 0.9954999999999998 | 0.0010844011831079516 |
+| Epoch_10_batch_2999.pt | 0.9953333333333333 | 0.0009558139185602972 |
+| Epoch_9_batch_2999.pt | 0.9951666666666668 | 0.0009765775461803873 |
+| Epoch_6_batch_5999.pt | 0.9951666666666666 | 0.0011235415786753737 |
+| Epoch_8_batch_5999.pt | 0.9951666666666666 | 0.0012031337682059829 |
+| Epoch_13.pt | 0.9950000000000001 | 0.001290994448735802 |
+| Epoch_5_batch_5999.pt | 0.9949999999999999 | 0.001024393828588094 |
+| Epoch_7_batch_5999.pt | 0.9946666666666667 | 0.0009558139185602963 |
+| Epoch_9_batch_5999.pt | 0.9945000000000002 | 0.0011666666666666696 |
+| Epoch_8.pt | 0.9945 | 0.0009953596037316037 |
+| Epoch_4_batch_5999.pt | 0.9943333333333333 | 0.0007535922203472535 |
+| Epoch_4.pt | 0.9943333333333333 | 0.0011706281947614185 |
+| Epoch_7.pt | 0.9943333333333332 | 0.001515353521887323 |
+| Epoch_3_batch_2999.pt | 0.9941666666666666 | 0.0011453071182271274 |
+| Epoch_7_batch_2999.pt | 0.9941666666666664 | 0.001001542020962221 |
+| Epoch_3_batch_5999.pt | 0.9940000000000001 | 0.001295767087743398 |
+| Epoch_6_batch_2999.pt | 0.994 | 0.001143958904554111 |
+| Epoch_5.pt | 0.9938333333333332 | 0.0011399046960379573 |
+| Epoch_6.pt | 0.9935 | 0.0010957268290731142 |
+| Epoch_8_batch_2999.pt | 0.9935 | 0.0010957268290731142 |
+| Epoch_5_batch_2999.pt | 0.9934999999999998 | 0.0009444444444444439 |
+| Epoch_4_batch_2999.pt | 0.993 | 0.0010482201257840634 |
+| Epoch_9.pt | 0.993 | 0.0011863420280034786 |
+| Epoch_2_batch_2999.pt | 0.9928333333333332 | 0.0011124991330278184 |
+| Epoch_3.pt | 0.9928333333333332 | 0.0016111111111111122 |
+| Epoch_2_batch_5999.pt | 0.9926666666666666 | 0.0011706281947614146 |
+| Epoch_1_batch_5999.pt | 0.9918333333333335 | 0.0019317042945237479 |
+| Epoch_1.pt | 0.9913333333333334 | 0.0010772621905369556 |
+| Epoch_2.pt | 0.9908333333333333 | 0.0013888888888888935 |
+| Epoch_1_batch_2999.pt | 0.9868333333333332 | 0.0013482956777235134 |
+| Epoch_0_batch_5999.pt | 0.9765 | 0.001979057014506314 |
+| Epoch_0.pt | 0.976 | 0.0017950549357114982 |
+| Epoch_0_batch_2999.pt | 0.9386666666666666 | 0.004653420353638203 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..caecfb0058691100f4bfec7ac0a101239e894428
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ReXNet_1/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.9316817891872893 |
+| 2 | 0.9479932827368942 |
+| 3 | 0.954515276566385 |
+| 4 | 0.9583750992618821 |
+| 5 | 0.9611544319616752 |
+| 6 | 0.9630615618938516 |
+| 7 | 0.9649296379707617 |
+| 8 | 0.9661923792910424 |
+| 9 | 0.9673835218766679 |
+| 10 | 0.968392413137717 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ReXNet_1/log.log b/bob/bio/facexzoo/models/backbones/ReXNet_1/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..b8ae09dbdd9fbe23a87977295bff816022e11c27
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ReXNet_1/log.log
@@ -0,0 +1,655 @@
+INFO 2021-03-18 11:39:26 train.py: 176] Start optimization.
+INFO 2021-03-18 11:39:26 train.py: 177] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='ReXNet', batch_size=512, data_root='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-rexnet', train_file='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f346e0b2198>)
+backbone param:
+{'input_ch': 16, 'final_ch': 180, 'width_mult': 1.0, 'depth_mult': 1.0, 'use_se': 0, 'se_ratio': 12, 'out_h': 7, 'out_w': 7, 'feat_dim': 512, 'dropout_ratio': 0.2}
+head param:
+{'feat_dim': 512, 'num_class': 360232, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2021-03-18 11:40:12 train.py: 78] Epoch 0, iter 0/6416, lr 0.100000, loss 17.872231
+INFO 2021-03-18 11:52:14 train.py: 78] Epoch 0, iter 200/6416, lr 0.100000, loss 17.330293
+INFO 2021-03-18 12:01:20 train.py: 78] Epoch 0, iter 400/6416, lr 0.100000, loss 16.307696
+INFO 2021-03-18 12:09:15 train.py: 78] Epoch 0, iter 600/6416, lr 0.100000, loss 15.892715
+INFO 2021-03-18 12:16:20 train.py: 78] Epoch 0, iter 800/6416, lr 0.100000, loss 15.584927
+INFO 2021-03-18 12:23:23 train.py: 78] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.252976
+INFO 2021-03-18 12:30:58 train.py: 78] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.878432
+INFO 2021-03-18 12:37:59 train.py: 78] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.490505
+INFO 2021-03-18 12:45:01 train.py: 78] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.062121
+INFO 2021-03-18 12:52:13 train.py: 78] Epoch 0, iter 1800/6416, lr 0.100000, loss 13.671984
+INFO 2021-03-18 12:59:31 train.py: 78] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.298539
+INFO 2021-03-18 13:06:15 train.py: 78] Epoch 0, iter 2200/6416, lr 0.100000, loss 12.974956
+INFO 2021-03-18 13:13:22 train.py: 78] Epoch 0, iter 2400/6416, lr 0.100000, loss 12.781531
+INFO 2021-03-18 13:20:10 train.py: 78] Epoch 0, iter 2600/6416, lr 0.100000, loss 12.746571
+INFO 2021-03-18 13:27:18 train.py: 78] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.873792
+INFO 2021-03-18 13:33:50 train.py: 91] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-03-18 13:33:52 train.py: 78] Epoch 0, iter 3000/6416, lr 0.100000, loss 13.168525
+INFO 2021-03-18 13:40:36 train.py: 78] Epoch 0, iter 3200/6416, lr 0.100000, loss 13.498302
+INFO 2021-03-18 13:47:26 train.py: 78] Epoch 0, iter 3400/6416, lr 0.100000, loss 13.925669
+INFO 2021-03-18 13:54:27 train.py: 78] Epoch 0, iter 3600/6416, lr 0.100000, loss 14.302528
+INFO 2021-03-18 14:02:03 train.py: 78] Epoch 0, iter 3800/6416, lr 0.100000, loss 14.625612
+INFO 2021-03-18 14:08:35 train.py: 78] Epoch 0, iter 4000/6416, lr 0.100000, loss 14.907151
+INFO 2021-03-18 14:15:22 train.py: 78] Epoch 0, iter 4200/6416, lr 0.100000, loss 15.126410
+INFO 2021-03-18 14:22:07 train.py: 78] Epoch 0, iter 4400/6416, lr 0.100000, loss 15.264146
+INFO 2021-03-18 14:28:41 train.py: 78] Epoch 0, iter 4600/6416, lr 0.100000, loss 15.343200
+INFO 2021-03-18 14:35:23 train.py: 78] Epoch 0, iter 4800/6416, lr 0.100000, loss 15.363958
+INFO 2021-03-18 14:41:51 train.py: 78] Epoch 0, iter 5000/6416, lr 0.100000, loss 15.306645
+INFO 2021-03-18 14:48:22 train.py: 78] Epoch 0, iter 5200/6416, lr 0.100000, loss 15.231745
+INFO 2021-03-18 14:54:53 train.py: 78] Epoch 0, iter 5400/6416, lr 0.100000, loss 15.087625
+INFO 2021-03-18 15:01:30 train.py: 78] Epoch 0, iter 5600/6416, lr 0.100000, loss 14.953604
+INFO 2021-03-18 15:08:05 train.py: 78] Epoch 0, iter 5800/6416, lr 0.100000, loss 14.763418
+INFO 2021-03-18 15:14:32 train.py: 91] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-03-18 15:14:33 train.py: 78] Epoch 0, iter 6000/6416, lr 0.100000, loss 14.569331
+INFO 2021-03-18 15:21:25 train.py: 78] Epoch 0, iter 6200/6416, lr 0.100000, loss 14.341868
+INFO 2021-03-18 15:27:53 train.py: 78] Epoch 0, iter 6400/6416, lr 0.100000, loss 14.107172
+INFO 2021-03-18 15:28:21 train.py: 96] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-03-18 15:28:24 train.py: 78] Epoch 1, iter 0/6416, lr 0.100000, loss 13.995048
+INFO 2021-03-18 15:32:21 train.py: 78] Epoch 1, iter 200/6416, lr 0.100000, loss 13.674845
+INFO 2021-03-18 15:36:17 train.py: 78] Epoch 1, iter 400/6416, lr 0.100000, loss 13.393356
+INFO 2021-03-18 15:40:14 train.py: 78] Epoch 1, iter 600/6416, lr 0.100000, loss 13.224054
+INFO 2021-03-18 15:44:10 train.py: 78] Epoch 1, iter 800/6416, lr 0.100000, loss 13.035218
+INFO 2021-03-18 15:48:07 train.py: 78] Epoch 1, iter 1000/6416, lr 0.100000, loss 12.846864
+INFO 2021-03-18 15:52:03 train.py: 78] Epoch 1, iter 1200/6416, lr 0.100000, loss 12.629327
+INFO 2021-03-18 15:55:59 train.py: 78] Epoch 1, iter 1400/6416, lr 0.100000, loss 12.444055
+INFO 2021-03-18 15:59:55 train.py: 78] Epoch 1, iter 1600/6416, lr 0.100000, loss 12.247348
+INFO 2021-03-18 16:03:50 train.py: 78] Epoch 1, iter 1800/6416, lr 0.100000, loss 12.076169
+INFO 2021-03-18 16:07:46 train.py: 78] Epoch 1, iter 2000/6416, lr 0.100000, loss 11.897474
+INFO 2021-03-18 16:11:42 train.py: 78] Epoch 1, iter 2200/6416, lr 0.100000, loss 11.723761
+INFO 2021-03-18 16:15:37 train.py: 78] Epoch 1, iter 2400/6416, lr 0.100000, loss 11.576442
+INFO 2021-03-18 16:19:33 train.py: 78] Epoch 1, iter 2600/6416, lr 0.100000, loss 11.432501
+INFO 2021-03-18 16:23:28 train.py: 78] Epoch 1, iter 2800/6416, lr 0.100000, loss 11.260628
+INFO 2021-03-18 16:27:24 train.py: 91] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-03-18 16:27:25 train.py: 78] Epoch 1, iter 3000/6416, lr 0.100000, loss 11.128048
+INFO 2021-03-18 16:31:20 train.py: 78] Epoch 1, iter 3200/6416, lr 0.100000, loss 10.988125
+INFO 2021-03-18 16:35:16 train.py: 78] Epoch 1, iter 3400/6416, lr 0.100000, loss 10.874905
+INFO 2021-03-18 16:39:11 train.py: 78] Epoch 1, iter 3600/6416, lr 0.100000, loss 10.751579
+INFO 2021-03-18 16:43:06 train.py: 78] Epoch 1, iter 3800/6416, lr 0.100000, loss 10.639918
+INFO 2021-03-18 16:47:02 train.py: 78] Epoch 1, iter 4000/6416, lr 0.100000, loss 10.527626
+INFO 2021-03-18 16:50:57 train.py: 78] Epoch 1, iter 4200/6416, lr 0.100000, loss 10.454046
+INFO 2021-03-18 16:54:52 train.py: 78] Epoch 1, iter 4400/6416, lr 0.100000, loss 10.343908
+INFO 2021-03-18 16:58:48 train.py: 78] Epoch 1, iter 4600/6416, lr 0.100000, loss 10.226095
+INFO 2021-03-18 17:02:43 train.py: 78] Epoch 1, iter 4800/6416, lr 0.100000, loss 10.154646
+INFO 2021-03-18 17:06:38 train.py: 78] Epoch 1, iter 5000/6416, lr 0.100000, loss 10.052671
+INFO 2021-03-18 17:10:33 train.py: 78] Epoch 1, iter 5200/6416, lr 0.100000, loss 9.942473
+INFO 2021-03-18 17:14:29 train.py: 78] Epoch 1, iter 5400/6416, lr 0.100000, loss 9.867441
+INFO 2021-03-18 17:18:24 train.py: 78] Epoch 1, iter 5600/6416, lr 0.100000, loss 9.797150
+INFO 2021-03-18 17:22:19 train.py: 78] Epoch 1, iter 5800/6416, lr 0.100000, loss 9.728185
+INFO 2021-03-18 17:26:14 train.py: 91] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-03-18 17:26:15 train.py: 78] Epoch 1, iter 6000/6416, lr 0.100000, loss 9.683334
+INFO 2021-03-18 17:30:10 train.py: 78] Epoch 1, iter 6200/6416, lr 0.100000, loss 9.602051
+INFO 2021-03-18 17:34:05 train.py: 78] Epoch 1, iter 6400/6416, lr 0.100000, loss 9.534742
+INFO 2021-03-18 17:34:24 train.py: 96] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-03-18 17:34:26 train.py: 78] Epoch 2, iter 0/6416, lr 0.100000, loss 9.490457
+INFO 2021-03-18 17:38:21 train.py: 78] Epoch 2, iter 200/6416, lr 0.100000, loss 8.940673
+INFO 2021-03-18 17:42:16 train.py: 78] Epoch 2, iter 400/6416, lr 0.100000, loss 8.956057
+INFO 2021-03-18 17:46:11 train.py: 78] Epoch 2, iter 600/6416, lr 0.100000, loss 8.996033
+INFO 2021-03-18 17:50:06 train.py: 78] Epoch 2, iter 800/6416, lr 0.100000, loss 9.022861
+INFO 2021-03-18 17:54:01 train.py: 78] Epoch 2, iter 1000/6416, lr 0.100000, loss 9.022265
+INFO 2021-03-18 17:57:56 train.py: 78] Epoch 2, iter 1200/6416, lr 0.100000, loss 9.001086
+INFO 2021-03-18 18:01:51 train.py: 78] Epoch 2, iter 1400/6416, lr 0.100000, loss 8.966400
+INFO 2021-03-18 18:05:46 train.py: 78] Epoch 2, iter 1600/6416, lr 0.100000, loss 8.965331
+INFO 2021-03-18 18:09:41 train.py: 78] Epoch 2, iter 1800/6416, lr 0.100000, loss 8.966231
+INFO 2021-03-18 18:13:36 train.py: 78] Epoch 2, iter 2000/6416, lr 0.100000, loss 8.892111
+INFO 2021-03-18 18:17:31 train.py: 78] Epoch 2, iter 2200/6416, lr 0.100000, loss 8.859752
+INFO 2021-03-18 18:21:26 train.py: 78] Epoch 2, iter 2400/6416, lr 0.100000, loss 8.835646
+INFO 2021-03-18 18:25:21 train.py: 78] Epoch 2, iter 2600/6416, lr 0.100000, loss 8.792579
+INFO 2021-03-18 18:29:16 train.py: 78] Epoch 2, iter 2800/6416, lr 0.100000, loss 8.774919
+INFO 2021-03-18 18:33:10 train.py: 91] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-03-18 18:33:12 train.py: 78] Epoch 2, iter 3000/6416, lr 0.100000, loss 8.719857
+INFO 2021-03-18 18:37:07 train.py: 78] Epoch 2, iter 3200/6416, lr 0.100000, loss 8.699471
+INFO 2021-03-18 18:41:02 train.py: 78] Epoch 2, iter 3400/6416, lr 0.100000, loss 8.645844
+INFO 2021-03-18 18:44:57 train.py: 78] Epoch 2, iter 3600/6416, lr 0.100000, loss 8.623665
+INFO 2021-03-18 18:48:52 train.py: 78] Epoch 2, iter 3800/6416, lr 0.100000, loss 8.595744
+INFO 2021-03-18 18:52:47 train.py: 78] Epoch 2, iter 4000/6416, lr 0.100000, loss 8.580572
+INFO 2021-03-18 18:56:42 train.py: 78] Epoch 2, iter 4200/6416, lr 0.100000, loss 8.537626
+INFO 2021-03-18 19:00:37 train.py: 78] Epoch 2, iter 4400/6416, lr 0.100000, loss 8.494640
+INFO 2021-03-18 19:04:32 train.py: 78] Epoch 2, iter 4600/6416, lr 0.100000, loss 8.476546
+INFO 2021-03-18 19:08:27 train.py: 78] Epoch 2, iter 4800/6416, lr 0.100000, loss 8.440517
+INFO 2021-03-18 19:12:22 train.py: 78] Epoch 2, iter 5000/6416, lr 0.100000, loss 8.398935
+INFO 2021-03-18 19:16:17 train.py: 78] Epoch 2, iter 5200/6416, lr 0.100000, loss 8.372356
+INFO 2021-03-18 19:20:12 train.py: 78] Epoch 2, iter 5400/6416, lr 0.100000, loss 8.310381
+INFO 2021-03-18 19:24:07 train.py: 78] Epoch 2, iter 5600/6416, lr 0.100000, loss 8.307603
+INFO 2021-03-18 19:28:02 train.py: 78] Epoch 2, iter 5800/6416, lr 0.100000, loss 8.299147
+INFO 2021-03-18 19:31:57 train.py: 91] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-03-18 19:31:58 train.py: 78] Epoch 2, iter 6000/6416, lr 0.100000, loss 8.296972
+INFO 2021-03-18 19:35:53 train.py: 78] Epoch 2, iter 6200/6416, lr 0.100000, loss 8.231998
+INFO 2021-03-18 19:39:48 train.py: 78] Epoch 2, iter 6400/6416, lr 0.100000, loss 8.213281
+INFO 2021-03-18 19:40:06 train.py: 96] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-03-18 19:40:09 train.py: 78] Epoch 3, iter 0/6416, lr 0.100000, loss 8.167140
+INFO 2021-03-18 19:44:04 train.py: 78] Epoch 3, iter 200/6416, lr 0.100000, loss 7.664231
+INFO 2021-03-18 19:47:58 train.py: 78] Epoch 3, iter 400/6416, lr 0.100000, loss 7.704569
+INFO 2021-03-18 19:51:53 train.py: 78] Epoch 3, iter 600/6416, lr 0.100000, loss 7.746656
+INFO 2021-03-18 19:55:48 train.py: 78] Epoch 3, iter 800/6416, lr 0.100000, loss 7.816969
+INFO 2021-03-18 19:59:43 train.py: 78] Epoch 3, iter 1000/6416, lr 0.100000, loss 7.838074
+INFO 2021-03-18 20:03:38 train.py: 78] Epoch 3, iter 1200/6416, lr 0.100000, loss 7.895765
+INFO 2021-03-18 20:07:32 train.py: 78] Epoch 3, iter 1400/6416, lr 0.100000, loss 7.918070
+INFO 2021-03-18 20:11:27 train.py: 78] Epoch 3, iter 1600/6416, lr 0.100000, loss 7.914514
+INFO 2021-03-18 20:15:22 train.py: 78] Epoch 3, iter 1800/6416, lr 0.100000, loss 7.889445
+INFO 2021-03-18 20:19:17 train.py: 78] Epoch 3, iter 2000/6416, lr 0.100000, loss 7.896461
+INFO 2021-03-18 20:23:12 train.py: 78] Epoch 3, iter 2200/6416, lr 0.100000, loss 7.905736
+INFO 2021-03-18 20:27:07 train.py: 78] Epoch 3, iter 2400/6416, lr 0.100000, loss 7.851352
+INFO 2021-03-18 20:31:02 train.py: 78] Epoch 3, iter 2600/6416, lr 0.100000, loss 7.853050
+INFO 2021-03-18 20:34:57 train.py: 78] Epoch 3, iter 2800/6416, lr 0.100000, loss 7.843649
+INFO 2021-03-18 20:38:51 train.py: 91] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-03-18 20:38:53 train.py: 78] Epoch 3, iter 3000/6416, lr 0.100000, loss 7.829684
+INFO 2021-03-18 20:42:48 train.py: 78] Epoch 3, iter 3200/6416, lr 0.100000, loss 7.818173
+INFO 2021-03-18 20:46:42 train.py: 78] Epoch 3, iter 3400/6416, lr 0.100000, loss 7.805731
+INFO 2021-03-18 20:50:37 train.py: 78] Epoch 3, iter 3600/6416, lr 0.100000, loss 7.795118
+INFO 2021-03-18 20:54:32 train.py: 78] Epoch 3, iter 3800/6416, lr 0.100000, loss 7.778923
+INFO 2021-03-18 20:58:27 train.py: 78] Epoch 3, iter 4000/6416, lr 0.100000, loss 7.767253
+INFO 2021-03-18 21:02:22 train.py: 78] Epoch 3, iter 4200/6416, lr 0.100000, loss 7.701257
+INFO 2021-03-18 21:06:17 train.py: 78] Epoch 3, iter 4400/6416, lr 0.100000, loss 7.735630
+INFO 2021-03-18 21:10:12 train.py: 78] Epoch 3, iter 4600/6416, lr 0.100000, loss 7.724208
+INFO 2021-03-18 21:14:07 train.py: 78] Epoch 3, iter 4800/6416, lr 0.100000, loss 7.650901
+INFO 2021-03-18 21:18:02 train.py: 78] Epoch 3, iter 5000/6416, lr 0.100000, loss 7.685625
+INFO 2021-03-18 21:21:57 train.py: 78] Epoch 3, iter 5200/6416, lr 0.100000, loss 7.654489
+INFO 2021-03-18 21:25:52 train.py: 78] Epoch 3, iter 5400/6416, lr 0.100000, loss 7.639582
+INFO 2021-03-18 21:29:47 train.py: 78] Epoch 3, iter 5600/6416, lr 0.100000, loss 7.627606
+INFO 2021-03-18 21:33:42 train.py: 78] Epoch 3, iter 5800/6416, lr 0.100000, loss 7.633270
+INFO 2021-03-18 21:37:36 train.py: 91] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-03-18 21:37:37 train.py: 78] Epoch 3, iter 6000/6416, lr 0.100000, loss 7.614877
+INFO 2021-03-18 21:41:32 train.py: 78] Epoch 3, iter 6200/6416, lr 0.100000, loss 7.603628
+INFO 2021-03-18 21:45:27 train.py: 78] Epoch 3, iter 6400/6416, lr 0.100000, loss 7.583062
+INFO 2021-03-18 21:45:45 train.py: 96] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-03-18 21:45:48 train.py: 78] Epoch 4, iter 0/6416, lr 0.100000, loss 7.698452
+INFO 2021-03-18 21:49:42 train.py: 78] Epoch 4, iter 200/6416, lr 0.100000, loss 7.083843
+INFO 2021-03-18 21:53:37 train.py: 78] Epoch 4, iter 400/6416, lr 0.100000, loss 7.060181
+INFO 2021-03-18 21:57:32 train.py: 78] Epoch 4, iter 600/6416, lr 0.100000, loss 7.098379
+INFO 2021-03-18 22:01:26 train.py: 78] Epoch 4, iter 800/6416, lr 0.100000, loss 7.210135
+INFO 2021-03-18 22:05:21 train.py: 78] Epoch 4, iter 1000/6416, lr 0.100000, loss 7.270411
+INFO 2021-03-18 22:09:16 train.py: 78] Epoch 4, iter 1200/6416, lr 0.100000, loss 7.281436
+INFO 2021-03-18 22:13:11 train.py: 78] Epoch 4, iter 1400/6416, lr 0.100000, loss 7.315881
+INFO 2021-03-18 22:17:06 train.py: 78] Epoch 4, iter 1600/6416, lr 0.100000, loss 7.334917
+INFO 2021-03-18 22:21:00 train.py: 78] Epoch 4, iter 1800/6416, lr 0.100000, loss 7.330238
+INFO 2021-03-18 22:24:55 train.py: 78] Epoch 4, iter 2000/6416, lr 0.100000, loss 7.331642
+INFO 2021-03-18 22:28:50 train.py: 78] Epoch 4, iter 2200/6416, lr 0.100000, loss 7.333812
+INFO 2021-03-18 22:32:45 train.py: 78] Epoch 4, iter 2400/6416, lr 0.100000, loss 7.326923
+INFO 2021-03-18 22:36:40 train.py: 78] Epoch 4, iter 2600/6416, lr 0.100000, loss 7.342534
+INFO 2021-03-18 22:40:35 train.py: 78] Epoch 4, iter 2800/6416, lr 0.100000, loss 7.337175
+INFO 2021-03-18 22:44:29 train.py: 91] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-03-18 22:44:30 train.py: 78] Epoch 4, iter 3000/6416, lr 0.100000, loss 7.328304
+INFO 2021-03-18 22:48:25 train.py: 78] Epoch 4, iter 3200/6416, lr 0.100000, loss 7.337928
+INFO 2021-03-18 22:52:20 train.py: 78] Epoch 4, iter 3400/6416, lr 0.100000, loss 7.301091
+INFO 2021-03-18 22:56:15 train.py: 78] Epoch 4, iter 3600/6416, lr 0.100000, loss 7.319077
+INFO 2021-03-18 23:00:10 train.py: 78] Epoch 4, iter 3800/6416, lr 0.100000, loss 7.280443
+INFO 2021-03-18 23:04:05 train.py: 78] Epoch 4, iter 4000/6416, lr 0.100000, loss 7.249261
+INFO 2021-03-18 23:08:00 train.py: 78] Epoch 4, iter 4200/6416, lr 0.100000, loss 7.292250
+INFO 2021-03-18 23:11:54 train.py: 78] Epoch 4, iter 4400/6416, lr 0.100000, loss 7.262050
+INFO 2021-03-18 23:15:49 train.py: 78] Epoch 4, iter 4600/6416, lr 0.100000, loss 7.279816
+INFO 2021-03-18 23:19:44 train.py: 78] Epoch 4, iter 4800/6416, lr 0.100000, loss 7.253316
+INFO 2021-03-18 23:23:39 train.py: 78] Epoch 4, iter 5000/6416, lr 0.100000, loss 7.232114
+INFO 2021-03-18 23:27:34 train.py: 78] Epoch 4, iter 5200/6416, lr 0.100000, loss 7.211947
+INFO 2021-03-18 23:31:29 train.py: 78] Epoch 4, iter 5400/6416, lr 0.100000, loss 7.227394
+INFO 2021-03-18 23:35:23 train.py: 78] Epoch 4, iter 5600/6416, lr 0.100000, loss 7.235565
+INFO 2021-03-18 23:39:18 train.py: 78] Epoch 4, iter 5800/6416, lr 0.100000, loss 7.165600
+INFO 2021-03-18 23:43:13 train.py: 91] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-03-18 23:43:14 train.py: 78] Epoch 4, iter 6000/6416, lr 0.100000, loss 7.185674
+INFO 2021-03-18 23:47:09 train.py: 78] Epoch 4, iter 6200/6416, lr 0.100000, loss 7.184557
+INFO 2021-03-18 23:51:04 train.py: 78] Epoch 4, iter 6400/6416, lr 0.100000, loss 7.185744
+INFO 2021-03-18 23:51:22 train.py: 96] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-03-18 23:51:25 train.py: 78] Epoch 5, iter 0/6416, lr 0.100000, loss 7.265677
+INFO 2021-03-18 23:55:19 train.py: 78] Epoch 5, iter 200/6416, lr 0.100000, loss 6.615274
+INFO 2021-03-18 23:59:14 train.py: 78] Epoch 5, iter 400/6416, lr 0.100000, loss 6.651365
+INFO 2021-03-19 00:03:08 train.py: 78] Epoch 5, iter 600/6416, lr 0.100000, loss 6.788537
+INFO 2021-03-19 00:07:03 train.py: 78] Epoch 5, iter 800/6416, lr 0.100000, loss 6.810198
+INFO 2021-03-19 00:10:58 train.py: 78] Epoch 5, iter 1000/6416, lr 0.100000, loss 6.833654
+INFO 2021-03-19 00:14:52 train.py: 78] Epoch 5, iter 1200/6416, lr 0.100000, loss 6.886417
+INFO 2021-03-19 00:18:47 train.py: 78] Epoch 5, iter 1400/6416, lr 0.100000, loss 6.952357
+INFO 2021-03-19 00:22:42 train.py: 78] Epoch 5, iter 1600/6416, lr 0.100000, loss 6.914136
+INFO 2021-03-19 00:26:37 train.py: 78] Epoch 5, iter 1800/6416, lr 0.100000, loss 6.959671
+INFO 2021-03-19 00:30:31 train.py: 78] Epoch 5, iter 2000/6416, lr 0.100000, loss 6.963630
+INFO 2021-03-19 00:34:26 train.py: 78] Epoch 5, iter 2200/6416, lr 0.100000, loss 6.959732
+INFO 2021-03-19 00:38:21 train.py: 78] Epoch 5, iter 2400/6416, lr 0.100000, loss 7.016838
+INFO 2021-03-19 00:42:15 train.py: 78] Epoch 5, iter 2600/6416, lr 0.100000, loss 6.957831
+INFO 2021-03-19 00:46:10 train.py: 78] Epoch 5, iter 2800/6416, lr 0.100000, loss 6.989584
+INFO 2021-03-19 00:50:05 train.py: 91] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-03-19 00:50:06 train.py: 78] Epoch 5, iter 3000/6416, lr 0.100000, loss 6.970374
+INFO 2021-03-19 00:54:01 train.py: 78] Epoch 5, iter 3200/6416, lr 0.100000, loss 6.987554
+INFO 2021-03-19 00:57:55 train.py: 78] Epoch 5, iter 3400/6416, lr 0.100000, loss 6.962708
+INFO 2021-03-19 01:01:50 train.py: 78] Epoch 5, iter 3600/6416, lr 0.100000, loss 6.948819
+INFO 2021-03-19 01:05:45 train.py: 78] Epoch 5, iter 3800/6416, lr 0.100000, loss 6.970038
+INFO 2021-03-19 01:09:39 train.py: 78] Epoch 5, iter 4000/6416, lr 0.100000, loss 6.931672
+INFO 2021-03-19 01:13:34 train.py: 78] Epoch 5, iter 4200/6416, lr 0.100000, loss 6.949328
+INFO 2021-03-19 01:17:29 train.py: 78] Epoch 5, iter 4400/6416, lr 0.100000, loss 6.948502
+INFO 2021-03-19 01:21:23 train.py: 78] Epoch 5, iter 4600/6416, lr 0.100000, loss 6.941133
+INFO 2021-03-19 01:25:18 train.py: 78] Epoch 5, iter 4800/6416, lr 0.100000, loss 6.923984
+INFO 2021-03-19 01:29:13 train.py: 78] Epoch 5, iter 5000/6416, lr 0.100000, loss 6.932709
+INFO 2021-03-19 01:33:08 train.py: 78] Epoch 5, iter 5200/6416, lr 0.100000, loss 6.919337
+INFO 2021-03-19 01:37:02 train.py: 78] Epoch 5, iter 5400/6416, lr 0.100000, loss 6.885154
+INFO 2021-03-19 01:40:57 train.py: 78] Epoch 5, iter 5600/6416, lr 0.100000, loss 6.883781
+INFO 2021-03-19 01:44:52 train.py: 78] Epoch 5, iter 5800/6416, lr 0.100000, loss 6.885192
+INFO 2021-03-19 01:48:46 train.py: 91] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-03-19 01:48:48 train.py: 78] Epoch 5, iter 6000/6416, lr 0.100000, loss 6.893103
+INFO 2021-03-19 01:52:42 train.py: 78] Epoch 5, iter 6200/6416, lr 0.100000, loss 6.876412
+INFO 2021-03-19 01:56:37 train.py: 78] Epoch 5, iter 6400/6416, lr 0.100000, loss 6.848571
+INFO 2021-03-19 01:56:55 train.py: 96] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-03-19 01:56:57 train.py: 78] Epoch 6, iter 0/6416, lr 0.100000, loss 6.841245
+INFO 2021-03-19 02:00:52 train.py: 78] Epoch 6, iter 200/6416, lr 0.100000, loss 6.348651
+INFO 2021-03-19 02:04:46 train.py: 78] Epoch 6, iter 400/6416, lr 0.100000, loss 6.370475
+INFO 2021-03-19 02:08:41 train.py: 78] Epoch 6, iter 600/6416, lr 0.100000, loss 6.461171
+INFO 2021-03-19 02:12:36 train.py: 78] Epoch 6, iter 800/6416, lr 0.100000, loss 6.501721
+INFO 2021-03-19 02:16:30 train.py: 78] Epoch 6, iter 1000/6416, lr 0.100000, loss 6.552796
+INFO 2021-03-19 02:20:25 train.py: 78] Epoch 6, iter 1200/6416, lr 0.100000, loss 6.638028
+INFO 2021-03-19 02:24:19 train.py: 78] Epoch 6, iter 1400/6416, lr 0.100000, loss 6.676625
+INFO 2021-03-19 02:28:14 train.py: 78] Epoch 6, iter 1600/6416, lr 0.100000, loss 6.687038
+INFO 2021-03-19 02:32:09 train.py: 78] Epoch 6, iter 1800/6416, lr 0.100000, loss 6.689456
+INFO 2021-03-19 02:36:03 train.py: 78] Epoch 6, iter 2000/6416, lr 0.100000, loss 6.692840
+INFO 2021-03-19 02:39:58 train.py: 78] Epoch 6, iter 2200/6416, lr 0.100000, loss 6.733830
+INFO 2021-03-19 02:43:52 train.py: 78] Epoch 6, iter 2400/6416, lr 0.100000, loss 6.734686
+INFO 2021-03-19 02:47:47 train.py: 78] Epoch 6, iter 2600/6416, lr 0.100000, loss 6.711833
+INFO 2021-03-19 02:51:42 train.py: 78] Epoch 6, iter 2800/6416, lr 0.100000, loss 6.727327
+INFO 2021-03-19 02:55:36 train.py: 91] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-03-19 02:55:37 train.py: 78] Epoch 6, iter 3000/6416, lr 0.100000, loss 6.721668
+INFO 2021-03-19 02:59:32 train.py: 78] Epoch 6, iter 3200/6416, lr 0.100000, loss 6.744344
+INFO 2021-03-19 03:03:27 train.py: 78] Epoch 6, iter 3400/6416, lr 0.100000, loss 6.736803
+INFO 2021-03-19 03:07:21 train.py: 78] Epoch 6, iter 3600/6416, lr 0.100000, loss 6.695277
+INFO 2021-03-19 03:11:16 train.py: 78] Epoch 6, iter 3800/6416, lr 0.100000, loss 6.749279
+INFO 2021-03-19 03:15:10 train.py: 78] Epoch 6, iter 4000/6416, lr 0.100000, loss 6.743833
+INFO 2021-03-19 03:19:05 train.py: 78] Epoch 6, iter 4200/6416, lr 0.100000, loss 6.732289
+INFO 2021-03-19 03:23:00 train.py: 78] Epoch 6, iter 4400/6416, lr 0.100000, loss 6.690273
+INFO 2021-03-19 03:26:54 train.py: 78] Epoch 6, iter 4600/6416, lr 0.100000, loss 6.737407
+INFO 2021-03-19 03:30:49 train.py: 78] Epoch 6, iter 4800/6416, lr 0.100000, loss 6.706612
+INFO 2021-03-19 03:34:43 train.py: 78] Epoch 6, iter 5000/6416, lr 0.100000, loss 6.689826
+INFO 2021-03-19 03:38:38 train.py: 78] Epoch 6, iter 5200/6416, lr 0.100000, loss 6.689431
+INFO 2021-03-19 03:42:33 train.py: 78] Epoch 6, iter 5400/6416, lr 0.100000, loss 6.720154
+INFO 2021-03-19 03:46:27 train.py: 78] Epoch 6, iter 5600/6416, lr 0.100000, loss 6.671153
+INFO 2021-03-19 03:50:22 train.py: 78] Epoch 6, iter 5800/6416, lr 0.100000, loss 6.679764
+INFO 2021-03-19 03:54:17 train.py: 91] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-03-19 03:54:18 train.py: 78] Epoch 6, iter 6000/6416, lr 0.100000, loss 6.695298
+INFO 2021-03-19 03:58:12 train.py: 78] Epoch 6, iter 6200/6416, lr 0.100000, loss 6.662221
+INFO 2021-03-19 04:02:07 train.py: 78] Epoch 6, iter 6400/6416, lr 0.100000, loss 6.645985
+INFO 2021-03-19 04:02:25 train.py: 96] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-03-19 04:02:28 train.py: 78] Epoch 7, iter 0/6416, lr 0.100000, loss 6.651794
+INFO 2021-03-19 04:06:22 train.py: 78] Epoch 7, iter 200/6416, lr 0.100000, loss 6.115563
+INFO 2021-03-19 04:10:17 train.py: 78] Epoch 7, iter 400/6416, lr 0.100000, loss 6.167348
+INFO 2021-03-19 04:14:11 train.py: 78] Epoch 7, iter 600/6416, lr 0.100000, loss 6.246819
+INFO 2021-03-19 04:18:06 train.py: 78] Epoch 7, iter 800/6416, lr 0.100000, loss 6.333862
+INFO 2021-03-19 04:22:00 train.py: 78] Epoch 7, iter 1000/6416, lr 0.100000, loss 6.376914
+INFO 2021-03-19 04:25:55 train.py: 78] Epoch 7, iter 1200/6416, lr 0.100000, loss 6.412811
+INFO 2021-03-19 04:29:49 train.py: 78] Epoch 7, iter 1400/6416, lr 0.100000, loss 6.468878
+INFO 2021-03-19 04:33:44 train.py: 78] Epoch 7, iter 1600/6416, lr 0.100000, loss 6.478902
+INFO 2021-03-19 04:37:38 train.py: 78] Epoch 7, iter 1800/6416, lr 0.100000, loss 6.498023
+INFO 2021-03-19 04:41:33 train.py: 78] Epoch 7, iter 2000/6416, lr 0.100000, loss 6.543498
+INFO 2021-03-19 04:45:28 train.py: 78] Epoch 7, iter 2200/6416, lr 0.100000, loss 6.515108
+INFO 2021-03-19 04:49:22 train.py: 78] Epoch 7, iter 2400/6416, lr 0.100000, loss 6.505551
+INFO 2021-03-19 04:53:17 train.py: 78] Epoch 7, iter 2600/6416, lr 0.100000, loss 6.531743
+INFO 2021-03-19 04:57:11 train.py: 78] Epoch 7, iter 2800/6416, lr 0.100000, loss 6.546517
+INFO 2021-03-19 05:01:06 train.py: 91] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-03-19 05:01:07 train.py: 78] Epoch 7, iter 3000/6416, lr 0.100000, loss 6.559822
+INFO 2021-03-19 05:05:02 train.py: 78] Epoch 7, iter 3200/6416, lr 0.100000, loss 6.550284
+INFO 2021-03-19 05:08:56 train.py: 78] Epoch 7, iter 3400/6416, lr 0.100000, loss 6.514897
+INFO 2021-03-19 05:12:51 train.py: 78] Epoch 7, iter 3600/6416, lr 0.100000, loss 6.534035
+INFO 2021-03-19 05:16:45 train.py: 78] Epoch 7, iter 3800/6416, lr 0.100000, loss 6.569572
+INFO 2021-03-19 05:20:40 train.py: 78] Epoch 7, iter 4000/6416, lr 0.100000, loss 6.550297
+INFO 2021-03-19 05:24:35 train.py: 78] Epoch 7, iter 4200/6416, lr 0.100000, loss 6.536845
+INFO 2021-03-19 05:28:29 train.py: 78] Epoch 7, iter 4400/6416, lr 0.100000, loss 6.529620
+INFO 2021-03-19 05:32:24 train.py: 78] Epoch 7, iter 4600/6416, lr 0.100000, loss 6.505814
+INFO 2021-03-19 05:36:18 train.py: 78] Epoch 7, iter 4800/6416, lr 0.100000, loss 6.561263
+INFO 2021-03-19 05:40:13 train.py: 78] Epoch 7, iter 5000/6416, lr 0.100000, loss 6.532819
+INFO 2021-03-19 05:44:08 train.py: 78] Epoch 7, iter 5200/6416, lr 0.100000, loss 6.526186
+INFO 2021-03-19 05:48:02 train.py: 78] Epoch 7, iter 5400/6416, lr 0.100000, loss 6.530271
+INFO 2021-03-19 05:51:57 train.py: 78] Epoch 7, iter 5600/6416, lr 0.100000, loss 6.528420
+INFO 2021-03-19 05:55:52 train.py: 78] Epoch 7, iter 5800/6416, lr 0.100000, loss 6.534301
+INFO 2021-03-19 05:59:46 train.py: 91] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-03-19 05:59:47 train.py: 78] Epoch 7, iter 6000/6416, lr 0.100000, loss 6.497143
+INFO 2021-03-19 06:03:42 train.py: 78] Epoch 7, iter 6200/6416, lr 0.100000, loss 6.509710
+INFO 2021-03-19 06:07:37 train.py: 78] Epoch 7, iter 6400/6416, lr 0.100000, loss 6.506533
+INFO 2021-03-19 06:07:55 train.py: 96] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-03-19 06:07:58 train.py: 78] Epoch 8, iter 0/6416, lr 0.100000, loss 6.506368
+INFO 2021-03-19 06:11:52 train.py: 78] Epoch 8, iter 200/6416, lr 0.100000, loss 5.978776
+INFO 2021-03-19 06:15:47 train.py: 78] Epoch 8, iter 400/6416, lr 0.100000, loss 6.018659
+INFO 2021-03-19 06:19:41 train.py: 78] Epoch 8, iter 600/6416, lr 0.100000, loss 6.092167
+INFO 2021-03-19 06:23:36 train.py: 78] Epoch 8, iter 800/6416, lr 0.100000, loss 6.187115
+INFO 2021-03-19 06:27:30 train.py: 78] Epoch 8, iter 1000/6416, lr 0.100000, loss 6.237979
+INFO 2021-03-19 06:31:25 train.py: 78] Epoch 8, iter 1200/6416, lr 0.100000, loss 6.290860
+INFO 2021-03-19 06:35:19 train.py: 78] Epoch 8, iter 1400/6416, lr 0.100000, loss 6.340390
+INFO 2021-03-19 06:39:14 train.py: 78] Epoch 8, iter 1600/6416, lr 0.100000, loss 6.336167
+INFO 2021-03-19 06:43:08 train.py: 78] Epoch 8, iter 1800/6416, lr 0.100000, loss 6.347223
+INFO 2021-03-19 06:47:03 train.py: 78] Epoch 8, iter 2000/6416, lr 0.100000, loss 6.363110
+INFO 2021-03-19 06:50:57 train.py: 78] Epoch 8, iter 2200/6416, lr 0.100000, loss 6.362799
+INFO 2021-03-19 06:54:52 train.py: 78] Epoch 8, iter 2400/6416, lr 0.100000, loss 6.425506
+INFO 2021-03-19 06:58:46 train.py: 78] Epoch 8, iter 2600/6416, lr 0.100000, loss 6.402283
+INFO 2021-03-19 07:02:41 train.py: 78] Epoch 8, iter 2800/6416, lr 0.100000, loss 6.411157
+INFO 2021-03-19 07:06:36 train.py: 91] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-03-19 07:06:37 train.py: 78] Epoch 8, iter 3000/6416, lr 0.100000, loss 6.409878
+INFO 2021-03-19 07:10:31 train.py: 78] Epoch 8, iter 3200/6416, lr 0.100000, loss 6.386874
+INFO 2021-03-19 07:14:26 train.py: 78] Epoch 8, iter 3400/6416, lr 0.100000, loss 6.372492
+INFO 2021-03-19 07:18:21 train.py: 78] Epoch 8, iter 3600/6416, lr 0.100000, loss 6.398647
+INFO 2021-03-19 07:22:15 train.py: 78] Epoch 8, iter 3800/6416, lr 0.100000, loss 6.398147
+INFO 2021-03-19 07:26:10 train.py: 78] Epoch 8, iter 4000/6416, lr 0.100000, loss 6.405181
+INFO 2021-03-19 07:30:04 train.py: 78] Epoch 8, iter 4200/6416, lr 0.100000, loss 6.390733
+INFO 2021-03-19 07:33:59 train.py: 78] Epoch 8, iter 4400/6416, lr 0.100000, loss 6.378577
+INFO 2021-03-19 07:37:53 train.py: 78] Epoch 8, iter 4600/6416, lr 0.100000, loss 6.424184
+INFO 2021-03-19 07:41:48 train.py: 78] Epoch 8, iter 4800/6416, lr 0.100000, loss 6.366973
+INFO 2021-03-19 07:45:43 train.py: 78] Epoch 8, iter 5000/6416, lr 0.100000, loss 6.397382
+INFO 2021-03-19 07:49:37 train.py: 78] Epoch 8, iter 5200/6416, lr 0.100000, loss 6.421305
+INFO 2021-03-19 07:53:32 train.py: 78] Epoch 8, iter 5400/6416, lr 0.100000, loss 6.383141
+INFO 2021-03-19 07:57:26 train.py: 78] Epoch 8, iter 5600/6416, lr 0.100000, loss 6.399861
+INFO 2021-03-19 08:01:21 train.py: 78] Epoch 8, iter 5800/6416, lr 0.100000, loss 6.379353
+INFO 2021-03-19 08:05:16 train.py: 91] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-03-19 08:05:17 train.py: 78] Epoch 8, iter 6000/6416, lr 0.100000, loss 6.407344
+INFO 2021-03-19 08:09:11 train.py: 78] Epoch 8, iter 6200/6416, lr 0.100000, loss 6.344912
+INFO 2021-03-19 08:13:06 train.py: 78] Epoch 8, iter 6400/6416, lr 0.100000, loss 6.383652
+INFO 2021-03-19 08:13:24 train.py: 96] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-03-19 08:13:27 train.py: 78] Epoch 9, iter 0/6416, lr 0.100000, loss 6.358603
+INFO 2021-03-19 08:17:21 train.py: 78] Epoch 9, iter 200/6416, lr 0.100000, loss 5.887477
+INFO 2021-03-19 08:21:16 train.py: 78] Epoch 9, iter 400/6416, lr 0.100000, loss 5.904647
+INFO 2021-03-19 08:25:10 train.py: 78] Epoch 9, iter 600/6416, lr 0.100000, loss 5.980715
+INFO 2021-03-19 08:29:05 train.py: 78] Epoch 9, iter 800/6416, lr 0.100000, loss 6.070103
+INFO 2021-03-19 08:32:59 train.py: 78] Epoch 9, iter 1000/6416, lr 0.100000, loss 6.121906
+INFO 2021-03-19 08:36:54 train.py: 78] Epoch 9, iter 1200/6416, lr 0.100000, loss 6.163510
+INFO 2021-03-19 08:40:48 train.py: 78] Epoch 9, iter 1400/6416, lr 0.100000, loss 6.187649
+INFO 2021-03-19 08:44:43 train.py: 78] Epoch 9, iter 1600/6416, lr 0.100000, loss 6.237368
+INFO 2021-03-19 08:48:37 train.py: 78] Epoch 9, iter 1800/6416, lr 0.100000, loss 6.251701
+INFO 2021-03-19 08:52:32 train.py: 78] Epoch 9, iter 2000/6416, lr 0.100000, loss 6.246961
+INFO 2021-03-19 08:56:27 train.py: 78] Epoch 9, iter 2200/6416, lr 0.100000, loss 6.222371
+INFO 2021-03-19 09:00:21 train.py: 78] Epoch 9, iter 2400/6416, lr 0.100000, loss 6.266408
+INFO 2021-03-19 09:04:16 train.py: 78] Epoch 9, iter 2600/6416, lr 0.100000, loss 6.266867
+INFO 2021-03-19 09:08:10 train.py: 78] Epoch 9, iter 2800/6416, lr 0.100000, loss 6.306914
+INFO 2021-03-19 09:12:05 train.py: 91] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-03-19 09:12:06 train.py: 78] Epoch 9, iter 3000/6416, lr 0.100000, loss 6.305083
+INFO 2021-03-19 09:16:01 train.py: 78] Epoch 9, iter 3200/6416, lr 0.100000, loss 6.315077
+INFO 2021-03-19 09:19:55 train.py: 78] Epoch 9, iter 3400/6416, lr 0.100000, loss 6.299579
+INFO 2021-03-19 09:23:50 train.py: 78] Epoch 9, iter 3600/6416, lr 0.100000, loss 6.314183
+INFO 2021-03-19 09:27:44 train.py: 78] Epoch 9, iter 3800/6416, lr 0.100000, loss 6.289157
+INFO 2021-03-19 09:31:39 train.py: 78] Epoch 9, iter 4000/6416, lr 0.100000, loss 6.267283
+INFO 2021-03-19 09:35:34 train.py: 78] Epoch 9, iter 4200/6416, lr 0.100000, loss 6.293458
+INFO 2021-03-19 09:39:28 train.py: 78] Epoch 9, iter 4400/6416, lr 0.100000, loss 6.295358
+INFO 2021-03-19 09:43:23 train.py: 78] Epoch 9, iter 4600/6416, lr 0.100000, loss 6.290014
+INFO 2021-03-19 09:47:17 train.py: 78] Epoch 9, iter 4800/6416, lr 0.100000, loss 6.296590
+INFO 2021-03-19 09:51:12 train.py: 78] Epoch 9, iter 5000/6416, lr 0.100000, loss 6.309894
+INFO 2021-03-19 09:55:07 train.py: 78] Epoch 9, iter 5200/6416, lr 0.100000, loss 6.262912
+INFO 2021-03-19 09:59:01 train.py: 78] Epoch 9, iter 5400/6416, lr 0.100000, loss 6.285177
+INFO 2021-03-19 10:02:56 train.py: 78] Epoch 9, iter 5600/6416, lr 0.100000, loss 6.298389
+INFO 2021-03-19 10:06:50 train.py: 78] Epoch 9, iter 5800/6416, lr 0.100000, loss 6.303705
+INFO 2021-03-19 10:10:45 train.py: 91] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-03-19 10:10:46 train.py: 78] Epoch 9, iter 6000/6416, lr 0.100000, loss 6.300152
+INFO 2021-03-19 10:14:41 train.py: 78] Epoch 9, iter 6200/6416, lr 0.100000, loss 6.279227
+INFO 2021-03-19 10:18:36 train.py: 78] Epoch 9, iter 6400/6416, lr 0.100000, loss 6.283215
+INFO 2021-03-19 10:18:54 train.py: 96] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-03-19 10:18:56 train.py: 78] Epoch 10, iter 0/6416, lr 0.010000, loss 6.306887
+INFO 2021-03-19 10:22:51 train.py: 78] Epoch 10, iter 200/6416, lr 0.010000, loss 5.102814
+INFO 2021-03-19 10:26:45 train.py: 78] Epoch 10, iter 400/6416, lr 0.010000, loss 4.828291
+INFO 2021-03-19 10:30:39 train.py: 78] Epoch 10, iter 600/6416, lr 0.010000, loss 4.736236
+INFO 2021-03-19 10:34:34 train.py: 78] Epoch 10, iter 800/6416, lr 0.010000, loss 4.687695
+INFO 2021-03-19 10:38:28 train.py: 78] Epoch 10, iter 1000/6416, lr 0.010000, loss 4.635845
+INFO 2021-03-19 10:42:22 train.py: 78] Epoch 10, iter 1200/6416, lr 0.010000, loss 4.567644
+INFO 2021-03-19 10:46:16 train.py: 78] Epoch 10, iter 1400/6416, lr 0.010000, loss 4.536621
+INFO 2021-03-19 10:50:11 train.py: 78] Epoch 10, iter 1600/6416, lr 0.010000, loss 4.498202
+INFO 2021-03-19 10:54:05 train.py: 78] Epoch 10, iter 1800/6416, lr 0.010000, loss 4.469244
+INFO 2021-03-19 10:57:59 train.py: 78] Epoch 10, iter 2000/6416, lr 0.010000, loss 4.422997
+INFO 2021-03-19 11:01:54 train.py: 78] Epoch 10, iter 2200/6416, lr 0.010000, loss 4.400673
+INFO 2021-03-19 11:05:48 train.py: 78] Epoch 10, iter 2400/6416, lr 0.010000, loss 4.371461
+INFO 2021-03-19 11:09:42 train.py: 78] Epoch 10, iter 2600/6416, lr 0.010000, loss 4.347874
+INFO 2021-03-19 11:13:37 train.py: 78] Epoch 10, iter 2800/6416, lr 0.010000, loss 4.324255
+INFO 2021-03-19 11:17:31 train.py: 91] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-03-19 11:17:32 train.py: 78] Epoch 10, iter 3000/6416, lr 0.010000, loss 4.321841
+INFO 2021-03-19 11:21:26 train.py: 78] Epoch 10, iter 3200/6416, lr 0.010000, loss 4.306121
+INFO 2021-03-19 11:25:21 train.py: 78] Epoch 10, iter 3400/6416, lr 0.010000, loss 4.278456
+INFO 2021-03-19 11:29:15 train.py: 78] Epoch 10, iter 3600/6416, lr 0.010000, loss 4.242358
+INFO 2021-03-19 11:33:09 train.py: 78] Epoch 10, iter 3800/6416, lr 0.010000, loss 4.220599
+INFO 2021-03-19 11:37:03 train.py: 78] Epoch 10, iter 4000/6416, lr 0.010000, loss 4.201499
+INFO 2021-03-19 11:40:58 train.py: 78] Epoch 10, iter 4200/6416, lr 0.010000, loss 4.191400
+INFO 2021-03-19 11:44:52 train.py: 78] Epoch 10, iter 4400/6416, lr 0.010000, loss 4.176634
+INFO 2021-03-19 11:48:46 train.py: 78] Epoch 10, iter 4600/6416, lr 0.010000, loss 4.182758
+INFO 2021-03-19 11:52:41 train.py: 78] Epoch 10, iter 4800/6416, lr 0.010000, loss 4.156792
+INFO 2021-03-19 11:56:35 train.py: 78] Epoch 10, iter 5000/6416, lr 0.010000, loss 4.138895
+INFO 2021-03-19 12:00:29 train.py: 78] Epoch 10, iter 5200/6416, lr 0.010000, loss 4.141925
+INFO 2021-03-19 12:04:24 train.py: 78] Epoch 10, iter 5400/6416, lr 0.010000, loss 4.122506
+INFO 2021-03-19 12:08:18 train.py: 78] Epoch 10, iter 5600/6416, lr 0.010000, loss 4.109429
+INFO 2021-03-19 12:12:12 train.py: 78] Epoch 10, iter 5800/6416, lr 0.010000, loss 4.111880
+INFO 2021-03-19 12:16:06 train.py: 91] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-03-19 12:16:07 train.py: 78] Epoch 10, iter 6000/6416, lr 0.010000, loss 4.079940
+INFO 2021-03-19 12:20:02 train.py: 78] Epoch 10, iter 6200/6416, lr 0.010000, loss 4.060515
+INFO 2021-03-19 12:23:56 train.py: 78] Epoch 10, iter 6400/6416, lr 0.010000, loss 4.070621
+INFO 2021-03-19 12:24:14 train.py: 96] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-03-19 12:24:17 train.py: 78] Epoch 11, iter 0/6416, lr 0.010000, loss 4.059079
+INFO 2021-03-19 12:28:11 train.py: 78] Epoch 11, iter 200/6416, lr 0.010000, loss 3.739686
+INFO 2021-03-19 12:32:05 train.py: 78] Epoch 11, iter 400/6416, lr 0.010000, loss 3.740610
+INFO 2021-03-19 12:35:59 train.py: 78] Epoch 11, iter 600/6416, lr 0.010000, loss 3.729643
+INFO 2021-03-19 12:39:54 train.py: 78] Epoch 11, iter 800/6416, lr 0.010000, loss 3.723222
+INFO 2021-03-19 12:43:48 train.py: 78] Epoch 11, iter 1000/6416, lr 0.010000, loss 3.733978
+INFO 2021-03-19 12:47:42 train.py: 78] Epoch 11, iter 1200/6416, lr 0.010000, loss 3.747483
+INFO 2021-03-19 12:51:37 train.py: 78] Epoch 11, iter 1400/6416, lr 0.010000, loss 3.749420
+INFO 2021-03-19 12:55:31 train.py: 78] Epoch 11, iter 1600/6416, lr 0.010000, loss 3.729718
+INFO 2021-03-19 12:59:25 train.py: 78] Epoch 11, iter 1800/6416, lr 0.010000, loss 3.767851
+INFO 2021-03-19 13:03:19 train.py: 78] Epoch 11, iter 2000/6416, lr 0.010000, loss 3.753646
+INFO 2021-03-19 13:07:14 train.py: 78] Epoch 11, iter 2200/6416, lr 0.010000, loss 3.774915
+INFO 2021-03-19 13:11:08 train.py: 78] Epoch 11, iter 2400/6416, lr 0.010000, loss 3.765724
+INFO 2021-03-19 13:15:02 train.py: 78] Epoch 11, iter 2600/6416, lr 0.010000, loss 3.743829
+INFO 2021-03-19 13:18:57 train.py: 78] Epoch 11, iter 2800/6416, lr 0.010000, loss 3.740179
+INFO 2021-03-19 13:22:51 train.py: 91] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-03-19 13:22:52 train.py: 78] Epoch 11, iter 3000/6416, lr 0.010000, loss 3.741612
+INFO 2021-03-19 13:26:46 train.py: 78] Epoch 11, iter 3200/6416, lr 0.010000, loss 3.752897
+INFO 2021-03-19 13:30:40 train.py: 78] Epoch 11, iter 3400/6416, lr 0.010000, loss 3.741632
+INFO 2021-03-19 13:34:35 train.py: 78] Epoch 11, iter 3600/6416, lr 0.010000, loss 3.757295
+INFO 2021-03-19 13:38:29 train.py: 78] Epoch 11, iter 3800/6416, lr 0.010000, loss 3.761692
+INFO 2021-03-19 13:42:23 train.py: 78] Epoch 11, iter 4000/6416, lr 0.010000, loss 3.730823
+INFO 2021-03-19 13:46:18 train.py: 78] Epoch 11, iter 4200/6416, lr 0.010000, loss 3.735064
+INFO 2021-03-19 13:50:12 train.py: 78] Epoch 11, iter 4400/6416, lr 0.010000, loss 3.756033
+INFO 2021-03-19 13:54:06 train.py: 78] Epoch 11, iter 4600/6416, lr 0.010000, loss 3.765811
+INFO 2021-03-19 13:58:16 train.py: 78] Epoch 11, iter 4800/6416, lr 0.010000, loss 3.754437
+INFO 2021-03-19 14:02:27 train.py: 78] Epoch 11, iter 5000/6416, lr 0.010000, loss 3.749697
+INFO 2021-03-19 14:06:37 train.py: 78] Epoch 11, iter 5200/6416, lr 0.010000, loss 3.751701
+INFO 2021-03-19 14:10:32 train.py: 78] Epoch 11, iter 5400/6416, lr 0.010000, loss 3.737948
+INFO 2021-03-19 14:14:26 train.py: 78] Epoch 11, iter 5600/6416, lr 0.010000, loss 3.755978
+INFO 2021-03-19 14:18:20 train.py: 78] Epoch 11, iter 5800/6416, lr 0.010000, loss 3.796466
+INFO 2021-03-19 14:22:14 train.py: 91] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-03-19 14:22:15 train.py: 78] Epoch 11, iter 6000/6416, lr 0.010000, loss 3.760121
+INFO 2021-03-19 14:26:10 train.py: 78] Epoch 11, iter 6200/6416, lr 0.010000, loss 3.757607
+INFO 2021-03-19 14:30:04 train.py: 78] Epoch 11, iter 6400/6416, lr 0.010000, loss 3.758687
+INFO 2021-03-19 14:30:22 train.py: 96] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-03-19 14:30:25 train.py: 78] Epoch 12, iter 0/6416, lr 0.010000, loss 3.784179
+INFO 2021-03-19 14:34:19 train.py: 78] Epoch 12, iter 200/6416, lr 0.010000, loss 3.431424
+INFO 2021-03-19 14:38:13 train.py: 78] Epoch 12, iter 400/6416, lr 0.010000, loss 3.455622
+INFO 2021-03-19 14:42:08 train.py: 78] Epoch 12, iter 600/6416, lr 0.010000, loss 3.464733
+INFO 2021-03-19 14:46:02 train.py: 78] Epoch 12, iter 800/6416, lr 0.010000, loss 3.476458
+INFO 2021-03-19 14:49:56 train.py: 78] Epoch 12, iter 1000/6416, lr 0.010000, loss 3.457887
+INFO 2021-03-19 14:53:50 train.py: 78] Epoch 12, iter 1200/6416, lr 0.010000, loss 3.479228
+INFO 2021-03-19 14:57:44 train.py: 78] Epoch 12, iter 1400/6416, lr 0.010000, loss 3.492333
+INFO 2021-03-19 15:01:39 train.py: 78] Epoch 12, iter 1600/6416, lr 0.010000, loss 3.498967
+INFO 2021-03-19 15:05:33 train.py: 78] Epoch 12, iter 1800/6416, lr 0.010000, loss 3.533412
+INFO 2021-03-19 15:09:27 train.py: 78] Epoch 12, iter 2000/6416, lr 0.010000, loss 3.508884
+INFO 2021-03-19 15:13:22 train.py: 78] Epoch 12, iter 2200/6416, lr 0.010000, loss 3.538386
+INFO 2021-03-19 15:17:16 train.py: 78] Epoch 12, iter 2400/6416, lr 0.010000, loss 3.548407
+INFO 2021-03-19 15:21:10 train.py: 78] Epoch 12, iter 2600/6416, lr 0.010000, loss 3.561202
+INFO 2021-03-19 15:25:04 train.py: 78] Epoch 12, iter 2800/6416, lr 0.010000, loss 3.553840
+INFO 2021-03-19 15:28:59 train.py: 91] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-03-19 15:29:00 train.py: 78] Epoch 12, iter 3000/6416, lr 0.010000, loss 3.550627
+INFO 2021-03-19 15:32:54 train.py: 78] Epoch 12, iter 3200/6416, lr 0.010000, loss 3.560952
+INFO 2021-03-19 15:36:48 train.py: 78] Epoch 12, iter 3400/6416, lr 0.010000, loss 3.582431
+INFO 2021-03-19 15:40:43 train.py: 78] Epoch 12, iter 3600/6416, lr 0.010000, loss 3.597199
+INFO 2021-03-19 15:44:37 train.py: 78] Epoch 12, iter 3800/6416, lr 0.010000, loss 3.581864
+INFO 2021-03-19 15:48:31 train.py: 78] Epoch 12, iter 4000/6416, lr 0.010000, loss 3.574441
+INFO 2021-03-19 15:52:26 train.py: 78] Epoch 12, iter 4200/6416, lr 0.010000, loss 3.607729
+INFO 2021-03-19 15:56:20 train.py: 78] Epoch 12, iter 4400/6416, lr 0.010000, loss 3.621575
+INFO 2021-03-19 16:00:14 train.py: 78] Epoch 12, iter 4600/6416, lr 0.010000, loss 3.609787
+INFO 2021-03-19 16:04:08 train.py: 78] Epoch 12, iter 4800/6416, lr 0.010000, loss 3.636674
+INFO 2021-03-19 16:08:03 train.py: 78] Epoch 12, iter 5000/6416, lr 0.010000, loss 3.609125
+INFO 2021-03-19 16:11:57 train.py: 78] Epoch 12, iter 5200/6416, lr 0.010000, loss 3.610963
+INFO 2021-03-19 16:15:51 train.py: 78] Epoch 12, iter 5400/6416, lr 0.010000, loss 3.622922
+INFO 2021-03-19 16:19:45 train.py: 78] Epoch 12, iter 5600/6416, lr 0.010000, loss 3.618367
+INFO 2021-03-19 16:23:40 train.py: 78] Epoch 12, iter 5800/6416, lr 0.010000, loss 3.600346
+INFO 2021-03-19 16:27:34 train.py: 91] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-03-19 16:27:35 train.py: 78] Epoch 12, iter 6000/6416, lr 0.010000, loss 3.637814
+INFO 2021-03-19 16:31:29 train.py: 78] Epoch 12, iter 6200/6416, lr 0.010000, loss 3.666515
+INFO 2021-03-19 16:35:24 train.py: 78] Epoch 12, iter 6400/6416, lr 0.010000, loss 3.650005
+INFO 2021-03-19 16:35:42 train.py: 96] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-03-19 16:35:44 train.py: 78] Epoch 13, iter 0/6416, lr 0.001000, loss 3.629887
+INFO 2021-03-19 16:39:39 train.py: 78] Epoch 13, iter 200/6416, lr 0.001000, loss 3.229679
+INFO 2021-03-19 16:43:33 train.py: 78] Epoch 13, iter 400/6416, lr 0.001000, loss 3.187297
+INFO 2021-03-19 16:47:27 train.py: 78] Epoch 13, iter 600/6416, lr 0.001000, loss 3.196512
+INFO 2021-03-19 16:51:22 train.py: 78] Epoch 13, iter 800/6416, lr 0.001000, loss 3.208044
+INFO 2021-03-19 16:55:16 train.py: 78] Epoch 13, iter 1000/6416, lr 0.001000, loss 3.185062
+INFO 2021-03-19 16:59:10 train.py: 78] Epoch 13, iter 1200/6416, lr 0.001000, loss 3.181594
+INFO 2021-03-19 17:03:04 train.py: 78] Epoch 13, iter 1400/6416, lr 0.001000, loss 3.165161
+INFO 2021-03-19 17:06:59 train.py: 78] Epoch 13, iter 1600/6416, lr 0.001000, loss 3.188835
+INFO 2021-03-19 17:10:53 train.py: 78] Epoch 13, iter 1800/6416, lr 0.001000, loss 3.178862
+INFO 2021-03-19 17:14:47 train.py: 78] Epoch 13, iter 2000/6416, lr 0.001000, loss 3.177287
+INFO 2021-03-19 17:18:41 train.py: 78] Epoch 13, iter 2200/6416, lr 0.001000, loss 3.165604
+INFO 2021-03-19 17:22:36 train.py: 78] Epoch 13, iter 2400/6416, lr 0.001000, loss 3.169563
+INFO 2021-03-19 17:26:30 train.py: 78] Epoch 13, iter 2600/6416, lr 0.001000, loss 3.193548
+INFO 2021-03-19 17:30:24 train.py: 78] Epoch 13, iter 2800/6416, lr 0.001000, loss 3.155197
+INFO 2021-03-19 17:34:18 train.py: 91] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-03-19 17:34:19 train.py: 78] Epoch 13, iter 3000/6416, lr 0.001000, loss 3.176681
+INFO 2021-03-19 17:38:14 train.py: 78] Epoch 13, iter 3200/6416, lr 0.001000, loss 3.171748
+INFO 2021-03-19 17:42:08 train.py: 78] Epoch 13, iter 3400/6416, lr 0.001000, loss 3.173824
+INFO 2021-03-19 17:46:02 train.py: 78] Epoch 13, iter 3600/6416, lr 0.001000, loss 3.200136
+INFO 2021-03-19 17:49:56 train.py: 78] Epoch 13, iter 3800/6416, lr 0.001000, loss 3.152063
+INFO 2021-03-19 17:53:51 train.py: 78] Epoch 13, iter 4000/6416, lr 0.001000, loss 3.167090
+INFO 2021-03-19 17:57:45 train.py: 78] Epoch 13, iter 4200/6416, lr 0.001000, loss 3.182108
+INFO 2021-03-19 18:01:39 train.py: 78] Epoch 13, iter 4400/6416, lr 0.001000, loss 3.145477
+INFO 2021-03-19 18:05:33 train.py: 78] Epoch 13, iter 4600/6416, lr 0.001000, loss 3.177578
+INFO 2021-03-19 18:09:28 train.py: 78] Epoch 13, iter 4800/6416, lr 0.001000, loss 3.179698
+INFO 2021-03-19 18:13:22 train.py: 78] Epoch 13, iter 5000/6416, lr 0.001000, loss 3.200265
+INFO 2021-03-19 18:17:16 train.py: 78] Epoch 13, iter 5200/6416, lr 0.001000, loss 3.165959
+INFO 2021-03-19 18:21:10 train.py: 78] Epoch 13, iter 5400/6416, lr 0.001000, loss 3.182200
+INFO 2021-03-19 18:25:05 train.py: 78] Epoch 13, iter 5600/6416, lr 0.001000, loss 3.184430
+INFO 2021-03-19 18:28:59 train.py: 78] Epoch 13, iter 5800/6416, lr 0.001000, loss 3.184106
+INFO 2021-03-19 18:32:53 train.py: 91] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-03-19 18:32:54 train.py: 78] Epoch 13, iter 6000/6416, lr 0.001000, loss 3.174152
+INFO 2021-03-19 18:36:49 train.py: 78] Epoch 13, iter 6200/6416, lr 0.001000, loss 3.174772
+INFO 2021-03-19 18:40:43 train.py: 78] Epoch 13, iter 6400/6416, lr 0.001000, loss 3.170197
+INFO 2021-03-19 18:41:01 train.py: 96] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-03-19 18:41:04 train.py: 78] Epoch 14, iter 0/6416, lr 0.001000, loss 3.216188
+INFO 2021-03-19 18:44:58 train.py: 78] Epoch 14, iter 200/6416, lr 0.001000, loss 3.115940
+INFO 2021-03-19 18:48:52 train.py: 78] Epoch 14, iter 400/6416, lr 0.001000, loss 3.110386
+INFO 2021-03-19 18:52:46 train.py: 78] Epoch 14, iter 600/6416, lr 0.001000, loss 3.112229
+INFO 2021-03-19 18:56:41 train.py: 78] Epoch 14, iter 800/6416, lr 0.001000, loss 3.135627
+INFO 2021-03-19 19:00:35 train.py: 78] Epoch 14, iter 1000/6416, lr 0.001000, loss 3.108733
+INFO 2021-03-19 19:04:29 train.py: 78] Epoch 14, iter 1200/6416, lr 0.001000, loss 3.128385
+INFO 2021-03-19 19:08:23 train.py: 78] Epoch 14, iter 1400/6416, lr 0.001000, loss 3.140274
+INFO 2021-03-19 19:12:17 train.py: 78] Epoch 14, iter 1600/6416, lr 0.001000, loss 3.130934
+INFO 2021-03-19 19:16:12 train.py: 78] Epoch 14, iter 1800/6416, lr 0.001000, loss 3.144010
+INFO 2021-03-19 19:20:06 train.py: 78] Epoch 14, iter 2000/6416, lr 0.001000, loss 3.135949
+INFO 2021-03-19 19:24:00 train.py: 78] Epoch 14, iter 2200/6416, lr 0.001000, loss 3.142621
+INFO 2021-03-19 19:27:54 train.py: 78] Epoch 14, iter 2400/6416, lr 0.001000, loss 3.136556
+INFO 2021-03-19 19:31:49 train.py: 78] Epoch 14, iter 2600/6416, lr 0.001000, loss 3.139964
+INFO 2021-03-19 19:35:43 train.py: 78] Epoch 14, iter 2800/6416, lr 0.001000, loss 3.142671
+INFO 2021-03-19 19:39:37 train.py: 91] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-03-19 19:39:38 train.py: 78] Epoch 14, iter 3000/6416, lr 0.001000, loss 3.147926
+INFO 2021-03-19 19:43:32 train.py: 78] Epoch 14, iter 3200/6416, lr 0.001000, loss 3.132282
+INFO 2021-03-19 19:47:27 train.py: 78] Epoch 14, iter 3400/6416, lr 0.001000, loss 3.138195
+INFO 2021-03-19 19:51:21 train.py: 78] Epoch 14, iter 3600/6416, lr 0.001000, loss 3.132258
+INFO 2021-03-19 19:55:15 train.py: 78] Epoch 14, iter 3800/6416, lr 0.001000, loss 3.140412
+INFO 2021-03-19 19:59:09 train.py: 78] Epoch 14, iter 4000/6416, lr 0.001000, loss 3.136316
+INFO 2021-03-19 20:03:03 train.py: 78] Epoch 14, iter 4200/6416, lr 0.001000, loss 3.147257
+INFO 2021-03-19 20:06:58 train.py: 78] Epoch 14, iter 4400/6416, lr 0.001000, loss 3.159377
+INFO 2021-03-19 20:10:52 train.py: 78] Epoch 14, iter 4600/6416, lr 0.001000, loss 3.138995
+INFO 2021-03-19 20:14:46 train.py: 78] Epoch 14, iter 4800/6416, lr 0.001000, loss 3.145203
+INFO 2021-03-19 20:18:40 train.py: 78] Epoch 14, iter 5000/6416, lr 0.001000, loss 3.139713
+INFO 2021-03-19 20:22:34 train.py: 78] Epoch 14, iter 5200/6416, lr 0.001000, loss 3.161223
+INFO 2021-03-19 20:26:29 train.py: 78] Epoch 14, iter 5400/6416, lr 0.001000, loss 3.154851
+INFO 2021-03-19 20:30:23 train.py: 78] Epoch 14, iter 5600/6416, lr 0.001000, loss 3.157712
+INFO 2021-03-19 20:34:17 train.py: 78] Epoch 14, iter 5800/6416, lr 0.001000, loss 3.130871
+INFO 2021-03-19 20:38:11 train.py: 91] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-03-19 20:38:12 train.py: 78] Epoch 14, iter 6000/6416, lr 0.001000, loss 3.145180
+INFO 2021-03-19 20:42:07 train.py: 78] Epoch 14, iter 6200/6416, lr 0.001000, loss 3.152377
+INFO 2021-03-19 20:46:01 train.py: 78] Epoch 14, iter 6400/6416, lr 0.001000, loss 3.140218
+INFO 2021-03-19 20:46:19 train.py: 96] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-03-19 20:46:22 train.py: 78] Epoch 15, iter 0/6416, lr 0.001000, loss 3.197044
+INFO 2021-03-19 20:50:16 train.py: 78] Epoch 15, iter 200/6416, lr 0.001000, loss 3.105113
+INFO 2021-03-19 20:54:10 train.py: 78] Epoch 15, iter 400/6416, lr 0.001000, loss 3.101389
+INFO 2021-03-19 20:58:04 train.py: 78] Epoch 15, iter 600/6416, lr 0.001000, loss 3.095628
+INFO 2021-03-19 21:01:59 train.py: 78] Epoch 15, iter 800/6416, lr 0.001000, loss 3.094118
+INFO 2021-03-19 21:05:53 train.py: 78] Epoch 15, iter 1000/6416, lr 0.001000, loss 3.100649
+INFO 2021-03-19 21:09:47 train.py: 78] Epoch 15, iter 1200/6416, lr 0.001000, loss 3.116633
+INFO 2021-03-19 21:13:41 train.py: 78] Epoch 15, iter 1400/6416, lr 0.001000, loss 3.104960
+INFO 2021-03-19 21:17:35 train.py: 78] Epoch 15, iter 1600/6416, lr 0.001000, loss 3.080331
+INFO 2021-03-19 21:21:30 train.py: 78] Epoch 15, iter 1800/6416, lr 0.001000, loss 3.128272
+INFO 2021-03-19 21:25:24 train.py: 78] Epoch 15, iter 2000/6416, lr 0.001000, loss 3.109825
+INFO 2021-03-19 21:29:18 train.py: 78] Epoch 15, iter 2200/6416, lr 0.001000, loss 3.122021
+INFO 2021-03-19 21:33:12 train.py: 78] Epoch 15, iter 2400/6416, lr 0.001000, loss 3.113386
+INFO 2021-03-19 21:37:07 train.py: 78] Epoch 15, iter 2600/6416, lr 0.001000, loss 3.087826
+INFO 2021-03-19 21:41:01 train.py: 78] Epoch 15, iter 2800/6416, lr 0.001000, loss 3.115891
+INFO 2021-03-19 21:44:55 train.py: 91] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-03-19 21:44:56 train.py: 78] Epoch 15, iter 3000/6416, lr 0.001000, loss 3.137241
+INFO 2021-03-19 21:48:51 train.py: 78] Epoch 15, iter 3200/6416, lr 0.001000, loss 3.104972
+INFO 2021-03-19 21:52:45 train.py: 78] Epoch 15, iter 3400/6416, lr 0.001000, loss 3.112512
+INFO 2021-03-19 21:56:39 train.py: 78] Epoch 15, iter 3600/6416, lr 0.001000, loss 3.113448
+INFO 2021-03-19 22:00:34 train.py: 78] Epoch 15, iter 3800/6416, lr 0.001000, loss 3.098038
+INFO 2021-03-19 22:04:28 train.py: 78] Epoch 15, iter 4000/6416, lr 0.001000, loss 3.118554
+INFO 2021-03-19 22:08:22 train.py: 78] Epoch 15, iter 4200/6416, lr 0.001000, loss 3.123196
+INFO 2021-03-19 22:12:16 train.py: 78] Epoch 15, iter 4400/6416, lr 0.001000, loss 3.115676
+INFO 2021-03-19 22:16:11 train.py: 78] Epoch 15, iter 4600/6416, lr 0.001000, loss 3.113284
+INFO 2021-03-19 22:20:05 train.py: 78] Epoch 15, iter 4800/6416, lr 0.001000, loss 3.121991
+INFO 2021-03-19 22:23:59 train.py: 78] Epoch 15, iter 5000/6416, lr 0.001000, loss 3.131621
+INFO 2021-03-19 22:27:53 train.py: 78] Epoch 15, iter 5200/6416, lr 0.001000, loss 3.132818
+INFO 2021-03-19 22:31:48 train.py: 78] Epoch 15, iter 5400/6416, lr 0.001000, loss 3.140174
+INFO 2021-03-19 22:35:42 train.py: 78] Epoch 15, iter 5600/6416, lr 0.001000, loss 3.130641
+INFO 2021-03-19 22:39:36 train.py: 78] Epoch 15, iter 5800/6416, lr 0.001000, loss 3.134250
+INFO 2021-03-19 22:43:30 train.py: 91] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-03-19 22:43:32 train.py: 78] Epoch 15, iter 6000/6416, lr 0.001000, loss 3.118152
+INFO 2021-03-19 22:47:26 train.py: 78] Epoch 15, iter 6200/6416, lr 0.001000, loss 3.137207
+INFO 2021-03-19 22:51:20 train.py: 78] Epoch 15, iter 6400/6416, lr 0.001000, loss 3.138534
+INFO 2021-03-19 22:51:38 train.py: 96] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-03-19 22:51:41 train.py: 78] Epoch 16, iter 0/6416, lr 0.000100, loss 3.155837
+INFO 2021-03-19 22:55:52 train.py: 78] Epoch 16, iter 200/6416, lr 0.000100, loss 3.052112
+INFO 2021-03-19 23:00:03 train.py: 78] Epoch 16, iter 400/6416, lr 0.000100, loss 3.036359
+INFO 2021-03-19 23:04:13 train.py: 78] Epoch 16, iter 600/6416, lr 0.000100, loss 3.068549
+INFO 2021-03-19 23:08:07 train.py: 78] Epoch 16, iter 800/6416, lr 0.000100, loss 3.071267
+INFO 2021-03-19 23:12:01 train.py: 78] Epoch 16, iter 1000/6416, lr 0.000100, loss 3.083179
+INFO 2021-03-19 23:15:55 train.py: 78] Epoch 16, iter 1200/6416, lr 0.000100, loss 3.093343
+INFO 2021-03-19 23:19:49 train.py: 78] Epoch 16, iter 1400/6416, lr 0.000100, loss 3.046113
+INFO 2021-03-19 23:23:44 train.py: 78] Epoch 16, iter 1600/6416, lr 0.000100, loss 3.063368
+INFO 2021-03-19 23:27:38 train.py: 78] Epoch 16, iter 1800/6416, lr 0.000100, loss 3.091686
+INFO 2021-03-19 23:31:32 train.py: 78] Epoch 16, iter 2000/6416, lr 0.000100, loss 3.071781
+INFO 2021-03-19 23:35:27 train.py: 78] Epoch 16, iter 2200/6416, lr 0.000100, loss 3.051654
+INFO 2021-03-19 23:39:21 train.py: 78] Epoch 16, iter 2400/6416, lr 0.000100, loss 3.035913
+INFO 2021-03-19 23:43:15 train.py: 78] Epoch 16, iter 2600/6416, lr 0.000100, loss 3.068752
+INFO 2021-03-19 23:47:09 train.py: 78] Epoch 16, iter 2800/6416, lr 0.000100, loss 3.060263
+INFO 2021-03-19 23:51:04 train.py: 91] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-03-19 23:51:05 train.py: 78] Epoch 16, iter 3000/6416, lr 0.000100, loss 3.072987
+INFO 2021-03-19 23:54:59 train.py: 78] Epoch 16, iter 3200/6416, lr 0.000100, loss 3.052553
+INFO 2021-03-19 23:58:53 train.py: 78] Epoch 16, iter 3400/6416, lr 0.000100, loss 3.071509
+INFO 2021-03-20 00:02:48 train.py: 78] Epoch 16, iter 3600/6416, lr 0.000100, loss 3.077967
+INFO 2021-03-20 00:06:42 train.py: 78] Epoch 16, iter 3800/6416, lr 0.000100, loss 3.064053
+INFO 2021-03-20 00:10:36 train.py: 78] Epoch 16, iter 4000/6416, lr 0.000100, loss 3.090127
+INFO 2021-03-20 00:14:30 train.py: 78] Epoch 16, iter 4200/6416, lr 0.000100, loss 3.100970
+INFO 2021-03-20 00:18:25 train.py: 78] Epoch 16, iter 4400/6416, lr 0.000100, loss 3.061124
+INFO 2021-03-20 00:22:19 train.py: 78] Epoch 16, iter 4600/6416, lr 0.000100, loss 3.074622
+INFO 2021-03-20 00:26:13 train.py: 78] Epoch 16, iter 4800/6416, lr 0.000100, loss 3.075948
+INFO 2021-03-20 00:30:07 train.py: 78] Epoch 16, iter 5000/6416, lr 0.000100, loss 3.073800
+INFO 2021-03-20 00:34:02 train.py: 78] Epoch 16, iter 5200/6416, lr 0.000100, loss 3.060539
+INFO 2021-03-20 00:37:56 train.py: 78] Epoch 16, iter 5400/6416, lr 0.000100, loss 3.074296
+INFO 2021-03-20 00:41:50 train.py: 78] Epoch 16, iter 5600/6416, lr 0.000100, loss 3.073779
+INFO 2021-03-20 00:45:44 train.py: 78] Epoch 16, iter 5800/6416, lr 0.000100, loss 3.063759
+INFO 2021-03-20 00:49:39 train.py: 91] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-03-20 00:49:40 train.py: 78] Epoch 16, iter 6000/6416, lr 0.000100, loss 3.066763
+INFO 2021-03-20 00:53:34 train.py: 78] Epoch 16, iter 6200/6416, lr 0.000100, loss 3.068334
+INFO 2021-03-20 00:57:28 train.py: 78] Epoch 16, iter 6400/6416, lr 0.000100, loss 3.059346
+INFO 2021-03-20 00:57:46 train.py: 96] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-03-20 00:57:49 train.py: 78] Epoch 17, iter 0/6416, lr 0.000100, loss 3.114026
+INFO 2021-03-20 01:01:43 train.py: 78] Epoch 17, iter 200/6416, lr 0.000100, loss 3.049200
+INFO 2021-03-20 01:05:38 train.py: 78] Epoch 17, iter 400/6416, lr 0.000100, loss 3.062540
+INFO 2021-03-20 01:09:32 train.py: 78] Epoch 17, iter 600/6416, lr 0.000100, loss 3.053811
+INFO 2021-03-20 01:13:26 train.py: 78] Epoch 17, iter 800/6416, lr 0.000100, loss 3.063780
+INFO 2021-03-20 01:17:20 train.py: 78] Epoch 17, iter 1000/6416, lr 0.000100, loss 3.074239
+INFO 2021-03-20 01:21:15 train.py: 78] Epoch 17, iter 1200/6416, lr 0.000100, loss 3.037428
+INFO 2021-03-20 01:25:09 train.py: 78] Epoch 17, iter 1400/6416, lr 0.000100, loss 3.060609
+INFO 2021-03-20 01:29:03 train.py: 78] Epoch 17, iter 1600/6416, lr 0.000100, loss 3.064894
+INFO 2021-03-20 01:32:58 train.py: 78] Epoch 17, iter 1800/6416, lr 0.000100, loss 3.062991
+INFO 2021-03-20 01:36:52 train.py: 78] Epoch 17, iter 2000/6416, lr 0.000100, loss 3.065635
+INFO 2021-03-20 01:40:46 train.py: 78] Epoch 17, iter 2200/6416, lr 0.000100, loss 3.061597
+INFO 2021-03-20 01:44:40 train.py: 78] Epoch 17, iter 2400/6416, lr 0.000100, loss 3.089202
+INFO 2021-03-20 01:48:35 train.py: 78] Epoch 17, iter 2600/6416, lr 0.000100, loss 3.064418
+INFO 2021-03-20 01:52:29 train.py: 78] Epoch 17, iter 2800/6416, lr 0.000100, loss 3.071868
+INFO 2021-03-20 01:56:23 train.py: 91] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-03-20 01:56:24 train.py: 78] Epoch 17, iter 3000/6416, lr 0.000100, loss 3.059083
+INFO 2021-03-20 02:00:18 train.py: 78] Epoch 17, iter 3200/6416, lr 0.000100, loss 3.092387
+INFO 2021-03-20 02:04:13 train.py: 78] Epoch 17, iter 3400/6416, lr 0.000100, loss 3.052788
+INFO 2021-03-20 02:08:07 train.py: 78] Epoch 17, iter 3600/6416, lr 0.000100, loss 3.051660
+INFO 2021-03-20 02:12:01 train.py: 78] Epoch 17, iter 3800/6416, lr 0.000100, loss 3.052828
+INFO 2021-03-20 02:15:55 train.py: 78] Epoch 17, iter 4000/6416, lr 0.000100, loss 3.070680
+INFO 2021-03-20 02:19:50 train.py: 78] Epoch 17, iter 4200/6416, lr 0.000100, loss 3.059567
+INFO 2021-03-20 02:23:44 train.py: 78] Epoch 17, iter 4400/6416, lr 0.000100, loss 3.081448
+INFO 2021-03-20 02:27:38 train.py: 78] Epoch 17, iter 4600/6416, lr 0.000100, loss 3.052200
+INFO 2021-03-20 02:31:33 train.py: 78] Epoch 17, iter 4800/6416, lr 0.000100, loss 3.069250
+INFO 2021-03-20 02:35:27 train.py: 78] Epoch 17, iter 5000/6416, lr 0.000100, loss 3.058106
+INFO 2021-03-20 02:39:21 train.py: 78] Epoch 17, iter 5200/6416, lr 0.000100, loss 3.076459
+INFO 2021-03-20 02:43:15 train.py: 78] Epoch 17, iter 5400/6416, lr 0.000100, loss 3.055265
+INFO 2021-03-20 02:47:10 train.py: 78] Epoch 17, iter 5600/6416, lr 0.000100, loss 3.058003
+INFO 2021-03-20 02:51:04 train.py: 78] Epoch 17, iter 5800/6416, lr 0.000100, loss 3.076422
+INFO 2021-03-20 02:54:58 train.py: 91] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-03-20 02:54:59 train.py: 78] Epoch 17, iter 6000/6416, lr 0.000100, loss 3.052855
+INFO 2021-03-20 02:58:53 train.py: 78] Epoch 17, iter 6200/6416, lr 0.000100, loss 3.080742
+INFO 2021-03-20 03:02:48 train.py: 78] Epoch 17, iter 6400/6416, lr 0.000100, loss 3.068288
+INFO 2021-03-20 03:03:06 train.py: 96] Save checkpoint Epoch_17.pt to disk...
+INFO 2021-03-20 03:03:06 train.py: 179] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_A0/.gitkeep b/bob/bio/facexzoo/models/backbones/RepVGG_A0/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..860eb1250c8442e45b35282ab7a778d6c8cad910
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+---------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+---------------------+-----------------------+
+| Epoch_11_batch_5999.pt | 0.9696666666666666 | 0.0025676044462869677 |
+| Epoch_17.pt | 0.9688333333333334 | 0.0024222477062879615 |
+| Epoch_15_batch_5999.pt | 0.9686666666666668 | 0.002673602092336881 |
+| Epoch_17_batch_2999.pt | 0.9683333333333334 | 0.002593915006650837 |
+| Epoch_13_batch_5999.pt | 0.9678333333333334 | 0.0023313483620397042 |
+| Epoch_17_batch_5999.pt | 0.9678333333333334 | 0.0023180717827805623 |
+| Epoch_15_batch_2999.pt | 0.9678333333333334 | 0.002318071782780565 |
+| Epoch_14.pt | 0.9678333333333333 | 0.0026063786901644255 |
+| Epoch_13.pt | 0.9676666666666668 | 0.0027464904654018346 |
+| Epoch_14_batch_5999.pt | 0.9676666666666668 | 0.002824058894919737 |
+| Epoch_15.pt | 0.9676666666666666 | 0.0027011657291429354 |
+| Epoch_10_batch_5999.pt | 0.967 | 0.003051006715054659 |
+| Epoch_16_batch_5999.pt | 0.9668333333333333 | 0.0023888888888888913 |
+| Epoch_16_batch_2999.pt | 0.9664999999999999 | 0.0027380492289670213 |
+| Epoch_11.pt | 0.9663333333333333 | 0.0031111111111111114 |
+| Epoch_10_batch_2999.pt | 0.9658333333333333 | 0.0028571979712497305 |
+| Epoch_14_batch_2999.pt | 0.9658333333333333 | 0.0029000851413640456 |
+| Epoch_16.pt | 0.9656666666666667 | 0.0030041124076879526 |
+| Epoch_13_batch_2999.pt | 0.9654999999999999 | 0.00345205252953466 |
+| Epoch_12_batch_2999.pt | 0.9654999999999999 | 0.002664929990046991 |
+| Epoch_12_batch_5999.pt | 0.9654999999999999 | 0.002606378690164421 |
+| Epoch_11_batch_2999.pt | 0.9653333333333334 | 0.003581502546952489 |
+| Epoch_12.pt | 0.9651666666666667 | 0.002477678124553087 |
+| Epoch_10.pt | 0.9641666666666667 | 0.0031254629286745 |
+| Epoch_7_batch_5999.pt | 0.9548333333333332 | 0.00293394353922465 |
+| Epoch_6_batch_5999.pt | 0.9543333333333333 | 0.003212629398844664 |
+| Epoch_8_batch_5999.pt | 0.9538333333333334 | 0.0029922740021100518 |
+| Epoch_9_batch_5999.pt | 0.9530000000000001 | 0.004200235149207952 |
+| Epoch_8.pt | 0.953 | 0.003081205471969347 |
+| Epoch_9_batch_2999.pt | 0.9523333333333334 | 0.0038904758666692407 |
+| Epoch_5_batch_5999.pt | 0.9515 | 0.0038123418809550558 |
+| Epoch_5_batch_2999.pt | 0.951 | 0.0019907192074632117 |
+| Epoch_4_batch_2999.pt | 0.9501666666666667 | 0.004078322700496209 |
+| Epoch_8_batch_2999.pt | 0.9498333333333333 | 0.0036637193541772324 |
+| Epoch_7_batch_2999.pt | 0.9496666666666667 | 0.0039031484600556203 |
+| Epoch_7.pt | 0.9491666666666667 | 0.004895513188902268 |
+| Epoch_4_batch_5999.pt | 0.9471666666666666 | 0.004759979769642654 |
+| Epoch_3_batch_5999.pt | 0.9466666666666667 | 0.004416579314300394 |
+| Epoch_6.pt | 0.9465 | 0.004100963451991244 |
+| Epoch_6_batch_2999.pt | 0.9455 | 0.0038574122970755445 |
+| Epoch_3_batch_2999.pt | 0.9445 | 0.004805154125815264 |
+| Epoch_5.pt | 0.943 | 0.004546060565661946 |
+| Epoch_2_batch_2999.pt | 0.9385 | 0.004597704741377906 |
+| Epoch_2_batch_5999.pt | 0.9359999999999999 | 0.004976487928124147 |
+| Epoch_3.pt | 0.9336666666666666 | 0.0034854193647462475 |
+| Epoch_2.pt | 0.9321666666666667 | 0.004090413363155542 |
+| Epoch_1_batch_5999.pt | 0.9218333333333332 | 0.005829098992273884 |
+| Epoch_1_batch_2999.pt | 0.9021666666666667 | 0.006455211312297978 |
+| Epoch_0.pt | 0.8541666666666666 | 0.006162661361922171 |
+| Epoch_0_batch_5999.pt | 0.8468333333333333 | 0.007159772739341133 |
+| Epoch_0_batch_2999.pt | 0.71 | 0.006270644915343315 |
+| Epoch_4.pt | 0.6346666666666667 | 0.009777777777777778 |
+| Epoch_1.pt | 0.6015 | 0.007438845323911927 |
+| Epoch_9.pt | 0.49700000000000005 | 0.003918931575232977 |
++------------------------+---------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..8c11c896ca843734f70aca1c65e858872b23a5e4
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13.pt | 0.9488333333333333 | 0.003201562118716419 |
+| Epoch_17_batch_5999.pt | 0.9488333333333333 | 0.0031919072049384927 |
+| Epoch_14_batch_5999.pt | 0.9488333333333333 | 0.003162765628331171 |
+| Epoch_16_batch_5999.pt | 0.9484999999999999 | 0.002880864924992029 |
+| Epoch_17_batch_2999.pt | 0.9483333333333333 | 0.003142696805273541 |
+| Epoch_15.pt | 0.9481666666666667 | 0.0029234049148717453 |
+| Epoch_14_batch_2999.pt | 0.9481666666666667 | 0.002912828161981845 |
+| Epoch_15_batch_2999.pt | 0.9481666666666667 | 0.003037319318408135 |
+| Epoch_17.pt | 0.9480000000000001 | 0.0030205060486818156 |
+| Epoch_15_batch_5999.pt | 0.9480000000000001 | 0.002841491522787644 |
+| Epoch_13_batch_5999.pt | 0.9478333333333333 | 0.0028979558568322165 |
+| Epoch_16_batch_2999.pt | 0.9478333333333333 | 0.003043410205511624 |
+| Epoch_12_batch_5999.pt | 0.9473333333333332 | 0.0025239592648001225 |
+| Epoch_13_batch_2999.pt | 0.9471666666666667 | 0.0029297326385411553 |
+| Epoch_11.pt | 0.9471666666666666 | 0.0030025709148603693 |
+| Epoch_14.pt | 0.9470000000000001 | 0.0029165343885348177 |
+| Epoch_16.pt | 0.9470000000000001 | 0.0028631330503833563 |
+| Epoch_12_batch_2999.pt | 0.9468333333333334 | 0.0034556270089075554 |
+| Epoch_10_batch_5999.pt | 0.9465 | 0.0027938424357067033 |
+| Epoch_12.pt | 0.9463333333333332 | 0.0024944382578492917 |
+| Epoch_11_batch_5999.pt | 0.9461666666666666 | 0.0033888888888888905 |
+| Epoch_10.pt | 0.946 | 0.0035763282087624615 |
+| Epoch_11_batch_2999.pt | 0.9455 | 0.003352261075899019 |
+| Epoch_10_batch_2999.pt | 0.9438333333333334 | 0.0030434102055116306 |
+| Epoch_7_batch_5999.pt | 0.9366666666666668 | 0.0031817380140614104 |
+| Epoch_9_batch_5999.pt | 0.9355 | 0.0039287638240526985 |
+| Epoch_8_batch_5999.pt | 0.9343333333333333 | 0.003834540872608308 |
+| Epoch_8_batch_2999.pt | 0.9338333333333333 | 0.002929732638541159 |
+| Epoch_6_batch_5999.pt | 0.9331666666666667 | 0.003107636949099329 |
+| Epoch_7_batch_2999.pt | 0.9321666666666667 | 0.0041129875597510236 |
+| Epoch_9_batch_2999.pt | 0.9321666666666667 | 0.003370624736026117 |
+| Epoch_6_batch_2999.pt | 0.9321666666666666 | 0.0031333530338230127 |
+| Epoch_8.pt | 0.9318333333333333 | 0.004085883557377088 |
+| Epoch_5_batch_2999.pt | 0.9313333333333332 | 0.003911047979288457 |
+| Epoch_4_batch_2999.pt | 0.9301666666666666 | 0.0032626770800057788 |
+| Epoch_3_batch_5999.pt | 0.9295 | 0.003920899997465308 |
+| Epoch_5_batch_5999.pt | 0.9291666666666666 | 0.0029000851413640352 |
+| Epoch_4_batch_5999.pt | 0.9279999999999999 | 0.0030408738185342264 |
+| Epoch_3_batch_2999.pt | 0.9271666666666667 | 0.0022505143445032366 |
+| Epoch_7.pt | 0.9269999999999999 | 0.0038393672318157773 |
+| Epoch_5.pt | 0.9233333333333332 | 0.0029710537682490906 |
+| Epoch_6.pt | 0.9233333333333332 | 0.003975231959999624 |
+| Epoch_2_batch_5999.pt | 0.9219999999999999 | 0.003368334753605362 |
+| Epoch_2_batch_2999.pt | 0.9191666666666668 | 0.0034179085367271166 |
+| Epoch_2.pt | 0.9188333333333333 | 0.002981941533404362 |
+| Epoch_3.pt | 0.9135 | 0.0035000000000000005 |
+| Epoch_1_batch_5999.pt | 0.9133333333333334 | 0.0032107073948444156 |
+| Epoch_1_batch_2999.pt | 0.9018333333333333 | 0.004299368314210191 |
+| Epoch_0.pt | 0.8636666666666667 | 0.0038554114606438837 |
+| Epoch_0_batch_5999.pt | 0.8628333333333333 | 0.004779392492000697 |
+| Epoch_4.pt | 0.8011666666666667 | 0.004635144863025807 |
+| Epoch_0_batch_2999.pt | 0.7271666666666666 | 0.00758511375111039 |
+| Epoch_9.pt | 0.5111666666666668 | 0.0057951128835712394 |
+| Epoch_1.pt | 0.5055 | 0.0022777777777777783 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3623abc793dbbd4c3ed33183feb49b3f4428b96f
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.8543333333333333 | 0.005704665623552635 |
+| Epoch_13_batch_2999.pt | 0.852 | 0.006160407233612938 |
+| Epoch_17_batch_2999.pt | 0.852 | 0.006115150180363115 |
+| Epoch_14.pt | 0.8513333333333334 | 0.006150378890265803 |
+| Epoch_16.pt | 0.8511666666666666 | 0.005305773701158132 |
+| Epoch_16_batch_5999.pt | 0.851 | 0.006162410943938684 |
+| Epoch_11.pt | 0.8505 | 0.005432242562630949 |
+| Epoch_15_batch_2999.pt | 0.8503333333333334 | 0.006195378604200292 |
+| Epoch_16_batch_2999.pt | 0.85 | 0.0059524867715461645 |
+| Epoch_15_batch_5999.pt | 0.8496666666666666 | 0.006534807090014498 |
+| Epoch_17_batch_5999.pt | 0.8494999999999999 | 0.00569302150243265 |
+| Epoch_11_batch_2999.pt | 0.8491666666666665 | 0.005450393575096828 |
+| Epoch_13_batch_5999.pt | 0.8484999999999999 | 0.006404329168520728 |
+| Epoch_14_batch_2999.pt | 0.8484999999999999 | 0.007002865374744126 |
+| Epoch_14_batch_5999.pt | 0.8483333333333333 | 0.005910860479231282 |
+| Epoch_13.pt | 0.8481666666666667 | 0.0056078098099166624 |
+| Epoch_17.pt | 0.8476666666666668 | 0.005785251576419918 |
+| Epoch_11_batch_5999.pt | 0.8473333333333333 | 0.005869990640832128 |
+| Epoch_12_batch_2999.pt | 0.8461666666666667 | 0.006319917564706943 |
+| Epoch_10.pt | 0.8458333333333334 | 0.005529103693576472 |
+| Epoch_12.pt | 0.8453333333333333 | 0.0054421763983370924 |
+| Epoch_12_batch_5999.pt | 0.845 | 0.00564921058617163 |
+| Epoch_10_batch_5999.pt | 0.8441666666666666 | 0.005623199300253408 |
+| Epoch_10_batch_2999.pt | 0.8428333333333333 | 0.0056000992054704764 |
+| Epoch_7_batch_5999.pt | 0.8201666666666666 | 0.00636565824862375 |
+| Epoch_7_batch_2999.pt | 0.8195 | 0.006334551539760054 |
+| Epoch_3_batch_2999.pt | 0.8181666666666667 | 0.006026945667091587 |
+| Epoch_8_batch_2999.pt | 0.8178333333333334 | 0.007625695677635006 |
+| Epoch_9_batch_2999.pt | 0.8166666666666668 | 0.0059421075364982 |
+| Epoch_9_batch_5999.pt | 0.8160000000000001 | 0.005123174170399334 |
+| Epoch_8_batch_5999.pt | 0.8160000000000001 | 0.007028162044830432 |
+| Epoch_6_batch_5999.pt | 0.8158333333333333 | 0.006957765002294951 |
+| Epoch_3_batch_5999.pt | 0.8106666666666668 | 0.00739202105710218 |
+| Epoch_5_batch_2999.pt | 0.8105 | 0.007474442049452075 |
+| Epoch_8.pt | 0.8101666666666667 | 0.006011562932290483 |
+| Epoch_6_batch_2999.pt | 0.8095000000000001 | 0.006634412716945294 |
+| Epoch_4_batch_5999.pt | 0.8086666666666666 | 0.007440297352348085 |
+| Epoch_5_batch_5999.pt | 0.8081666666666667 | 0.007155460661896663 |
+| Epoch_4_batch_2999.pt | 0.8063333333333332 | 0.007109374788020616 |
+| Epoch_2_batch_2999.pt | 0.8049999999999999 | 0.007005289007176944 |
+| Epoch_7.pt | 0.8 | 0.0059886724347608285 |
+| Epoch_2_batch_5999.pt | 0.7981666666666667 | 0.006612978261077342 |
+| Epoch_5.pt | 0.7976666666666666 | 0.006560261964055785 |
+| Epoch_6.pt | 0.7928333333333335 | 0.006305249625471317 |
+| Epoch_3.pt | 0.7906666666666666 | 0.007197393218089618 |
+| Epoch_2.pt | 0.7905 | 0.006726812023536861 |
+| Epoch_1_batch_5999.pt | 0.7896666666666666 | 0.005565546571719193 |
+| Epoch_1_batch_2999.pt | 0.7665 | 0.006788015946786453 |
+| Epoch_0.pt | 0.7178333333333333 | 0.008621141112061326 |
+| Epoch_0_batch_5999.pt | 0.7135 | 0.00945571223267505 |
+| Epoch_4.pt | 0.6858333333333334 | 0.008865420408099428 |
+| Epoch_0_batch_2999.pt | 0.6245 | 0.0071666666666666675 |
+| Epoch_9.pt | 0.5003333333333334 | 0.007859884081701065 |
+| Epoch_1.pt | 0.4973333333333332 | 0.007520342781785655 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..47eda47b800a91db1c4b72547dc20c1f46ebe7be
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11_batch_2999.pt | 0.9976666666666667 | 0.0008314794192831014 |
+| Epoch_14_batch_2999.pt | 0.9976666666666667 | 0.0008314794192831014 |
+| Epoch_12_batch_5999.pt | 0.9976666666666667 | 0.0008314794192831014 |
+| Epoch_13_batch_2999.pt | 0.9974999999999999 | 0.0007954345035153557 |
+| Epoch_13_batch_5999.pt | 0.9974999999999999 | 0.0009378857231185614 |
+| Epoch_15.pt | 0.9974999999999999 | 0.0007954345035153557 |
+| Epoch_17.pt | 0.9974999999999999 | 0.0009378857231185614 |
+| Epoch_14.pt | 0.9974999999999999 | 0.0009378857231185614 |
+| Epoch_17_batch_5999.pt | 0.9974999999999999 | 0.0009378857231185614 |
+| Epoch_15_batch_2999.pt | 0.9974999999999999 | 0.0009378857231185614 |
+| Epoch_11_batch_5999.pt | 0.9974999999999999 | 0.0008695819912499218 |
+| Epoch_14_batch_5999.pt | 0.9974999999999999 | 0.0007954345035153557 |
+| Epoch_17_batch_2999.pt | 0.9973333333333333 | 0.0009686442096757043 |
+| Epoch_16_batch_5999.pt | 0.9973333333333333 | 0.0010599324460188297 |
+| Epoch_10_batch_2999.pt | 0.9973333333333333 | 0.000831479419283101 |
+| Epoch_16_batch_2999.pt | 0.9973333333333333 | 0.0010599324460188297 |
+| Epoch_16.pt | 0.9973333333333333 | 0.000831479419283101 |
+| Epoch_15_batch_5999.pt | 0.9971666666666668 | 0.001084401183107953 |
+| Epoch_10_batch_5999.pt | 0.9971666666666668 | 0.0008258927081843643 |
+| Epoch_12_batch_2999.pt | 0.9971666666666665 | 0.0009312808119022322 |
+| Epoch_12.pt | 0.9970000000000001 | 0.0011600340565456147 |
+| Epoch_13.pt | 0.9969999999999999 | 0.0010482201257840677 |
+| Epoch_11.pt | 0.9968333333333333 | 0.0010957268290731099 |
+| Epoch_10.pt | 0.9966666666666667 | 0.0007453559924999334 |
+| Epoch_7_batch_5999.pt | 0.9961666666666666 | 0.000611111111111113 |
+| Epoch_9_batch_2999.pt | 0.9958333333333332 | 0.0012484558363469042 |
+| Epoch_8_batch_5999.pt | 0.9951666666666668 | 0.0010076865081787238 |
+| Epoch_8_batch_2999.pt | 0.9951666666666668 | 0.0011772011166898395 |
+| Epoch_5_batch_2999.pt | 0.9951666666666666 | 0.0009111788592698187 |
+| Epoch_8.pt | 0.9951666666666666 | 0.0007222222222222221 |
+| Epoch_7_batch_2999.pt | 0.9950000000000001 | 0.0009938079899999047 |
+| Epoch_5_batch_5999.pt | 0.9948333333333335 | 0.0010378634273483023 |
+| Epoch_7.pt | 0.9948333333333332 | 0.001203133768205988 |
+| Epoch_4_batch_5999.pt | 0.9946666666666667 | 0.0009558139185602859 |
+| Epoch_9_batch_5999.pt | 0.9946666666666666 | 0.0011331154474650677 |
+| Epoch_3_batch_2999.pt | 0.9944999999999998 | 0.001112499133027825 |
+| Epoch_6_batch_2999.pt | 0.9943333333333333 | 0.001247219128924641 |
+| Epoch_5.pt | 0.9943333333333332 | 0.0010599324460188284 |
+| Epoch_6_batch_5999.pt | 0.9941666666666666 | 0.0007556372504853021 |
+| Epoch_2.pt | 0.9936666666666666 | 0.001160034056545618 |
+| Epoch_6.pt | 0.9936666666666666 | 0.0011863420280034795 |
+| Epoch_3.pt | 0.9933333333333334 | 0.0011653431646335057 |
+| Epoch_3_batch_5999.pt | 0.9928333333333332 | 0.0010555555555555516 |
+| Epoch_2_batch_5999.pt | 0.9925 | 0.0010318986456114916 |
+| Epoch_4_batch_2999.pt | 0.9925 | 0.0015163715626618003 |
+| Epoch_2_batch_2999.pt | 0.992 | 0.0009229582069909026 |
+| Epoch_1_batch_5999.pt | 0.991 | 0.0014740554623801818 |
+| Epoch_1_batch_2999.pt | 0.9861666666666669 | 0.0018600743380870212 |
+| Epoch_0_batch_5999.pt | 0.9730000000000001 | 0.002905932629027115 |
+| Epoch_0.pt | 0.9730000000000001 | 0.002419060117453028 |
+| Epoch_4.pt | 0.9675 | 0.0022804861946415325 |
+| Epoch_0_batch_2999.pt | 0.9318333333333333 | 0.003655285366576881 |
+| Epoch_1.pt | 0.6048333333333333 | 0.006514229439096016 |
+| Epoch_9.pt | 0.5583333333333333 | 0.006004113816047244 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..fc5b1b81ce4787207c2eb7049a04ae84b7336026
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_A0/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.9440032805238424 |
+| 2 | 0.9563247718603954 |
+| 3 | 0.9617272218389159 |
+| 4 | 0.9648385123084734 |
+| 5 | 0.9669409115169819 |
+| 6 | 0.9687373888592369 |
+| 7 | 0.9701823815040942 |
+| 8 | 0.9713279612585756 |
+| 9 | 0.9723043076402359 |
+| 10 | 0.9729812411315204 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_A0/log.log b/bob/bio/facexzoo/models/backbones/RepVGG_A0/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..e2f69db5a227c725349652451e3e986f4725154d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_A0/log.log
@@ -0,0 +1,655 @@
+INFO 2021-09-18 23:45:55 train.py: 180] Start optimization.
+INFO 2021-09-18 23:45:55 train.py: 181] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='RepVGG', batch_size=512, data_root='/export2/wj_data/FaceX-Zoo/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-repvgg', train_file='/export2/wj_data/FaceX-Zoo/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7ff0d6840e80>)
+backbone param:
+{'block_stage1': 2, 'block_stage2': 4, 'block_stage3': 14, 'block_stage4': 1, 'width_stage1': 0.75, 'width_stage2': 0.75, 'width_stage3': 0.75, 'width_stage4': 2.5, 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2021-09-18 23:46:17 train.py: 82] Epoch 0, iter 0/6416, lr 0.100000, loss 16.353008
+INFO 2021-09-18 23:48:49 train.py: 82] Epoch 0, iter 200/6416, lr 0.100000, loss 15.731858
+INFO 2021-09-18 23:52:29 train.py: 82] Epoch 0, iter 400/6416, lr 0.100000, loss 15.398714
+INFO 2021-09-18 23:55:58 train.py: 82] Epoch 0, iter 600/6416, lr 0.100000, loss 15.288831
+INFO 2021-09-18 23:59:52 train.py: 82] Epoch 0, iter 800/6416, lr 0.100000, loss 15.139525
+INFO 2021-09-19 00:03:46 train.py: 82] Epoch 0, iter 1000/6416, lr 0.100000, loss 14.951804
+INFO 2021-09-19 00:07:37 train.py: 82] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.688585
+INFO 2021-09-19 00:11:27 train.py: 82] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.384985
+INFO 2021-09-19 00:15:35 train.py: 82] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.068007
+INFO 2021-09-19 00:19:20 train.py: 82] Epoch 0, iter 1800/6416, lr 0.100000, loss 13.735417
+INFO 2021-09-19 00:22:50 train.py: 82] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.378092
+INFO 2021-09-19 00:26:46 train.py: 82] Epoch 0, iter 2200/6416, lr 0.100000, loss 13.019949
+INFO 2021-09-19 00:30:35 train.py: 82] Epoch 0, iter 2400/6416, lr 0.100000, loss 12.640061
+INFO 2021-09-19 00:34:28 train.py: 82] Epoch 0, iter 2600/6416, lr 0.100000, loss 12.300912
+INFO 2021-09-19 00:38:23 train.py: 82] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.022937
+INFO 2021-09-19 00:42:14 train.py: 95] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-09-19 00:42:14 train.py: 82] Epoch 0, iter 3000/6416, lr 0.100000, loss 11.875123
+INFO 2021-09-19 00:45:50 train.py: 82] Epoch 0, iter 3200/6416, lr 0.100000, loss 11.820335
+INFO 2021-09-19 00:49:47 train.py: 82] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.878041
+INFO 2021-09-19 00:53:36 train.py: 82] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.044552
+INFO 2021-09-19 00:57:38 train.py: 82] Epoch 0, iter 3800/6416, lr 0.100000, loss 12.299685
+INFO 2021-09-19 01:01:13 train.py: 82] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.550010
+INFO 2021-09-19 01:05:16 train.py: 82] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.808782
+INFO 2021-09-19 01:08:58 train.py: 82] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.042086
+INFO 2021-09-19 01:12:37 train.py: 82] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.175337
+INFO 2021-09-19 01:16:21 train.py: 82] Epoch 0, iter 4800/6416, lr 0.100000, loss 13.318335
+INFO 2021-09-19 01:20:08 train.py: 82] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.428427
+INFO 2021-09-19 01:23:32 train.py: 82] Epoch 0, iter 5200/6416, lr 0.100000, loss 13.422494
+INFO 2021-09-19 01:27:30 train.py: 82] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.452501
+INFO 2021-09-19 01:31:27 train.py: 82] Epoch 0, iter 5600/6416, lr 0.100000, loss 13.369404
+INFO 2021-09-19 01:35:01 train.py: 82] Epoch 0, iter 5800/6416, lr 0.100000, loss 13.261045
+INFO 2021-09-19 01:39:14 train.py: 95] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-09-19 01:39:14 train.py: 82] Epoch 0, iter 6000/6416, lr 0.100000, loss 13.097086
+INFO 2021-09-19 01:42:54 train.py: 82] Epoch 0, iter 6200/6416, lr 0.100000, loss 12.957289
+INFO 2021-09-19 01:46:18 train.py: 82] Epoch 0, iter 6400/6416, lr 0.100000, loss 12.764238
+INFO 2021-09-19 01:46:39 train.py: 100] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-09-19 01:46:40 train.py: 82] Epoch 1, iter 0/6416, lr 0.100000, loss 12.618020
+INFO 2021-09-19 01:47:53 train.py: 82] Epoch 1, iter 200/6416, lr 0.100000, loss 12.325811
+INFO 2021-09-19 01:49:04 train.py: 82] Epoch 1, iter 400/6416, lr 0.100000, loss 12.102500
+INFO 2021-09-19 01:50:15 train.py: 82] Epoch 1, iter 600/6416, lr 0.100000, loss 11.914946
+INFO 2021-09-19 01:51:26 train.py: 82] Epoch 1, iter 800/6416, lr 0.100000, loss 11.784229
+INFO 2021-09-19 01:52:37 train.py: 82] Epoch 1, iter 1000/6416, lr 0.100000, loss 11.561483
+INFO 2021-09-19 01:53:47 train.py: 82] Epoch 1, iter 1200/6416, lr 0.100000, loss 11.398685
+INFO 2021-09-19 01:54:57 train.py: 82] Epoch 1, iter 1400/6416, lr 0.100000, loss 11.213444
+INFO 2021-09-19 01:56:08 train.py: 82] Epoch 1, iter 1600/6416, lr 0.100000, loss 11.016647
+INFO 2021-09-19 01:57:18 train.py: 82] Epoch 1, iter 1800/6416, lr 0.100000, loss 10.871447
+INFO 2021-09-19 01:58:29 train.py: 82] Epoch 1, iter 2000/6416, lr 0.100000, loss 10.704226
+INFO 2021-09-19 01:59:39 train.py: 82] Epoch 1, iter 2200/6416, lr 0.100000, loss 10.543172
+INFO 2021-09-19 02:00:50 train.py: 82] Epoch 1, iter 2400/6416, lr 0.100000, loss 10.367924
+INFO 2021-09-19 02:02:00 train.py: 82] Epoch 1, iter 2600/6416, lr 0.100000, loss 10.193261
+INFO 2021-09-19 02:03:11 train.py: 82] Epoch 1, iter 2800/6416, lr 0.100000, loss 10.093986
+INFO 2021-09-19 02:04:23 train.py: 95] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-09-19 02:04:23 train.py: 82] Epoch 1, iter 3000/6416, lr 0.100000, loss 9.908404
+INFO 2021-09-19 02:05:34 train.py: 82] Epoch 1, iter 3200/6416, lr 0.100000, loss 9.804338
+INFO 2021-09-19 02:06:44 train.py: 82] Epoch 1, iter 3400/6416, lr 0.100000, loss 9.675282
+INFO 2021-09-19 02:07:55 train.py: 82] Epoch 1, iter 3600/6416, lr 0.100000, loss 9.550288
+INFO 2021-09-19 02:09:06 train.py: 82] Epoch 1, iter 3800/6416, lr 0.100000, loss 9.452527
+INFO 2021-09-19 02:10:17 train.py: 82] Epoch 1, iter 4000/6416, lr 0.100000, loss 9.356773
+INFO 2021-09-19 02:11:28 train.py: 82] Epoch 1, iter 4200/6416, lr 0.100000, loss 9.235131
+INFO 2021-09-19 02:12:39 train.py: 82] Epoch 1, iter 4400/6416, lr 0.100000, loss 9.129302
+INFO 2021-09-19 02:13:50 train.py: 82] Epoch 1, iter 4600/6416, lr 0.100000, loss 9.039162
+INFO 2021-09-19 02:15:01 train.py: 82] Epoch 1, iter 4800/6416, lr 0.100000, loss 8.975147
+INFO 2021-09-19 02:16:12 train.py: 82] Epoch 1, iter 5000/6416, lr 0.100000, loss 8.865282
+INFO 2021-09-19 02:17:22 train.py: 82] Epoch 1, iter 5200/6416, lr 0.100000, loss 8.826673
+INFO 2021-09-19 02:18:33 train.py: 82] Epoch 1, iter 5400/6416, lr 0.100000, loss 8.713062
+INFO 2021-09-19 02:19:44 train.py: 82] Epoch 1, iter 5600/6416, lr 0.100000, loss 8.662784
+INFO 2021-09-19 02:20:55 train.py: 82] Epoch 1, iter 5800/6416, lr 0.100000, loss 8.612908
+INFO 2021-09-19 02:22:07 train.py: 95] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-09-19 02:22:07 train.py: 82] Epoch 1, iter 6000/6416, lr 0.100000, loss 8.511197
+INFO 2021-09-19 02:23:18 train.py: 82] Epoch 1, iter 6200/6416, lr 0.100000, loss 8.440990
+INFO 2021-09-19 02:24:29 train.py: 82] Epoch 1, iter 6400/6416, lr 0.100000, loss 8.394171
+INFO 2021-09-19 02:24:36 train.py: 100] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-09-19 02:24:37 train.py: 82] Epoch 2, iter 0/6416, lr 0.100000, loss 8.303778
+INFO 2021-09-19 02:25:48 train.py: 82] Epoch 2, iter 200/6416, lr 0.100000, loss 7.826875
+INFO 2021-09-19 02:26:59 train.py: 82] Epoch 2, iter 400/6416, lr 0.100000, loss 7.754892
+INFO 2021-09-19 02:28:09 train.py: 82] Epoch 2, iter 600/6416, lr 0.100000, loss 7.789363
+INFO 2021-09-19 02:29:19 train.py: 82] Epoch 2, iter 800/6416, lr 0.100000, loss 7.848291
+INFO 2021-09-19 02:30:29 train.py: 82] Epoch 2, iter 1000/6416, lr 0.100000, loss 7.857591
+INFO 2021-09-19 02:31:39 train.py: 82] Epoch 2, iter 1200/6416, lr 0.100000, loss 7.864572
+INFO 2021-09-19 02:32:49 train.py: 82] Epoch 2, iter 1400/6416, lr 0.100000, loss 7.843249
+INFO 2021-09-19 02:34:00 train.py: 82] Epoch 2, iter 1600/6416, lr 0.100000, loss 7.827703
+INFO 2021-09-19 02:35:09 train.py: 82] Epoch 2, iter 1800/6416, lr 0.100000, loss 7.804352
+INFO 2021-09-19 02:36:19 train.py: 82] Epoch 2, iter 2000/6416, lr 0.100000, loss 7.781755
+INFO 2021-09-19 02:37:29 train.py: 82] Epoch 2, iter 2200/6416, lr 0.100000, loss 7.746896
+INFO 2021-09-19 02:38:39 train.py: 82] Epoch 2, iter 2400/6416, lr 0.100000, loss 7.713318
+INFO 2021-09-19 02:39:49 train.py: 82] Epoch 2, iter 2600/6416, lr 0.100000, loss 7.728170
+INFO 2021-09-19 02:41:00 train.py: 82] Epoch 2, iter 2800/6416, lr 0.100000, loss 7.683657
+INFO 2021-09-19 02:42:11 train.py: 95] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-09-19 02:42:12 train.py: 82] Epoch 2, iter 3000/6416, lr 0.100000, loss 7.664109
+INFO 2021-09-19 02:43:22 train.py: 82] Epoch 2, iter 3200/6416, lr 0.100000, loss 7.604242
+INFO 2021-09-19 02:44:33 train.py: 82] Epoch 2, iter 3400/6416, lr 0.100000, loss 7.596752
+INFO 2021-09-19 02:45:43 train.py: 82] Epoch 2, iter 3600/6416, lr 0.100000, loss 7.569513
+INFO 2021-09-19 02:46:54 train.py: 82] Epoch 2, iter 3800/6416, lr 0.100000, loss 7.536128
+INFO 2021-09-19 02:48:04 train.py: 82] Epoch 2, iter 4000/6416, lr 0.100000, loss 7.497804
+INFO 2021-09-19 02:49:15 train.py: 82] Epoch 2, iter 4200/6416, lr 0.100000, loss 7.486746
+INFO 2021-09-19 02:50:25 train.py: 82] Epoch 2, iter 4400/6416, lr 0.100000, loss 7.450782
+INFO 2021-09-19 02:51:36 train.py: 82] Epoch 2, iter 4600/6416, lr 0.100000, loss 7.432486
+INFO 2021-09-19 02:52:46 train.py: 82] Epoch 2, iter 4800/6416, lr 0.100000, loss 7.430892
+INFO 2021-09-19 02:53:57 train.py: 82] Epoch 2, iter 5000/6416, lr 0.100000, loss 7.352362
+INFO 2021-09-19 02:55:07 train.py: 82] Epoch 2, iter 5200/6416, lr 0.100000, loss 7.311288
+INFO 2021-09-19 02:56:18 train.py: 82] Epoch 2, iter 5400/6416, lr 0.100000, loss 7.339102
+INFO 2021-09-19 02:57:28 train.py: 82] Epoch 2, iter 5600/6416, lr 0.100000, loss 7.269108
+INFO 2021-09-19 02:58:39 train.py: 82] Epoch 2, iter 5800/6416, lr 0.100000, loss 7.225426
+INFO 2021-09-19 02:59:51 train.py: 95] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-09-19 02:59:51 train.py: 82] Epoch 2, iter 6000/6416, lr 0.100000, loss 7.286758
+INFO 2021-09-19 03:01:02 train.py: 82] Epoch 2, iter 6200/6416, lr 0.100000, loss 7.246595
+INFO 2021-09-19 03:02:12 train.py: 82] Epoch 2, iter 6400/6416, lr 0.100000, loss 7.206277
+INFO 2021-09-19 03:02:20 train.py: 100] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-09-19 03:02:21 train.py: 82] Epoch 3, iter 0/6416, lr 0.100000, loss 7.128949
+INFO 2021-09-19 03:03:33 train.py: 82] Epoch 3, iter 200/6416, lr 0.100000, loss 6.581805
+INFO 2021-09-19 03:04:44 train.py: 82] Epoch 3, iter 400/6416, lr 0.100000, loss 6.584225
+INFO 2021-09-19 03:05:54 train.py: 82] Epoch 3, iter 600/6416, lr 0.100000, loss 6.645071
+INFO 2021-09-19 03:07:04 train.py: 82] Epoch 3, iter 800/6416, lr 0.100000, loss 6.736682
+INFO 2021-09-19 03:08:14 train.py: 82] Epoch 3, iter 1000/6416, lr 0.100000, loss 6.768208
+INFO 2021-09-19 03:09:24 train.py: 82] Epoch 3, iter 1200/6416, lr 0.100000, loss 6.847166
+INFO 2021-09-19 03:10:35 train.py: 82] Epoch 3, iter 1400/6416, lr 0.100000, loss 6.866416
+INFO 2021-09-19 03:11:45 train.py: 82] Epoch 3, iter 1600/6416, lr 0.100000, loss 6.847016
+INFO 2021-09-19 03:12:55 train.py: 82] Epoch 3, iter 1800/6416, lr 0.100000, loss 6.855584
+INFO 2021-09-19 03:14:06 train.py: 82] Epoch 3, iter 2000/6416, lr 0.100000, loss 6.812305
+INFO 2021-09-19 03:15:16 train.py: 82] Epoch 3, iter 2200/6416, lr 0.100000, loss 6.839051
+INFO 2021-09-19 03:16:26 train.py: 82] Epoch 3, iter 2400/6416, lr 0.100000, loss 6.869794
+INFO 2021-09-19 03:17:37 train.py: 82] Epoch 3, iter 2600/6416, lr 0.100000, loss 6.836675
+INFO 2021-09-19 03:18:47 train.py: 82] Epoch 3, iter 2800/6416, lr 0.100000, loss 6.818048
+INFO 2021-09-19 03:19:59 train.py: 95] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-09-19 03:19:59 train.py: 82] Epoch 3, iter 3000/6416, lr 0.100000, loss 6.847676
+INFO 2021-09-19 03:21:10 train.py: 82] Epoch 3, iter 3200/6416, lr 0.100000, loss 6.802153
+INFO 2021-09-19 03:22:20 train.py: 82] Epoch 3, iter 3400/6416, lr 0.100000, loss 6.790721
+INFO 2021-09-19 03:23:31 train.py: 82] Epoch 3, iter 3600/6416, lr 0.100000, loss 6.801906
+INFO 2021-09-19 03:24:42 train.py: 82] Epoch 3, iter 3800/6416, lr 0.100000, loss 6.782529
+INFO 2021-09-19 03:25:52 train.py: 82] Epoch 3, iter 4000/6416, lr 0.100000, loss 6.766145
+INFO 2021-09-19 03:27:03 train.py: 82] Epoch 3, iter 4200/6416, lr 0.100000, loss 6.785294
+INFO 2021-09-19 03:28:13 train.py: 82] Epoch 3, iter 4400/6416, lr 0.100000, loss 6.773862
+INFO 2021-09-19 03:29:23 train.py: 82] Epoch 3, iter 4600/6416, lr 0.100000, loss 6.738414
+INFO 2021-09-19 03:30:33 train.py: 82] Epoch 3, iter 4800/6416, lr 0.100000, loss 6.738544
+INFO 2021-09-19 03:31:44 train.py: 82] Epoch 3, iter 5000/6416, lr 0.100000, loss 6.713134
+INFO 2021-09-19 03:32:54 train.py: 82] Epoch 3, iter 5200/6416, lr 0.100000, loss 6.728994
+INFO 2021-09-19 03:34:05 train.py: 82] Epoch 3, iter 5400/6416, lr 0.100000, loss 6.698849
+INFO 2021-09-19 03:35:15 train.py: 82] Epoch 3, iter 5600/6416, lr 0.100000, loss 6.671636
+INFO 2021-09-19 03:36:26 train.py: 82] Epoch 3, iter 5800/6416, lr 0.100000, loss 6.682286
+INFO 2021-09-19 03:37:37 train.py: 95] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-09-19 03:37:38 train.py: 82] Epoch 3, iter 6000/6416, lr 0.100000, loss 6.650610
+INFO 2021-09-19 03:38:49 train.py: 82] Epoch 3, iter 6200/6416, lr 0.100000, loss 6.635904
+INFO 2021-09-19 03:39:59 train.py: 82] Epoch 3, iter 6400/6416, lr 0.100000, loss 6.612389
+INFO 2021-09-19 03:40:07 train.py: 100] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-09-19 03:40:08 train.py: 82] Epoch 4, iter 0/6416, lr 0.100000, loss 6.507708
+INFO 2021-09-19 03:41:20 train.py: 82] Epoch 4, iter 200/6416, lr 0.100000, loss 6.112202
+INFO 2021-09-19 03:42:30 train.py: 82] Epoch 4, iter 400/6416, lr 0.100000, loss 6.069352
+INFO 2021-09-19 03:43:40 train.py: 82] Epoch 4, iter 600/6416, lr 0.100000, loss 6.144737
+INFO 2021-09-19 03:44:50 train.py: 82] Epoch 4, iter 800/6416, lr 0.100000, loss 6.188641
+INFO 2021-09-19 03:46:00 train.py: 82] Epoch 4, iter 1000/6416, lr 0.100000, loss 6.255131
+INFO 2021-09-19 03:47:11 train.py: 82] Epoch 4, iter 1200/6416, lr 0.100000, loss 6.280993
+INFO 2021-09-19 03:48:21 train.py: 82] Epoch 4, iter 1400/6416, lr 0.100000, loss 6.317935
+INFO 2021-09-19 03:49:31 train.py: 82] Epoch 4, iter 1600/6416, lr 0.100000, loss 6.309303
+INFO 2021-09-19 03:50:41 train.py: 82] Epoch 4, iter 1800/6416, lr 0.100000, loss 6.352117
+INFO 2021-09-19 03:51:51 train.py: 82] Epoch 4, iter 2000/6416, lr 0.100000, loss 6.355208
+INFO 2021-09-19 03:53:01 train.py: 82] Epoch 4, iter 2200/6416, lr 0.100000, loss 6.379259
+INFO 2021-09-19 03:54:12 train.py: 82] Epoch 4, iter 2400/6416, lr 0.100000, loss 6.392918
+INFO 2021-09-19 03:55:22 train.py: 82] Epoch 4, iter 2600/6416, lr 0.100000, loss 6.390889
+INFO 2021-09-19 03:56:32 train.py: 82] Epoch 4, iter 2800/6416, lr 0.100000, loss 6.382627
+INFO 2021-09-19 03:57:44 train.py: 95] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-09-19 03:57:44 train.py: 82] Epoch 4, iter 3000/6416, lr 0.100000, loss 6.359970
+INFO 2021-09-19 03:58:56 train.py: 82] Epoch 4, iter 3200/6416, lr 0.100000, loss 6.383309
+INFO 2021-09-19 04:00:06 train.py: 82] Epoch 4, iter 3400/6416, lr 0.100000, loss 6.387433
+INFO 2021-09-19 04:01:17 train.py: 82] Epoch 4, iter 3600/6416, lr 0.100000, loss 6.395325
+INFO 2021-09-19 04:02:28 train.py: 82] Epoch 4, iter 3800/6416, lr 0.100000, loss 6.368671
+INFO 2021-09-19 04:03:39 train.py: 82] Epoch 4, iter 4000/6416, lr 0.100000, loss 6.366275
+INFO 2021-09-19 04:04:50 train.py: 82] Epoch 4, iter 4200/6416, lr 0.100000, loss 6.349015
+INFO 2021-09-19 04:06:01 train.py: 82] Epoch 4, iter 4400/6416, lr 0.100000, loss 6.357250
+INFO 2021-09-19 04:07:12 train.py: 82] Epoch 4, iter 4600/6416, lr 0.100000, loss 6.359817
+INFO 2021-09-19 04:08:23 train.py: 82] Epoch 4, iter 4800/6416, lr 0.100000, loss 6.328359
+INFO 2021-09-19 04:09:34 train.py: 82] Epoch 4, iter 5000/6416, lr 0.100000, loss 6.354290
+INFO 2021-09-19 04:10:45 train.py: 82] Epoch 4, iter 5200/6416, lr 0.100000, loss 6.340010
+INFO 2021-09-19 04:11:56 train.py: 82] Epoch 4, iter 5400/6416, lr 0.100000, loss 6.299857
+INFO 2021-09-19 04:13:07 train.py: 82] Epoch 4, iter 5600/6416, lr 0.100000, loss 6.309415
+INFO 2021-09-19 04:14:17 train.py: 82] Epoch 4, iter 5800/6416, lr 0.100000, loss 6.294510
+INFO 2021-09-19 04:15:29 train.py: 95] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-09-19 04:15:30 train.py: 82] Epoch 4, iter 6000/6416, lr 0.100000, loss 6.297518
+INFO 2021-09-19 04:16:41 train.py: 82] Epoch 4, iter 6200/6416, lr 0.100000, loss 6.246137
+INFO 2021-09-19 04:17:52 train.py: 82] Epoch 4, iter 6400/6416, lr 0.100000, loss 6.279172
+INFO 2021-09-19 04:17:59 train.py: 100] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-09-19 04:18:01 train.py: 82] Epoch 5, iter 0/6416, lr 0.100000, loss 6.301932
+INFO 2021-09-19 04:19:12 train.py: 82] Epoch 5, iter 200/6416, lr 0.100000, loss 5.684716
+INFO 2021-09-19 04:20:22 train.py: 82] Epoch 5, iter 400/6416, lr 0.100000, loss 5.695194
+INFO 2021-09-19 04:21:33 train.py: 82] Epoch 5, iter 600/6416, lr 0.100000, loss 5.792845
+INFO 2021-09-19 04:22:43 train.py: 82] Epoch 5, iter 800/6416, lr 0.100000, loss 5.843522
+INFO 2021-09-19 04:23:53 train.py: 82] Epoch 5, iter 1000/6416, lr 0.100000, loss 5.896922
+INFO 2021-09-19 04:25:02 train.py: 82] Epoch 5, iter 1200/6416, lr 0.100000, loss 5.954686
+INFO 2021-09-19 04:26:12 train.py: 82] Epoch 5, iter 1400/6416, lr 0.100000, loss 5.999677
+INFO 2021-09-19 04:27:22 train.py: 82] Epoch 5, iter 1600/6416, lr 0.100000, loss 6.030915
+INFO 2021-09-19 04:28:32 train.py: 82] Epoch 5, iter 1800/6416, lr 0.100000, loss 6.053623
+INFO 2021-09-19 04:29:42 train.py: 82] Epoch 5, iter 2000/6416, lr 0.100000, loss 6.061182
+INFO 2021-09-19 04:30:52 train.py: 82] Epoch 5, iter 2200/6416, lr 0.100000, loss 6.063276
+INFO 2021-09-19 04:32:02 train.py: 82] Epoch 5, iter 2400/6416, lr 0.100000, loss 6.085247
+INFO 2021-09-19 04:33:12 train.py: 82] Epoch 5, iter 2600/6416, lr 0.100000, loss 6.089018
+INFO 2021-09-19 04:34:22 train.py: 82] Epoch 5, iter 2800/6416, lr 0.100000, loss 6.077935
+INFO 2021-09-19 04:35:34 train.py: 95] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-09-19 04:35:35 train.py: 82] Epoch 5, iter 3000/6416, lr 0.100000, loss 6.100785
+INFO 2021-09-19 04:36:45 train.py: 82] Epoch 5, iter 3200/6416, lr 0.100000, loss 6.105000
+INFO 2021-09-19 04:37:56 train.py: 82] Epoch 5, iter 3400/6416, lr 0.100000, loss 6.063684
+INFO 2021-09-19 04:39:06 train.py: 82] Epoch 5, iter 3600/6416, lr 0.100000, loss 6.120810
+INFO 2021-09-19 04:40:17 train.py: 82] Epoch 5, iter 3800/6416, lr 0.100000, loss 6.074528
+INFO 2021-09-19 04:41:27 train.py: 82] Epoch 5, iter 4000/6416, lr 0.100000, loss 6.047152
+INFO 2021-09-19 04:42:37 train.py: 82] Epoch 5, iter 4200/6416, lr 0.100000, loss 6.074724
+INFO 2021-09-19 04:43:48 train.py: 82] Epoch 5, iter 4400/6416, lr 0.100000, loss 6.064087
+INFO 2021-09-19 04:44:58 train.py: 82] Epoch 5, iter 4600/6416, lr 0.100000, loss 6.069888
+INFO 2021-09-19 04:46:09 train.py: 82] Epoch 5, iter 4800/6416, lr 0.100000, loss 6.087017
+INFO 2021-09-19 04:47:19 train.py: 82] Epoch 5, iter 5000/6416, lr 0.100000, loss 6.041490
+INFO 2021-09-19 04:48:29 train.py: 82] Epoch 5, iter 5200/6416, lr 0.100000, loss 6.070181
+INFO 2021-09-19 04:49:40 train.py: 82] Epoch 5, iter 5400/6416, lr 0.100000, loss 6.035474
+INFO 2021-09-19 04:50:49 train.py: 82] Epoch 5, iter 5600/6416, lr 0.100000, loss 6.050062
+INFO 2021-09-19 04:52:00 train.py: 82] Epoch 5, iter 5800/6416, lr 0.100000, loss 6.055520
+INFO 2021-09-19 04:53:11 train.py: 95] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-09-19 04:53:12 train.py: 82] Epoch 5, iter 6000/6416, lr 0.100000, loss 6.058685
+INFO 2021-09-19 04:54:22 train.py: 82] Epoch 5, iter 6200/6416, lr 0.100000, loss 6.050825
+INFO 2021-09-19 04:55:32 train.py: 82] Epoch 5, iter 6400/6416, lr 0.100000, loss 6.054288
+INFO 2021-09-19 04:55:39 train.py: 100] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-09-19 04:55:41 train.py: 82] Epoch 6, iter 0/6416, lr 0.100000, loss 6.046018
+INFO 2021-09-19 04:56:52 train.py: 82] Epoch 6, iter 200/6416, lr 0.100000, loss 5.513563
+INFO 2021-09-19 04:58:02 train.py: 82] Epoch 6, iter 400/6416, lr 0.100000, loss 5.459075
+INFO 2021-09-19 04:59:13 train.py: 82] Epoch 6, iter 600/6416, lr 0.100000, loss 5.533977
+INFO 2021-09-19 05:00:23 train.py: 82] Epoch 6, iter 800/6416, lr 0.100000, loss 5.622576
+INFO 2021-09-19 05:01:33 train.py: 82] Epoch 6, iter 1000/6416, lr 0.100000, loss 5.693591
+INFO 2021-09-19 05:02:44 train.py: 82] Epoch 6, iter 1200/6416, lr 0.100000, loss 5.736691
+INFO 2021-09-19 05:03:54 train.py: 82] Epoch 6, iter 1400/6416, lr 0.100000, loss 5.789540
+INFO 2021-09-19 05:05:04 train.py: 82] Epoch 6, iter 1600/6416, lr 0.100000, loss 5.796751
+INFO 2021-09-19 05:06:14 train.py: 82] Epoch 6, iter 1800/6416, lr 0.100000, loss 5.809555
+INFO 2021-09-19 05:07:25 train.py: 82] Epoch 6, iter 2000/6416, lr 0.100000, loss 5.840025
+INFO 2021-09-19 05:08:35 train.py: 82] Epoch 6, iter 2200/6416, lr 0.100000, loss 5.816417
+INFO 2021-09-19 05:09:46 train.py: 82] Epoch 6, iter 2400/6416, lr 0.100000, loss 5.868867
+INFO 2021-09-19 05:10:56 train.py: 82] Epoch 6, iter 2600/6416, lr 0.100000, loss 5.883127
+INFO 2021-09-19 05:12:07 train.py: 82] Epoch 6, iter 2800/6416, lr 0.100000, loss 5.887144
+INFO 2021-09-19 05:13:18 train.py: 95] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-09-19 05:13:19 train.py: 82] Epoch 6, iter 3000/6416, lr 0.100000, loss 5.866807
+INFO 2021-09-19 05:14:29 train.py: 82] Epoch 6, iter 3200/6416, lr 0.100000, loss 5.884314
+INFO 2021-09-19 05:15:40 train.py: 82] Epoch 6, iter 3400/6416, lr 0.100000, loss 5.867385
+INFO 2021-09-19 05:16:50 train.py: 82] Epoch 6, iter 3600/6416, lr 0.100000, loss 5.894966
+INFO 2021-09-19 05:18:01 train.py: 82] Epoch 6, iter 3800/6416, lr 0.100000, loss 5.879976
+INFO 2021-09-19 05:19:12 train.py: 82] Epoch 6, iter 4000/6416, lr 0.100000, loss 5.874800
+INFO 2021-09-19 05:20:22 train.py: 82] Epoch 6, iter 4200/6416, lr 0.100000, loss 5.871449
+INFO 2021-09-19 05:21:33 train.py: 82] Epoch 6, iter 4400/6416, lr 0.100000, loss 5.916727
+INFO 2021-09-19 05:22:44 train.py: 82] Epoch 6, iter 4600/6416, lr 0.100000, loss 5.909396
+INFO 2021-09-19 05:23:55 train.py: 82] Epoch 6, iter 4800/6416, lr 0.100000, loss 5.878181
+INFO 2021-09-19 05:25:06 train.py: 82] Epoch 6, iter 5000/6416, lr 0.100000, loss 5.872513
+INFO 2021-09-19 05:26:17 train.py: 82] Epoch 6, iter 5200/6416, lr 0.100000, loss 5.869475
+INFO 2021-09-19 05:27:27 train.py: 82] Epoch 6, iter 5400/6416, lr 0.100000, loss 5.900326
+INFO 2021-09-19 05:28:38 train.py: 82] Epoch 6, iter 5600/6416, lr 0.100000, loss 5.877262
+INFO 2021-09-19 05:29:49 train.py: 82] Epoch 6, iter 5800/6416, lr 0.100000, loss 5.860929
+INFO 2021-09-19 05:31:01 train.py: 95] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-09-19 05:31:01 train.py: 82] Epoch 6, iter 6000/6416, lr 0.100000, loss 5.850844
+INFO 2021-09-19 05:32:12 train.py: 82] Epoch 6, iter 6200/6416, lr 0.100000, loss 5.822910
+INFO 2021-09-19 05:33:22 train.py: 82] Epoch 6, iter 6400/6416, lr 0.100000, loss 5.860385
+INFO 2021-09-19 05:33:29 train.py: 100] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-09-19 05:33:31 train.py: 82] Epoch 7, iter 0/6416, lr 0.100000, loss 5.900809
+INFO 2021-09-19 05:34:42 train.py: 82] Epoch 7, iter 200/6416, lr 0.100000, loss 5.303712
+INFO 2021-09-19 05:35:52 train.py: 82] Epoch 7, iter 400/6416, lr 0.100000, loss 5.308420
+INFO 2021-09-19 05:37:03 train.py: 82] Epoch 7, iter 600/6416, lr 0.100000, loss 5.406546
+INFO 2021-09-19 05:38:13 train.py: 82] Epoch 7, iter 800/6416, lr 0.100000, loss 5.441214
+INFO 2021-09-19 05:39:23 train.py: 82] Epoch 7, iter 1000/6416, lr 0.100000, loss 5.521501
+INFO 2021-09-19 05:40:32 train.py: 82] Epoch 7, iter 1200/6416, lr 0.100000, loss 5.527788
+INFO 2021-09-19 05:41:42 train.py: 82] Epoch 7, iter 1400/6416, lr 0.100000, loss 5.596058
+INFO 2021-09-19 05:42:52 train.py: 82] Epoch 7, iter 1600/6416, lr 0.100000, loss 5.715074
+INFO 2021-09-19 05:44:02 train.py: 82] Epoch 7, iter 1800/6416, lr 0.100000, loss 5.660813
+INFO 2021-09-19 05:45:12 train.py: 82] Epoch 7, iter 2000/6416, lr 0.100000, loss 5.705364
+INFO 2021-09-19 05:46:22 train.py: 82] Epoch 7, iter 2200/6416, lr 0.100000, loss 5.725795
+INFO 2021-09-19 05:47:31 train.py: 82] Epoch 7, iter 2400/6416, lr 0.100000, loss 5.720624
+INFO 2021-09-19 05:48:41 train.py: 82] Epoch 7, iter 2600/6416, lr 0.100000, loss 5.712527
+INFO 2021-09-19 05:49:51 train.py: 82] Epoch 7, iter 2800/6416, lr 0.100000, loss 5.734387
+INFO 2021-09-19 05:51:02 train.py: 95] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-09-19 05:51:02 train.py: 82] Epoch 7, iter 3000/6416, lr 0.100000, loss 5.729689
+INFO 2021-09-19 05:52:12 train.py: 82] Epoch 7, iter 3200/6416, lr 0.100000, loss 5.734296
+INFO 2021-09-19 05:53:22 train.py: 82] Epoch 7, iter 3400/6416, lr 0.100000, loss 5.717445
+INFO 2021-09-19 05:54:32 train.py: 82] Epoch 7, iter 3600/6416, lr 0.100000, loss 5.751255
+INFO 2021-09-19 05:55:42 train.py: 82] Epoch 7, iter 3800/6416, lr 0.100000, loss 5.737377
+INFO 2021-09-19 05:56:52 train.py: 82] Epoch 7, iter 4000/6416, lr 0.100000, loss 5.735906
+INFO 2021-09-19 05:58:01 train.py: 82] Epoch 7, iter 4200/6416, lr 0.100000, loss 5.757196
+INFO 2021-09-19 05:59:11 train.py: 82] Epoch 7, iter 4400/6416, lr 0.100000, loss 5.735972
+INFO 2021-09-19 06:00:21 train.py: 82] Epoch 7, iter 4600/6416, lr 0.100000, loss 5.732712
+INFO 2021-09-19 06:01:30 train.py: 82] Epoch 7, iter 4800/6416, lr 0.100000, loss 5.735229
+INFO 2021-09-19 06:02:41 train.py: 82] Epoch 7, iter 5000/6416, lr 0.100000, loss 5.750493
+INFO 2021-09-19 06:03:51 train.py: 82] Epoch 7, iter 5200/6416, lr 0.100000, loss 5.755233
+INFO 2021-09-19 06:05:01 train.py: 82] Epoch 7, iter 5400/6416, lr 0.100000, loss 5.745013
+INFO 2021-09-19 06:06:11 train.py: 82] Epoch 7, iter 5600/6416, lr 0.100000, loss 5.735132
+INFO 2021-09-19 06:07:21 train.py: 82] Epoch 7, iter 5800/6416, lr 0.100000, loss 5.723073
+INFO 2021-09-19 06:08:32 train.py: 95] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-09-19 06:08:33 train.py: 82] Epoch 7, iter 6000/6416, lr 0.100000, loss 5.722929
+INFO 2021-09-19 06:09:43 train.py: 82] Epoch 7, iter 6200/6416, lr 0.100000, loss 5.747271
+INFO 2021-09-19 06:10:52 train.py: 82] Epoch 7, iter 6400/6416, lr 0.100000, loss 5.672806
+INFO 2021-09-19 06:10:59 train.py: 100] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-09-19 06:11:01 train.py: 82] Epoch 8, iter 0/6416, lr 0.100000, loss 5.689348
+INFO 2021-09-19 06:12:11 train.py: 82] Epoch 8, iter 200/6416, lr 0.100000, loss 5.139348
+INFO 2021-09-19 06:13:22 train.py: 82] Epoch 8, iter 400/6416, lr 0.100000, loss 5.120587
+INFO 2021-09-19 06:14:32 train.py: 82] Epoch 8, iter 600/6416, lr 0.100000, loss 5.272039
+INFO 2021-09-19 06:15:42 train.py: 82] Epoch 8, iter 800/6416, lr 0.100000, loss 5.316755
+INFO 2021-09-19 06:16:52 train.py: 82] Epoch 8, iter 1000/6416, lr 0.100000, loss 5.359534
+INFO 2021-09-19 06:18:02 train.py: 82] Epoch 8, iter 1200/6416, lr 0.100000, loss 5.427262
+INFO 2021-09-19 06:19:12 train.py: 82] Epoch 8, iter 1400/6416, lr 0.100000, loss 5.492239
+INFO 2021-09-19 06:20:22 train.py: 82] Epoch 8, iter 1600/6416, lr 0.100000, loss 5.485799
+INFO 2021-09-19 06:21:32 train.py: 82] Epoch 8, iter 1800/6416, lr 0.100000, loss 5.513596
+INFO 2021-09-19 06:22:42 train.py: 82] Epoch 8, iter 2000/6416, lr 0.100000, loss 5.530857
+INFO 2021-09-19 06:23:52 train.py: 82] Epoch 8, iter 2200/6416, lr 0.100000, loss 5.560761
+INFO 2021-09-19 06:25:02 train.py: 82] Epoch 8, iter 2400/6416, lr 0.100000, loss 5.567073
+INFO 2021-09-19 06:26:12 train.py: 82] Epoch 8, iter 2600/6416, lr 0.100000, loss 5.558556
+INFO 2021-09-19 06:27:22 train.py: 82] Epoch 8, iter 2800/6416, lr 0.100000, loss 5.589994
+INFO 2021-09-19 06:28:33 train.py: 95] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-09-19 06:28:33 train.py: 82] Epoch 8, iter 3000/6416, lr 0.100000, loss 5.581484
+INFO 2021-09-19 06:29:44 train.py: 82] Epoch 8, iter 3200/6416, lr 0.100000, loss 5.609827
+INFO 2021-09-19 06:30:54 train.py: 82] Epoch 8, iter 3400/6416, lr 0.100000, loss 5.623524
+INFO 2021-09-19 06:32:04 train.py: 82] Epoch 8, iter 3600/6416, lr 0.100000, loss 5.612797
+INFO 2021-09-19 06:33:14 train.py: 82] Epoch 8, iter 3800/6416, lr 0.100000, loss 5.624250
+INFO 2021-09-19 06:34:24 train.py: 82] Epoch 8, iter 4000/6416, lr 0.100000, loss 5.628313
+INFO 2021-09-19 06:35:34 train.py: 82] Epoch 8, iter 4200/6416, lr 0.100000, loss 5.629772
+INFO 2021-09-19 06:36:44 train.py: 82] Epoch 8, iter 4400/6416, lr 0.100000, loss 5.593678
+INFO 2021-09-19 06:37:55 train.py: 82] Epoch 8, iter 4600/6416, lr 0.100000, loss 5.600658
+INFO 2021-09-19 06:39:05 train.py: 82] Epoch 8, iter 4800/6416, lr 0.100000, loss 5.590963
+INFO 2021-09-19 06:40:15 train.py: 82] Epoch 8, iter 5000/6416, lr 0.100000, loss 5.596903
+INFO 2021-09-19 06:41:24 train.py: 82] Epoch 8, iter 5200/6416, lr 0.100000, loss 5.605292
+INFO 2021-09-19 06:42:35 train.py: 82] Epoch 8, iter 5400/6416, lr 0.100000, loss 5.584203
+INFO 2021-09-19 06:43:45 train.py: 82] Epoch 8, iter 5600/6416, lr 0.100000, loss 5.630040
+INFO 2021-09-19 06:44:55 train.py: 82] Epoch 8, iter 5800/6416, lr 0.100000, loss 5.641527
+INFO 2021-09-19 06:46:07 train.py: 95] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-09-19 06:46:07 train.py: 82] Epoch 8, iter 6000/6416, lr 0.100000, loss 5.579319
+INFO 2021-09-19 06:47:17 train.py: 82] Epoch 8, iter 6200/6416, lr 0.100000, loss 5.588082
+INFO 2021-09-19 06:48:27 train.py: 82] Epoch 8, iter 6400/6416, lr 0.100000, loss 5.612361
+INFO 2021-09-19 06:48:34 train.py: 100] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-09-19 06:48:36 train.py: 82] Epoch 9, iter 0/6416, lr 0.100000, loss 5.660765
+INFO 2021-09-19 06:49:46 train.py: 82] Epoch 9, iter 200/6416, lr 0.100000, loss 5.054737
+INFO 2021-09-19 06:50:56 train.py: 82] Epoch 9, iter 400/6416, lr 0.100000, loss 5.013385
+INFO 2021-09-19 06:52:06 train.py: 82] Epoch 9, iter 600/6416, lr 0.100000, loss 5.127389
+INFO 2021-09-19 06:53:16 train.py: 82] Epoch 9, iter 800/6416, lr 0.100000, loss 5.228197
+INFO 2021-09-19 06:54:26 train.py: 82] Epoch 9, iter 1000/6416, lr 0.100000, loss 5.265991
+INFO 2021-09-19 06:55:35 train.py: 82] Epoch 9, iter 1200/6416, lr 0.100000, loss 5.321754
+INFO 2021-09-19 06:56:46 train.py: 82] Epoch 9, iter 1400/6416, lr 0.100000, loss 5.329900
+INFO 2021-09-19 06:57:56 train.py: 82] Epoch 9, iter 1600/6416, lr 0.100000, loss 5.378988
+INFO 2021-09-19 06:59:06 train.py: 82] Epoch 9, iter 1800/6416, lr 0.100000, loss 5.380778
+INFO 2021-09-19 07:00:15 train.py: 82] Epoch 9, iter 2000/6416, lr 0.100000, loss 5.445940
+INFO 2021-09-19 07:01:25 train.py: 82] Epoch 9, iter 2200/6416, lr 0.100000, loss 5.484453
+INFO 2021-09-19 07:02:35 train.py: 82] Epoch 9, iter 2400/6416, lr 0.100000, loss 5.484496
+INFO 2021-09-19 07:03:45 train.py: 82] Epoch 9, iter 2600/6416, lr 0.100000, loss 5.500321
+INFO 2021-09-19 07:04:55 train.py: 82] Epoch 9, iter 2800/6416, lr 0.100000, loss 5.466910
+INFO 2021-09-19 07:06:07 train.py: 95] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-09-19 07:06:07 train.py: 82] Epoch 9, iter 3000/6416, lr 0.100000, loss 5.516197
+INFO 2021-09-19 07:07:17 train.py: 82] Epoch 9, iter 3200/6416, lr 0.100000, loss 5.525093
+INFO 2021-09-19 07:08:27 train.py: 82] Epoch 9, iter 3400/6416, lr 0.100000, loss 5.508861
+INFO 2021-09-19 07:09:37 train.py: 82] Epoch 9, iter 3600/6416, lr 0.100000, loss 5.473113
+INFO 2021-09-19 07:10:47 train.py: 82] Epoch 9, iter 3800/6416, lr 0.100000, loss 5.498541
+INFO 2021-09-19 07:11:57 train.py: 82] Epoch 9, iter 4000/6416, lr 0.100000, loss 5.509012
+INFO 2021-09-19 07:13:08 train.py: 82] Epoch 9, iter 4200/6416, lr 0.100000, loss 5.495477
+INFO 2021-09-19 07:14:18 train.py: 82] Epoch 9, iter 4400/6416, lr 0.100000, loss 5.479660
+INFO 2021-09-19 07:15:28 train.py: 82] Epoch 9, iter 4600/6416, lr 0.100000, loss 5.527653
+INFO 2021-09-19 07:16:39 train.py: 82] Epoch 9, iter 4800/6416, lr 0.100000, loss 5.513913
+INFO 2021-09-19 07:17:49 train.py: 82] Epoch 9, iter 5000/6416, lr 0.100000, loss 5.516574
+INFO 2021-09-19 07:18:59 train.py: 82] Epoch 9, iter 5200/6416, lr 0.100000, loss 5.549802
+INFO 2021-09-19 07:20:09 train.py: 82] Epoch 9, iter 5400/6416, lr 0.100000, loss 5.558803
+INFO 2021-09-19 07:21:20 train.py: 82] Epoch 9, iter 5600/6416, lr 0.100000, loss 5.499242
+INFO 2021-09-19 07:22:30 train.py: 82] Epoch 9, iter 5800/6416, lr 0.100000, loss 5.509315
+INFO 2021-09-19 07:23:41 train.py: 95] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-09-19 07:23:41 train.py: 82] Epoch 9, iter 6000/6416, lr 0.100000, loss 5.523698
+INFO 2021-09-19 07:24:52 train.py: 82] Epoch 9, iter 6200/6416, lr 0.100000, loss 5.465705
+INFO 2021-09-19 07:26:02 train.py: 82] Epoch 9, iter 6400/6416, lr 0.100000, loss 5.497083
+INFO 2021-09-19 07:26:09 train.py: 100] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-09-19 07:26:11 train.py: 82] Epoch 10, iter 0/6416, lr 0.010000, loss 5.432965
+INFO 2021-09-19 07:27:21 train.py: 82] Epoch 10, iter 200/6416, lr 0.010000, loss 4.352161
+INFO 2021-09-19 07:28:31 train.py: 82] Epoch 10, iter 400/6416, lr 0.010000, loss 4.064512
+INFO 2021-09-19 07:29:40 train.py: 82] Epoch 10, iter 600/6416, lr 0.010000, loss 3.970171
+INFO 2021-09-19 07:30:50 train.py: 82] Epoch 10, iter 800/6416, lr 0.010000, loss 3.887686
+INFO 2021-09-19 07:32:00 train.py: 82] Epoch 10, iter 1000/6416, lr 0.010000, loss 3.853635
+INFO 2021-09-19 07:33:09 train.py: 82] Epoch 10, iter 1200/6416, lr 0.010000, loss 3.802465
+INFO 2021-09-19 07:34:19 train.py: 82] Epoch 10, iter 1400/6416, lr 0.010000, loss 3.778143
+INFO 2021-09-19 07:35:29 train.py: 82] Epoch 10, iter 1600/6416, lr 0.010000, loss 3.721591
+INFO 2021-09-19 07:36:39 train.py: 82] Epoch 10, iter 1800/6416, lr 0.010000, loss 3.718536
+INFO 2021-09-19 07:37:48 train.py: 82] Epoch 10, iter 2000/6416, lr 0.010000, loss 3.645258
+INFO 2021-09-19 07:38:58 train.py: 82] Epoch 10, iter 2200/6416, lr 0.010000, loss 3.630993
+INFO 2021-09-19 07:40:08 train.py: 82] Epoch 10, iter 2400/6416, lr 0.010000, loss 3.595042
+INFO 2021-09-19 07:41:18 train.py: 82] Epoch 10, iter 2600/6416, lr 0.010000, loss 3.597292
+INFO 2021-09-19 07:42:28 train.py: 82] Epoch 10, iter 2800/6416, lr 0.010000, loss 3.544022
+INFO 2021-09-19 07:43:39 train.py: 95] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-09-19 07:43:39 train.py: 82] Epoch 10, iter 3000/6416, lr 0.010000, loss 3.547099
+INFO 2021-09-19 07:44:49 train.py: 82] Epoch 10, iter 3200/6416, lr 0.010000, loss 3.520643
+INFO 2021-09-19 07:45:59 train.py: 82] Epoch 10, iter 3400/6416, lr 0.010000, loss 3.497581
+INFO 2021-09-19 07:47:09 train.py: 82] Epoch 10, iter 3600/6416, lr 0.010000, loss 3.448916
+INFO 2021-09-19 07:48:18 train.py: 82] Epoch 10, iter 3800/6416, lr 0.010000, loss 3.445067
+INFO 2021-09-19 07:49:29 train.py: 82] Epoch 10, iter 4000/6416, lr 0.010000, loss 3.460121
+INFO 2021-09-19 07:50:39 train.py: 82] Epoch 10, iter 4200/6416, lr 0.010000, loss 3.419068
+INFO 2021-09-19 07:51:48 train.py: 82] Epoch 10, iter 4400/6416, lr 0.010000, loss 3.389921
+INFO 2021-09-19 07:52:58 train.py: 82] Epoch 10, iter 4600/6416, lr 0.010000, loss 3.432742
+INFO 2021-09-19 07:54:08 train.py: 82] Epoch 10, iter 4800/6416, lr 0.010000, loss 3.401298
+INFO 2021-09-19 07:55:18 train.py: 82] Epoch 10, iter 5000/6416, lr 0.010000, loss 3.360272
+INFO 2021-09-19 07:56:28 train.py: 82] Epoch 10, iter 5200/6416, lr 0.010000, loss 3.364981
+INFO 2021-09-19 07:57:38 train.py: 82] Epoch 10, iter 5400/6416, lr 0.010000, loss 3.355302
+INFO 2021-09-19 07:58:47 train.py: 82] Epoch 10, iter 5600/6416, lr 0.010000, loss 3.317275
+INFO 2021-09-19 07:59:58 train.py: 82] Epoch 10, iter 5800/6416, lr 0.010000, loss 3.311607
+INFO 2021-09-19 08:01:09 train.py: 95] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-09-19 08:01:09 train.py: 82] Epoch 10, iter 6000/6416, lr 0.010000, loss 3.330219
+INFO 2021-09-19 08:02:19 train.py: 82] Epoch 10, iter 6200/6416, lr 0.010000, loss 3.289577
+INFO 2021-09-19 08:03:29 train.py: 82] Epoch 10, iter 6400/6416, lr 0.010000, loss 3.298895
+INFO 2021-09-19 08:03:36 train.py: 100] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-09-19 08:03:38 train.py: 82] Epoch 11, iter 0/6416, lr 0.010000, loss 3.216603
+INFO 2021-09-19 08:04:48 train.py: 82] Epoch 11, iter 200/6416, lr 0.010000, loss 2.958152
+INFO 2021-09-19 08:05:58 train.py: 82] Epoch 11, iter 400/6416, lr 0.010000, loss 2.955029
+INFO 2021-09-19 08:07:07 train.py: 82] Epoch 11, iter 600/6416, lr 0.010000, loss 2.951972
+INFO 2021-09-19 08:08:17 train.py: 82] Epoch 11, iter 800/6416, lr 0.010000, loss 2.940289
+INFO 2021-09-19 08:09:27 train.py: 82] Epoch 11, iter 1000/6416, lr 0.010000, loss 2.948962
+INFO 2021-09-19 08:10:37 train.py: 82] Epoch 11, iter 1200/6416, lr 0.010000, loss 2.962738
+INFO 2021-09-19 08:11:47 train.py: 82] Epoch 11, iter 1400/6416, lr 0.010000, loss 2.962874
+INFO 2021-09-19 08:12:58 train.py: 82] Epoch 11, iter 1600/6416, lr 0.010000, loss 2.984128
+INFO 2021-09-19 08:14:08 train.py: 82] Epoch 11, iter 1800/6416, lr 0.010000, loss 2.991319
+INFO 2021-09-19 08:15:18 train.py: 82] Epoch 11, iter 2000/6416, lr 0.010000, loss 2.961891
+INFO 2021-09-19 08:16:28 train.py: 82] Epoch 11, iter 2200/6416, lr 0.010000, loss 2.947934
+INFO 2021-09-19 08:17:38 train.py: 82] Epoch 11, iter 2400/6416, lr 0.010000, loss 2.942955
+INFO 2021-09-19 08:18:49 train.py: 82] Epoch 11, iter 2600/6416, lr 0.010000, loss 2.948131
+INFO 2021-09-19 08:19:59 train.py: 82] Epoch 11, iter 2800/6416, lr 0.010000, loss 2.962273
+INFO 2021-09-19 08:21:11 train.py: 95] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-09-19 08:21:12 train.py: 82] Epoch 11, iter 3000/6416, lr 0.010000, loss 2.938132
+INFO 2021-09-19 08:22:22 train.py: 82] Epoch 11, iter 3200/6416, lr 0.010000, loss 2.972156
+INFO 2021-09-19 08:23:33 train.py: 82] Epoch 11, iter 3400/6416, lr 0.010000, loss 2.978951
+INFO 2021-09-19 08:24:44 train.py: 82] Epoch 11, iter 3600/6416, lr 0.010000, loss 2.975123
+INFO 2021-09-19 08:25:54 train.py: 82] Epoch 11, iter 3800/6416, lr 0.010000, loss 2.976828
+INFO 2021-09-19 08:27:05 train.py: 82] Epoch 11, iter 4000/6416, lr 0.010000, loss 2.977453
+INFO 2021-09-19 08:28:16 train.py: 82] Epoch 11, iter 4200/6416, lr 0.010000, loss 2.983852
+INFO 2021-09-19 08:29:26 train.py: 82] Epoch 11, iter 4400/6416, lr 0.010000, loss 2.948758
+INFO 2021-09-19 08:30:37 train.py: 82] Epoch 11, iter 4600/6416, lr 0.010000, loss 2.947861
+INFO 2021-09-19 08:31:47 train.py: 82] Epoch 11, iter 4800/6416, lr 0.010000, loss 2.943177
+INFO 2021-09-19 08:32:57 train.py: 82] Epoch 11, iter 5000/6416, lr 0.010000, loss 2.963660
+INFO 2021-09-19 08:34:08 train.py: 82] Epoch 11, iter 5200/6416, lr 0.010000, loss 2.960780
+INFO 2021-09-19 08:35:18 train.py: 82] Epoch 11, iter 5400/6416, lr 0.010000, loss 2.955101
+INFO 2021-09-19 08:36:29 train.py: 82] Epoch 11, iter 5600/6416, lr 0.010000, loss 2.963177
+INFO 2021-09-19 08:37:39 train.py: 82] Epoch 11, iter 5800/6416, lr 0.010000, loss 2.951466
+INFO 2021-09-19 08:38:51 train.py: 95] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-09-19 08:38:51 train.py: 82] Epoch 11, iter 6000/6416, lr 0.010000, loss 2.960908
+INFO 2021-09-19 08:40:02 train.py: 82] Epoch 11, iter 6200/6416, lr 0.010000, loss 2.986121
+INFO 2021-09-19 08:41:12 train.py: 82] Epoch 11, iter 6400/6416, lr 0.010000, loss 2.946486
+INFO 2021-09-19 08:41:19 train.py: 100] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-09-19 08:41:21 train.py: 82] Epoch 12, iter 0/6416, lr 0.010000, loss 2.942438
+INFO 2021-09-19 08:42:31 train.py: 82] Epoch 12, iter 200/6416, lr 0.010000, loss 2.644667
+INFO 2021-09-19 08:43:41 train.py: 82] Epoch 12, iter 400/6416, lr 0.010000, loss 2.637221
+INFO 2021-09-19 08:44:50 train.py: 82] Epoch 12, iter 600/6416, lr 0.010000, loss 2.631070
+INFO 2021-09-19 08:46:00 train.py: 82] Epoch 12, iter 800/6416, lr 0.010000, loss 2.656773
+INFO 2021-09-19 08:47:10 train.py: 82] Epoch 12, iter 1000/6416, lr 0.010000, loss 2.644617
+INFO 2021-09-19 08:48:20 train.py: 82] Epoch 12, iter 1200/6416, lr 0.010000, loss 2.670818
+INFO 2021-09-19 08:49:30 train.py: 82] Epoch 12, iter 1400/6416, lr 0.010000, loss 2.695057
+INFO 2021-09-19 08:50:40 train.py: 82] Epoch 12, iter 1600/6416, lr 0.010000, loss 2.681815
+INFO 2021-09-19 08:51:50 train.py: 82] Epoch 12, iter 1800/6416, lr 0.010000, loss 2.692749
+INFO 2021-09-19 08:53:00 train.py: 82] Epoch 12, iter 2000/6416, lr 0.010000, loss 2.702770
+INFO 2021-09-19 08:54:10 train.py: 82] Epoch 12, iter 2200/6416, lr 0.010000, loss 2.705044
+INFO 2021-09-19 08:55:21 train.py: 82] Epoch 12, iter 2400/6416, lr 0.010000, loss 2.726312
+INFO 2021-09-19 08:56:31 train.py: 82] Epoch 12, iter 2600/6416, lr 0.010000, loss 2.735580
+INFO 2021-09-19 08:57:41 train.py: 82] Epoch 12, iter 2800/6416, lr 0.010000, loss 2.729822
+INFO 2021-09-19 08:58:52 train.py: 95] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-09-19 08:58:53 train.py: 82] Epoch 12, iter 3000/6416, lr 0.010000, loss 2.733575
+INFO 2021-09-19 09:00:03 train.py: 82] Epoch 12, iter 3200/6416, lr 0.010000, loss 2.732999
+INFO 2021-09-19 09:01:13 train.py: 82] Epoch 12, iter 3400/6416, lr 0.010000, loss 2.777878
+INFO 2021-09-19 09:02:24 train.py: 82] Epoch 12, iter 3600/6416, lr 0.010000, loss 2.769897
+INFO 2021-09-19 09:03:35 train.py: 82] Epoch 12, iter 3800/6416, lr 0.010000, loss 2.767126
+INFO 2021-09-19 09:04:45 train.py: 82] Epoch 12, iter 4000/6416, lr 0.010000, loss 2.756580
+INFO 2021-09-19 09:05:55 train.py: 82] Epoch 12, iter 4200/6416, lr 0.010000, loss 2.773906
+INFO 2021-09-19 09:07:05 train.py: 82] Epoch 12, iter 4400/6416, lr 0.010000, loss 2.785596
+INFO 2021-09-19 09:08:16 train.py: 82] Epoch 12, iter 4600/6416, lr 0.010000, loss 2.805755
+INFO 2021-09-19 09:09:27 train.py: 82] Epoch 12, iter 4800/6416, lr 0.010000, loss 2.792646
+INFO 2021-09-19 09:10:37 train.py: 82] Epoch 12, iter 5000/6416, lr 0.010000, loss 2.785725
+INFO 2021-09-19 09:11:47 train.py: 82] Epoch 12, iter 5200/6416, lr 0.010000, loss 2.791236
+INFO 2021-09-19 09:12:58 train.py: 82] Epoch 12, iter 5400/6416, lr 0.010000, loss 2.805673
+INFO 2021-09-19 09:14:08 train.py: 82] Epoch 12, iter 5600/6416, lr 0.010000, loss 2.823273
+INFO 2021-09-19 09:15:18 train.py: 82] Epoch 12, iter 5800/6416, lr 0.010000, loss 2.808817
+INFO 2021-09-19 09:16:29 train.py: 95] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-09-19 09:16:30 train.py: 82] Epoch 12, iter 6000/6416, lr 0.010000, loss 2.821896
+INFO 2021-09-19 09:17:41 train.py: 82] Epoch 12, iter 6200/6416, lr 0.010000, loss 2.825202
+INFO 2021-09-19 09:18:51 train.py: 82] Epoch 12, iter 6400/6416, lr 0.010000, loss 2.815725
+INFO 2021-09-19 09:18:58 train.py: 100] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-09-19 09:19:00 train.py: 82] Epoch 13, iter 0/6416, lr 0.001000, loss 2.800413
+INFO 2021-09-19 09:20:10 train.py: 82] Epoch 13, iter 200/6416, lr 0.001000, loss 2.413482
+INFO 2021-09-19 09:21:20 train.py: 82] Epoch 13, iter 400/6416, lr 0.001000, loss 2.380512
+INFO 2021-09-19 09:22:30 train.py: 82] Epoch 13, iter 600/6416, lr 0.001000, loss 2.390611
+INFO 2021-09-19 09:23:40 train.py: 82] Epoch 13, iter 800/6416, lr 0.001000, loss 2.357303
+INFO 2021-09-19 09:24:50 train.py: 82] Epoch 13, iter 1000/6416, lr 0.001000, loss 2.386323
+INFO 2021-09-19 09:25:59 train.py: 82] Epoch 13, iter 1200/6416, lr 0.001000, loss 2.369664
+INFO 2021-09-19 09:27:09 train.py: 82] Epoch 13, iter 1400/6416, lr 0.001000, loss 2.366166
+INFO 2021-09-19 09:28:20 train.py: 82] Epoch 13, iter 1600/6416, lr 0.001000, loss 2.361037
+INFO 2021-09-19 09:29:30 train.py: 82] Epoch 13, iter 1800/6416, lr 0.001000, loss 2.376281
+INFO 2021-09-19 09:30:39 train.py: 82] Epoch 13, iter 2000/6416, lr 0.001000, loss 2.378462
+INFO 2021-09-19 09:31:49 train.py: 82] Epoch 13, iter 2200/6416, lr 0.001000, loss 2.376838
+INFO 2021-09-19 09:32:59 train.py: 82] Epoch 13, iter 2400/6416, lr 0.001000, loss 2.365319
+INFO 2021-09-19 09:34:09 train.py: 82] Epoch 13, iter 2600/6416, lr 0.001000, loss 2.356722
+INFO 2021-09-19 09:35:19 train.py: 82] Epoch 13, iter 2800/6416, lr 0.001000, loss 2.378973
+INFO 2021-09-19 09:36:31 train.py: 95] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-09-19 09:36:31 train.py: 82] Epoch 13, iter 3000/6416, lr 0.001000, loss 2.352463
+INFO 2021-09-19 09:37:41 train.py: 82] Epoch 13, iter 3200/6416, lr 0.001000, loss 2.382218
+INFO 2021-09-19 09:38:52 train.py: 82] Epoch 13, iter 3400/6416, lr 0.001000, loss 2.367465
+INFO 2021-09-19 09:40:02 train.py: 82] Epoch 13, iter 3600/6416, lr 0.001000, loss 2.350621
+INFO 2021-09-19 09:41:12 train.py: 82] Epoch 13, iter 3800/6416, lr 0.001000, loss 2.369132
+INFO 2021-09-19 09:42:22 train.py: 82] Epoch 13, iter 4000/6416, lr 0.001000, loss 2.378563
+INFO 2021-09-19 09:43:32 train.py: 82] Epoch 13, iter 4200/6416, lr 0.001000, loss 2.363354
+INFO 2021-09-19 09:44:42 train.py: 82] Epoch 13, iter 4400/6416, lr 0.001000, loss 2.363441
+INFO 2021-09-19 09:45:53 train.py: 82] Epoch 13, iter 4600/6416, lr 0.001000, loss 2.346006
+INFO 2021-09-19 09:47:03 train.py: 82] Epoch 13, iter 4800/6416, lr 0.001000, loss 2.380742
+INFO 2021-09-19 09:48:14 train.py: 82] Epoch 13, iter 5000/6416, lr 0.001000, loss 2.403296
+INFO 2021-09-19 09:49:24 train.py: 82] Epoch 13, iter 5200/6416, lr 0.001000, loss 2.363068
+INFO 2021-09-19 09:50:35 train.py: 82] Epoch 13, iter 5400/6416, lr 0.001000, loss 2.362031
+INFO 2021-09-19 09:51:45 train.py: 82] Epoch 13, iter 5600/6416, lr 0.001000, loss 2.369125
+INFO 2021-09-19 09:52:55 train.py: 82] Epoch 13, iter 5800/6416, lr 0.001000, loss 2.369220
+INFO 2021-09-19 09:54:07 train.py: 95] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-09-19 09:54:07 train.py: 82] Epoch 13, iter 6000/6416, lr 0.001000, loss 2.373340
+INFO 2021-09-19 09:55:18 train.py: 82] Epoch 13, iter 6200/6416, lr 0.001000, loss 2.378530
+INFO 2021-09-19 09:56:28 train.py: 82] Epoch 13, iter 6400/6416, lr 0.001000, loss 2.381798
+INFO 2021-09-19 09:56:35 train.py: 100] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-09-19 09:56:37 train.py: 82] Epoch 14, iter 0/6416, lr 0.001000, loss 2.369519
+INFO 2021-09-19 09:57:47 train.py: 82] Epoch 14, iter 200/6416, lr 0.001000, loss 2.315587
+INFO 2021-09-19 09:58:56 train.py: 82] Epoch 14, iter 400/6416, lr 0.001000, loss 2.311979
+INFO 2021-09-19 10:00:06 train.py: 82] Epoch 14, iter 600/6416, lr 0.001000, loss 2.327739
+INFO 2021-09-19 10:01:16 train.py: 82] Epoch 14, iter 800/6416, lr 0.001000, loss 2.318597
+INFO 2021-09-19 10:02:25 train.py: 82] Epoch 14, iter 1000/6416, lr 0.001000, loss 2.322992
+INFO 2021-09-19 10:03:35 train.py: 82] Epoch 14, iter 1200/6416, lr 0.001000, loss 2.321007
+INFO 2021-09-19 10:04:45 train.py: 82] Epoch 14, iter 1400/6416, lr 0.001000, loss 2.312466
+INFO 2021-09-19 10:05:55 train.py: 82] Epoch 14, iter 1600/6416, lr 0.001000, loss 2.324459
+INFO 2021-09-19 10:07:04 train.py: 82] Epoch 14, iter 1800/6416, lr 0.001000, loss 2.345706
+INFO 2021-09-19 10:08:14 train.py: 82] Epoch 14, iter 2000/6416, lr 0.001000, loss 2.341147
+INFO 2021-09-19 10:09:24 train.py: 82] Epoch 14, iter 2200/6416, lr 0.001000, loss 2.329970
+INFO 2021-09-19 10:10:33 train.py: 82] Epoch 14, iter 2400/6416, lr 0.001000, loss 2.320496
+INFO 2021-09-19 10:11:43 train.py: 82] Epoch 14, iter 2600/6416, lr 0.001000, loss 2.319579
+INFO 2021-09-19 10:12:53 train.py: 82] Epoch 14, iter 2800/6416, lr 0.001000, loss 2.321475
+INFO 2021-09-19 10:14:04 train.py: 95] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-09-19 10:14:05 train.py: 82] Epoch 14, iter 3000/6416, lr 0.001000, loss 2.323610
+INFO 2021-09-19 10:15:15 train.py: 82] Epoch 14, iter 3200/6416, lr 0.001000, loss 2.338683
+INFO 2021-09-19 10:16:25 train.py: 82] Epoch 14, iter 3400/6416, lr 0.001000, loss 2.335192
+INFO 2021-09-19 10:17:36 train.py: 82] Epoch 14, iter 3600/6416, lr 0.001000, loss 2.332269
+INFO 2021-09-19 10:18:46 train.py: 82] Epoch 14, iter 3800/6416, lr 0.001000, loss 2.347203
+INFO 2021-09-19 10:19:57 train.py: 82] Epoch 14, iter 4000/6416, lr 0.001000, loss 2.330586
+INFO 2021-09-19 10:21:08 train.py: 82] Epoch 14, iter 4200/6416, lr 0.001000, loss 2.356861
+INFO 2021-09-19 10:22:19 train.py: 82] Epoch 14, iter 4400/6416, lr 0.001000, loss 2.344010
+INFO 2021-09-19 10:23:30 train.py: 82] Epoch 14, iter 4600/6416, lr 0.001000, loss 2.324589
+INFO 2021-09-19 10:24:41 train.py: 82] Epoch 14, iter 4800/6416, lr 0.001000, loss 2.326964
+INFO 2021-09-19 10:25:51 train.py: 82] Epoch 14, iter 5000/6416, lr 0.001000, loss 2.329856
+INFO 2021-09-19 10:27:02 train.py: 82] Epoch 14, iter 5200/6416, lr 0.001000, loss 2.355338
+INFO 2021-09-19 10:28:13 train.py: 82] Epoch 14, iter 5400/6416, lr 0.001000, loss 2.325544
+INFO 2021-09-19 10:29:23 train.py: 82] Epoch 14, iter 5600/6416, lr 0.001000, loss 2.330749
+INFO 2021-09-19 10:30:34 train.py: 82] Epoch 14, iter 5800/6416, lr 0.001000, loss 2.343659
+INFO 2021-09-19 10:31:46 train.py: 95] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-09-19 10:31:46 train.py: 82] Epoch 14, iter 6000/6416, lr 0.001000, loss 2.320386
+INFO 2021-09-19 10:32:57 train.py: 82] Epoch 14, iter 6200/6416, lr 0.001000, loss 2.339257
+INFO 2021-09-19 10:34:07 train.py: 82] Epoch 14, iter 6400/6416, lr 0.001000, loss 2.343378
+INFO 2021-09-19 10:34:14 train.py: 100] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-09-19 10:34:16 train.py: 82] Epoch 15, iter 0/6416, lr 0.001000, loss 2.403512
+INFO 2021-09-19 10:35:27 train.py: 82] Epoch 15, iter 200/6416, lr 0.001000, loss 2.302820
+INFO 2021-09-19 10:36:37 train.py: 82] Epoch 15, iter 400/6416, lr 0.001000, loss 2.306578
+INFO 2021-09-19 10:37:47 train.py: 82] Epoch 15, iter 600/6416, lr 0.001000, loss 2.298548
+INFO 2021-09-19 10:38:57 train.py: 82] Epoch 15, iter 800/6416, lr 0.001000, loss 2.304025
+INFO 2021-09-19 10:40:07 train.py: 82] Epoch 15, iter 1000/6416, lr 0.001000, loss 2.274814
+INFO 2021-09-19 10:41:17 train.py: 82] Epoch 15, iter 1200/6416, lr 0.001000, loss 2.295869
+INFO 2021-09-19 10:42:27 train.py: 82] Epoch 15, iter 1400/6416, lr 0.001000, loss 2.310575
+INFO 2021-09-19 10:43:38 train.py: 82] Epoch 15, iter 1600/6416, lr 0.001000, loss 2.293905
+INFO 2021-09-19 10:44:48 train.py: 82] Epoch 15, iter 1800/6416, lr 0.001000, loss 2.303342
+INFO 2021-09-19 10:45:58 train.py: 82] Epoch 15, iter 2000/6416, lr 0.001000, loss 2.303691
+INFO 2021-09-19 10:47:08 train.py: 82] Epoch 15, iter 2200/6416, lr 0.001000, loss 2.307822
+INFO 2021-09-19 10:48:18 train.py: 82] Epoch 15, iter 2400/6416, lr 0.001000, loss 2.299887
+INFO 2021-09-19 10:49:29 train.py: 82] Epoch 15, iter 2600/6416, lr 0.001000, loss 2.288764
+INFO 2021-09-19 10:50:39 train.py: 82] Epoch 15, iter 2800/6416, lr 0.001000, loss 2.287246
+INFO 2021-09-19 10:51:51 train.py: 95] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-09-19 10:51:51 train.py: 82] Epoch 15, iter 3000/6416, lr 0.001000, loss 2.303929
+INFO 2021-09-19 10:53:02 train.py: 82] Epoch 15, iter 3200/6416, lr 0.001000, loss 2.303544
+INFO 2021-09-19 10:54:13 train.py: 82] Epoch 15, iter 3400/6416, lr 0.001000, loss 2.308096
+INFO 2021-09-19 10:55:24 train.py: 82] Epoch 15, iter 3600/6416, lr 0.001000, loss 2.304996
+INFO 2021-09-19 10:56:35 train.py: 82] Epoch 15, iter 3800/6416, lr 0.001000, loss 2.297510
+INFO 2021-09-19 10:57:46 train.py: 82] Epoch 15, iter 4000/6416, lr 0.001000, loss 2.300384
+INFO 2021-09-19 10:58:57 train.py: 82] Epoch 15, iter 4200/6416, lr 0.001000, loss 2.314665
+INFO 2021-09-19 11:00:08 train.py: 82] Epoch 15, iter 4400/6416, lr 0.001000, loss 2.310723
+INFO 2021-09-19 11:01:19 train.py: 82] Epoch 15, iter 4600/6416, lr 0.001000, loss 2.316207
+INFO 2021-09-19 11:02:30 train.py: 82] Epoch 15, iter 4800/6416, lr 0.001000, loss 2.299171
+INFO 2021-09-19 11:03:42 train.py: 82] Epoch 15, iter 5000/6416, lr 0.001000, loss 2.309225
+INFO 2021-09-19 11:04:53 train.py: 82] Epoch 15, iter 5200/6416, lr 0.001000, loss 2.313352
+INFO 2021-09-19 11:06:04 train.py: 82] Epoch 15, iter 5400/6416, lr 0.001000, loss 2.322332
+INFO 2021-09-19 11:07:14 train.py: 82] Epoch 15, iter 5600/6416, lr 0.001000, loss 2.323402
+INFO 2021-09-19 11:08:26 train.py: 82] Epoch 15, iter 5800/6416, lr 0.001000, loss 2.304920
+INFO 2021-09-19 11:09:38 train.py: 95] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-09-19 11:09:38 train.py: 82] Epoch 15, iter 6000/6416, lr 0.001000, loss 2.312304
+INFO 2021-09-19 11:10:49 train.py: 82] Epoch 15, iter 6200/6416, lr 0.001000, loss 2.309592
+INFO 2021-09-19 11:12:00 train.py: 82] Epoch 15, iter 6400/6416, lr 0.001000, loss 2.333553
+INFO 2021-09-19 11:12:07 train.py: 100] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-09-19 11:12:09 train.py: 82] Epoch 16, iter 0/6416, lr 0.000100, loss 2.303236
+INFO 2021-09-19 11:13:19 train.py: 82] Epoch 16, iter 200/6416, lr 0.000100, loss 2.278505
+INFO 2021-09-19 11:14:29 train.py: 82] Epoch 16, iter 400/6416, lr 0.000100, loss 2.254318
+INFO 2021-09-19 11:15:39 train.py: 82] Epoch 16, iter 600/6416, lr 0.000100, loss 2.252762
+INFO 2021-09-19 11:16:49 train.py: 82] Epoch 16, iter 800/6416, lr 0.000100, loss 2.250345
+INFO 2021-09-19 11:17:59 train.py: 82] Epoch 16, iter 1000/6416, lr 0.000100, loss 2.254362
+INFO 2021-09-19 11:19:08 train.py: 82] Epoch 16, iter 1200/6416, lr 0.000100, loss 2.269996
+INFO 2021-09-19 11:20:18 train.py: 82] Epoch 16, iter 1400/6416, lr 0.000100, loss 2.247288
+INFO 2021-09-19 11:21:28 train.py: 82] Epoch 16, iter 1600/6416, lr 0.000100, loss 2.246930
+INFO 2021-09-19 11:22:38 train.py: 82] Epoch 16, iter 1800/6416, lr 0.000100, loss 2.252600
+INFO 2021-09-19 11:23:48 train.py: 82] Epoch 16, iter 2000/6416, lr 0.000100, loss 2.267004
+INFO 2021-09-19 11:24:58 train.py: 82] Epoch 16, iter 2200/6416, lr 0.000100, loss 2.250348
+INFO 2021-09-19 11:26:08 train.py: 82] Epoch 16, iter 2400/6416, lr 0.000100, loss 2.253333
+INFO 2021-09-19 11:27:19 train.py: 82] Epoch 16, iter 2600/6416, lr 0.000100, loss 2.263728
+INFO 2021-09-19 11:28:29 train.py: 82] Epoch 16, iter 2800/6416, lr 0.000100, loss 2.264098
+INFO 2021-09-19 11:29:41 train.py: 95] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-09-19 11:29:41 train.py: 82] Epoch 16, iter 3000/6416, lr 0.000100, loss 2.257215
+INFO 2021-09-19 11:30:52 train.py: 82] Epoch 16, iter 3200/6416, lr 0.000100, loss 2.266333
+INFO 2021-09-19 11:32:03 train.py: 82] Epoch 16, iter 3400/6416, lr 0.000100, loss 2.267763
+INFO 2021-09-19 11:33:14 train.py: 82] Epoch 16, iter 3600/6416, lr 0.000100, loss 2.261790
+INFO 2021-09-19 11:34:24 train.py: 82] Epoch 16, iter 3800/6416, lr 0.000100, loss 2.254827
+INFO 2021-09-19 11:35:36 train.py: 82] Epoch 16, iter 4000/6416, lr 0.000100, loss 2.276496
+INFO 2021-09-19 11:36:47 train.py: 82] Epoch 16, iter 4200/6416, lr 0.000100, loss 2.259896
+INFO 2021-09-19 11:37:58 train.py: 82] Epoch 16, iter 4400/6416, lr 0.000100, loss 2.259406
+INFO 2021-09-19 11:39:08 train.py: 82] Epoch 16, iter 4600/6416, lr 0.000100, loss 2.278588
+INFO 2021-09-19 11:40:19 train.py: 82] Epoch 16, iter 4800/6416, lr 0.000100, loss 2.256832
+INFO 2021-09-19 11:41:29 train.py: 82] Epoch 16, iter 5000/6416, lr 0.000100, loss 2.256046
+INFO 2021-09-19 11:42:39 train.py: 82] Epoch 16, iter 5200/6416, lr 0.000100, loss 2.262844
+INFO 2021-09-19 11:43:50 train.py: 82] Epoch 16, iter 5400/6416, lr 0.000100, loss 2.265277
+INFO 2021-09-19 11:45:00 train.py: 82] Epoch 16, iter 5600/6416, lr 0.000100, loss 2.248440
+INFO 2021-09-19 11:46:11 train.py: 82] Epoch 16, iter 5800/6416, lr 0.000100, loss 2.257869
+INFO 2021-09-19 11:47:23 train.py: 95] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-09-19 11:47:23 train.py: 82] Epoch 16, iter 6000/6416, lr 0.000100, loss 2.267195
+INFO 2021-09-19 11:48:35 train.py: 82] Epoch 16, iter 6200/6416, lr 0.000100, loss 2.265269
+INFO 2021-09-19 11:49:45 train.py: 82] Epoch 16, iter 6400/6416, lr 0.000100, loss 2.255789
+INFO 2021-09-19 11:49:53 train.py: 100] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-09-19 11:49:55 train.py: 82] Epoch 17, iter 0/6416, lr 0.000100, loss 2.233435
+INFO 2021-09-19 11:51:05 train.py: 82] Epoch 17, iter 200/6416, lr 0.000100, loss 2.243726
+INFO 2021-09-19 11:52:15 train.py: 82] Epoch 17, iter 400/6416, lr 0.000100, loss 2.255259
+INFO 2021-09-19 11:53:25 train.py: 82] Epoch 17, iter 600/6416, lr 0.000100, loss 2.243763
+INFO 2021-09-19 11:54:35 train.py: 82] Epoch 17, iter 800/6416, lr 0.000100, loss 2.244552
+INFO 2021-09-19 11:55:44 train.py: 82] Epoch 17, iter 1000/6416, lr 0.000100, loss 2.258126
+INFO 2021-09-19 11:56:54 train.py: 82] Epoch 17, iter 1200/6416, lr 0.000100, loss 2.248700
+INFO 2021-09-19 11:58:04 train.py: 82] Epoch 17, iter 1400/6416, lr 0.000100, loss 2.251632
+INFO 2021-09-19 11:59:14 train.py: 82] Epoch 17, iter 1600/6416, lr 0.000100, loss 2.259832
+INFO 2021-09-19 12:00:23 train.py: 82] Epoch 17, iter 1800/6416, lr 0.000100, loss 2.249055
+INFO 2021-09-19 12:01:33 train.py: 82] Epoch 17, iter 2000/6416, lr 0.000100, loss 2.261270
+INFO 2021-09-19 12:02:43 train.py: 82] Epoch 17, iter 2200/6416, lr 0.000100, loss 2.250452
+INFO 2021-09-19 12:03:53 train.py: 82] Epoch 17, iter 2400/6416, lr 0.000100, loss 2.257999
+INFO 2021-09-19 12:05:03 train.py: 82] Epoch 17, iter 2600/6416, lr 0.000100, loss 2.267271
+INFO 2021-09-19 12:06:13 train.py: 82] Epoch 17, iter 2800/6416, lr 0.000100, loss 2.263096
+INFO 2021-09-19 12:07:24 train.py: 95] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-09-19 12:07:25 train.py: 82] Epoch 17, iter 3000/6416, lr 0.000100, loss 2.267387
+INFO 2021-09-19 12:08:35 train.py: 82] Epoch 17, iter 3200/6416, lr 0.000100, loss 2.267748
+INFO 2021-09-19 12:09:45 train.py: 82] Epoch 17, iter 3400/6416, lr 0.000100, loss 2.263820
+INFO 2021-09-19 12:10:56 train.py: 82] Epoch 17, iter 3600/6416, lr 0.000100, loss 2.266134
+INFO 2021-09-19 12:12:06 train.py: 82] Epoch 17, iter 3800/6416, lr 0.000100, loss 2.248134
+INFO 2021-09-19 12:13:17 train.py: 82] Epoch 17, iter 4000/6416, lr 0.000100, loss 2.255861
+INFO 2021-09-19 12:14:27 train.py: 82] Epoch 17, iter 4200/6416, lr 0.000100, loss 2.249989
+INFO 2021-09-19 12:15:38 train.py: 82] Epoch 17, iter 4400/6416, lr 0.000100, loss 2.250293
+INFO 2021-09-19 12:16:48 train.py: 82] Epoch 17, iter 4600/6416, lr 0.000100, loss 2.255407
+INFO 2021-09-19 12:17:58 train.py: 82] Epoch 17, iter 4800/6416, lr 0.000100, loss 2.252727
+INFO 2021-09-19 12:19:09 train.py: 82] Epoch 17, iter 5000/6416, lr 0.000100, loss 2.259792
+INFO 2021-09-19 12:20:19 train.py: 82] Epoch 17, iter 5200/6416, lr 0.000100, loss 2.243176
+INFO 2021-09-19 12:21:30 train.py: 82] Epoch 17, iter 5400/6416, lr 0.000100, loss 2.245568
+INFO 2021-09-19 12:22:40 train.py: 82] Epoch 17, iter 5600/6416, lr 0.000100, loss 2.263195
+INFO 2021-09-19 12:23:50 train.py: 82] Epoch 17, iter 5800/6416, lr 0.000100, loss 2.232738
+INFO 2021-09-19 12:25:02 train.py: 95] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-09-19 12:25:02 train.py: 82] Epoch 17, iter 6000/6416, lr 0.000100, loss 2.255250
+INFO 2021-09-19 12:26:13 train.py: 82] Epoch 17, iter 6200/6416, lr 0.000100, loss 2.260690
+INFO 2021-09-19 12:27:24 train.py: 82] Epoch 17, iter 6400/6416, lr 0.000100, loss 2.268025
+INFO 2021-09-19 12:27:31 train.py: 100] Save checkpoint Epoch_17.pt to disk...
+INFO 2021-09-19 12:27:31 train.py: 183] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B0/.gitkeep b/bob/bio/facexzoo/models/backbones/RepVGG_B0/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..88b05cb4c553a40bca708b02bdf9c20b53a3c0d6
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9756666666666666 | 0.0019907192074632087 |
+| Epoch_15_batch_2999.pt | 0.9755 | 0.0019571016615342785 |
+| Epoch_12_batch_5999.pt | 0.9753333333333334 | 0.002469567863432542 |
+| Epoch_12.pt | 0.975 | 0.001956312984628778 |
+| Epoch_14_batch_5999.pt | 0.9743333333333333 | 0.0020964402515681337 |
+| Epoch_16.pt | 0.9743333333333333 | 0.0018459164139817941 |
+| Epoch_13_batch_5999.pt | 0.974 | 0.002111111111111113 |
+| Epoch_17_batch_2999.pt | 0.974 | 0.0021256807188565485 |
+| Epoch_13_batch_2999.pt | 0.9739999999999999 | 0.0023985592383247703 |
+| Epoch_14.pt | 0.9738333333333333 | 0.0021808651584878285 |
+| Epoch_16_batch_2999.pt | 0.9736666666666667 | 0.002105255035721823 |
+| Epoch_14_batch_2999.pt | 0.9736666666666667 | 0.0020457725155024445 |
+| Epoch_16_batch_5999.pt | 0.9735000000000001 | 0.0021723571995574567 |
+| Epoch_15_batch_5999.pt | 0.9733333333333333 | 0.0020184335693983254 |
+| Epoch_12_batch_2999.pt | 0.9733333333333333 | 0.002496911672693804 |
+| Epoch_11_batch_2999.pt | 0.9731666666666667 | 0.002085369375458154 |
+| Epoch_17_batch_5999.pt | 0.9730000000000001 | 0.002045772515502444 |
+| Epoch_11.pt | 0.9730000000000001 | 0.0020905430802474223 |
+| Epoch_13.pt | 0.9729999999999999 | 0.002163102481547977 |
+| Epoch_17.pt | 0.9720000000000001 | 0.002119864892037658 |
+| Epoch_10_batch_5999.pt | 0.9713333333333335 | 0.002328036315528548 |
+| Epoch_11_batch_5999.pt | 0.9705 | 0.0025098571106836696 |
+| Epoch_10.pt | 0.9691666666666665 | 0.0025367666792957557 |
+| Epoch_10_batch_2999.pt | 0.9678333333333333 | 0.002108916988455812 |
+| Epoch_9_batch_5999.pt | 0.9658333333333335 | 0.003576759689515523 |
+| Epoch_7_batch_2999.pt | 0.9650000000000001 | 0.0027888667551135907 |
+| Epoch_8_batch_5999.pt | 0.9621666666666666 | 0.0021523745142011676 |
+| Epoch_9_batch_2999.pt | 0.9620000000000001 | 0.004133572275052949 |
+| Epoch_7.pt | 0.9616666666666667 | 0.0028436630871266095 |
+| Epoch_6_batch_2999.pt | 0.9611666666666666 | 0.0027894200468073817 |
+| Epoch_8_batch_2999.pt | 0.9598333333333334 | 0.004212342241614973 |
+| Epoch_6_batch_5999.pt | 0.9594999999999999 | 0.0029402485827553847 |
+| Epoch_6.pt | 0.9586666666666666 | 0.0039267993437938475 |
+| Epoch_5_batch_2999.pt | 0.958 | 0.0036413265795942106 |
+| Epoch_7_batch_5999.pt | 0.9578333333333333 | 0.0035490391674854386 |
+| Epoch_4.pt | 0.9563333333333333 | 0.003581502546952496 |
+| Epoch_4_batch_5999.pt | 0.9555 | 0.003452052529534665 |
+| Epoch_4_batch_2999.pt | 0.9546666666666666 | 0.00318948890986829 |
+| Epoch_8.pt | 0.9541666666666668 | 0.003559459662142519 |
+| Epoch_9.pt | 0.9538333333333332 | 0.0052941267247562 |
+| Epoch_5_batch_5999.pt | 0.953 | 0.002567604446286967 |
+| Epoch_3_batch_5999.pt | 0.9525000000000002 | 0.0033907098932593536 |
+| Epoch_3_batch_2999.pt | 0.9518333333333333 | 0.003374285475346714 |
+| Epoch_2_batch_5999.pt | 0.9484999999999999 | 0.0033096380019125523 |
+| Epoch_5.pt | 0.9481666666666667 | 0.004256077716732179 |
+| Epoch_3.pt | 0.9448333333333332 | 0.004342227339926476 |
+| Epoch_2.pt | 0.9436666666666665 | 0.002786652489774326 |
+| Epoch_2_batch_2999.pt | 0.9398333333333333 | 0.004756087716849883 |
+| Epoch_1_batch_5999.pt | 0.9305 | 0.0053923242200212 |
+| Epoch_1.pt | 0.9286666666666665 | 0.005315943872846968 |
+| Epoch_1_batch_2999.pt | 0.909 | 0.007106769507976614 |
+| Epoch_0_batch_5999.pt | 0.8563333333333333 | 0.008072969683218984 |
+| Epoch_0.pt | 0.8521666666666666 | 0.0075279315277451695 |
+| Epoch_0_batch_2999.pt | 0.7011666666666667 | 0.005703854014373438 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..817d69401e4b13cdb6afb7a88a2a53add20a5d0d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9516666666666668 | 0.003591828861165467 |
+| Epoch_13_batch_2999.pt | 0.9514999999999999 | 0.0031569050322611944 |
+| Epoch_15_batch_5999.pt | 0.9514999999999999 | 0.0029757248312982402 |
+| Epoch_14_batch_2999.pt | 0.9513333333333334 | 0.003170076137263794 |
+| Epoch_16.pt | 0.9510000000000002 | 0.0037449554547450497 |
+| Epoch_17.pt | 0.951 | 0.0038506052113696544 |
+| Epoch_14_batch_5999.pt | 0.951 | 0.003344425987398313 |
+| Epoch_11.pt | 0.9506666666666668 | 0.003497794719710483 |
+| Epoch_11_batch_2999.pt | 0.9506666666666665 | 0.003593547028621378 |
+| Epoch_14.pt | 0.9503333333333334 | 0.0037498971179302674 |
+| Epoch_16_batch_2999.pt | 0.9503333333333334 | 0.0035468643776694125 |
+| Epoch_11_batch_5999.pt | 0.9503333333333334 | 0.003449816599168893 |
+| Epoch_13_batch_5999.pt | 0.9501666666666667 | 0.0036972629182166297 |
+| Epoch_13.pt | 0.9501666666666667 | 0.0030373193184081377 |
+| Epoch_15_batch_2999.pt | 0.9501666666666667 | 0.0035438174297371208 |
+| Epoch_17_batch_2999.pt | 0.9496666666666668 | 0.0035294178165041316 |
+| Epoch_10.pt | 0.9496666666666667 | 0.0029376231258671725 |
+| Epoch_17_batch_5999.pt | 0.9496666666666667 | 0.003965904066127721 |
+| Epoch_10_batch_5999.pt | 0.9495000000000001 | 0.003287180487219337 |
+| Epoch_12_batch_2999.pt | 0.9495000000000001 | 0.0034964708839021236 |
+| Epoch_15.pt | 0.9493333333333334 | 0.0037531879453454563 |
+| Epoch_12.pt | 0.9488333333333333 | 0.0031822229981377063 |
+| Epoch_10_batch_2999.pt | 0.9486666666666667 | 0.0031011746082117496 |
+| Epoch_12_batch_5999.pt | 0.9478333333333333 | 0.002940248582755387 |
+| Epoch_9_batch_2999.pt | 0.9421666666666667 | 0.003370624736026112 |
+| Epoch_8_batch_2999.pt | 0.9410000000000001 | 0.0031249691356500446 |
+| Epoch_6_batch_5999.pt | 0.9395 | 0.002699451247390298 |
+| Epoch_8_batch_5999.pt | 0.9395 | 0.0031234872881934603 |
+| Epoch_8.pt | 0.9393333333333332 | 0.003279189902971933 |
+| Epoch_7_batch_5999.pt | 0.9385000000000001 | 0.0037716453494430415 |
+| Epoch_9_batch_5999.pt | 0.9381666666666666 | 0.003963179294891509 |
+| Epoch_5_batch_2999.pt | 0.9375 | 0.003381595065229599 |
+| Epoch_4_batch_2999.pt | 0.9373333333333335 | 0.003969015799887048 |
+| Epoch_9.pt | 0.9373333333333334 | 0.003532914021241031 |
+| Epoch_6.pt | 0.9368333333333334 | 0.003629865827537694 |
+| Epoch_7_batch_2999.pt | 0.9366666666666668 | 0.003990729999126217 |
+| Epoch_7.pt | 0.9364999999999999 | 0.003473444229409012 |
+| Epoch_6_batch_2999.pt | 0.9364999999999999 | 0.0030271406057389458 |
+| Epoch_5_batch_5999.pt | 0.9363333333333334 | 0.0038393672318157795 |
+| Epoch_3_batch_5999.pt | 0.9355 | 0.0033522610758990198 |
+| Epoch_5.pt | 0.9350000000000002 | 0.0034694433324435523 |
+| Epoch_4_batch_5999.pt | 0.9339999999999999 | 0.0037531879453454467 |
+| Epoch_3_batch_2999.pt | 0.9315000000000001 | 0.004523259504296639 |
+| Epoch_2_batch_5999.pt | 0.9293333333333333 | 0.0027464904654018406 |
+| Epoch_4.pt | 0.9281666666666666 | 0.003612820108419801 |
+| Epoch_2.pt | 0.9261666666666667 | 0.0031627656283311698 |
+| Epoch_3.pt | 0.9258333333333335 | 0.003593976442141308 |
+| Epoch_2_batch_2999.pt | 0.9243333333333335 | 0.0034444444444444414 |
+| Epoch_1_batch_5999.pt | 0.9223333333333334 | 0.003144660377352199 |
+| Epoch_1.pt | 0.9108333333333333 | 0.003712258638286249 |
+| Epoch_1_batch_2999.pt | 0.9086666666666667 | 0.004706181907677198 |
+| Epoch_0_batch_5999.pt | 0.8735000000000002 | 0.0043422273399264695 |
+| Epoch_0.pt | 0.8675 | 0.006962199524735146 |
+| Epoch_0_batch_2999.pt | 0.725 | 0.0079115480528524 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..581a07f5de70e2524e0afacd77a80a115b84284a
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.8676666666666668 | 0.0059014540999508865 |
+| Epoch_17_batch_5999.pt | 0.8661666666666668 | 0.005848129383311438 |
+| Epoch_17_batch_2999.pt | 0.866 | 0.005753152651555382 |
+| Epoch_16.pt | 0.8648333333333333 | 0.00574912768477118 |
+| Epoch_14_batch_5999.pt | 0.8644999999999999 | 0.006373411203615397 |
+| Epoch_17.pt | 0.8643333333333333 | 0.005467073155618902 |
+| Epoch_16_batch_5999.pt | 0.8640000000000001 | 0.006142344427148998 |
+| Epoch_16_batch_2999.pt | 0.8638333333333333 | 0.005968280352862987 |
+| Epoch_15_batch_5999.pt | 0.8634999999999999 | 0.005813192743797783 |
+| Epoch_15_batch_2999.pt | 0.8625 | 0.006432220206905504 |
+| Epoch_14_batch_2999.pt | 0.8623333333333333 | 0.005806552280201741 |
+| Epoch_11_batch_5999.pt | 0.8619999999999999 | 0.00597215762238964 |
+| Epoch_13_batch_5999.pt | 0.8618333333333335 | 0.005360175508458787 |
+| Epoch_13.pt | 0.8615 | 0.005045778090959797 |
+| Epoch_15.pt | 0.861 | 0.006071589374361893 |
+| Epoch_12_batch_2999.pt | 0.8605 | 0.0060504565688349506 |
+| Epoch_13_batch_2999.pt | 0.8601666666666666 | 0.0051547053896727184 |
+| Epoch_11_batch_2999.pt | 0.8598333333333334 | 0.005990991179149073 |
+| Epoch_12_batch_5999.pt | 0.8598333333333332 | 0.005678908345800272 |
+| Epoch_10.pt | 0.8576666666666668 | 0.005551109331909688 |
+| Epoch_10_batch_2999.pt | 0.8568333333333333 | 0.005876559420041964 |
+| Epoch_12.pt | 0.8568333333333333 | 0.006148622248565571 |
+| Epoch_11.pt | 0.8566666666666668 | 0.006309898162000302 |
+| Epoch_10_batch_5999.pt | 0.8563333333333333 | 0.00475186772896513 |
+| Epoch_6_batch_5999.pt | 0.8294999999999998 | 0.006339422024517554 |
+| Epoch_9_batch_2999.pt | 0.8288333333333334 | 0.006493348923853817 |
+| Epoch_9_batch_5999.pt | 0.8241666666666667 | 0.006948887467576154 |
+| Epoch_4_batch_5999.pt | 0.8241666666666667 | 0.007963070628040447 |
+| Epoch_6_batch_2999.pt | 0.8240000000000001 | 0.006522515609528522 |
+| Epoch_7_batch_2999.pt | 0.8238333333333333 | 0.006407220078614626 |
+| Epoch_5_batch_5999.pt | 0.8238333333333333 | 0.0058322750362757554 |
+| Epoch_8_batch_5999.pt | 0.8234999999999999 | 0.006737814790282389 |
+| Epoch_7.pt | 0.8216666666666667 | 0.007294510027631041 |
+| Epoch_4_batch_2999.pt | 0.8203333333333334 | 0.00694688845881804 |
+| Epoch_5_batch_2999.pt | 0.8200000000000001 | 0.005510931896727663 |
+| Epoch_3_batch_5999.pt | 0.8195 | 0.006639063224117319 |
+| Epoch_4.pt | 0.8185 | 0.006575534527396817 |
+| Epoch_8_batch_2999.pt | 0.8185 | 0.006945333276451722 |
+| Epoch_7_batch_5999.pt | 0.8168333333333333 | 0.007384710598072473 |
+| Epoch_8.pt | 0.8166666666666667 | 0.007027283689263063 |
+| Epoch_6.pt | 0.8156666666666667 | 0.007337540880488467 |
+| Epoch_2_batch_5999.pt | 0.8148333333333333 | 0.006864877508730691 |
+| Epoch_3.pt | 0.8130000000000001 | 0.006185406959874121 |
+| Epoch_9.pt | 0.8105 | 0.006146614043319366 |
+| Epoch_3_batch_2999.pt | 0.8103333333333333 | 0.0064060156912887 |
+| Epoch_2_batch_2999.pt | 0.8076666666666668 | 0.007137105267889657 |
+| Epoch_5.pt | 0.8030000000000002 | 0.007895146188218004 |
+| Epoch_2.pt | 0.8011666666666667 | 0.00644085148883551 |
+| Epoch_1_batch_5999.pt | 0.8011666666666667 | 0.00767008780351178 |
+| Epoch_1.pt | 0.7813333333333334 | 0.005799105933644391 |
+| Epoch_1_batch_2999.pt | 0.7763333333333333 | 0.006665740676431394 |
+| Epoch_0_batch_5999.pt | 0.7313333333333334 | 0.008439325934114774 |
+| Epoch_0.pt | 0.7311666666666666 | 0.006055555555555557 |
+| Epoch_0_batch_2999.pt | 0.6271666666666667 | 0.008784280291914278 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7c32bab445e1490a24a8ef687a04b009ff6ff92a
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9971666666666665 | 0.0006596856715021067 |
+| Epoch_15.pt | 0.9969999999999999 | 0.0006938886664887116 |
+| Epoch_15_batch_5999.pt | 0.9969999999999999 | 0.0007370277311900913 |
+| Epoch_12_batch_2999.pt | 0.9969999999999999 | 0.0006938886664887116 |
+| Epoch_13_batch_2999.pt | 0.9968333333333333 | 0.0007637626158259759 |
+| Epoch_13.pt | 0.9968333333333333 | 0.0008407081083567512 |
+| Epoch_14_batch_2999.pt | 0.9968333333333333 | 0.0008407081083567512 |
+| Epoch_14.pt | 0.9966666666666667 | 0.0007856742013183834 |
+| Epoch_10_batch_2999.pt | 0.9966666666666667 | 0.0008240220541217444 |
+| Epoch_11_batch_2999.pt | 0.9966666666666665 | 0.0006573421981221823 |
+| Epoch_17_batch_2999.pt | 0.9964999999999999 | 0.001095726829073116 |
+| Epoch_12.pt | 0.9964999999999999 | 0.0007637626158259764 |
+| Epoch_11_batch_5999.pt | 0.9964999999999999 | 0.0008031573497111666 |
+| Epoch_11.pt | 0.9964999999999999 | 0.0005241100628920327 |
+| Epoch_16_batch_5999.pt | 0.9963333333333333 | 0.0011055415967851372 |
+| Epoch_16_batch_2999.pt | 0.9963333333333333 | 0.0011055415967851372 |
+| Epoch_10_batch_5999.pt | 0.9963333333333333 | 0.0008888888888888872 |
+| Epoch_17_batch_5999.pt | 0.9963333333333333 | 0.0011055415967851372 |
+| Epoch_15_batch_2999.pt | 0.9963333333333333 | 0.0011055415967851372 |
+| Epoch_12_batch_5999.pt | 0.9963333333333333 | 0.0007370277311900846 |
+| Epoch_14_batch_5999.pt | 0.9963333333333333 | 0.0008164965809277232 |
+| Epoch_16.pt | 0.9963333333333333 | 0.001160034056545621 |
+| Epoch_9_batch_2999.pt | 0.9961666666666668 | 0.0008624541497922222 |
+| Epoch_10.pt | 0.9961666666666666 | 0.0009638528651609659 |
+| Epoch_8_batch_5999.pt | 0.9961666666666666 | 0.0012921892610681153 |
+| Epoch_9_batch_5999.pt | 0.9958333333333332 | 0.0010318986456114817 |
+| Epoch_13_batch_5999.pt | 0.9956666666666667 | 0.000968644209675708 |
+| Epoch_8.pt | 0.9956666666666665 | 0.0008314794192830935 |
+| Epoch_7_batch_2999.pt | 0.9955 | 0.0015723301886761032 |
+| Epoch_4.pt | 0.9955 | 0.0007876359377087703 |
+| Epoch_6.pt | 0.9955 | 0.0009953596037316041 |
+| Epoch_5_batch_2999.pt | 0.9953333333333333 | 0.000987577157479508 |
+| Epoch_5_batch_5999.pt | 0.9953333333333333 | 0.000987577157479508 |
+| Epoch_7.pt | 0.9951666666666666 | 0.0006309898162000319 |
+| Epoch_3_batch_5999.pt | 0.9951666666666666 | 0.0011235415786753752 |
+| Epoch_6_batch_2999.pt | 0.9951666666666666 | 0.0009444444444444425 |
+| Epoch_6_batch_5999.pt | 0.9949999999999999 | 0.0009938079899999047 |
+| Epoch_4_batch_2999.pt | 0.9949999999999999 | 0.0008958064164776189 |
+| Epoch_8_batch_2999.pt | 0.9949999999999999 | 0.0012171612389003702 |
+| Epoch_7_batch_5999.pt | 0.9948333333333332 | 0.0009444444444444378 |
+| Epoch_9.pt | 0.9946666666666666 | 0.0009229582069909011 |
+| Epoch_2.pt | 0.9943333333333333 | 0.0010304020550550815 |
+| Epoch_5.pt | 0.9943333333333333 | 0.0009026709338484367 |
+| Epoch_3_batch_2999.pt | 0.9943333333333332 | 0.0009686442096757094 |
+| Epoch_4_batch_5999.pt | 0.9943333333333332 | 0.0006666666666666666 |
+| Epoch_2_batch_2999.pt | 0.9938333333333332 | 0.0010258991840344088 |
+| Epoch_3.pt | 0.9936666666666667 | 0.0011863420280034797 |
+| Epoch_2_batch_5999.pt | 0.9931666666666666 | 0.0011506841765115585 |
+| Epoch_1_batch_5999.pt | 0.9931666666666666 | 0.0013709958532503398 |
+| Epoch_1.pt | 0.9918333333333333 | 0.0009444444444444498 |
+| Epoch_1_batch_2999.pt | 0.9890000000000001 | 0.0015947444549341513 |
+| Epoch_0_batch_5999.pt | 0.9786666666666666 | 0.0019372884193514085 |
+| Epoch_0.pt | 0.9773333333333334 | 0.0022933074933944716 |
+| Epoch_0_batch_2999.pt | 0.9361666666666668 | 0.004157768276015036 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ce0f126bd43f4745c0eae874c8345c89eff884fe
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B0/accu_files/accu_megaface.txt
@@ -0,0 +1,8 @@
+INFO 2021-09-19 14:29:18 megaface_evaluator.py: 144] Rank 1 accuracy: 0.957483.
+INFO 2021-09-19 14:30:18 megaface_evaluator.py: 144] Rank 2 accuracy: 0.968047.
+INFO 2021-09-19 14:31:15 megaface_evaluator.py: 144] Rank 3 accuracy: 0.972122.
+INFO 2021-09-19 14:32:12 megaface_evaluator.py: 144] Rank 4 accuracy: 0.974459.
+INFO 2021-09-19 14:33:13 megaface_evaluator.py: 144] Rank 5 accuracy: 0.976014.
+INFO 2021-09-19 14:34:11 megaface_evaluator.py: 144] Rank 6 accuracy: 0.977433.
+INFO 2021-09-19 14:35:10 megaface_evaluator.py: 144] Rank 7 accuracy: 0.978455.
+INFO 2021-09-19 14:36:10 megaface_evaluator.py: 144] Rank 8 accuracy: 0.979288.
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B0/log.log b/bob/bio/facexzoo/models/backbones/RepVGG_B0/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..57c1f70cb0ad042c054256162896e410686573ef
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B0/log.log
@@ -0,0 +1,655 @@
+INFO 2021-09-16 19:41:47 train.py: 180] Start optimization.
+INFO 2021-09-16 19:41:47 train.py: 181] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='RepVGG', batch_size=512, data_root='/export2/wj_data/FaceX-Zoo/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-repvgg', train_file='/export2/wj_data/FaceX-Zoo/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7fc626371dd8>)
+backbone param:
+{'block_stage1': 4, 'block_stage2': 6, 'block_stage3': 16, 'block_stage4': 1, 'width_stage1': 1, 'width_stage2': 1, 'width_stage3': 1, 'width_stage4': 2.5, 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2021-09-16 19:42:14 train.py: 82] Epoch 0, iter 0/6416, lr 0.100000, loss 16.264084
+INFO 2021-09-16 19:43:50 train.py: 82] Epoch 0, iter 200/6416, lr 0.100000, loss 15.661278
+INFO 2021-09-16 19:45:26 train.py: 82] Epoch 0, iter 400/6416, lr 0.100000, loss 15.400107
+INFO 2021-09-16 19:47:02 train.py: 82] Epoch 0, iter 600/6416, lr 0.100000, loss 15.327783
+INFO 2021-09-16 19:48:38 train.py: 82] Epoch 0, iter 800/6416, lr 0.100000, loss 15.210372
+INFO 2021-09-16 19:50:14 train.py: 82] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.007365
+INFO 2021-09-16 19:51:50 train.py: 82] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.706685
+INFO 2021-09-16 19:53:25 train.py: 82] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.366101
+INFO 2021-09-16 19:55:02 train.py: 82] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.018750
+INFO 2021-09-16 19:56:38 train.py: 82] Epoch 0, iter 1800/6416, lr 0.100000, loss 13.681811
+INFO 2021-09-16 19:58:14 train.py: 82] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.310346
+INFO 2021-09-16 19:59:50 train.py: 82] Epoch 0, iter 2200/6416, lr 0.100000, loss 12.933504
+INFO 2021-09-16 20:01:26 train.py: 82] Epoch 0, iter 2400/6416, lr 0.100000, loss 12.542779
+INFO 2021-09-16 20:03:02 train.py: 82] Epoch 0, iter 2600/6416, lr 0.100000, loss 12.217304
+INFO 2021-09-16 20:04:37 train.py: 82] Epoch 0, iter 2800/6416, lr 0.100000, loss 11.967700
+INFO 2021-09-16 20:06:14 train.py: 95] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-09-16 20:06:15 train.py: 82] Epoch 0, iter 3000/6416, lr 0.100000, loss 11.815062
+INFO 2021-09-16 20:07:50 train.py: 82] Epoch 0, iter 3200/6416, lr 0.100000, loss 11.813564
+INFO 2021-09-16 20:09:26 train.py: 82] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.904291
+INFO 2021-09-16 20:11:01 train.py: 82] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.068402
+INFO 2021-09-16 20:12:36 train.py: 82] Epoch 0, iter 3800/6416, lr 0.100000, loss 12.323702
+INFO 2021-09-16 20:14:11 train.py: 82] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.622537
+INFO 2021-09-16 20:15:45 train.py: 82] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.852440
+INFO 2021-09-16 20:17:20 train.py: 82] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.061609
+INFO 2021-09-16 20:18:55 train.py: 82] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.278873
+INFO 2021-09-16 20:20:30 train.py: 82] Epoch 0, iter 4800/6416, lr 0.100000, loss 13.396634
+INFO 2021-09-16 20:22:04 train.py: 82] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.451684
+INFO 2021-09-16 20:23:39 train.py: 82] Epoch 0, iter 5200/6416, lr 0.100000, loss 13.481663
+INFO 2021-09-16 20:25:13 train.py: 82] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.411484
+INFO 2021-09-16 20:26:47 train.py: 82] Epoch 0, iter 5600/6416, lr 0.100000, loss 13.336041
+INFO 2021-09-16 20:28:21 train.py: 82] Epoch 0, iter 5800/6416, lr 0.100000, loss 13.219929
+INFO 2021-09-16 20:29:56 train.py: 95] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-09-16 20:29:57 train.py: 82] Epoch 0, iter 6000/6416, lr 0.100000, loss 13.057388
+INFO 2021-09-16 20:31:31 train.py: 82] Epoch 0, iter 6200/6416, lr 0.100000, loss 12.836571
+INFO 2021-09-16 20:33:05 train.py: 82] Epoch 0, iter 6400/6416, lr 0.100000, loss 12.648535
+INFO 2021-09-16 20:33:14 train.py: 100] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-09-16 20:33:16 train.py: 82] Epoch 1, iter 0/6416, lr 0.100000, loss 12.551365
+INFO 2021-09-16 20:34:49 train.py: 82] Epoch 1, iter 200/6416, lr 0.100000, loss 12.140407
+INFO 2021-09-16 20:36:23 train.py: 82] Epoch 1, iter 400/6416, lr 0.100000, loss 11.929699
+INFO 2021-09-16 20:37:57 train.py: 82] Epoch 1, iter 600/6416, lr 0.100000, loss 11.733131
+INFO 2021-09-16 20:39:31 train.py: 82] Epoch 1, iter 800/6416, lr 0.100000, loss 11.580440
+INFO 2021-09-16 20:41:04 train.py: 82] Epoch 1, iter 1000/6416, lr 0.100000, loss 11.342613
+INFO 2021-09-16 20:42:38 train.py: 82] Epoch 1, iter 1200/6416, lr 0.100000, loss 11.134010
+INFO 2021-09-16 20:44:11 train.py: 82] Epoch 1, iter 1400/6416, lr 0.100000, loss 10.938638
+INFO 2021-09-16 20:45:44 train.py: 82] Epoch 1, iter 1600/6416, lr 0.100000, loss 10.750035
+INFO 2021-09-16 20:47:17 train.py: 82] Epoch 1, iter 1800/6416, lr 0.100000, loss 10.549917
+INFO 2021-09-16 20:48:50 train.py: 82] Epoch 1, iter 2000/6416, lr 0.100000, loss 10.383498
+INFO 2021-09-16 20:50:24 train.py: 82] Epoch 1, iter 2200/6416, lr 0.100000, loss 10.238290
+INFO 2021-09-16 20:51:57 train.py: 82] Epoch 1, iter 2400/6416, lr 0.100000, loss 10.030859
+INFO 2021-09-16 20:53:30 train.py: 82] Epoch 1, iter 2600/6416, lr 0.100000, loss 9.890399
+INFO 2021-09-16 20:55:03 train.py: 82] Epoch 1, iter 2800/6416, lr 0.100000, loss 9.725298
+INFO 2021-09-16 20:56:38 train.py: 95] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-09-16 20:56:38 train.py: 82] Epoch 1, iter 3000/6416, lr 0.100000, loss 9.600577
+INFO 2021-09-16 20:58:11 train.py: 82] Epoch 1, iter 3200/6416, lr 0.100000, loss 9.447812
+INFO 2021-09-16 20:59:44 train.py: 82] Epoch 1, iter 3400/6416, lr 0.100000, loss 9.331855
+INFO 2021-09-16 21:01:17 train.py: 82] Epoch 1, iter 3600/6416, lr 0.100000, loss 9.206323
+INFO 2021-09-16 21:02:50 train.py: 82] Epoch 1, iter 3800/6416, lr 0.100000, loss 9.088237
+INFO 2021-09-16 21:04:23 train.py: 82] Epoch 1, iter 4000/6416, lr 0.100000, loss 9.000854
+INFO 2021-09-16 21:05:57 train.py: 82] Epoch 1, iter 4200/6416, lr 0.100000, loss 8.869513
+INFO 2021-09-16 21:07:30 train.py: 82] Epoch 1, iter 4400/6416, lr 0.100000, loss 8.769005
+INFO 2021-09-16 21:09:03 train.py: 82] Epoch 1, iter 4600/6416, lr 0.100000, loss 8.683575
+INFO 2021-09-16 21:10:36 train.py: 82] Epoch 1, iter 4800/6416, lr 0.100000, loss 8.584510
+INFO 2021-09-16 21:12:09 train.py: 82] Epoch 1, iter 5000/6416, lr 0.100000, loss 8.474653
+INFO 2021-09-16 21:13:42 train.py: 82] Epoch 1, iter 5200/6416, lr 0.100000, loss 8.406663
+INFO 2021-09-16 21:15:15 train.py: 82] Epoch 1, iter 5400/6416, lr 0.100000, loss 8.358311
+INFO 2021-09-16 21:16:48 train.py: 82] Epoch 1, iter 5600/6416, lr 0.100000, loss 8.238079
+INFO 2021-09-16 21:18:22 train.py: 82] Epoch 1, iter 5800/6416, lr 0.100000, loss 8.177464
+INFO 2021-09-16 21:19:56 train.py: 95] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-09-16 21:19:57 train.py: 82] Epoch 1, iter 6000/6416, lr 0.100000, loss 8.109052
+INFO 2021-09-16 21:21:30 train.py: 82] Epoch 1, iter 6200/6416, lr 0.100000, loss 8.034128
+INFO 2021-09-16 21:23:03 train.py: 82] Epoch 1, iter 6400/6416, lr 0.100000, loss 7.973383
+INFO 2021-09-16 21:23:12 train.py: 100] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-09-16 21:23:14 train.py: 82] Epoch 2, iter 0/6416, lr 0.100000, loss 7.866819
+INFO 2021-09-16 21:24:48 train.py: 82] Epoch 2, iter 200/6416, lr 0.100000, loss 7.309457
+INFO 2021-09-16 21:26:21 train.py: 82] Epoch 2, iter 400/6416, lr 0.100000, loss 7.256321
+INFO 2021-09-16 21:27:54 train.py: 82] Epoch 2, iter 600/6416, lr 0.100000, loss 7.348893
+INFO 2021-09-16 21:29:27 train.py: 82] Epoch 2, iter 800/6416, lr 0.100000, loss 7.359792
+INFO 2021-09-16 21:31:00 train.py: 82] Epoch 2, iter 1000/6416, lr 0.100000, loss 7.372995
+INFO 2021-09-16 21:32:33 train.py: 82] Epoch 2, iter 1200/6416, lr 0.100000, loss 7.381105
+INFO 2021-09-16 21:34:06 train.py: 82] Epoch 2, iter 1400/6416, lr 0.100000, loss 7.394651
+INFO 2021-09-16 21:35:38 train.py: 82] Epoch 2, iter 1600/6416, lr 0.100000, loss 7.331339
+INFO 2021-09-16 21:37:11 train.py: 82] Epoch 2, iter 1800/6416, lr 0.100000, loss 7.324999
+INFO 2021-09-16 21:38:43 train.py: 82] Epoch 2, iter 2000/6416, lr 0.100000, loss 7.316104
+INFO 2021-09-16 21:40:16 train.py: 82] Epoch 2, iter 2200/6416, lr 0.100000, loss 7.262855
+INFO 2021-09-16 21:41:48 train.py: 82] Epoch 2, iter 2400/6416, lr 0.100000, loss 7.285564
+INFO 2021-09-16 21:43:21 train.py: 82] Epoch 2, iter 2600/6416, lr 0.100000, loss 7.259609
+INFO 2021-09-16 21:44:53 train.py: 82] Epoch 2, iter 2800/6416, lr 0.100000, loss 7.182303
+INFO 2021-09-16 21:46:27 train.py: 95] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-09-16 21:46:27 train.py: 82] Epoch 2, iter 3000/6416, lr 0.100000, loss 7.176984
+INFO 2021-09-16 21:48:00 train.py: 82] Epoch 2, iter 3200/6416, lr 0.100000, loss 7.113152
+INFO 2021-09-16 21:49:32 train.py: 82] Epoch 2, iter 3400/6416, lr 0.100000, loss 7.103764
+INFO 2021-09-16 21:51:05 train.py: 82] Epoch 2, iter 3600/6416, lr 0.100000, loss 7.057729
+INFO 2021-09-16 21:52:38 train.py: 82] Epoch 2, iter 3800/6416, lr 0.100000, loss 7.048240
+INFO 2021-09-16 21:54:10 train.py: 82] Epoch 2, iter 4000/6416, lr 0.100000, loss 7.026205
+INFO 2021-09-16 21:55:43 train.py: 82] Epoch 2, iter 4200/6416, lr 0.100000, loss 6.988499
+INFO 2021-09-16 21:57:16 train.py: 82] Epoch 2, iter 4400/6416, lr 0.100000, loss 6.935102
+INFO 2021-09-16 21:58:49 train.py: 82] Epoch 2, iter 4600/6416, lr 0.100000, loss 6.912633
+INFO 2021-09-16 22:00:22 train.py: 82] Epoch 2, iter 4800/6416, lr 0.100000, loss 6.860900
+INFO 2021-09-16 22:01:54 train.py: 82] Epoch 2, iter 5000/6416, lr 0.100000, loss 6.872863
+INFO 2021-09-16 22:03:27 train.py: 82] Epoch 2, iter 5200/6416, lr 0.100000, loss 6.832867
+INFO 2021-09-16 22:05:00 train.py: 82] Epoch 2, iter 5400/6416, lr 0.100000, loss 6.811628
+INFO 2021-09-16 22:06:33 train.py: 82] Epoch 2, iter 5600/6416, lr 0.100000, loss 6.766112
+INFO 2021-09-16 22:08:05 train.py: 82] Epoch 2, iter 5800/6416, lr 0.100000, loss 6.734297
+INFO 2021-09-16 22:09:39 train.py: 95] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-09-16 22:09:40 train.py: 82] Epoch 2, iter 6000/6416, lr 0.100000, loss 6.710829
+INFO 2021-09-16 22:11:12 train.py: 82] Epoch 2, iter 6200/6416, lr 0.100000, loss 6.692142
+INFO 2021-09-16 22:12:45 train.py: 82] Epoch 2, iter 6400/6416, lr 0.100000, loss 6.646334
+INFO 2021-09-16 22:12:54 train.py: 100] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-09-16 22:12:56 train.py: 82] Epoch 3, iter 0/6416, lr 0.100000, loss 6.645002
+INFO 2021-09-16 22:14:29 train.py: 82] Epoch 3, iter 200/6416, lr 0.100000, loss 6.054503
+INFO 2021-09-16 22:16:02 train.py: 82] Epoch 3, iter 400/6416, lr 0.100000, loss 6.053902
+INFO 2021-09-16 22:17:35 train.py: 82] Epoch 3, iter 600/6416, lr 0.100000, loss 6.074604
+INFO 2021-09-16 22:19:07 train.py: 82] Epoch 3, iter 800/6416, lr 0.100000, loss 6.178117
+INFO 2021-09-16 22:20:39 train.py: 82] Epoch 3, iter 1000/6416, lr 0.100000, loss 6.263463
+INFO 2021-09-16 22:22:11 train.py: 82] Epoch 3, iter 1200/6416, lr 0.100000, loss 6.251473
+INFO 2021-09-16 22:23:43 train.py: 82] Epoch 3, iter 1400/6416, lr 0.100000, loss 6.265714
+INFO 2021-09-16 22:25:16 train.py: 82] Epoch 3, iter 1600/6416, lr 0.100000, loss 6.283561
+INFO 2021-09-16 22:26:48 train.py: 82] Epoch 3, iter 1800/6416, lr 0.100000, loss 6.292271
+INFO 2021-09-16 22:28:20 train.py: 82] Epoch 3, iter 2000/6416, lr 0.100000, loss 6.295875
+INFO 2021-09-16 22:29:53 train.py: 82] Epoch 3, iter 2200/6416, lr 0.100000, loss 6.287003
+INFO 2021-09-16 22:31:25 train.py: 82] Epoch 3, iter 2400/6416, lr 0.100000, loss 6.298549
+INFO 2021-09-16 22:32:57 train.py: 82] Epoch 3, iter 2600/6416, lr 0.100000, loss 6.283034
+INFO 2021-09-16 22:34:29 train.py: 82] Epoch 3, iter 2800/6416, lr 0.100000, loss 6.242559
+INFO 2021-09-16 22:36:02 train.py: 95] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-09-16 22:36:03 train.py: 82] Epoch 3, iter 3000/6416, lr 0.100000, loss 6.270848
+INFO 2021-09-16 22:37:35 train.py: 82] Epoch 3, iter 3200/6416, lr 0.100000, loss 6.222117
+INFO 2021-09-16 22:39:08 train.py: 82] Epoch 3, iter 3400/6416, lr 0.100000, loss 6.215120
+INFO 2021-09-16 22:40:40 train.py: 82] Epoch 3, iter 3600/6416, lr 0.100000, loss 6.200288
+INFO 2021-09-16 22:42:12 train.py: 82] Epoch 3, iter 3800/6416, lr 0.100000, loss 6.160233
+INFO 2021-09-16 22:43:45 train.py: 82] Epoch 3, iter 4000/6416, lr 0.100000, loss 6.177357
+INFO 2021-09-16 22:45:17 train.py: 82] Epoch 3, iter 4200/6416, lr 0.100000, loss 6.179563
+INFO 2021-09-16 22:46:49 train.py: 82] Epoch 3, iter 4400/6416, lr 0.100000, loss 6.201195
+INFO 2021-09-16 22:48:22 train.py: 82] Epoch 3, iter 4600/6416, lr 0.100000, loss 6.131716
+INFO 2021-09-16 22:49:54 train.py: 82] Epoch 3, iter 4800/6416, lr 0.100000, loss 6.110546
+INFO 2021-09-16 22:51:27 train.py: 82] Epoch 3, iter 5000/6416, lr 0.100000, loss 6.092837
+INFO 2021-09-16 22:52:59 train.py: 82] Epoch 3, iter 5200/6416, lr 0.100000, loss 6.096896
+INFO 2021-09-16 22:54:32 train.py: 82] Epoch 3, iter 5400/6416, lr 0.100000, loss 6.098371
+INFO 2021-09-16 22:56:05 train.py: 82] Epoch 3, iter 5600/6416, lr 0.100000, loss 6.097321
+INFO 2021-09-16 22:57:37 train.py: 82] Epoch 3, iter 5800/6416, lr 0.100000, loss 6.066154
+INFO 2021-09-16 22:59:11 train.py: 95] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-09-16 22:59:11 train.py: 82] Epoch 3, iter 6000/6416, lr 0.100000, loss 6.065526
+INFO 2021-09-16 23:00:43 train.py: 82] Epoch 3, iter 6200/6416, lr 0.100000, loss 6.010776
+INFO 2021-09-16 23:02:15 train.py: 82] Epoch 3, iter 6400/6416, lr 0.100000, loss 6.019417
+INFO 2021-09-16 23:02:24 train.py: 100] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-09-16 23:02:26 train.py: 82] Epoch 4, iter 0/6416, lr 0.100000, loss 5.888591
+INFO 2021-09-16 23:03:59 train.py: 82] Epoch 4, iter 200/6416, lr 0.100000, loss 5.430405
+INFO 2021-09-16 23:05:32 train.py: 82] Epoch 4, iter 400/6416, lr 0.100000, loss 5.414519
+INFO 2021-09-16 23:07:05 train.py: 82] Epoch 4, iter 600/6416, lr 0.100000, loss 5.466144
+INFO 2021-09-16 23:08:37 train.py: 82] Epoch 4, iter 800/6416, lr 0.100000, loss 5.586032
+INFO 2021-09-16 23:10:09 train.py: 82] Epoch 4, iter 1000/6416, lr 0.100000, loss 5.604651
+INFO 2021-09-16 23:11:40 train.py: 82] Epoch 4, iter 1200/6416, lr 0.100000, loss 5.655258
+INFO 2021-09-16 23:13:12 train.py: 82] Epoch 4, iter 1400/6416, lr 0.100000, loss 5.710244
+INFO 2021-09-16 23:14:44 train.py: 82] Epoch 4, iter 1600/6416, lr 0.100000, loss 5.723429
+INFO 2021-09-16 23:16:16 train.py: 82] Epoch 4, iter 1800/6416, lr 0.100000, loss 5.703756
+INFO 2021-09-16 23:17:47 train.py: 82] Epoch 4, iter 2000/6416, lr 0.100000, loss 5.744333
+INFO 2021-09-16 23:19:19 train.py: 82] Epoch 4, iter 2200/6416, lr 0.100000, loss 5.764536
+INFO 2021-09-16 23:20:51 train.py: 82] Epoch 4, iter 2400/6416, lr 0.100000, loss 5.775996
+INFO 2021-09-16 23:22:24 train.py: 82] Epoch 4, iter 2600/6416, lr 0.100000, loss 5.771736
+INFO 2021-09-16 23:23:56 train.py: 82] Epoch 4, iter 2800/6416, lr 0.100000, loss 5.751640
+INFO 2021-09-16 23:25:29 train.py: 95] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-09-16 23:25:30 train.py: 82] Epoch 4, iter 3000/6416, lr 0.100000, loss 5.751266
+INFO 2021-09-16 23:27:02 train.py: 82] Epoch 4, iter 3200/6416, lr 0.100000, loss 5.753745
+INFO 2021-09-16 23:28:35 train.py: 82] Epoch 4, iter 3400/6416, lr 0.100000, loss 5.740605
+INFO 2021-09-16 23:30:07 train.py: 82] Epoch 4, iter 3600/6416, lr 0.100000, loss 5.749350
+INFO 2021-09-16 23:31:39 train.py: 82] Epoch 4, iter 3800/6416, lr 0.100000, loss 5.726547
+INFO 2021-09-16 23:33:12 train.py: 82] Epoch 4, iter 4000/6416, lr 0.100000, loss 5.733989
+INFO 2021-09-16 23:34:44 train.py: 82] Epoch 4, iter 4200/6416, lr 0.100000, loss 5.709351
+INFO 2021-09-16 23:36:17 train.py: 82] Epoch 4, iter 4400/6416, lr 0.100000, loss 5.689769
+INFO 2021-09-16 23:37:50 train.py: 82] Epoch 4, iter 4600/6416, lr 0.100000, loss 5.697974
+INFO 2021-09-16 23:39:22 train.py: 82] Epoch 4, iter 4800/6416, lr 0.100000, loss 5.698458
+INFO 2021-09-16 23:40:54 train.py: 82] Epoch 4, iter 5000/6416, lr 0.100000, loss 5.657177
+INFO 2021-09-16 23:42:27 train.py: 82] Epoch 4, iter 5200/6416, lr 0.100000, loss 5.667414
+INFO 2021-09-16 23:43:59 train.py: 82] Epoch 4, iter 5400/6416, lr 0.100000, loss 5.684025
+INFO 2021-09-16 23:45:31 train.py: 82] Epoch 4, iter 5600/6416, lr 0.100000, loss 5.649860
+INFO 2021-09-16 23:47:04 train.py: 82] Epoch 4, iter 5800/6416, lr 0.100000, loss 5.644955
+INFO 2021-09-16 23:48:38 train.py: 95] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-09-16 23:48:38 train.py: 82] Epoch 4, iter 6000/6416, lr 0.100000, loss 5.605525
+INFO 2021-09-16 23:50:11 train.py: 82] Epoch 4, iter 6200/6416, lr 0.100000, loss 5.603739
+INFO 2021-09-16 23:51:43 train.py: 82] Epoch 4, iter 6400/6416, lr 0.100000, loss 5.586209
+INFO 2021-09-16 23:51:52 train.py: 100] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-09-16 23:51:54 train.py: 82] Epoch 5, iter 0/6416, lr 0.100000, loss 5.569964
+INFO 2021-09-16 23:53:27 train.py: 82] Epoch 5, iter 200/6416, lr 0.100000, loss 5.059988
+INFO 2021-09-16 23:55:01 train.py: 82] Epoch 5, iter 400/6416, lr 0.100000, loss 5.075252
+INFO 2021-09-16 23:56:33 train.py: 82] Epoch 5, iter 600/6416, lr 0.100000, loss 5.147770
+INFO 2021-09-16 23:58:05 train.py: 82] Epoch 5, iter 800/6416, lr 0.100000, loss 5.204382
+INFO 2021-09-16 23:59:37 train.py: 82] Epoch 5, iter 1000/6416, lr 0.100000, loss 5.264986
+INFO 2021-09-17 00:01:08 train.py: 82] Epoch 5, iter 1200/6416, lr 0.100000, loss 5.296878
+INFO 2021-09-17 00:02:41 train.py: 82] Epoch 5, iter 1400/6416, lr 0.100000, loss 5.301634
+INFO 2021-09-17 00:04:13 train.py: 82] Epoch 5, iter 1600/6416, lr 0.100000, loss 5.352289
+INFO 2021-09-17 00:05:45 train.py: 82] Epoch 5, iter 1800/6416, lr 0.100000, loss 5.393488
+INFO 2021-09-17 00:07:17 train.py: 82] Epoch 5, iter 2000/6416, lr 0.100000, loss 5.393345
+INFO 2021-09-17 00:08:49 train.py: 82] Epoch 5, iter 2200/6416, lr 0.100000, loss 5.417219
+INFO 2021-09-17 00:10:21 train.py: 82] Epoch 5, iter 2400/6416, lr 0.100000, loss 5.420899
+INFO 2021-09-17 00:11:53 train.py: 82] Epoch 5, iter 2600/6416, lr 0.100000, loss 5.427281
+INFO 2021-09-17 00:13:25 train.py: 82] Epoch 5, iter 2800/6416, lr 0.100000, loss 5.420875
+INFO 2021-09-17 00:14:58 train.py: 95] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-09-17 00:14:59 train.py: 82] Epoch 5, iter 3000/6416, lr 0.100000, loss 5.417903
+INFO 2021-09-17 00:16:31 train.py: 82] Epoch 5, iter 3200/6416, lr 0.100000, loss 5.412522
+INFO 2021-09-17 00:18:03 train.py: 82] Epoch 5, iter 3400/6416, lr 0.100000, loss 5.411889
+INFO 2021-09-17 00:19:36 train.py: 82] Epoch 5, iter 3600/6416, lr 0.100000, loss 5.422176
+INFO 2021-09-17 00:21:08 train.py: 82] Epoch 5, iter 3800/6416, lr 0.100000, loss 5.404112
+INFO 2021-09-17 00:22:40 train.py: 82] Epoch 5, iter 4000/6416, lr 0.100000, loss 5.405759
+INFO 2021-09-17 00:24:12 train.py: 82] Epoch 5, iter 4200/6416, lr 0.100000, loss 5.400752
+INFO 2021-09-17 00:25:45 train.py: 82] Epoch 5, iter 4400/6416, lr 0.100000, loss 5.389306
+INFO 2021-09-17 00:27:16 train.py: 82] Epoch 5, iter 4600/6416, lr 0.100000, loss 5.395241
+INFO 2021-09-17 00:28:49 train.py: 82] Epoch 5, iter 4800/6416, lr 0.100000, loss 5.378862
+INFO 2021-09-17 00:30:20 train.py: 82] Epoch 5, iter 5000/6416, lr 0.100000, loss 5.387434
+INFO 2021-09-17 00:31:53 train.py: 82] Epoch 5, iter 5200/6416, lr 0.100000, loss 5.367028
+INFO 2021-09-17 00:33:25 train.py: 82] Epoch 5, iter 5400/6416, lr 0.100000, loss 5.385673
+INFO 2021-09-17 00:34:57 train.py: 82] Epoch 5, iter 5600/6416, lr 0.100000, loss 5.383887
+INFO 2021-09-17 00:36:30 train.py: 82] Epoch 5, iter 5800/6416, lr 0.100000, loss 5.350840
+INFO 2021-09-17 00:38:03 train.py: 95] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-09-17 00:38:03 train.py: 82] Epoch 5, iter 6000/6416, lr 0.100000, loss 5.348172
+INFO 2021-09-17 00:39:36 train.py: 82] Epoch 5, iter 6200/6416, lr 0.100000, loss 5.358875
+INFO 2021-09-17 00:41:08 train.py: 82] Epoch 5, iter 6400/6416, lr 0.100000, loss 5.334617
+INFO 2021-09-17 00:41:17 train.py: 100] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-09-17 00:41:19 train.py: 82] Epoch 6, iter 0/6416, lr 0.100000, loss 5.331950
+INFO 2021-09-17 00:42:52 train.py: 82] Epoch 6, iter 200/6416, lr 0.100000, loss 4.825155
+INFO 2021-09-17 00:44:24 train.py: 82] Epoch 6, iter 400/6416, lr 0.100000, loss 4.771967
+INFO 2021-09-17 00:45:56 train.py: 82] Epoch 6, iter 600/6416, lr 0.100000, loss 4.850631
+INFO 2021-09-17 00:47:28 train.py: 82] Epoch 6, iter 800/6416, lr 0.100000, loss 4.940105
+INFO 2021-09-17 00:49:00 train.py: 82] Epoch 6, iter 1000/6416, lr 0.100000, loss 4.996032
+INFO 2021-09-17 00:50:32 train.py: 82] Epoch 6, iter 1200/6416, lr 0.100000, loss 5.031637
+INFO 2021-09-17 00:52:04 train.py: 82] Epoch 6, iter 1400/6416, lr 0.100000, loss 5.116790
+INFO 2021-09-17 00:53:35 train.py: 82] Epoch 6, iter 1600/6416, lr 0.100000, loss 5.092117
+INFO 2021-09-17 00:55:07 train.py: 82] Epoch 6, iter 1800/6416, lr 0.100000, loss 5.144563
+INFO 2021-09-17 00:56:39 train.py: 82] Epoch 6, iter 2000/6416, lr 0.100000, loss 5.162694
+INFO 2021-09-17 00:58:11 train.py: 82] Epoch 6, iter 2200/6416, lr 0.100000, loss 5.170516
+INFO 2021-09-17 00:59:43 train.py: 82] Epoch 6, iter 2400/6416, lr 0.100000, loss 5.199002
+INFO 2021-09-17 01:01:15 train.py: 82] Epoch 6, iter 2600/6416, lr 0.100000, loss 5.161922
+INFO 2021-09-17 01:02:47 train.py: 82] Epoch 6, iter 2800/6416, lr 0.100000, loss 5.158922
+INFO 2021-09-17 01:04:20 train.py: 95] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-09-17 01:04:20 train.py: 82] Epoch 6, iter 3000/6416, lr 0.100000, loss 5.188388
+INFO 2021-09-17 01:05:53 train.py: 82] Epoch 6, iter 3200/6416, lr 0.100000, loss 5.188697
+INFO 2021-09-17 01:07:25 train.py: 82] Epoch 6, iter 3400/6416, lr 0.100000, loss 5.201222
+INFO 2021-09-17 01:08:58 train.py: 82] Epoch 6, iter 3600/6416, lr 0.100000, loss 5.198324
+INFO 2021-09-17 01:10:30 train.py: 82] Epoch 6, iter 3800/6416, lr 0.100000, loss 5.219489
+INFO 2021-09-17 01:12:03 train.py: 82] Epoch 6, iter 4000/6416, lr 0.100000, loss 5.214616
+INFO 2021-09-17 01:13:35 train.py: 82] Epoch 6, iter 4200/6416, lr 0.100000, loss 5.184642
+INFO 2021-09-17 01:15:08 train.py: 82] Epoch 6, iter 4400/6416, lr 0.100000, loss 5.221622
+INFO 2021-09-17 01:16:41 train.py: 82] Epoch 6, iter 4600/6416, lr 0.100000, loss 5.183578
+INFO 2021-09-17 01:18:13 train.py: 82] Epoch 6, iter 4800/6416, lr 0.100000, loss 5.195134
+INFO 2021-09-17 01:19:46 train.py: 82] Epoch 6, iter 5000/6416, lr 0.100000, loss 5.233120
+INFO 2021-09-17 01:21:19 train.py: 82] Epoch 6, iter 5200/6416, lr 0.100000, loss 5.198898
+INFO 2021-09-17 01:22:52 train.py: 82] Epoch 6, iter 5400/6416, lr 0.100000, loss 5.187761
+INFO 2021-09-17 01:24:24 train.py: 82] Epoch 6, iter 5600/6416, lr 0.100000, loss 5.161280
+INFO 2021-09-17 01:25:57 train.py: 82] Epoch 6, iter 5800/6416, lr 0.100000, loss 5.214184
+INFO 2021-09-17 01:27:31 train.py: 95] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-09-17 01:27:31 train.py: 82] Epoch 6, iter 6000/6416, lr 0.100000, loss 5.141961
+INFO 2021-09-17 01:29:04 train.py: 82] Epoch 6, iter 6200/6416, lr 0.100000, loss 5.141327
+INFO 2021-09-17 01:30:37 train.py: 82] Epoch 6, iter 6400/6416, lr 0.100000, loss 5.140300
+INFO 2021-09-17 01:30:45 train.py: 100] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-09-17 01:30:47 train.py: 82] Epoch 7, iter 0/6416, lr 0.100000, loss 5.092330
+INFO 2021-09-17 01:32:21 train.py: 82] Epoch 7, iter 200/6416, lr 0.100000, loss 4.605258
+INFO 2021-09-17 01:33:53 train.py: 82] Epoch 7, iter 400/6416, lr 0.100000, loss 4.606459
+INFO 2021-09-17 01:35:25 train.py: 82] Epoch 7, iter 600/6416, lr 0.100000, loss 4.684873
+INFO 2021-09-17 01:36:57 train.py: 82] Epoch 7, iter 800/6416, lr 0.100000, loss 4.770544
+INFO 2021-09-17 01:38:29 train.py: 82] Epoch 7, iter 1000/6416, lr 0.100000, loss 4.797153
+INFO 2021-09-17 01:40:01 train.py: 82] Epoch 7, iter 1200/6416, lr 0.100000, loss 4.847652
+INFO 2021-09-17 01:41:33 train.py: 82] Epoch 7, iter 1400/6416, lr 0.100000, loss 4.910492
+INFO 2021-09-17 01:43:06 train.py: 82] Epoch 7, iter 1600/6416, lr 0.100000, loss 4.924279
+INFO 2021-09-17 01:44:38 train.py: 82] Epoch 7, iter 1800/6416, lr 0.100000, loss 4.948795
+INFO 2021-09-17 01:46:10 train.py: 82] Epoch 7, iter 2000/6416, lr 0.100000, loss 4.955816
+INFO 2021-09-17 01:47:42 train.py: 82] Epoch 7, iter 2200/6416, lr 0.100000, loss 5.003122
+INFO 2021-09-17 01:49:15 train.py: 82] Epoch 7, iter 2400/6416, lr 0.100000, loss 4.991083
+INFO 2021-09-17 01:50:47 train.py: 82] Epoch 7, iter 2600/6416, lr 0.100000, loss 5.028439
+INFO 2021-09-17 01:52:19 train.py: 82] Epoch 7, iter 2800/6416, lr 0.100000, loss 5.015730
+INFO 2021-09-17 01:53:53 train.py: 95] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-09-17 01:53:54 train.py: 82] Epoch 7, iter 3000/6416, lr 0.100000, loss 5.048214
+INFO 2021-09-17 01:55:26 train.py: 82] Epoch 7, iter 3200/6416, lr 0.100000, loss 4.987984
+INFO 2021-09-17 01:56:58 train.py: 82] Epoch 7, iter 3400/6416, lr 0.100000, loss 5.035593
+INFO 2021-09-17 01:58:31 train.py: 82] Epoch 7, iter 3600/6416, lr 0.100000, loss 5.003642
+INFO 2021-09-17 02:00:04 train.py: 82] Epoch 7, iter 3800/6416, lr 0.100000, loss 5.051496
+INFO 2021-09-17 02:01:36 train.py: 82] Epoch 7, iter 4000/6416, lr 0.100000, loss 5.021516
+INFO 2021-09-17 02:03:09 train.py: 82] Epoch 7, iter 4200/6416, lr 0.100000, loss 5.041113
+INFO 2021-09-17 02:04:41 train.py: 82] Epoch 7, iter 4400/6416, lr 0.100000, loss 5.012691
+INFO 2021-09-17 02:06:14 train.py: 82] Epoch 7, iter 4600/6416, lr 0.100000, loss 5.020426
+INFO 2021-09-17 02:07:46 train.py: 82] Epoch 7, iter 4800/6416, lr 0.100000, loss 5.013516
+INFO 2021-09-17 02:09:18 train.py: 82] Epoch 7, iter 5000/6416, lr 0.100000, loss 5.049451
+INFO 2021-09-17 02:10:50 train.py: 82] Epoch 7, iter 5200/6416, lr 0.100000, loss 5.008332
+INFO 2021-09-17 02:12:22 train.py: 82] Epoch 7, iter 5400/6416, lr 0.100000, loss 4.989865
+INFO 2021-09-17 02:13:54 train.py: 82] Epoch 7, iter 5600/6416, lr 0.100000, loss 5.055245
+INFO 2021-09-17 02:15:26 train.py: 82] Epoch 7, iter 5800/6416, lr 0.100000, loss 5.037957
+INFO 2021-09-17 02:16:59 train.py: 95] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-09-17 02:17:00 train.py: 82] Epoch 7, iter 6000/6416, lr 0.100000, loss 4.979538
+INFO 2021-09-17 02:18:32 train.py: 82] Epoch 7, iter 6200/6416, lr 0.100000, loss 5.011980
+INFO 2021-09-17 02:20:04 train.py: 82] Epoch 7, iter 6400/6416, lr 0.100000, loss 4.998070
+INFO 2021-09-17 02:20:13 train.py: 100] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-09-17 02:20:15 train.py: 82] Epoch 8, iter 0/6416, lr 0.100000, loss 4.929272
+INFO 2021-09-17 02:21:48 train.py: 82] Epoch 8, iter 200/6416, lr 0.100000, loss 4.475036
+INFO 2021-09-17 02:23:20 train.py: 82] Epoch 8, iter 400/6416, lr 0.100000, loss 4.467852
+INFO 2021-09-17 02:24:51 train.py: 82] Epoch 8, iter 600/6416, lr 0.100000, loss 4.529003
+INFO 2021-09-17 02:26:24 train.py: 82] Epoch 8, iter 800/6416, lr 0.100000, loss 4.605097
+INFO 2021-09-17 02:27:55 train.py: 82] Epoch 8, iter 1000/6416, lr 0.100000, loss 4.688232
+INFO 2021-09-17 02:29:27 train.py: 82] Epoch 8, iter 1200/6416, lr 0.100000, loss 4.681448
+INFO 2021-09-17 02:30:59 train.py: 82] Epoch 8, iter 1400/6416, lr 0.100000, loss 4.733311
+INFO 2021-09-17 02:32:31 train.py: 82] Epoch 8, iter 1600/6416, lr 0.100000, loss 4.803483
+INFO 2021-09-17 02:34:03 train.py: 82] Epoch 8, iter 1800/6416, lr 0.100000, loss 4.808274
+INFO 2021-09-17 02:35:34 train.py: 82] Epoch 8, iter 2000/6416, lr 0.100000, loss 4.823631
+INFO 2021-09-17 02:37:06 train.py: 82] Epoch 8, iter 2200/6416, lr 0.100000, loss 4.834197
+INFO 2021-09-17 02:38:38 train.py: 82] Epoch 8, iter 2400/6416, lr 0.100000, loss 4.858599
+INFO 2021-09-17 02:40:11 train.py: 82] Epoch 8, iter 2600/6416, lr 0.100000, loss 4.900997
+INFO 2021-09-17 02:41:43 train.py: 82] Epoch 8, iter 2800/6416, lr 0.100000, loss 4.887682
+INFO 2021-09-17 02:43:16 train.py: 95] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-09-17 02:43:17 train.py: 82] Epoch 8, iter 3000/6416, lr 0.100000, loss 4.881066
+INFO 2021-09-17 02:44:49 train.py: 82] Epoch 8, iter 3200/6416, lr 0.100000, loss 4.887744
+INFO 2021-09-17 02:46:22 train.py: 82] Epoch 8, iter 3400/6416, lr 0.100000, loss 4.892846
+INFO 2021-09-17 02:47:54 train.py: 82] Epoch 8, iter 3600/6416, lr 0.100000, loss 4.915452
+INFO 2021-09-17 02:49:26 train.py: 82] Epoch 8, iter 3800/6416, lr 0.100000, loss 4.919376
+INFO 2021-09-17 02:50:58 train.py: 82] Epoch 8, iter 4000/6416, lr 0.100000, loss 4.872629
+INFO 2021-09-17 02:52:31 train.py: 82] Epoch 8, iter 4200/6416, lr 0.100000, loss 4.895602
+INFO 2021-09-17 02:54:03 train.py: 82] Epoch 8, iter 4400/6416, lr 0.100000, loss 4.877744
+INFO 2021-09-17 02:55:36 train.py: 82] Epoch 8, iter 4600/6416, lr 0.100000, loss 4.900730
+INFO 2021-09-17 02:57:08 train.py: 82] Epoch 8, iter 4800/6416, lr 0.100000, loss 4.903669
+INFO 2021-09-17 02:58:41 train.py: 82] Epoch 8, iter 5000/6416, lr 0.100000, loss 4.902806
+INFO 2021-09-17 03:00:13 train.py: 82] Epoch 8, iter 5200/6416, lr 0.100000, loss 4.873756
+INFO 2021-09-17 03:01:45 train.py: 82] Epoch 8, iter 5400/6416, lr 0.100000, loss 4.888844
+INFO 2021-09-17 03:03:18 train.py: 82] Epoch 8, iter 5600/6416, lr 0.100000, loss 4.901817
+INFO 2021-09-17 03:04:50 train.py: 82] Epoch 8, iter 5800/6416, lr 0.100000, loss 4.921406
+INFO 2021-09-17 03:06:23 train.py: 95] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-09-17 03:06:24 train.py: 82] Epoch 8, iter 6000/6416, lr 0.100000, loss 4.888860
+INFO 2021-09-17 03:07:56 train.py: 82] Epoch 8, iter 6200/6416, lr 0.100000, loss 4.890106
+INFO 2021-09-17 03:09:29 train.py: 82] Epoch 8, iter 6400/6416, lr 0.100000, loss 4.892730
+INFO 2021-09-17 03:09:38 train.py: 100] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-09-17 03:09:40 train.py: 82] Epoch 9, iter 0/6416, lr 0.100000, loss 4.880278
+INFO 2021-09-17 03:11:13 train.py: 82] Epoch 9, iter 200/6416, lr 0.100000, loss 4.410804
+INFO 2021-09-17 03:12:45 train.py: 82] Epoch 9, iter 400/6416, lr 0.100000, loss 4.336949
+INFO 2021-09-17 03:14:17 train.py: 82] Epoch 9, iter 600/6416, lr 0.100000, loss 4.400163
+INFO 2021-09-17 03:15:48 train.py: 82] Epoch 9, iter 800/6416, lr 0.100000, loss 4.471257
+INFO 2021-09-17 03:17:20 train.py: 82] Epoch 9, iter 1000/6416, lr 0.100000, loss 4.553718
+INFO 2021-09-17 03:18:52 train.py: 82] Epoch 9, iter 1200/6416, lr 0.100000, loss 4.603666
+INFO 2021-09-17 03:20:24 train.py: 82] Epoch 9, iter 1400/6416, lr 0.100000, loss 4.638796
+INFO 2021-09-17 03:21:56 train.py: 82] Epoch 9, iter 1600/6416, lr 0.100000, loss 4.664761
+INFO 2021-09-17 03:23:28 train.py: 82] Epoch 9, iter 1800/6416, lr 0.100000, loss 4.692238
+INFO 2021-09-17 03:25:00 train.py: 82] Epoch 9, iter 2000/6416, lr 0.100000, loss 4.725207
+INFO 2021-09-17 03:26:32 train.py: 82] Epoch 9, iter 2200/6416, lr 0.100000, loss 4.717188
+INFO 2021-09-17 03:28:04 train.py: 82] Epoch 9, iter 2400/6416, lr 0.100000, loss 4.759216
+INFO 2021-09-17 03:29:36 train.py: 82] Epoch 9, iter 2600/6416, lr 0.100000, loss 4.769640
+INFO 2021-09-17 03:31:08 train.py: 82] Epoch 9, iter 2800/6416, lr 0.100000, loss 4.790487
+INFO 2021-09-17 03:32:41 train.py: 95] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-09-17 03:32:42 train.py: 82] Epoch 9, iter 3000/6416, lr 0.100000, loss 4.794496
+INFO 2021-09-17 03:34:14 train.py: 82] Epoch 9, iter 3200/6416, lr 0.100000, loss 4.799691
+INFO 2021-09-17 03:35:47 train.py: 82] Epoch 9, iter 3400/6416, lr 0.100000, loss 4.792707
+INFO 2021-09-17 03:37:19 train.py: 82] Epoch 9, iter 3600/6416, lr 0.100000, loss 4.776469
+INFO 2021-09-17 03:38:51 train.py: 82] Epoch 9, iter 3800/6416, lr 0.100000, loss 4.774457
+INFO 2021-09-17 03:40:24 train.py: 82] Epoch 9, iter 4000/6416, lr 0.100000, loss 4.768630
+INFO 2021-09-17 03:41:56 train.py: 82] Epoch 9, iter 4200/6416, lr 0.100000, loss 4.754323
+INFO 2021-09-17 03:43:29 train.py: 82] Epoch 9, iter 4400/6416, lr 0.100000, loss 4.778079
+INFO 2021-09-17 03:45:01 train.py: 82] Epoch 9, iter 4600/6416, lr 0.100000, loss 4.810099
+INFO 2021-09-17 03:46:34 train.py: 82] Epoch 9, iter 4800/6416, lr 0.100000, loss 4.824768
+INFO 2021-09-17 03:48:06 train.py: 82] Epoch 9, iter 5000/6416, lr 0.100000, loss 4.788132
+INFO 2021-09-17 03:49:39 train.py: 82] Epoch 9, iter 5200/6416, lr 0.100000, loss 4.796879
+INFO 2021-09-17 03:51:11 train.py: 82] Epoch 9, iter 5400/6416, lr 0.100000, loss 4.790698
+INFO 2021-09-17 03:52:44 train.py: 82] Epoch 9, iter 5600/6416, lr 0.100000, loss 4.786067
+INFO 2021-09-17 03:54:16 train.py: 82] Epoch 9, iter 5800/6416, lr 0.100000, loss 4.789535
+INFO 2021-09-17 03:55:50 train.py: 95] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-09-17 03:55:51 train.py: 82] Epoch 9, iter 6000/6416, lr 0.100000, loss 4.785182
+INFO 2021-09-17 03:57:24 train.py: 82] Epoch 9, iter 6200/6416, lr 0.100000, loss 4.799157
+INFO 2021-09-17 03:58:56 train.py: 82] Epoch 9, iter 6400/6416, lr 0.100000, loss 4.783874
+INFO 2021-09-17 03:59:05 train.py: 100] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-09-17 03:59:07 train.py: 82] Epoch 10, iter 0/6416, lr 0.010000, loss 4.803516
+INFO 2021-09-17 04:00:39 train.py: 82] Epoch 10, iter 200/6416, lr 0.010000, loss 3.593149
+INFO 2021-09-17 04:02:12 train.py: 82] Epoch 10, iter 400/6416, lr 0.010000, loss 3.343426
+INFO 2021-09-17 04:03:43 train.py: 82] Epoch 10, iter 600/6416, lr 0.010000, loss 3.258991
+INFO 2021-09-17 04:05:15 train.py: 82] Epoch 10, iter 800/6416, lr 0.010000, loss 3.168543
+INFO 2021-09-17 04:06:47 train.py: 82] Epoch 10, iter 1000/6416, lr 0.010000, loss 3.107442
+INFO 2021-09-17 04:08:19 train.py: 82] Epoch 10, iter 1200/6416, lr 0.010000, loss 3.077062
+INFO 2021-09-17 04:09:51 train.py: 82] Epoch 10, iter 1400/6416, lr 0.010000, loss 3.056982
+INFO 2021-09-17 04:11:23 train.py: 82] Epoch 10, iter 1600/6416, lr 0.010000, loss 3.009703
+INFO 2021-09-17 04:12:55 train.py: 82] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.979670
+INFO 2021-09-17 04:14:26 train.py: 82] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.961445
+INFO 2021-09-17 04:15:58 train.py: 82] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.922332
+INFO 2021-09-17 04:17:30 train.py: 82] Epoch 10, iter 2400/6416, lr 0.010000, loss 2.879532
+INFO 2021-09-17 04:19:02 train.py: 82] Epoch 10, iter 2600/6416, lr 0.010000, loss 2.905262
+INFO 2021-09-17 04:20:34 train.py: 82] Epoch 10, iter 2800/6416, lr 0.010000, loss 2.866600
+INFO 2021-09-17 04:22:06 train.py: 95] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-09-17 04:22:07 train.py: 82] Epoch 10, iter 3000/6416, lr 0.010000, loss 2.855847
+INFO 2021-09-17 04:23:39 train.py: 82] Epoch 10, iter 3200/6416, lr 0.010000, loss 2.841305
+INFO 2021-09-17 04:25:11 train.py: 82] Epoch 10, iter 3400/6416, lr 0.010000, loss 2.802454
+INFO 2021-09-17 04:26:44 train.py: 82] Epoch 10, iter 3600/6416, lr 0.010000, loss 2.763495
+INFO 2021-09-17 04:28:16 train.py: 82] Epoch 10, iter 3800/6416, lr 0.010000, loss 2.774854
+INFO 2021-09-17 04:29:49 train.py: 82] Epoch 10, iter 4000/6416, lr 0.010000, loss 2.752714
+INFO 2021-09-17 04:31:21 train.py: 82] Epoch 10, iter 4200/6416, lr 0.010000, loss 2.738505
+INFO 2021-09-17 04:32:54 train.py: 82] Epoch 10, iter 4400/6416, lr 0.010000, loss 2.714128
+INFO 2021-09-17 04:34:26 train.py: 82] Epoch 10, iter 4600/6416, lr 0.010000, loss 2.715667
+INFO 2021-09-17 04:35:59 train.py: 82] Epoch 10, iter 4800/6416, lr 0.010000, loss 2.692296
+INFO 2021-09-17 04:37:31 train.py: 82] Epoch 10, iter 5000/6416, lr 0.010000, loss 2.686138
+INFO 2021-09-17 04:39:04 train.py: 82] Epoch 10, iter 5200/6416, lr 0.010000, loss 2.674896
+INFO 2021-09-17 04:40:37 train.py: 82] Epoch 10, iter 5400/6416, lr 0.010000, loss 2.647006
+INFO 2021-09-17 04:42:10 train.py: 82] Epoch 10, iter 5600/6416, lr 0.010000, loss 2.635188
+INFO 2021-09-17 04:43:42 train.py: 82] Epoch 10, iter 5800/6416, lr 0.010000, loss 2.647114
+INFO 2021-09-17 04:45:16 train.py: 95] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-09-17 04:45:17 train.py: 82] Epoch 10, iter 6000/6416, lr 0.010000, loss 2.614934
+INFO 2021-09-17 04:46:50 train.py: 82] Epoch 10, iter 6200/6416, lr 0.010000, loss 2.614308
+INFO 2021-09-17 04:48:22 train.py: 82] Epoch 10, iter 6400/6416, lr 0.010000, loss 2.619571
+INFO 2021-09-17 04:48:31 train.py: 100] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-09-17 04:48:33 train.py: 82] Epoch 11, iter 0/6416, lr 0.010000, loss 2.533707
+INFO 2021-09-17 04:50:06 train.py: 82] Epoch 11, iter 200/6416, lr 0.010000, loss 2.318053
+INFO 2021-09-17 04:51:38 train.py: 82] Epoch 11, iter 400/6416, lr 0.010000, loss 2.288318
+INFO 2021-09-17 04:53:11 train.py: 82] Epoch 11, iter 600/6416, lr 0.010000, loss 2.287960
+INFO 2021-09-17 04:54:43 train.py: 82] Epoch 11, iter 800/6416, lr 0.010000, loss 2.289226
+INFO 2021-09-17 04:56:16 train.py: 82] Epoch 11, iter 1000/6416, lr 0.010000, loss 2.303832
+INFO 2021-09-17 04:57:48 train.py: 82] Epoch 11, iter 1200/6416, lr 0.010000, loss 2.293429
+INFO 2021-09-17 04:59:19 train.py: 82] Epoch 11, iter 1400/6416, lr 0.010000, loss 2.291608
+INFO 2021-09-17 05:00:51 train.py: 82] Epoch 11, iter 1600/6416, lr 0.010000, loss 2.280869
+INFO 2021-09-17 05:02:23 train.py: 82] Epoch 11, iter 1800/6416, lr 0.010000, loss 2.298780
+INFO 2021-09-17 05:03:55 train.py: 82] Epoch 11, iter 2000/6416, lr 0.010000, loss 2.286786
+INFO 2021-09-17 05:05:27 train.py: 82] Epoch 11, iter 2200/6416, lr 0.010000, loss 2.281992
+INFO 2021-09-17 05:06:59 train.py: 82] Epoch 11, iter 2400/6416, lr 0.010000, loss 2.283318
+INFO 2021-09-17 05:08:32 train.py: 82] Epoch 11, iter 2600/6416, lr 0.010000, loss 2.300605
+INFO 2021-09-17 05:10:04 train.py: 82] Epoch 11, iter 2800/6416, lr 0.010000, loss 2.303029
+INFO 2021-09-17 05:11:37 train.py: 95] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-09-17 05:11:38 train.py: 82] Epoch 11, iter 3000/6416, lr 0.010000, loss 2.279977
+INFO 2021-09-17 05:13:10 train.py: 82] Epoch 11, iter 3200/6416, lr 0.010000, loss 2.305846
+INFO 2021-09-17 05:14:43 train.py: 82] Epoch 11, iter 3400/6416, lr 0.010000, loss 2.288871
+INFO 2021-09-17 05:16:15 train.py: 82] Epoch 11, iter 3600/6416, lr 0.010000, loss 2.273834
+INFO 2021-09-17 05:17:48 train.py: 82] Epoch 11, iter 3800/6416, lr 0.010000, loss 2.301581
+INFO 2021-09-17 05:19:20 train.py: 82] Epoch 11, iter 4000/6416, lr 0.010000, loss 2.293179
+INFO 2021-09-17 05:20:53 train.py: 82] Epoch 11, iter 4200/6416, lr 0.010000, loss 2.298645
+INFO 2021-09-17 05:22:25 train.py: 82] Epoch 11, iter 4400/6416, lr 0.010000, loss 2.276742
+INFO 2021-09-17 05:23:58 train.py: 82] Epoch 11, iter 4600/6416, lr 0.010000, loss 2.254408
+INFO 2021-09-17 05:25:30 train.py: 82] Epoch 11, iter 4800/6416, lr 0.010000, loss 2.279991
+INFO 2021-09-17 05:27:02 train.py: 82] Epoch 11, iter 5000/6416, lr 0.010000, loss 2.291928
+INFO 2021-09-17 05:28:35 train.py: 82] Epoch 11, iter 5200/6416, lr 0.010000, loss 2.271676
+INFO 2021-09-17 05:30:07 train.py: 82] Epoch 11, iter 5400/6416, lr 0.010000, loss 2.287489
+INFO 2021-09-17 05:31:40 train.py: 82] Epoch 11, iter 5600/6416, lr 0.010000, loss 2.288821
+INFO 2021-09-17 05:33:12 train.py: 82] Epoch 11, iter 5800/6416, lr 0.010000, loss 2.262747
+INFO 2021-09-17 05:34:46 train.py: 95] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-09-17 05:34:46 train.py: 82] Epoch 11, iter 6000/6416, lr 0.010000, loss 2.285672
+INFO 2021-09-17 05:36:19 train.py: 82] Epoch 11, iter 6200/6416, lr 0.010000, loss 2.293871
+INFO 2021-09-17 05:37:52 train.py: 82] Epoch 11, iter 6400/6416, lr 0.010000, loss 2.262752
+INFO 2021-09-17 05:38:00 train.py: 100] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-09-17 05:38:02 train.py: 82] Epoch 12, iter 0/6416, lr 0.010000, loss 2.300846
+INFO 2021-09-17 05:39:36 train.py: 82] Epoch 12, iter 200/6416, lr 0.010000, loss 1.984294
+INFO 2021-09-17 05:41:08 train.py: 82] Epoch 12, iter 400/6416, lr 0.010000, loss 1.978422
+INFO 2021-09-17 05:42:40 train.py: 82] Epoch 12, iter 600/6416, lr 0.010000, loss 1.984881
+INFO 2021-09-17 05:44:12 train.py: 82] Epoch 12, iter 800/6416, lr 0.010000, loss 1.998398
+INFO 2021-09-17 05:45:43 train.py: 82] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.996305
+INFO 2021-09-17 05:47:16 train.py: 82] Epoch 12, iter 1200/6416, lr 0.010000, loss 2.009561
+INFO 2021-09-17 05:48:48 train.py: 82] Epoch 12, iter 1400/6416, lr 0.010000, loss 2.030360
+INFO 2021-09-17 05:50:20 train.py: 82] Epoch 12, iter 1600/6416, lr 0.010000, loss 2.028437
+INFO 2021-09-17 05:51:52 train.py: 82] Epoch 12, iter 1800/6416, lr 0.010000, loss 2.030260
+INFO 2021-09-17 05:53:24 train.py: 82] Epoch 12, iter 2000/6416, lr 0.010000, loss 2.044406
+INFO 2021-09-17 05:54:56 train.py: 82] Epoch 12, iter 2200/6416, lr 0.010000, loss 2.059969
+INFO 2021-09-17 05:56:29 train.py: 82] Epoch 12, iter 2400/6416, lr 0.010000, loss 2.056551
+INFO 2021-09-17 05:58:01 train.py: 82] Epoch 12, iter 2600/6416, lr 0.010000, loss 2.047188
+INFO 2021-09-17 05:59:33 train.py: 82] Epoch 12, iter 2800/6416, lr 0.010000, loss 2.048483
+INFO 2021-09-17 06:01:07 train.py: 95] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-09-17 06:01:07 train.py: 82] Epoch 12, iter 3000/6416, lr 0.010000, loss 2.070303
+INFO 2021-09-17 06:02:40 train.py: 82] Epoch 12, iter 3200/6416, lr 0.010000, loss 2.072240
+INFO 2021-09-17 06:04:12 train.py: 82] Epoch 12, iter 3400/6416, lr 0.010000, loss 2.072334
+INFO 2021-09-17 06:05:45 train.py: 82] Epoch 12, iter 3600/6416, lr 0.010000, loss 2.078040
+INFO 2021-09-17 06:07:17 train.py: 82] Epoch 12, iter 3800/6416, lr 0.010000, loss 2.095011
+INFO 2021-09-17 06:08:50 train.py: 82] Epoch 12, iter 4000/6416, lr 0.010000, loss 2.096343
+INFO 2021-09-17 06:10:23 train.py: 82] Epoch 12, iter 4200/6416, lr 0.010000, loss 2.092617
+INFO 2021-09-17 06:11:55 train.py: 82] Epoch 12, iter 4400/6416, lr 0.010000, loss 2.104593
+INFO 2021-09-17 06:13:29 train.py: 82] Epoch 12, iter 4600/6416, lr 0.010000, loss 2.109142
+INFO 2021-09-17 06:15:01 train.py: 82] Epoch 12, iter 4800/6416, lr 0.010000, loss 2.107329
+INFO 2021-09-17 06:16:34 train.py: 82] Epoch 12, iter 5000/6416, lr 0.010000, loss 2.108961
+INFO 2021-09-17 06:18:06 train.py: 82] Epoch 12, iter 5200/6416, lr 0.010000, loss 2.112416
+INFO 2021-09-17 06:19:39 train.py: 82] Epoch 12, iter 5400/6416, lr 0.010000, loss 2.114941
+INFO 2021-09-17 06:21:12 train.py: 82] Epoch 12, iter 5600/6416, lr 0.010000, loss 2.096043
+INFO 2021-09-17 06:22:44 train.py: 82] Epoch 12, iter 5800/6416, lr 0.010000, loss 2.127511
+INFO 2021-09-17 06:24:18 train.py: 95] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-09-17 06:24:18 train.py: 82] Epoch 12, iter 6000/6416, lr 0.010000, loss 2.120107
+INFO 2021-09-17 06:25:51 train.py: 82] Epoch 12, iter 6200/6416, lr 0.010000, loss 2.121091
+INFO 2021-09-17 06:27:24 train.py: 82] Epoch 12, iter 6400/6416, lr 0.010000, loss 2.144610
+INFO 2021-09-17 06:27:32 train.py: 100] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-09-17 06:27:34 train.py: 82] Epoch 13, iter 0/6416, lr 0.001000, loss 2.096312
+INFO 2021-09-17 06:29:07 train.py: 82] Epoch 13, iter 200/6416, lr 0.001000, loss 1.787402
+INFO 2021-09-17 06:30:39 train.py: 82] Epoch 13, iter 400/6416, lr 0.001000, loss 1.739275
+INFO 2021-09-17 06:32:10 train.py: 82] Epoch 13, iter 600/6416, lr 0.001000, loss 1.741618
+INFO 2021-09-17 06:33:42 train.py: 82] Epoch 13, iter 800/6416, lr 0.001000, loss 1.744287
+INFO 2021-09-17 06:35:14 train.py: 82] Epoch 13, iter 1000/6416, lr 0.001000, loss 1.741477
+INFO 2021-09-17 06:36:46 train.py: 82] Epoch 13, iter 1200/6416, lr 0.001000, loss 1.734860
+INFO 2021-09-17 06:38:17 train.py: 82] Epoch 13, iter 1400/6416, lr 0.001000, loss 1.732881
+INFO 2021-09-17 06:39:49 train.py: 82] Epoch 13, iter 1600/6416, lr 0.001000, loss 1.721535
+INFO 2021-09-17 06:41:21 train.py: 82] Epoch 13, iter 1800/6416, lr 0.001000, loss 1.722999
+INFO 2021-09-17 06:42:53 train.py: 82] Epoch 13, iter 2000/6416, lr 0.001000, loss 1.737704
+INFO 2021-09-17 06:44:24 train.py: 82] Epoch 13, iter 2200/6416, lr 0.001000, loss 1.729291
+INFO 2021-09-17 06:45:56 train.py: 82] Epoch 13, iter 2400/6416, lr 0.001000, loss 1.717288
+INFO 2021-09-17 06:47:28 train.py: 82] Epoch 13, iter 2600/6416, lr 0.001000, loss 1.728385
+INFO 2021-09-17 06:49:00 train.py: 82] Epoch 13, iter 2800/6416, lr 0.001000, loss 1.724939
+INFO 2021-09-17 06:50:32 train.py: 95] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-09-17 06:50:33 train.py: 82] Epoch 13, iter 3000/6416, lr 0.001000, loss 1.711869
+INFO 2021-09-17 06:52:05 train.py: 82] Epoch 13, iter 3200/6416, lr 0.001000, loss 1.724575
+INFO 2021-09-17 06:53:37 train.py: 82] Epoch 13, iter 3400/6416, lr 0.001000, loss 1.728718
+INFO 2021-09-17 06:55:09 train.py: 82] Epoch 13, iter 3600/6416, lr 0.001000, loss 1.712458
+INFO 2021-09-17 06:56:42 train.py: 82] Epoch 13, iter 3800/6416, lr 0.001000, loss 1.731049
+INFO 2021-09-17 06:58:14 train.py: 82] Epoch 13, iter 4000/6416, lr 0.001000, loss 1.739248
+INFO 2021-09-17 06:59:46 train.py: 82] Epoch 13, iter 4200/6416, lr 0.001000, loss 1.722661
+INFO 2021-09-17 07:01:19 train.py: 82] Epoch 13, iter 4400/6416, lr 0.001000, loss 1.716428
+INFO 2021-09-17 07:02:51 train.py: 82] Epoch 13, iter 4600/6416, lr 0.001000, loss 1.733585
+INFO 2021-09-17 07:04:23 train.py: 82] Epoch 13, iter 4800/6416, lr 0.001000, loss 1.727239
+INFO 2021-09-17 07:05:56 train.py: 82] Epoch 13, iter 5000/6416, lr 0.001000, loss 1.727764
+INFO 2021-09-17 07:07:28 train.py: 82] Epoch 13, iter 5200/6416, lr 0.001000, loss 1.710731
+INFO 2021-09-17 07:09:01 train.py: 82] Epoch 13, iter 5400/6416, lr 0.001000, loss 1.722069
+INFO 2021-09-17 07:10:33 train.py: 82] Epoch 13, iter 5600/6416, lr 0.001000, loss 1.717373
+INFO 2021-09-17 07:12:05 train.py: 82] Epoch 13, iter 5800/6416, lr 0.001000, loss 1.716422
+INFO 2021-09-17 07:13:38 train.py: 95] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-09-17 07:13:39 train.py: 82] Epoch 13, iter 6000/6416, lr 0.001000, loss 1.733602
+INFO 2021-09-17 07:15:11 train.py: 82] Epoch 13, iter 6200/6416, lr 0.001000, loss 1.735715
+INFO 2021-09-17 07:16:44 train.py: 82] Epoch 13, iter 6400/6416, lr 0.001000, loss 1.735755
+INFO 2021-09-17 07:16:53 train.py: 100] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-09-17 07:16:55 train.py: 82] Epoch 14, iter 0/6416, lr 0.001000, loss 1.704220
+INFO 2021-09-17 07:18:27 train.py: 82] Epoch 14, iter 200/6416, lr 0.001000, loss 1.680985
+INFO 2021-09-17 07:19:59 train.py: 82] Epoch 14, iter 400/6416, lr 0.001000, loss 1.668753
+INFO 2021-09-17 07:21:31 train.py: 82] Epoch 14, iter 600/6416, lr 0.001000, loss 1.675919
+INFO 2021-09-17 07:23:03 train.py: 82] Epoch 14, iter 800/6416, lr 0.001000, loss 1.664963
+INFO 2021-09-17 07:24:35 train.py: 82] Epoch 14, iter 1000/6416, lr 0.001000, loss 1.668047
+INFO 2021-09-17 07:26:07 train.py: 82] Epoch 14, iter 1200/6416, lr 0.001000, loss 1.676194
+INFO 2021-09-17 07:27:39 train.py: 82] Epoch 14, iter 1400/6416, lr 0.001000, loss 1.677905
+INFO 2021-09-17 07:29:11 train.py: 82] Epoch 14, iter 1600/6416, lr 0.001000, loss 1.685143
+INFO 2021-09-17 07:30:42 train.py: 82] Epoch 14, iter 1800/6416, lr 0.001000, loss 1.684014
+INFO 2021-09-17 07:32:14 train.py: 82] Epoch 14, iter 2000/6416, lr 0.001000, loss 1.689739
+INFO 2021-09-17 07:33:47 train.py: 82] Epoch 14, iter 2200/6416, lr 0.001000, loss 1.684034
+INFO 2021-09-17 07:35:19 train.py: 82] Epoch 14, iter 2400/6416, lr 0.001000, loss 1.690418
+INFO 2021-09-17 07:36:51 train.py: 82] Epoch 14, iter 2600/6416, lr 0.001000, loss 1.693889
+INFO 2021-09-17 07:38:22 train.py: 82] Epoch 14, iter 2800/6416, lr 0.001000, loss 1.696419
+INFO 2021-09-17 07:39:55 train.py: 95] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-09-17 07:39:56 train.py: 82] Epoch 14, iter 3000/6416, lr 0.001000, loss 1.700553
+INFO 2021-09-17 07:41:28 train.py: 82] Epoch 14, iter 3200/6416, lr 0.001000, loss 1.687815
+INFO 2021-09-17 07:43:01 train.py: 82] Epoch 14, iter 3400/6416, lr 0.001000, loss 1.697562
+INFO 2021-09-17 07:44:33 train.py: 82] Epoch 14, iter 3600/6416, lr 0.001000, loss 1.672227
+INFO 2021-09-17 07:46:05 train.py: 82] Epoch 14, iter 3800/6416, lr 0.001000, loss 1.693138
+INFO 2021-09-17 07:47:37 train.py: 82] Epoch 14, iter 4000/6416, lr 0.001000, loss 1.682881
+INFO 2021-09-17 07:49:10 train.py: 82] Epoch 14, iter 4200/6416, lr 0.001000, loss 1.693437
+INFO 2021-09-17 07:50:42 train.py: 82] Epoch 14, iter 4400/6416, lr 0.001000, loss 1.699042
+INFO 2021-09-17 07:52:14 train.py: 82] Epoch 14, iter 4600/6416, lr 0.001000, loss 1.688267
+INFO 2021-09-17 07:53:46 train.py: 82] Epoch 14, iter 4800/6416, lr 0.001000, loss 1.700087
+INFO 2021-09-17 07:55:19 train.py: 82] Epoch 14, iter 5000/6416, lr 0.001000, loss 1.688774
+INFO 2021-09-17 07:56:51 train.py: 82] Epoch 14, iter 5200/6416, lr 0.001000, loss 1.688839
+INFO 2021-09-17 07:58:24 train.py: 82] Epoch 14, iter 5400/6416, lr 0.001000, loss 1.707735
+INFO 2021-09-17 07:59:56 train.py: 82] Epoch 14, iter 5600/6416, lr 0.001000, loss 1.702186
+INFO 2021-09-17 08:01:28 train.py: 82] Epoch 14, iter 5800/6416, lr 0.001000, loss 1.699786
+INFO 2021-09-17 08:03:02 train.py: 95] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-09-17 08:03:02 train.py: 82] Epoch 14, iter 6000/6416, lr 0.001000, loss 1.692826
+INFO 2021-09-17 08:04:34 train.py: 82] Epoch 14, iter 6200/6416, lr 0.001000, loss 1.693920
+INFO 2021-09-17 08:06:06 train.py: 82] Epoch 14, iter 6400/6416, lr 0.001000, loss 1.694143
+INFO 2021-09-17 08:06:15 train.py: 100] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-09-17 08:06:17 train.py: 82] Epoch 15, iter 0/6416, lr 0.001000, loss 1.687596
+INFO 2021-09-17 08:07:49 train.py: 82] Epoch 15, iter 200/6416, lr 0.001000, loss 1.636696
+INFO 2021-09-17 08:09:21 train.py: 82] Epoch 15, iter 400/6416, lr 0.001000, loss 1.646974
+INFO 2021-09-17 08:10:53 train.py: 82] Epoch 15, iter 600/6416, lr 0.001000, loss 1.652427
+INFO 2021-09-17 08:12:25 train.py: 82] Epoch 15, iter 800/6416, lr 0.001000, loss 1.637968
+INFO 2021-09-17 08:13:57 train.py: 82] Epoch 15, iter 1000/6416, lr 0.001000, loss 1.658541
+INFO 2021-09-17 08:15:30 train.py: 82] Epoch 15, iter 1200/6416, lr 0.001000, loss 1.664379
+INFO 2021-09-17 08:17:02 train.py: 82] Epoch 15, iter 1400/6416, lr 0.001000, loss 1.680450
+INFO 2021-09-17 08:18:34 train.py: 82] Epoch 15, iter 1600/6416, lr 0.001000, loss 1.643123
+INFO 2021-09-17 08:20:06 train.py: 82] Epoch 15, iter 1800/6416, lr 0.001000, loss 1.639892
+INFO 2021-09-17 08:21:38 train.py: 82] Epoch 15, iter 2000/6416, lr 0.001000, loss 1.662015
+INFO 2021-09-17 08:23:10 train.py: 82] Epoch 15, iter 2200/6416, lr 0.001000, loss 1.682518
+INFO 2021-09-17 08:24:43 train.py: 82] Epoch 15, iter 2400/6416, lr 0.001000, loss 1.661900
+INFO 2021-09-17 08:26:14 train.py: 82] Epoch 15, iter 2600/6416, lr 0.001000, loss 1.660998
+INFO 2021-09-17 08:27:47 train.py: 82] Epoch 15, iter 2800/6416, lr 0.001000, loss 1.657256
+INFO 2021-09-17 08:29:20 train.py: 95] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-09-17 08:29:20 train.py: 82] Epoch 15, iter 3000/6416, lr 0.001000, loss 1.675822
+INFO 2021-09-17 08:30:53 train.py: 82] Epoch 15, iter 3200/6416, lr 0.001000, loss 1.654904
+INFO 2021-09-17 08:32:25 train.py: 82] Epoch 15, iter 3400/6416, lr 0.001000, loss 1.676090
+INFO 2021-09-17 08:33:58 train.py: 82] Epoch 15, iter 3600/6416, lr 0.001000, loss 1.662378
+INFO 2021-09-17 08:35:30 train.py: 82] Epoch 15, iter 3800/6416, lr 0.001000, loss 1.660323
+INFO 2021-09-17 08:37:03 train.py: 82] Epoch 15, iter 4000/6416, lr 0.001000, loss 1.661360
+INFO 2021-09-17 08:38:35 train.py: 82] Epoch 15, iter 4200/6416, lr 0.001000, loss 1.674782
+INFO 2021-09-17 08:40:07 train.py: 82] Epoch 15, iter 4400/6416, lr 0.001000, loss 1.649986
+INFO 2021-09-17 08:41:39 train.py: 82] Epoch 15, iter 4600/6416, lr 0.001000, loss 1.672945
+INFO 2021-09-17 08:43:12 train.py: 82] Epoch 15, iter 4800/6416, lr 0.001000, loss 1.650549
+INFO 2021-09-17 08:44:45 train.py: 82] Epoch 15, iter 5000/6416, lr 0.001000, loss 1.665742
+INFO 2021-09-17 08:46:17 train.py: 82] Epoch 15, iter 5200/6416, lr 0.001000, loss 1.669584
+INFO 2021-09-17 08:47:49 train.py: 82] Epoch 15, iter 5400/6416, lr 0.001000, loss 1.671357
+INFO 2021-09-17 08:49:22 train.py: 82] Epoch 15, iter 5600/6416, lr 0.001000, loss 1.680265
+INFO 2021-09-17 08:50:55 train.py: 82] Epoch 15, iter 5800/6416, lr 0.001000, loss 1.679076
+INFO 2021-09-17 08:52:28 train.py: 95] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-09-17 08:52:29 train.py: 82] Epoch 15, iter 6000/6416, lr 0.001000, loss 1.673102
+INFO 2021-09-17 08:54:01 train.py: 82] Epoch 15, iter 6200/6416, lr 0.001000, loss 1.680908
+INFO 2021-09-17 08:55:34 train.py: 82] Epoch 15, iter 6400/6416, lr 0.001000, loss 1.681875
+INFO 2021-09-17 08:55:43 train.py: 100] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-09-17 08:55:44 train.py: 82] Epoch 16, iter 0/6416, lr 0.000100, loss 1.765504
+INFO 2021-09-17 08:57:17 train.py: 82] Epoch 16, iter 200/6416, lr 0.000100, loss 1.620723
+INFO 2021-09-17 08:58:49 train.py: 82] Epoch 16, iter 400/6416, lr 0.000100, loss 1.627695
+INFO 2021-09-17 09:00:21 train.py: 82] Epoch 16, iter 600/6416, lr 0.000100, loss 1.612941
+INFO 2021-09-17 09:01:53 train.py: 82] Epoch 16, iter 800/6416, lr 0.000100, loss 1.612128
+INFO 2021-09-17 09:03:26 train.py: 82] Epoch 16, iter 1000/6416, lr 0.000100, loss 1.623295
+INFO 2021-09-17 09:04:58 train.py: 82] Epoch 16, iter 1200/6416, lr 0.000100, loss 1.621130
+INFO 2021-09-17 09:06:30 train.py: 82] Epoch 16, iter 1400/6416, lr 0.000100, loss 1.619956
+INFO 2021-09-17 09:08:02 train.py: 82] Epoch 16, iter 1600/6416, lr 0.000100, loss 1.623024
+INFO 2021-09-17 09:09:34 train.py: 82] Epoch 16, iter 1800/6416, lr 0.000100, loss 1.609937
+INFO 2021-09-17 09:11:06 train.py: 82] Epoch 16, iter 2000/6416, lr 0.000100, loss 1.615599
+INFO 2021-09-17 09:12:39 train.py: 82] Epoch 16, iter 2200/6416, lr 0.000100, loss 1.615302
+INFO 2021-09-17 09:14:11 train.py: 82] Epoch 16, iter 2400/6416, lr 0.000100, loss 1.623361
+INFO 2021-09-17 09:15:43 train.py: 82] Epoch 16, iter 2600/6416, lr 0.000100, loss 1.617204
+INFO 2021-09-17 09:17:16 train.py: 82] Epoch 16, iter 2800/6416, lr 0.000100, loss 1.625166
+INFO 2021-09-17 09:18:49 train.py: 95] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-09-17 09:18:50 train.py: 82] Epoch 16, iter 3000/6416, lr 0.000100, loss 1.612472
+INFO 2021-09-17 09:20:22 train.py: 82] Epoch 16, iter 3200/6416, lr 0.000100, loss 1.628349
+INFO 2021-09-17 09:21:55 train.py: 82] Epoch 16, iter 3400/6416, lr 0.000100, loss 1.625452
+INFO 2021-09-17 09:23:27 train.py: 82] Epoch 16, iter 3600/6416, lr 0.000100, loss 1.632724
+INFO 2021-09-17 09:25:00 train.py: 82] Epoch 16, iter 3800/6416, lr 0.000100, loss 1.628694
+INFO 2021-09-17 09:26:32 train.py: 82] Epoch 16, iter 4000/6416, lr 0.000100, loss 1.618106
+INFO 2021-09-17 09:28:05 train.py: 82] Epoch 16, iter 4200/6416, lr 0.000100, loss 1.622793
+INFO 2021-09-17 09:29:37 train.py: 82] Epoch 16, iter 4400/6416, lr 0.000100, loss 1.631416
+INFO 2021-09-17 09:31:10 train.py: 82] Epoch 16, iter 4600/6416, lr 0.000100, loss 1.611960
+INFO 2021-09-17 09:32:42 train.py: 82] Epoch 16, iter 4800/6416, lr 0.000100, loss 1.633533
+INFO 2021-09-17 09:34:15 train.py: 82] Epoch 16, iter 5000/6416, lr 0.000100, loss 1.627026
+INFO 2021-09-17 09:35:47 train.py: 82] Epoch 16, iter 5200/6416, lr 0.000100, loss 1.635878
+INFO 2021-09-17 09:37:20 train.py: 82] Epoch 16, iter 5400/6416, lr 0.000100, loss 1.610844
+INFO 2021-09-17 09:38:53 train.py: 82] Epoch 16, iter 5600/6416, lr 0.000100, loss 1.626107
+INFO 2021-09-17 09:40:26 train.py: 82] Epoch 16, iter 5800/6416, lr 0.000100, loss 1.621249
+INFO 2021-09-17 09:41:59 train.py: 95] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-09-17 09:42:00 train.py: 82] Epoch 16, iter 6000/6416, lr 0.000100, loss 1.609684
+INFO 2021-09-17 09:43:33 train.py: 82] Epoch 16, iter 6200/6416, lr 0.000100, loss 1.619737
+INFO 2021-09-17 09:45:06 train.py: 82] Epoch 16, iter 6400/6416, lr 0.000100, loss 1.641055
+INFO 2021-09-17 09:45:14 train.py: 100] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-09-17 09:45:16 train.py: 82] Epoch 17, iter 0/6416, lr 0.000100, loss 1.653002
+INFO 2021-09-17 09:46:48 train.py: 82] Epoch 17, iter 200/6416, lr 0.000100, loss 1.621332
+INFO 2021-09-17 09:48:21 train.py: 82] Epoch 17, iter 400/6416, lr 0.000100, loss 1.614825
+INFO 2021-09-17 09:49:53 train.py: 82] Epoch 17, iter 600/6416, lr 0.000100, loss 1.615034
+INFO 2021-09-17 09:51:25 train.py: 82] Epoch 17, iter 800/6416, lr 0.000100, loss 1.617731
+INFO 2021-09-17 09:52:57 train.py: 82] Epoch 17, iter 1000/6416, lr 0.000100, loss 1.607662
+INFO 2021-09-17 09:54:30 train.py: 82] Epoch 17, iter 1200/6416, lr 0.000100, loss 1.616919
+INFO 2021-09-17 09:56:02 train.py: 82] Epoch 17, iter 1400/6416, lr 0.000100, loss 1.610678
+INFO 2021-09-17 09:57:35 train.py: 82] Epoch 17, iter 1600/6416, lr 0.000100, loss 1.614414
+INFO 2021-09-17 09:59:07 train.py: 82] Epoch 17, iter 1800/6416, lr 0.000100, loss 1.626630
+INFO 2021-09-17 10:00:39 train.py: 82] Epoch 17, iter 2000/6416, lr 0.000100, loss 1.626345
+INFO 2021-09-17 10:02:11 train.py: 82] Epoch 17, iter 2200/6416, lr 0.000100, loss 1.631727
+INFO 2021-09-17 10:03:43 train.py: 82] Epoch 17, iter 2400/6416, lr 0.000100, loss 1.632018
+INFO 2021-09-17 10:05:15 train.py: 82] Epoch 17, iter 2600/6416, lr 0.000100, loss 1.619205
+INFO 2021-09-17 10:06:47 train.py: 82] Epoch 17, iter 2800/6416, lr 0.000100, loss 1.613809
+INFO 2021-09-17 10:08:20 train.py: 95] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-09-17 10:08:21 train.py: 82] Epoch 17, iter 3000/6416, lr 0.000100, loss 1.618722
+INFO 2021-09-17 10:09:53 train.py: 82] Epoch 17, iter 3200/6416, lr 0.000100, loss 1.614290
+INFO 2021-09-17 10:11:26 train.py: 82] Epoch 17, iter 3400/6416, lr 0.000100, loss 1.614336
+INFO 2021-09-17 10:12:58 train.py: 82] Epoch 17, iter 3600/6416, lr 0.000100, loss 1.621089
+INFO 2021-09-17 10:14:30 train.py: 82] Epoch 17, iter 3800/6416, lr 0.000100, loss 1.617981
+INFO 2021-09-17 10:16:02 train.py: 82] Epoch 17, iter 4000/6416, lr 0.000100, loss 1.619380
+INFO 2021-09-17 10:17:34 train.py: 82] Epoch 17, iter 4200/6416, lr 0.000100, loss 1.606993
+INFO 2021-09-17 10:19:06 train.py: 82] Epoch 17, iter 4400/6416, lr 0.000100, loss 1.617127
+INFO 2021-09-17 10:20:39 train.py: 82] Epoch 17, iter 4600/6416, lr 0.000100, loss 1.632343
+INFO 2021-09-17 10:22:11 train.py: 82] Epoch 17, iter 4800/6416, lr 0.000100, loss 1.615973
+INFO 2021-09-17 10:23:44 train.py: 82] Epoch 17, iter 5000/6416, lr 0.000100, loss 1.599065
+INFO 2021-09-17 10:25:16 train.py: 82] Epoch 17, iter 5200/6416, lr 0.000100, loss 1.621521
+INFO 2021-09-17 10:26:48 train.py: 82] Epoch 17, iter 5400/6416, lr 0.000100, loss 1.600619
+INFO 2021-09-17 10:28:20 train.py: 82] Epoch 17, iter 5600/6416, lr 0.000100, loss 1.627305
+INFO 2021-09-17 10:29:52 train.py: 82] Epoch 17, iter 5800/6416, lr 0.000100, loss 1.615763
+INFO 2021-09-17 10:31:25 train.py: 95] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-09-17 10:31:26 train.py: 82] Epoch 17, iter 6000/6416, lr 0.000100, loss 1.602896
+INFO 2021-09-17 10:32:58 train.py: 82] Epoch 17, iter 6200/6416, lr 0.000100, loss 1.624158
+INFO 2021-09-17 10:34:31 train.py: 82] Epoch 17, iter 6400/6416, lr 0.000100, loss 1.625425
+INFO 2021-09-17 10:34:40 train.py: 100] Save checkpoint Epoch_17.pt to disk...
+INFO 2021-09-17 10:34:40 train.py: 183] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/.gitkeep b/bob/bio/facexzoo/models/backbones/RepVGG_B1/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3dddabb5e269e6c048cc8d30b277a39a51397caa
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9778333333333334 | 0.002357677241546989 |
+| Epoch_16_batch_2999.pt | 0.9776666666666667 | 0.002385656728175981 |
+| Epoch_14.pt | 0.9775 | 0.002399202542409903 |
+| Epoch_14_batch_5999.pt | 0.9775 | 0.002266911751455905 |
+| Epoch_16_batch_5999.pt | 0.9771666666666668 | 0.0024975296436654474 |
+| Epoch_15_batch_2999.pt | 0.9771666666666666 | 0.00249752964366545 |
+| Epoch_15_batch_5999.pt | 0.9770000000000001 | 0.0023147407395555123 |
+| Epoch_14_batch_2999.pt | 0.9770000000000001 | 0.0024695678634325353 |
+| Epoch_15.pt | 0.9763333333333334 | 0.002506780927261885 |
+| Epoch_12_batch_5999.pt | 0.9761666666666665 | 0.0022367580154663697 |
+| Epoch_13_batch_5999.pt | 0.976 | 0.0022933074933944734 |
+| Epoch_13.pt | 0.9758333333333334 | 0.0022256917360011204 |
+| Epoch_17_batch_5999.pt | 0.9758333333333334 | 0.002386303510546061 |
+| Epoch_16.pt | 0.9756666666666668 | 0.0021401511426953554 |
+| Epoch_12_batch_2999.pt | 0.9755 | 0.0018929694486000896 |
+| Epoch_17_batch_2999.pt | 0.9753333333333334 | 0.0024570382652773348 |
+| Epoch_12.pt | 0.9751666666666667 | 0.002427339141661492 |
+| Epoch_17.pt | 0.9746666666666666 | 0.0029059326290271146 |
+| Epoch_11_batch_5999.pt | 0.9744999999999997 | 0.0024975296436654405 |
+| Epoch_10.pt | 0.9743333333333334 | 0.0022933074933944755 |
+| Epoch_10_batch_2999.pt | 0.9743333333333333 | 0.0018790593916986412 |
+| Epoch_11_batch_2999.pt | 0.9739999999999999 | 0.002006163342807529 |
+| Epoch_11.pt | 0.9735000000000001 | 0.0023100692095879777 |
+| Epoch_10_batch_5999.pt | 0.9730000000000001 | 0.0019212907184211734 |
+| Epoch_9_batch_5999.pt | 0.9675 | 0.0019602531960419252 |
+| Epoch_7_batch_2999.pt | 0.9674999999999999 | 0.002584975583140416 |
+| Epoch_8_batch_5999.pt | 0.9666666666666666 | 0.0024969116726938057 |
+| Epoch_8_batch_2999.pt | 0.9663333333333334 | 0.002419060117453027 |
+| Epoch_5_batch_5999.pt | 0.9661666666666665 | 0.00271086064416797 |
+| Epoch_6_batch_5999.pt | 0.9658333333333335 | 0.003471666622215111 |
+| Epoch_9_batch_2999.pt | 0.9654999999999999 | 0.002641665206215286 |
+| Epoch_7_batch_5999.pt | 0.9631666666666666 | 0.0026229048075806414 |
+| Epoch_9.pt | 0.962 | 0.002852332811776326 |
+| Epoch_4_batch_5999.pt | 0.9618333333333334 | 0.0037305048809332317 |
+| Epoch_5_batch_2999.pt | 0.9615 | 0.002870131411797664 |
+| Epoch_6.pt | 0.9611666666666666 | 0.003865405285955329 |
+| Epoch_5.pt | 0.9603333333333335 | 0.004200235149207959 |
+| Epoch_7.pt | 0.9603333333333334 | 0.0026270200927859776 |
+| Epoch_6_batch_2999.pt | 0.9603333333333334 | 0.0029585615457098534 |
+| Epoch_4_batch_2999.pt | 0.9590000000000002 | 0.00319335730293332 |
+| Epoch_3_batch_5999.pt | 0.958 | 0.0040809705938237685 |
+| Epoch_3_batch_2999.pt | 0.9571666666666667 | 0.0035750334540435953 |
+| Epoch_3.pt | 0.9566666666666667 | 0.0032489314482696506 |
+| Epoch_2_batch_5999.pt | 0.9546666666666667 | 0.0021914536581462292 |
+| Epoch_8.pt | 0.9543333333333333 | 0.002608154354290112 |
+| Epoch_4.pt | 0.9543333333333333 | 0.0024870032539554936 |
+| Epoch_2.pt | 0.9518333333333333 | 0.0037879764297179485 |
+| Epoch_2_batch_2999.pt | 0.9501666666666667 | 0.002870131411797663 |
+| Epoch_1_batch_5999.pt | 0.942 | 0.004498284995554917 |
+| Epoch_1.pt | 0.9351666666666667 | 0.004978038187632844 |
+| Epoch_1_batch_2999.pt | 0.9244999999999999 | 0.005671294406651241 |
+| Epoch_0_batch_5999.pt | 0.8755 | 0.006378252014888725 |
+| Epoch_0_batch_2999.pt | 0.7336666666666666 | 0.004058218303944368 |
+| Epoch_0.pt | 0.6769999999999999 | 0.008317763232094622 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..8ea418b950b5a053fc9310506c01c26e0320b038
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.9549999999999998 | 0.0036004114991154746 |
+| Epoch_13_batch_5999.pt | 0.9546666666666667 | 0.003598696609044811 |
+| Epoch_16_batch_5999.pt | 0.9546666666666667 | 0.003675074535231386 |
+| Epoch_15_batch_2999.pt | 0.9545000000000001 | 0.003531603350069515 |
+| Epoch_13_batch_2999.pt | 0.9545 | 0.0036264630934150156 |
+| Epoch_14_batch_5999.pt | 0.9545 | 0.0033979841518606735 |
+| Epoch_15.pt | 0.9543333333333335 | 0.0033811386788228756 |
+| Epoch_16_batch_2999.pt | 0.9541666666666668 | 0.0032984283575105324 |
+| Epoch_12_batch_2999.pt | 0.9541666666666668 | 0.003515837184954836 |
+| Epoch_12_batch_5999.pt | 0.9541666666666668 | 0.003712258638286247 |
+| Epoch_17_batch_2999.pt | 0.9540000000000001 | 0.0035849479566448174 |
+| Epoch_17_batch_5999.pt | 0.9540000000000001 | 0.0036616126785075686 |
+| Epoch_17.pt | 0.9538333333333334 | 0.0036179422567033195 |
+| Epoch_12.pt | 0.9538333333333334 | 0.0036094013046179536 |
+| Epoch_15_batch_5999.pt | 0.9538333333333334 | 0.003514081022415214 |
+| Epoch_13.pt | 0.9536666666666667 | 0.0035030850600965414 |
+| Epoch_11_batch_2999.pt | 0.9535 | 0.0036298658275376924 |
+| Epoch_10_batch_5999.pt | 0.9535 | 0.003621352997273838 |
+| Epoch_11_batch_5999.pt | 0.9535 | 0.0036128201084197997 |
+| Epoch_16.pt | 0.9535 | 0.003346732329295342 |
+| Epoch_14.pt | 0.9533333333333334 | 0.003333333333333332 |
+| Epoch_11.pt | 0.953 | 0.0034047896546765674 |
+| Epoch_10.pt | 0.9525 | 0.0036025539637496652 |
+| Epoch_10_batch_2999.pt | 0.9521666666666666 | 0.0032777777777777796 |
+| Epoch_7_batch_5999.pt | 0.9476666666666664 | 0.004247003300803639 |
+| Epoch_5_batch_5999.pt | 0.9463333333333332 | 0.0040123266856150605 |
+| Epoch_6_batch_2999.pt | 0.9461666666666666 | 0.0038009907740216256 |
+| Epoch_8_batch_5999.pt | 0.9453333333333334 | 0.003520662115056636 |
+| Epoch_9_batch_5999.pt | 0.9451666666666666 | 0.0035870996571906472 |
+| Epoch_6_batch_5999.pt | 0.9448333333333332 | 0.003812341880955058 |
+| Epoch_8_batch_2999.pt | 0.9446666666666668 | 0.003807075933113735 |
+| Epoch_9_batch_2999.pt | 0.9436666666666668 | 0.0033314809667922144 |
+| Epoch_5_batch_2999.pt | 0.9431666666666667 | 0.003508807261030692 |
+| Epoch_7_batch_2999.pt | 0.9425000000000001 | 0.0033998002846209394 |
+| Epoch_4_batch_2999.pt | 0.9421666666666667 | 0.003857412297075546 |
+| Epoch_4_batch_5999.pt | 0.9408333333333333 | 0.003222701113838477 |
+| Epoch_7.pt | 0.9404999999999999 | 0.003073683598582898 |
+| Epoch_3_batch_2999.pt | 0.9395 | 0.0028004629246952306 |
+| Epoch_5.pt | 0.9385 | 0.0038845213569807433 |
+| Epoch_3.pt | 0.937 | 0.0034854193647462475 |
+| Epoch_3_batch_5999.pt | 0.9368333333333332 | 0.0029963970133692836 |
+| Epoch_8.pt | 0.9366666666666668 | 0.0033054388401430016 |
+| Epoch_6.pt | 0.9356666666666668 | 0.0038103173776627185 |
+| Epoch_9.pt | 0.9348333333333333 | 0.0029860788111948163 |
+| Epoch_2_batch_5999.pt | 0.9321666666666667 | 0.00370226824034358 |
+| Epoch_2_batch_2999.pt | 0.9316666666666666 | 0.002971053768249093 |
+| Epoch_4.pt | 0.9296666666666665 | 0.003798960221617499 |
+| Epoch_2.pt | 0.9285 | 0.00336512616020173 |
+| Epoch_1_batch_5999.pt | 0.9223333333333332 | 0.004895197949313491 |
+| Epoch_1.pt | 0.9209999999999999 | 0.004210510071443646 |
+| Epoch_1_batch_2999.pt | 0.9111666666666665 | 0.0037437190197403204 |
+| Epoch_0_batch_5999.pt | 0.8745 | 0.004661703710177372 |
+| Epoch_0_batch_2999.pt | 0.7233333333333334 | 0.007092858519335389 |
+| Epoch_0.pt | 0.7201666666666668 | 0.005383158465225451 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..698a717ef80dd1766d60e0546d107abbc14fdac3
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.8755000000000001 | 0.004701260666231113 |
+| Epoch_17_batch_5999.pt | 0.8753333333333334 | 0.004816381511487384 |
+| Epoch_15_batch_5999.pt | 0.8751666666666665 | 0.0049531758112633115 |
+| Epoch_15_batch_2999.pt | 0.8744999999999999 | 0.005061656879952487 |
+| Epoch_15.pt | 0.8743333333333334 | 0.005242278272037973 |
+| Epoch_13_batch_2999.pt | 0.8743333333333332 | 0.004970282054743333 |
+| Epoch_13.pt | 0.8741666666666668 | 0.005135509399282944 |
+| Epoch_17_batch_2999.pt | 0.874 | 0.00505647122354789 |
+| Epoch_16_batch_2999.pt | 0.8736666666666668 | 0.004803548071381615 |
+| Epoch_14.pt | 0.873 | 0.0051747248987533394 |
+| Epoch_14_batch_5999.pt | 0.873 | 0.004841946348777981 |
+| Epoch_12_batch_5999.pt | 0.8728333333333333 | 0.004714372560794834 |
+| Epoch_11.pt | 0.8723333333333334 | 0.005763872155263528 |
+| Epoch_12.pt | 0.8723333333333333 | 0.004812535072823933 |
+| Epoch_16.pt | 0.8721666666666668 | 0.004963135707330867 |
+| Epoch_10.pt | 0.8721666666666665 | 0.005363629230149916 |
+| Epoch_17.pt | 0.8718333333333333 | 0.004953175811263304 |
+| Epoch_16_batch_5999.pt | 0.8716666666666667 | 0.004594682917363401 |
+| Epoch_14_batch_2999.pt | 0.8716666666666665 | 0.004687782913273106 |
+| Epoch_11_batch_2999.pt | 0.8701666666666668 | 0.00529645816898448 |
+| Epoch_11_batch_5999.pt | 0.8699999999999999 | 0.0054489777045507665 |
+| Epoch_12_batch_2999.pt | 0.8686666666666667 | 0.0059202519132254315 |
+| Epoch_10_batch_2999.pt | 0.8683333333333334 | 0.006136311676215146 |
+| Epoch_10_batch_5999.pt | 0.8666666666666668 | 0.005392038026022631 |
+| Epoch_8_batch_5999.pt | 0.8493333333333334 | 0.006305983832738983 |
+| Epoch_7_batch_5999.pt | 0.8493333333333334 | 0.007201680188043908 |
+| Epoch_5_batch_5999.pt | 0.8451666666666666 | 0.006437975680615574 |
+| Epoch_9_batch_2999.pt | 0.8426666666666666 | 0.005617433182117574 |
+| Epoch_6_batch_5999.pt | 0.842 | 0.006155395104206462 |
+| Epoch_4_batch_5999.pt | 0.841 | 0.006939776208723197 |
+| Epoch_9_batch_5999.pt | 0.8401666666666667 | 0.005569150033781231 |
+| Epoch_6_batch_2999.pt | 0.8385 | 0.005871304984602847 |
+| Epoch_7_batch_2999.pt | 0.8380000000000001 | 0.006628594994788738 |
+| Epoch_8_batch_2999.pt | 0.8376666666666667 | 0.005080828162299409 |
+| Epoch_4_batch_2999.pt | 0.836 | 0.006764102780099742 |
+| Epoch_3_batch_5999.pt | 0.8358333333333334 | 0.005984805864400948 |
+| Epoch_5_batch_2999.pt | 0.8343333333333334 | 0.007032552177087404 |
+| Epoch_3.pt | 0.834 | 0.0071673126323867275 |
+| Epoch_5.pt | 0.8328333333333333 | 0.006440851488835508 |
+| Epoch_3_batch_2999.pt | 0.8311666666666667 | 0.006749714214343354 |
+| Epoch_2_batch_2999.pt | 0.827 | 0.0081642097589206 |
+| Epoch_2_batch_5999.pt | 0.8263333333333334 | 0.007268229529522799 |
+| Epoch_8.pt | 0.8261666666666667 | 0.005890199015690757 |
+| Epoch_9.pt | 0.8258333333333333 | 0.005694105680980364 |
+| Epoch_7.pt | 0.8246666666666667 | 0.008652409926191902 |
+| Epoch_2.pt | 0.8205000000000002 | 0.007829392231587923 |
+| Epoch_4.pt | 0.8191666666666666 | 0.007758905675670424 |
+| Epoch_1_batch_5999.pt | 0.8156666666666667 | 0.008791831225421492 |
+| Epoch_6.pt | 0.8121666666666666 | 0.0070406667052432545 |
+| Epoch_1.pt | 0.8023333333333333 | 0.007223076872508893 |
+| Epoch_1_batch_2999.pt | 0.7833333333333334 | 0.007490735018081407 |
+| Epoch_0_batch_5999.pt | 0.7365 | 0.00816666666666667 |
+| Epoch_0_batch_2999.pt | 0.6320000000000001 | 0.008314051767346854 |
+| Epoch_0.pt | 0.5715 | 0.008098353742170898 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..faa51c872277bc4d4a940a88ad7f08cf336d693a
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_13_batch_5999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_15.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_17_batch_2999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_14.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_12.pt | 0.9981666666666665 | 0.0007637626158259756 |
+| Epoch_15_batch_5999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_13.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_16_batch_2999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_14_batch_2999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_15_batch_2999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_14_batch_5999.pt | 0.9981666666666665 | 0.0007222222222222252 |
+| Epoch_16_batch_5999.pt | 0.9979999999999999 | 0.0007370277311900908 |
+| Epoch_17_batch_5999.pt | 0.9979999999999999 | 0.0007370277311900908 |
+| Epoch_12_batch_5999.pt | 0.9979999999999999 | 0.0007370277311900908 |
+| Epoch_16.pt | 0.9979999999999999 | 0.0007370277311900908 |
+| Epoch_17.pt | 0.9978333333333333 | 0.0007474235581707629 |
+| Epoch_12_batch_2999.pt | 0.9976666666666667 | 0.0007934920476158755 |
+| Epoch_11_batch_2999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_10_batch_5999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_10.pt | 0.9971666666666665 | 0.0008975274678557489 |
+| Epoch_11.pt | 0.9971666666666665 | 0.0009312808119022322 |
+| Epoch_10_batch_2999.pt | 0.9969999999999999 | 0.000853460638652065 |
+| Epoch_11_batch_5999.pt | 0.9969999999999999 | 0.0008888888888888863 |
+| Epoch_9_batch_2999.pt | 0.9968333333333333 | 0.000976577546180386 |
+| Epoch_8_batch_5999.pt | 0.9968333333333332 | 0.000722222222222217 |
+| Epoch_3.pt | 0.9966666666666665 | 0.0007856742013183833 |
+| Epoch_9_batch_5999.pt | 0.9964999999999999 | 0.0009444444444444461 |
+| Epoch_4_batch_5999.pt | 0.9964999999999999 | 0.0010076865081787277 |
+| Epoch_5.pt | 0.9964999999999999 | 0.0010378634273482956 |
+| Epoch_5_batch_2999.pt | 0.9961666666666666 | 0.000704920974469411 |
+| Epoch_7_batch_2999.pt | 0.9961666666666666 | 0.0012184284555256302 |
+| Epoch_5_batch_5999.pt | 0.9961666666666666 | 0.0010555555555555535 |
+| Epoch_3_batch_5999.pt | 0.9961666666666666 | 0.0009312808119022304 |
+| Epoch_4.pt | 0.9961666666666666 | 0.0011124991330278195 |
+| Epoch_9.pt | 0.9960000000000001 | 0.0011706281947614185 |
+| Epoch_4_batch_2999.pt | 0.9960000000000001 | 0.001222222222222224 |
+| Epoch_7_batch_5999.pt | 0.9960000000000001 | 0.0009686442096757033 |
+| Epoch_7.pt | 0.9958333333333333 | 0.0007556372504853008 |
+| Epoch_6_batch_5999.pt | 0.9958333333333332 | 0.0007954345035153545 |
+| Epoch_6_batch_2999.pt | 0.9958333333333332 | 0.0009043789220055362 |
+| Epoch_6.pt | 0.9958333333333332 | 0.0010015420209622187 |
+| Epoch_3_batch_2999.pt | 0.9955 | 0.0011928283640879926 |
+| Epoch_8_batch_2999.pt | 0.9955 | 0.0008624541497922252 |
+| Epoch_2.pt | 0.9953333333333333 | 0.0009229582069909011 |
+| Epoch_2_batch_5999.pt | 0.9953333333333333 | 0.0010183501544346308 |
+| Epoch_2_batch_2999.pt | 0.9948333333333335 | 0.0008407081083567555 |
+| Epoch_1_batch_5999.pt | 0.9946666666666667 | 0.001133115447465063 |
+| Epoch_8.pt | 0.9941666666666666 | 0.0013437096247164253 |
+| Epoch_1.pt | 0.9908333333333331 | 0.0009043789220055385 |
+| Epoch_1_batch_2999.pt | 0.9898333333333333 | 0.0010671873729054782 |
+| Epoch_0_batch_5999.pt | 0.9778333333333332 | 0.00240947204913349 |
+| Epoch_0_batch_2999.pt | 0.9386666666666666 | 0.003782676562344267 |
+| Epoch_0.pt | 0.8089999999999998 | 0.003761402417735725 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..06b2630e195db48a4554269ad4d5cde873ff2424
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.9673509769972792 |
+| 2 | 0.9760925316010779 |
+| 3 | 0.9795357798403999 |
+| 4 | 0.9814949815795982 |
+| 5 | 0.9828813934415559 |
+| 6 | 0.9837991590403166 |
+| 7 | 0.984528164338623 |
+| 8 | 0.9852311337334184 |
+| 9 | 0.9856542171654712 |
+| 10 | 0.9861423903563014 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_rfw.txt b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_rfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..4d6c2679f6854673227e2c1774072dc0b382b9f9
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B1/accu_rfw.txt
@@ -0,0 +1,27 @@
+RFW_African
++-------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++-------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9456666666666667 | 0.0024620577562400373 |
++-------------+--------------------+-----------------------+
+
+RFW_Asian
++-------------+--------------------+---------------------+
+| model_name | mean accuracy | standard error |
++-------------+--------------------+---------------------+
+| Epoch_17.pt | 0.9388333333333334 | 0.00398337594894297 |
++-------------+--------------------+---------------------+
+
+RFW_Caucasian
++-------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++-------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9873333333333333 | 0.0019751543149590183 |
++-------------+--------------------+-----------------------+
+
+RFW_Indian
++-------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++-------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9503333333333333 | 0.0030205060486818178 |
++-------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/RepVGG_B1/log.log b/bob/bio/facexzoo/models/backbones/RepVGG_B1/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..93ceb756515b0fbe32b8ea4722c7e606492c5f85
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/RepVGG_B1/log.log
@@ -0,0 +1,655 @@
+INFO 2021-09-17 11:09:50 train.py: 180] Start optimization.
+INFO 2021-09-17 11:09:50 train.py: 181] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='RepVGG', batch_size=512, data_root='/export2/wj_data/FaceX-Zoo/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-repvgg', train_file='/export2/wj_data/FaceX-Zoo/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f2dd0211da0>)
+backbone param:
+{'block_stage1': 4, 'block_stage2': 6, 'block_stage3': 16, 'block_stage4': 1, 'width_stage1': 2, 'width_stage2': 2, 'width_stage3': 2, 'width_stage4': 4, 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2021-09-17 11:10:15 train.py: 82] Epoch 0, iter 0/6416, lr 0.100000, loss 16.309374
+INFO 2021-09-17 11:13:13 train.py: 82] Epoch 0, iter 200/6416, lr 0.100000, loss 15.664051
+INFO 2021-09-17 11:16:11 train.py: 82] Epoch 0, iter 400/6416, lr 0.100000, loss 15.389331
+INFO 2021-09-17 11:19:09 train.py: 82] Epoch 0, iter 600/6416, lr 0.100000, loss 15.317583
+INFO 2021-09-17 11:22:08 train.py: 82] Epoch 0, iter 800/6416, lr 0.100000, loss 15.194844
+INFO 2021-09-17 11:25:06 train.py: 82] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.002481
+INFO 2021-09-17 11:28:04 train.py: 82] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.741223
+INFO 2021-09-17 11:31:02 train.py: 82] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.427274
+INFO 2021-09-17 11:34:01 train.py: 82] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.101495
+INFO 2021-09-17 11:37:04 train.py: 82] Epoch 0, iter 1800/6416, lr 0.100000, loss 13.756198
+INFO 2021-09-17 11:40:10 train.py: 82] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.395571
+INFO 2021-09-17 11:43:09 train.py: 82] Epoch 0, iter 2200/6416, lr 0.100000, loss 13.032782
+INFO 2021-09-17 11:46:08 train.py: 82] Epoch 0, iter 2400/6416, lr 0.100000, loss 12.640520
+INFO 2021-09-17 11:49:07 train.py: 82] Epoch 0, iter 2600/6416, lr 0.100000, loss 12.294255
+INFO 2021-09-17 11:52:06 train.py: 82] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.013902
+INFO 2021-09-17 11:55:07 train.py: 95] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-09-17 11:55:08 train.py: 82] Epoch 0, iter 3000/6416, lr 0.100000, loss 11.833890
+INFO 2021-09-17 11:58:07 train.py: 82] Epoch 0, iter 3200/6416, lr 0.100000, loss 11.812492
+INFO 2021-09-17 12:01:06 train.py: 82] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.873455
+INFO 2021-09-17 12:04:05 train.py: 82] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.082802
+INFO 2021-09-17 12:07:04 train.py: 82] Epoch 0, iter 3800/6416, lr 0.100000, loss 12.330923
+INFO 2021-09-17 12:10:02 train.py: 82] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.596411
+INFO 2021-09-17 12:13:01 train.py: 82] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.873820
+INFO 2021-09-17 12:15:59 train.py: 82] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.103076
+INFO 2021-09-17 12:18:56 train.py: 82] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.258172
+INFO 2021-09-17 12:21:54 train.py: 82] Epoch 0, iter 4800/6416, lr 0.100000, loss 13.374078
+INFO 2021-09-17 12:24:51 train.py: 82] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.423296
+INFO 2021-09-17 12:27:50 train.py: 82] Epoch 0, iter 5200/6416, lr 0.100000, loss 13.407871
+INFO 2021-09-17 12:30:48 train.py: 82] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.314967
+INFO 2021-09-17 12:33:45 train.py: 82] Epoch 0, iter 5600/6416, lr 0.100000, loss 13.199173
+INFO 2021-09-17 12:36:42 train.py: 82] Epoch 0, iter 5800/6416, lr 0.100000, loss 13.032544
+INFO 2021-09-17 12:39:42 train.py: 95] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-09-17 12:39:42 train.py: 82] Epoch 0, iter 6000/6416, lr 0.100000, loss 12.823961
+INFO 2021-09-17 12:42:40 train.py: 82] Epoch 0, iter 6200/6416, lr 0.100000, loss 12.596705
+INFO 2021-09-17 12:45:46 train.py: 82] Epoch 0, iter 6400/6416, lr 0.100000, loss 12.338798
+INFO 2021-09-17 12:46:02 train.py: 100] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-09-17 12:46:04 train.py: 82] Epoch 1, iter 0/6416, lr 0.100000, loss 12.174789
+INFO 2021-09-17 12:49:01 train.py: 82] Epoch 1, iter 200/6416, lr 0.100000, loss 11.949627
+INFO 2021-09-17 12:51:56 train.py: 82] Epoch 1, iter 400/6416, lr 0.100000, loss 11.617720
+INFO 2021-09-17 12:54:52 train.py: 82] Epoch 1, iter 600/6416, lr 0.100000, loss 11.385432
+INFO 2021-09-17 12:57:47 train.py: 82] Epoch 1, iter 800/6416, lr 0.100000, loss 11.121999
+INFO 2021-09-17 13:00:43 train.py: 82] Epoch 1, iter 1000/6416, lr 0.100000, loss 10.884434
+INFO 2021-09-17 13:03:39 train.py: 82] Epoch 1, iter 1200/6416, lr 0.100000, loss 10.693742
+INFO 2021-09-17 13:06:35 train.py: 82] Epoch 1, iter 1400/6416, lr 0.100000, loss 10.462524
+INFO 2021-09-17 13:09:30 train.py: 82] Epoch 1, iter 1600/6416, lr 0.100000, loss 10.286891
+INFO 2021-09-17 13:12:26 train.py: 82] Epoch 1, iter 1800/6416, lr 0.100000, loss 10.063639
+INFO 2021-09-17 13:15:22 train.py: 82] Epoch 1, iter 2000/6416, lr 0.100000, loss 9.896982
+INFO 2021-09-17 13:18:17 train.py: 82] Epoch 1, iter 2200/6416, lr 0.100000, loss 9.702913
+INFO 2021-09-17 13:21:13 train.py: 82] Epoch 1, iter 2400/6416, lr 0.100000, loss 9.494096
+INFO 2021-09-17 13:24:10 train.py: 82] Epoch 1, iter 2600/6416, lr 0.100000, loss 9.317434
+INFO 2021-09-17 13:27:06 train.py: 82] Epoch 1, iter 2800/6416, lr 0.100000, loss 9.201912
+INFO 2021-09-17 13:30:08 train.py: 95] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-09-17 13:30:09 train.py: 82] Epoch 1, iter 3000/6416, lr 0.100000, loss 9.021920
+INFO 2021-09-17 13:33:06 train.py: 82] Epoch 1, iter 3200/6416, lr 0.100000, loss 8.869344
+INFO 2021-09-17 13:36:11 train.py: 82] Epoch 1, iter 3400/6416, lr 0.100000, loss 8.739332
+INFO 2021-09-17 13:39:19 train.py: 82] Epoch 1, iter 3600/6416, lr 0.100000, loss 8.604727
+INFO 2021-09-17 13:42:21 train.py: 82] Epoch 1, iter 3800/6416, lr 0.100000, loss 8.512420
+INFO 2021-09-17 13:45:31 train.py: 82] Epoch 1, iter 4000/6416, lr 0.100000, loss 8.403469
+INFO 2021-09-17 13:48:42 train.py: 82] Epoch 1, iter 4200/6416, lr 0.100000, loss 8.249118
+INFO 2021-09-17 13:51:59 train.py: 82] Epoch 1, iter 4400/6416, lr 0.100000, loss 8.187804
+INFO 2021-09-17 13:55:15 train.py: 82] Epoch 1, iter 4600/6416, lr 0.100000, loss 8.050545
+INFO 2021-09-17 13:58:36 train.py: 82] Epoch 1, iter 4800/6416, lr 0.100000, loss 7.953538
+INFO 2021-09-17 14:01:48 train.py: 82] Epoch 1, iter 5000/6416, lr 0.100000, loss 7.863615
+INFO 2021-09-17 14:05:19 train.py: 82] Epoch 1, iter 5200/6416, lr 0.100000, loss 7.783763
+INFO 2021-09-17 14:08:45 train.py: 82] Epoch 1, iter 5400/6416, lr 0.100000, loss 7.709828
+INFO 2021-09-17 14:12:18 train.py: 82] Epoch 1, iter 5600/6416, lr 0.100000, loss 7.599424
+INFO 2021-09-17 14:16:07 train.py: 82] Epoch 1, iter 5800/6416, lr 0.100000, loss 7.491830
+INFO 2021-09-17 14:19:46 train.py: 95] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-09-17 14:19:47 train.py: 82] Epoch 1, iter 6000/6416, lr 0.100000, loss 7.464325
+INFO 2021-09-17 14:23:15 train.py: 82] Epoch 1, iter 6200/6416, lr 0.100000, loss 7.401899
+INFO 2021-09-17 14:27:09 train.py: 82] Epoch 1, iter 6400/6416, lr 0.100000, loss 7.315104
+INFO 2021-09-17 14:27:25 train.py: 100] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-09-17 14:27:27 train.py: 82] Epoch 2, iter 0/6416, lr 0.100000, loss 7.161765
+INFO 2021-09-17 14:30:23 train.py: 82] Epoch 2, iter 200/6416, lr 0.100000, loss 6.645709
+INFO 2021-09-17 14:33:18 train.py: 82] Epoch 2, iter 400/6416, lr 0.100000, loss 6.619753
+INFO 2021-09-17 14:36:13 train.py: 82] Epoch 2, iter 600/6416, lr 0.100000, loss 6.663946
+INFO 2021-09-17 14:39:08 train.py: 82] Epoch 2, iter 800/6416, lr 0.100000, loss 6.655832
+INFO 2021-09-17 14:42:04 train.py: 82] Epoch 2, iter 1000/6416, lr 0.100000, loss 6.705996
+INFO 2021-09-17 14:44:59 train.py: 82] Epoch 2, iter 1200/6416, lr 0.100000, loss 6.702095
+INFO 2021-09-17 14:47:54 train.py: 82] Epoch 2, iter 1400/6416, lr 0.100000, loss 6.694686
+INFO 2021-09-17 14:50:49 train.py: 82] Epoch 2, iter 1600/6416, lr 0.100000, loss 6.687016
+INFO 2021-09-17 14:53:44 train.py: 82] Epoch 2, iter 1800/6416, lr 0.100000, loss 6.620364
+INFO 2021-09-17 14:56:40 train.py: 82] Epoch 2, iter 2000/6416, lr 0.100000, loss 6.632934
+INFO 2021-09-17 14:59:39 train.py: 82] Epoch 2, iter 2200/6416, lr 0.100000, loss 6.588741
+INFO 2021-09-17 15:02:34 train.py: 82] Epoch 2, iter 2400/6416, lr 0.100000, loss 6.556752
+INFO 2021-09-17 15:05:30 train.py: 82] Epoch 2, iter 2600/6416, lr 0.100000, loss 6.550915
+INFO 2021-09-17 15:08:25 train.py: 82] Epoch 2, iter 2800/6416, lr 0.100000, loss 6.522386
+INFO 2021-09-17 15:11:23 train.py: 95] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-09-17 15:11:24 train.py: 82] Epoch 2, iter 3000/6416, lr 0.100000, loss 6.461958
+INFO 2021-09-17 15:14:26 train.py: 82] Epoch 2, iter 3200/6416, lr 0.100000, loss 6.447941
+INFO 2021-09-17 15:17:27 train.py: 82] Epoch 2, iter 3400/6416, lr 0.100000, loss 6.381988
+INFO 2021-09-17 15:20:26 train.py: 82] Epoch 2, iter 3600/6416, lr 0.100000, loss 6.406001
+INFO 2021-09-17 15:23:33 train.py: 82] Epoch 2, iter 3800/6416, lr 0.100000, loss 6.362502
+INFO 2021-09-17 15:26:38 train.py: 82] Epoch 2, iter 4000/6416, lr 0.100000, loss 6.354174
+INFO 2021-09-17 15:29:48 train.py: 82] Epoch 2, iter 4200/6416, lr 0.100000, loss 6.298320
+INFO 2021-09-17 15:33:11 train.py: 82] Epoch 2, iter 4400/6416, lr 0.100000, loss 6.246674
+INFO 2021-09-17 15:36:26 train.py: 82] Epoch 2, iter 4600/6416, lr 0.100000, loss 6.234659
+INFO 2021-09-17 15:39:42 train.py: 82] Epoch 2, iter 4800/6416, lr 0.100000, loss 6.224736
+INFO 2021-09-17 15:43:14 train.py: 82] Epoch 2, iter 5000/6416, lr 0.100000, loss 6.144226
+INFO 2021-09-17 15:46:41 train.py: 82] Epoch 2, iter 5200/6416, lr 0.100000, loss 6.116475
+INFO 2021-09-17 15:49:46 train.py: 82] Epoch 2, iter 5400/6416, lr 0.100000, loss 6.098807
+INFO 2021-09-17 15:53:24 train.py: 82] Epoch 2, iter 5600/6416, lr 0.100000, loss 6.070324
+INFO 2021-09-17 15:56:45 train.py: 82] Epoch 2, iter 5800/6416, lr 0.100000, loss 6.051608
+INFO 2021-09-17 15:59:59 train.py: 95] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-09-17 15:59:59 train.py: 82] Epoch 2, iter 6000/6416, lr 0.100000, loss 6.028819
+INFO 2021-09-17 16:03:42 train.py: 82] Epoch 2, iter 6200/6416, lr 0.100000, loss 5.992936
+INFO 2021-09-17 16:07:15 train.py: 82] Epoch 2, iter 6400/6416, lr 0.100000, loss 5.976271
+INFO 2021-09-17 16:07:30 train.py: 100] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-09-17 16:07:33 train.py: 82] Epoch 3, iter 0/6416, lr 0.100000, loss 5.908553
+INFO 2021-09-17 16:10:29 train.py: 82] Epoch 3, iter 200/6416, lr 0.100000, loss 5.294825
+INFO 2021-09-17 16:13:24 train.py: 82] Epoch 3, iter 400/6416, lr 0.100000, loss 5.298500
+INFO 2021-09-17 16:16:19 train.py: 82] Epoch 3, iter 600/6416, lr 0.100000, loss 5.369991
+INFO 2021-09-17 16:19:14 train.py: 82] Epoch 3, iter 800/6416, lr 0.100000, loss 5.461115
+INFO 2021-09-17 16:22:09 train.py: 82] Epoch 3, iter 1000/6416, lr 0.100000, loss 5.487593
+INFO 2021-09-17 16:25:04 train.py: 82] Epoch 3, iter 1200/6416, lr 0.100000, loss 5.504771
+INFO 2021-09-17 16:27:59 train.py: 82] Epoch 3, iter 1400/6416, lr 0.100000, loss 5.524527
+INFO 2021-09-17 16:30:55 train.py: 82] Epoch 3, iter 1600/6416, lr 0.100000, loss 5.573862
+INFO 2021-09-17 16:33:50 train.py: 82] Epoch 3, iter 1800/6416, lr 0.100000, loss 5.517027
+INFO 2021-09-17 16:36:45 train.py: 82] Epoch 3, iter 2000/6416, lr 0.100000, loss 5.554914
+INFO 2021-09-17 16:39:40 train.py: 82] Epoch 3, iter 2200/6416, lr 0.100000, loss 5.565343
+INFO 2021-09-17 16:42:35 train.py: 82] Epoch 3, iter 2400/6416, lr 0.100000, loss 5.545687
+INFO 2021-09-17 16:45:31 train.py: 82] Epoch 3, iter 2600/6416, lr 0.100000, loss 5.560186
+INFO 2021-09-17 16:48:26 train.py: 82] Epoch 3, iter 2800/6416, lr 0.100000, loss 5.519884
+INFO 2021-09-17 16:51:23 train.py: 95] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-09-17 16:51:24 train.py: 82] Epoch 3, iter 3000/6416, lr 0.100000, loss 5.537192
+INFO 2021-09-17 16:54:20 train.py: 82] Epoch 3, iter 3200/6416, lr 0.100000, loss 5.522058
+INFO 2021-09-17 16:57:16 train.py: 82] Epoch 3, iter 3400/6416, lr 0.100000, loss 5.485364
+INFO 2021-09-17 17:00:12 train.py: 82] Epoch 3, iter 3600/6416, lr 0.100000, loss 5.508276
+INFO 2021-09-17 17:03:09 train.py: 82] Epoch 3, iter 3800/6416, lr 0.100000, loss 5.463995
+INFO 2021-09-17 17:06:05 train.py: 82] Epoch 3, iter 4000/6416, lr 0.100000, loss 5.434506
+INFO 2021-09-17 17:09:01 train.py: 82] Epoch 3, iter 4200/6416, lr 0.100000, loss 5.444456
+INFO 2021-09-17 17:11:57 train.py: 82] Epoch 3, iter 4400/6416, lr 0.100000, loss 5.461675
+INFO 2021-09-17 17:14:53 train.py: 82] Epoch 3, iter 4600/6416, lr 0.100000, loss 5.423387
+INFO 2021-09-17 17:17:49 train.py: 82] Epoch 3, iter 4800/6416, lr 0.100000, loss 5.409078
+INFO 2021-09-17 17:20:46 train.py: 82] Epoch 3, iter 5000/6416, lr 0.100000, loss 5.374251
+INFO 2021-09-17 17:23:43 train.py: 82] Epoch 3, iter 5200/6416, lr 0.100000, loss 5.365706
+INFO 2021-09-17 17:26:39 train.py: 82] Epoch 3, iter 5400/6416, lr 0.100000, loss 5.337068
+INFO 2021-09-17 17:29:43 train.py: 82] Epoch 3, iter 5600/6416, lr 0.100000, loss 5.320861
+INFO 2021-09-17 17:32:49 train.py: 82] Epoch 3, iter 5800/6416, lr 0.100000, loss 5.312380
+INFO 2021-09-17 17:35:52 train.py: 95] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-09-17 17:35:53 train.py: 82] Epoch 3, iter 6000/6416, lr 0.100000, loss 5.317948
+INFO 2021-09-17 17:38:50 train.py: 82] Epoch 3, iter 6200/6416, lr 0.100000, loss 5.315086
+INFO 2021-09-17 17:41:47 train.py: 82] Epoch 3, iter 6400/6416, lr 0.100000, loss 5.248424
+INFO 2021-09-17 17:42:03 train.py: 100] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-09-17 17:42:05 train.py: 82] Epoch 4, iter 0/6416, lr 0.100000, loss 5.296052
+INFO 2021-09-17 17:45:01 train.py: 82] Epoch 4, iter 200/6416, lr 0.100000, loss 4.677725
+INFO 2021-09-17 17:47:56 train.py: 82] Epoch 4, iter 400/6416, lr 0.100000, loss 4.651695
+INFO 2021-09-17 17:50:52 train.py: 82] Epoch 4, iter 600/6416, lr 0.100000, loss 4.759351
+INFO 2021-09-17 17:53:47 train.py: 82] Epoch 4, iter 800/6416, lr 0.100000, loss 4.757572
+INFO 2021-09-17 17:56:42 train.py: 82] Epoch 4, iter 1000/6416, lr 0.100000, loss 4.822478
+INFO 2021-09-17 17:59:37 train.py: 82] Epoch 4, iter 1200/6416, lr 0.100000, loss 4.889204
+INFO 2021-09-17 18:02:32 train.py: 82] Epoch 4, iter 1400/6416, lr 0.100000, loss 4.921907
+INFO 2021-09-17 18:05:27 train.py: 82] Epoch 4, iter 1600/6416, lr 0.100000, loss 4.922673
+INFO 2021-09-17 18:08:22 train.py: 82] Epoch 4, iter 1800/6416, lr 0.100000, loss 4.958898
+INFO 2021-09-17 18:11:18 train.py: 82] Epoch 4, iter 2000/6416, lr 0.100000, loss 4.973729
+INFO 2021-09-17 18:14:13 train.py: 82] Epoch 4, iter 2200/6416, lr 0.100000, loss 4.955810
+INFO 2021-09-17 18:17:09 train.py: 82] Epoch 4, iter 2400/6416, lr 0.100000, loss 4.971030
+INFO 2021-09-17 18:20:05 train.py: 82] Epoch 4, iter 2600/6416, lr 0.100000, loss 4.999764
+INFO 2021-09-17 18:23:00 train.py: 82] Epoch 4, iter 2800/6416, lr 0.100000, loss 4.963907
+INFO 2021-09-17 18:25:58 train.py: 95] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-09-17 18:25:58 train.py: 82] Epoch 4, iter 3000/6416, lr 0.100000, loss 4.977689
+INFO 2021-09-17 18:28:55 train.py: 82] Epoch 4, iter 3200/6416, lr 0.100000, loss 4.990679
+INFO 2021-09-17 18:31:51 train.py: 82] Epoch 4, iter 3400/6416, lr 0.100000, loss 4.972846
+INFO 2021-09-17 18:34:48 train.py: 82] Epoch 4, iter 3600/6416, lr 0.100000, loss 4.988738
+INFO 2021-09-17 18:37:44 train.py: 82] Epoch 4, iter 3800/6416, lr 0.100000, loss 4.958032
+INFO 2021-09-17 18:40:41 train.py: 82] Epoch 4, iter 4000/6416, lr 0.100000, loss 4.949542
+INFO 2021-09-17 18:43:37 train.py: 82] Epoch 4, iter 4200/6416, lr 0.100000, loss 4.930051
+INFO 2021-09-17 18:46:34 train.py: 82] Epoch 4, iter 4400/6416, lr 0.100000, loss 4.935332
+INFO 2021-09-17 18:49:30 train.py: 82] Epoch 4, iter 4600/6416, lr 0.100000, loss 4.899679
+INFO 2021-09-17 18:52:26 train.py: 82] Epoch 4, iter 4800/6416, lr 0.100000, loss 4.919532
+INFO 2021-09-17 18:55:22 train.py: 82] Epoch 4, iter 5000/6416, lr 0.100000, loss 4.923083
+INFO 2021-09-17 18:58:19 train.py: 82] Epoch 4, iter 5200/6416, lr 0.100000, loss 4.884307
+INFO 2021-09-17 19:01:15 train.py: 82] Epoch 4, iter 5400/6416, lr 0.100000, loss 4.888160
+INFO 2021-09-17 19:04:11 train.py: 82] Epoch 4, iter 5600/6416, lr 0.100000, loss 4.838418
+INFO 2021-09-17 19:07:07 train.py: 82] Epoch 4, iter 5800/6416, lr 0.100000, loss 4.851031
+INFO 2021-09-17 19:10:05 train.py: 95] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-09-17 19:10:06 train.py: 82] Epoch 4, iter 6000/6416, lr 0.100000, loss 4.857647
+INFO 2021-09-17 19:13:02 train.py: 82] Epoch 4, iter 6200/6416, lr 0.100000, loss 4.836750
+INFO 2021-09-17 19:15:58 train.py: 82] Epoch 4, iter 6400/6416, lr 0.100000, loss 4.866601
+INFO 2021-09-17 19:16:14 train.py: 100] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-09-17 19:16:16 train.py: 82] Epoch 5, iter 0/6416, lr 0.100000, loss 4.732683
+INFO 2021-09-17 19:19:12 train.py: 82] Epoch 5, iter 200/6416, lr 0.100000, loss 4.270972
+INFO 2021-09-17 19:22:08 train.py: 82] Epoch 5, iter 400/6416, lr 0.100000, loss 4.234095
+INFO 2021-09-17 19:25:03 train.py: 82] Epoch 5, iter 600/6416, lr 0.100000, loss 4.300946
+INFO 2021-09-17 19:27:58 train.py: 82] Epoch 5, iter 800/6416, lr 0.100000, loss 4.382501
+INFO 2021-09-17 19:30:53 train.py: 82] Epoch 5, iter 1000/6416, lr 0.100000, loss 4.435821
+INFO 2021-09-17 19:33:48 train.py: 82] Epoch 5, iter 1200/6416, lr 0.100000, loss 4.439177
+INFO 2021-09-17 19:36:43 train.py: 82] Epoch 5, iter 1400/6416, lr 0.100000, loss 4.523221
+INFO 2021-09-17 19:39:38 train.py: 82] Epoch 5, iter 1600/6416, lr 0.100000, loss 4.576399
+INFO 2021-09-17 19:42:33 train.py: 82] Epoch 5, iter 1800/6416, lr 0.100000, loss 4.550761
+INFO 2021-09-17 19:45:28 train.py: 82] Epoch 5, iter 2000/6416, lr 0.100000, loss 4.570737
+INFO 2021-09-17 19:48:24 train.py: 82] Epoch 5, iter 2200/6416, lr 0.100000, loss 4.595881
+INFO 2021-09-17 19:51:19 train.py: 82] Epoch 5, iter 2400/6416, lr 0.100000, loss 4.597567
+INFO 2021-09-17 19:54:15 train.py: 82] Epoch 5, iter 2600/6416, lr 0.100000, loss 4.606163
+INFO 2021-09-17 19:57:10 train.py: 82] Epoch 5, iter 2800/6416, lr 0.100000, loss 4.604537
+INFO 2021-09-17 20:00:07 train.py: 95] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-09-17 20:00:08 train.py: 82] Epoch 5, iter 3000/6416, lr 0.100000, loss 4.634249
+INFO 2021-09-17 20:03:04 train.py: 82] Epoch 5, iter 3200/6416, lr 0.100000, loss 4.602003
+INFO 2021-09-17 20:06:00 train.py: 82] Epoch 5, iter 3400/6416, lr 0.100000, loss 4.618445
+INFO 2021-09-17 20:08:57 train.py: 82] Epoch 5, iter 3600/6416, lr 0.100000, loss 4.605092
+INFO 2021-09-17 20:11:53 train.py: 82] Epoch 5, iter 3800/6416, lr 0.100000, loss 4.617896
+INFO 2021-09-17 20:14:49 train.py: 82] Epoch 5, iter 4000/6416, lr 0.100000, loss 4.620143
+INFO 2021-09-17 20:17:45 train.py: 82] Epoch 5, iter 4200/6416, lr 0.100000, loss 4.595380
+INFO 2021-09-17 20:20:41 train.py: 82] Epoch 5, iter 4400/6416, lr 0.100000, loss 4.603217
+INFO 2021-09-17 20:23:38 train.py: 82] Epoch 5, iter 4600/6416, lr 0.100000, loss 4.610298
+INFO 2021-09-17 20:26:34 train.py: 82] Epoch 5, iter 4800/6416, lr 0.100000, loss 4.580344
+INFO 2021-09-17 20:29:31 train.py: 82] Epoch 5, iter 5000/6416, lr 0.100000, loss 4.554963
+INFO 2021-09-17 20:32:28 train.py: 82] Epoch 5, iter 5200/6416, lr 0.100000, loss 4.565137
+INFO 2021-09-17 20:35:25 train.py: 82] Epoch 5, iter 5400/6416, lr 0.100000, loss 4.554716
+INFO 2021-09-17 20:38:21 train.py: 82] Epoch 5, iter 5600/6416, lr 0.100000, loss 4.558520
+INFO 2021-09-17 20:41:18 train.py: 82] Epoch 5, iter 5800/6416, lr 0.100000, loss 4.587286
+INFO 2021-09-17 20:44:16 train.py: 95] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-09-17 20:44:17 train.py: 82] Epoch 5, iter 6000/6416, lr 0.100000, loss 4.557456
+INFO 2021-09-17 20:47:14 train.py: 82] Epoch 5, iter 6200/6416, lr 0.100000, loss 4.563951
+INFO 2021-09-17 20:50:11 train.py: 82] Epoch 5, iter 6400/6416, lr 0.100000, loss 4.545721
+INFO 2021-09-17 20:50:27 train.py: 100] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-09-17 20:50:29 train.py: 82] Epoch 6, iter 0/6416, lr 0.100000, loss 4.476425
+INFO 2021-09-17 20:53:26 train.py: 82] Epoch 6, iter 200/6416, lr 0.100000, loss 4.013819
+INFO 2021-09-17 20:56:22 train.py: 82] Epoch 6, iter 400/6416, lr 0.100000, loss 3.942493
+INFO 2021-09-17 20:59:18 train.py: 82] Epoch 6, iter 600/6416, lr 0.100000, loss 4.034682
+INFO 2021-09-17 21:02:13 train.py: 82] Epoch 6, iter 800/6416, lr 0.100000, loss 4.105159
+INFO 2021-09-17 21:05:08 train.py: 82] Epoch 6, iter 1000/6416, lr 0.100000, loss 4.133657
+INFO 2021-09-17 21:08:03 train.py: 82] Epoch 6, iter 1200/6416, lr 0.100000, loss 4.167027
+INFO 2021-09-17 21:10:59 train.py: 82] Epoch 6, iter 1400/6416, lr 0.100000, loss 4.235294
+INFO 2021-09-17 21:13:54 train.py: 82] Epoch 6, iter 1600/6416, lr 0.100000, loss 4.278386
+INFO 2021-09-17 21:16:50 train.py: 82] Epoch 6, iter 1800/6416, lr 0.100000, loss 4.279158
+INFO 2021-09-17 21:19:45 train.py: 82] Epoch 6, iter 2000/6416, lr 0.100000, loss 4.295179
+INFO 2021-09-17 21:22:41 train.py: 82] Epoch 6, iter 2200/6416, lr 0.100000, loss 4.324185
+INFO 2021-09-17 21:25:36 train.py: 82] Epoch 6, iter 2400/6416, lr 0.100000, loss 4.354933
+INFO 2021-09-17 21:28:31 train.py: 82] Epoch 6, iter 2600/6416, lr 0.100000, loss 4.362097
+INFO 2021-09-17 21:31:27 train.py: 82] Epoch 6, iter 2800/6416, lr 0.100000, loss 4.354158
+INFO 2021-09-17 21:34:24 train.py: 95] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-09-17 21:34:25 train.py: 82] Epoch 6, iter 3000/6416, lr 0.100000, loss 4.348253
+INFO 2021-09-17 21:37:21 train.py: 82] Epoch 6, iter 3200/6416, lr 0.100000, loss 4.363612
+INFO 2021-09-17 21:40:17 train.py: 82] Epoch 6, iter 3400/6416, lr 0.100000, loss 4.358525
+INFO 2021-09-17 21:43:14 train.py: 82] Epoch 6, iter 3600/6416, lr 0.100000, loss 4.381478
+INFO 2021-09-17 21:46:10 train.py: 82] Epoch 6, iter 3800/6416, lr 0.100000, loss 4.368137
+INFO 2021-09-17 21:49:06 train.py: 82] Epoch 6, iter 4000/6416, lr 0.100000, loss 4.400902
+INFO 2021-09-17 21:52:02 train.py: 82] Epoch 6, iter 4200/6416, lr 0.100000, loss 4.371290
+INFO 2021-09-17 21:54:58 train.py: 82] Epoch 6, iter 4400/6416, lr 0.100000, loss 4.356726
+INFO 2021-09-17 21:57:54 train.py: 82] Epoch 6, iter 4600/6416, lr 0.100000, loss 4.350521
+INFO 2021-09-17 22:00:51 train.py: 82] Epoch 6, iter 4800/6416, lr 0.100000, loss 4.362551
+INFO 2021-09-17 22:03:47 train.py: 82] Epoch 6, iter 5000/6416, lr 0.100000, loss 4.358347
+INFO 2021-09-17 22:06:43 train.py: 82] Epoch 6, iter 5200/6416, lr 0.100000, loss 4.348168
+INFO 2021-09-17 22:09:39 train.py: 82] Epoch 6, iter 5400/6416, lr 0.100000, loss 4.332241
+INFO 2021-09-17 22:12:36 train.py: 82] Epoch 6, iter 5600/6416, lr 0.100000, loss 4.323315
+INFO 2021-09-17 22:15:33 train.py: 82] Epoch 6, iter 5800/6416, lr 0.100000, loss 4.333842
+INFO 2021-09-17 22:18:31 train.py: 95] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-09-17 22:18:32 train.py: 82] Epoch 6, iter 6000/6416, lr 0.100000, loss 4.332040
+INFO 2021-09-17 22:21:28 train.py: 82] Epoch 6, iter 6200/6416, lr 0.100000, loss 4.339341
+INFO 2021-09-17 22:24:24 train.py: 82] Epoch 6, iter 6400/6416, lr 0.100000, loss 4.348284
+INFO 2021-09-17 22:24:40 train.py: 100] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-09-17 22:24:42 train.py: 82] Epoch 7, iter 0/6416, lr 0.100000, loss 4.334388
+INFO 2021-09-17 22:27:38 train.py: 82] Epoch 7, iter 200/6416, lr 0.100000, loss 3.841594
+INFO 2021-09-17 22:30:35 train.py: 82] Epoch 7, iter 400/6416, lr 0.100000, loss 3.762655
+INFO 2021-09-17 22:33:30 train.py: 82] Epoch 7, iter 600/6416, lr 0.100000, loss 3.832084
+INFO 2021-09-17 22:36:26 train.py: 82] Epoch 7, iter 800/6416, lr 0.100000, loss 3.852691
+INFO 2021-09-17 22:39:21 train.py: 82] Epoch 7, iter 1000/6416, lr 0.100000, loss 3.924323
+INFO 2021-09-17 22:42:16 train.py: 82] Epoch 7, iter 1200/6416, lr 0.100000, loss 3.960904
+INFO 2021-09-17 22:45:11 train.py: 82] Epoch 7, iter 1400/6416, lr 0.100000, loss 4.028336
+INFO 2021-09-17 22:48:07 train.py: 82] Epoch 7, iter 1600/6416, lr 0.100000, loss 4.045936
+INFO 2021-09-17 22:51:02 train.py: 82] Epoch 7, iter 1800/6416, lr 0.100000, loss 4.064921
+INFO 2021-09-17 22:53:57 train.py: 82] Epoch 7, iter 2000/6416, lr 0.100000, loss 4.106194
+INFO 2021-09-17 22:56:53 train.py: 82] Epoch 7, iter 2200/6416, lr 0.100000, loss 4.117006
+INFO 2021-09-17 22:59:48 train.py: 82] Epoch 7, iter 2400/6416, lr 0.100000, loss 4.135059
+INFO 2021-09-17 23:02:43 train.py: 82] Epoch 7, iter 2600/6416, lr 0.100000, loss 4.149813
+INFO 2021-09-17 23:05:38 train.py: 82] Epoch 7, iter 2800/6416, lr 0.100000, loss 4.153727
+INFO 2021-09-17 23:08:35 train.py: 95] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-09-17 23:08:36 train.py: 82] Epoch 7, iter 3000/6416, lr 0.100000, loss 4.172337
+INFO 2021-09-17 23:11:32 train.py: 82] Epoch 7, iter 3200/6416, lr 0.100000, loss 4.168360
+INFO 2021-09-17 23:14:28 train.py: 82] Epoch 7, iter 3400/6416, lr 0.100000, loss 4.172255
+INFO 2021-09-17 23:17:24 train.py: 82] Epoch 7, iter 3600/6416, lr 0.100000, loss 4.181818
+INFO 2021-09-17 23:20:20 train.py: 82] Epoch 7, iter 3800/6416, lr 0.100000, loss 4.179822
+INFO 2021-09-17 23:23:17 train.py: 82] Epoch 7, iter 4000/6416, lr 0.100000, loss 4.177024
+INFO 2021-09-17 23:26:13 train.py: 82] Epoch 7, iter 4200/6416, lr 0.100000, loss 4.207950
+INFO 2021-09-17 23:29:09 train.py: 82] Epoch 7, iter 4400/6416, lr 0.100000, loss 4.200332
+INFO 2021-09-17 23:32:06 train.py: 82] Epoch 7, iter 4600/6416, lr 0.100000, loss 4.198892
+INFO 2021-09-17 23:35:02 train.py: 82] Epoch 7, iter 4800/6416, lr 0.100000, loss 4.146377
+INFO 2021-09-17 23:37:58 train.py: 82] Epoch 7, iter 5000/6416, lr 0.100000, loss 4.174750
+INFO 2021-09-17 23:40:54 train.py: 82] Epoch 7, iter 5200/6416, lr 0.100000, loss 4.216462
+INFO 2021-09-17 23:43:51 train.py: 82] Epoch 7, iter 5400/6416, lr 0.100000, loss 4.143825
+INFO 2021-09-17 23:46:47 train.py: 82] Epoch 7, iter 5600/6416, lr 0.100000, loss 4.194038
+INFO 2021-09-17 23:49:43 train.py: 82] Epoch 7, iter 5800/6416, lr 0.100000, loss 4.176992
+INFO 2021-09-17 23:52:41 train.py: 95] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-09-17 23:52:42 train.py: 82] Epoch 7, iter 6000/6416, lr 0.100000, loss 4.174215
+INFO 2021-09-17 23:55:38 train.py: 82] Epoch 7, iter 6200/6416, lr 0.100000, loss 4.159925
+INFO 2021-09-17 23:58:34 train.py: 82] Epoch 7, iter 6400/6416, lr 0.100000, loss 4.164781
+INFO 2021-09-17 23:58:50 train.py: 100] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-09-17 23:58:52 train.py: 82] Epoch 8, iter 0/6416, lr 0.100000, loss 4.178131
+INFO 2021-09-18 00:01:48 train.py: 82] Epoch 8, iter 200/6416, lr 0.100000, loss 3.613638
+INFO 2021-09-18 00:04:44 train.py: 82] Epoch 8, iter 400/6416, lr 0.100000, loss 3.586572
+INFO 2021-09-18 00:07:40 train.py: 82] Epoch 8, iter 600/6416, lr 0.100000, loss 3.632357
+INFO 2021-09-18 00:10:36 train.py: 82] Epoch 8, iter 800/6416, lr 0.100000, loss 3.722838
+INFO 2021-09-18 00:13:31 train.py: 82] Epoch 8, iter 1000/6416, lr 0.100000, loss 3.771134
+INFO 2021-09-18 00:16:26 train.py: 82] Epoch 8, iter 1200/6416, lr 0.100000, loss 3.831397
+INFO 2021-09-18 00:19:21 train.py: 82] Epoch 8, iter 1400/6416, lr 0.100000, loss 3.871126
+INFO 2021-09-18 00:22:16 train.py: 82] Epoch 8, iter 1600/6416, lr 0.100000, loss 3.907012
+INFO 2021-09-18 00:25:11 train.py: 82] Epoch 8, iter 1800/6416, lr 0.100000, loss 3.927835
+INFO 2021-09-18 00:28:07 train.py: 82] Epoch 8, iter 2000/6416, lr 0.100000, loss 3.956889
+INFO 2021-09-18 00:31:03 train.py: 82] Epoch 8, iter 2200/6416, lr 0.100000, loss 3.945203
+INFO 2021-09-18 00:33:58 train.py: 82] Epoch 8, iter 2400/6416, lr 0.100000, loss 3.978433
+INFO 2021-09-18 00:36:54 train.py: 82] Epoch 8, iter 2600/6416, lr 0.100000, loss 3.969880
+INFO 2021-09-18 00:39:49 train.py: 82] Epoch 8, iter 2800/6416, lr 0.100000, loss 4.034084
+INFO 2021-09-18 00:42:47 train.py: 95] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-09-18 00:42:48 train.py: 82] Epoch 8, iter 3000/6416, lr 0.100000, loss 4.008741
+INFO 2021-09-18 00:45:44 train.py: 82] Epoch 8, iter 3200/6416, lr 0.100000, loss 4.036693
+INFO 2021-09-18 00:48:41 train.py: 82] Epoch 8, iter 3400/6416, lr 0.100000, loss 4.055786
+INFO 2021-09-18 00:51:38 train.py: 82] Epoch 8, iter 3600/6416, lr 0.100000, loss 4.019490
+INFO 2021-09-18 00:54:34 train.py: 82] Epoch 8, iter 3800/6416, lr 0.100000, loss 4.006062
+INFO 2021-09-18 00:57:31 train.py: 82] Epoch 8, iter 4000/6416, lr 0.100000, loss 4.043445
+INFO 2021-09-18 01:00:27 train.py: 82] Epoch 8, iter 4200/6416, lr 0.100000, loss 4.050289
+INFO 2021-09-18 01:03:24 train.py: 82] Epoch 8, iter 4400/6416, lr 0.100000, loss 4.050702
+INFO 2021-09-18 01:06:21 train.py: 82] Epoch 8, iter 4600/6416, lr 0.100000, loss 4.045620
+INFO 2021-09-18 01:09:17 train.py: 82] Epoch 8, iter 4800/6416, lr 0.100000, loss 4.041173
+INFO 2021-09-18 01:12:14 train.py: 82] Epoch 8, iter 5000/6416, lr 0.100000, loss 4.039389
+INFO 2021-09-18 01:15:10 train.py: 82] Epoch 8, iter 5200/6416, lr 0.100000, loss 4.014804
+INFO 2021-09-18 01:18:07 train.py: 82] Epoch 8, iter 5400/6416, lr 0.100000, loss 4.056807
+INFO 2021-09-18 01:21:04 train.py: 82] Epoch 8, iter 5600/6416, lr 0.100000, loss 4.025864
+INFO 2021-09-18 01:24:01 train.py: 82] Epoch 8, iter 5800/6416, lr 0.100000, loss 4.031175
+INFO 2021-09-18 01:26:59 train.py: 95] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-09-18 01:27:00 train.py: 82] Epoch 8, iter 6000/6416, lr 0.100000, loss 4.061855
+INFO 2021-09-18 01:29:56 train.py: 82] Epoch 8, iter 6200/6416, lr 0.100000, loss 4.026516
+INFO 2021-09-18 01:32:53 train.py: 82] Epoch 8, iter 6400/6416, lr 0.100000, loss 4.045800
+INFO 2021-09-18 01:33:09 train.py: 100] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-09-18 01:33:11 train.py: 82] Epoch 9, iter 0/6416, lr 0.100000, loss 3.968938
+INFO 2021-09-18 01:36:08 train.py: 82] Epoch 9, iter 200/6416, lr 0.100000, loss 3.493595
+INFO 2021-09-18 01:39:04 train.py: 82] Epoch 9, iter 400/6416, lr 0.100000, loss 3.436604
+INFO 2021-09-18 01:42:00 train.py: 82] Epoch 9, iter 600/6416, lr 0.100000, loss 3.537616
+INFO 2021-09-18 01:44:55 train.py: 82] Epoch 9, iter 800/6416, lr 0.100000, loss 3.600449
+INFO 2021-09-18 01:47:50 train.py: 82] Epoch 9, iter 1000/6416, lr 0.100000, loss 3.650454
+INFO 2021-09-18 01:50:45 train.py: 82] Epoch 9, iter 1200/6416, lr 0.100000, loss 3.693088
+INFO 2021-09-18 01:53:40 train.py: 82] Epoch 9, iter 1400/6416, lr 0.100000, loss 3.753376
+INFO 2021-09-18 01:56:36 train.py: 82] Epoch 9, iter 1600/6416, lr 0.100000, loss 3.759929
+INFO 2021-09-18 01:59:31 train.py: 82] Epoch 9, iter 1800/6416, lr 0.100000, loss 3.814349
+INFO 2021-09-18 02:02:26 train.py: 82] Epoch 9, iter 2000/6416, lr 0.100000, loss 3.809961
+INFO 2021-09-18 02:05:21 train.py: 82] Epoch 9, iter 2200/6416, lr 0.100000, loss 3.860529
+INFO 2021-09-18 02:08:16 train.py: 82] Epoch 9, iter 2400/6416, lr 0.100000, loss 3.853160
+INFO 2021-09-18 02:11:12 train.py: 82] Epoch 9, iter 2600/6416, lr 0.100000, loss 3.890652
+INFO 2021-09-18 02:14:07 train.py: 82] Epoch 9, iter 2800/6416, lr 0.100000, loss 3.889155
+INFO 2021-09-18 02:17:05 train.py: 95] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-09-18 02:17:06 train.py: 82] Epoch 9, iter 3000/6416, lr 0.100000, loss 3.883740
+INFO 2021-09-18 02:20:02 train.py: 82] Epoch 9, iter 3200/6416, lr 0.100000, loss 3.899819
+INFO 2021-09-18 02:22:58 train.py: 82] Epoch 9, iter 3400/6416, lr 0.100000, loss 3.915588
+INFO 2021-09-18 02:25:55 train.py: 82] Epoch 9, iter 3600/6416, lr 0.100000, loss 3.924188
+INFO 2021-09-18 02:28:52 train.py: 82] Epoch 9, iter 3800/6416, lr 0.100000, loss 3.898071
+INFO 2021-09-18 02:31:48 train.py: 82] Epoch 9, iter 4000/6416, lr 0.100000, loss 3.917155
+INFO 2021-09-18 02:34:44 train.py: 82] Epoch 9, iter 4200/6416, lr 0.100000, loss 3.916235
+INFO 2021-09-18 02:37:41 train.py: 82] Epoch 9, iter 4400/6416, lr 0.100000, loss 3.901233
+INFO 2021-09-18 02:40:37 train.py: 82] Epoch 9, iter 4600/6416, lr 0.100000, loss 3.910037
+INFO 2021-09-18 02:43:34 train.py: 82] Epoch 9, iter 4800/6416, lr 0.100000, loss 3.923329
+INFO 2021-09-18 02:46:30 train.py: 82] Epoch 9, iter 5000/6416, lr 0.100000, loss 3.924127
+INFO 2021-09-18 02:49:27 train.py: 82] Epoch 9, iter 5200/6416, lr 0.100000, loss 3.929169
+INFO 2021-09-18 02:52:23 train.py: 82] Epoch 9, iter 5400/6416, lr 0.100000, loss 3.924546
+INFO 2021-09-18 02:55:19 train.py: 82] Epoch 9, iter 5600/6416, lr 0.100000, loss 3.921665
+INFO 2021-09-18 02:58:15 train.py: 82] Epoch 9, iter 5800/6416, lr 0.100000, loss 3.948178
+INFO 2021-09-18 03:01:14 train.py: 95] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-09-18 03:01:14 train.py: 82] Epoch 9, iter 6000/6416, lr 0.100000, loss 3.949781
+INFO 2021-09-18 03:04:11 train.py: 82] Epoch 9, iter 6200/6416, lr 0.100000, loss 3.916651
+INFO 2021-09-18 03:07:08 train.py: 82] Epoch 9, iter 6400/6416, lr 0.100000, loss 3.922783
+INFO 2021-09-18 03:07:23 train.py: 100] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-09-18 03:07:26 train.py: 82] Epoch 10, iter 0/6416, lr 0.010000, loss 3.951655
+INFO 2021-09-18 03:10:22 train.py: 82] Epoch 10, iter 200/6416, lr 0.010000, loss 2.785600
+INFO 2021-09-18 03:13:18 train.py: 82] Epoch 10, iter 400/6416, lr 0.010000, loss 2.529166
+INFO 2021-09-18 03:16:14 train.py: 82] Epoch 10, iter 600/6416, lr 0.010000, loss 2.420989
+INFO 2021-09-18 03:19:10 train.py: 82] Epoch 10, iter 800/6416, lr 0.010000, loss 2.372029
+INFO 2021-09-18 03:22:05 train.py: 82] Epoch 10, iter 1000/6416, lr 0.010000, loss 2.299332
+INFO 2021-09-18 03:25:01 train.py: 82] Epoch 10, iter 1200/6416, lr 0.010000, loss 2.282172
+INFO 2021-09-18 03:27:56 train.py: 82] Epoch 10, iter 1400/6416, lr 0.010000, loss 2.220315
+INFO 2021-09-18 03:30:51 train.py: 82] Epoch 10, iter 1600/6416, lr 0.010000, loss 2.192235
+INFO 2021-09-18 03:33:47 train.py: 82] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.174173
+INFO 2021-09-18 03:36:42 train.py: 82] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.142446
+INFO 2021-09-18 03:39:38 train.py: 82] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.114646
+INFO 2021-09-18 03:42:33 train.py: 82] Epoch 10, iter 2400/6416, lr 0.010000, loss 2.089790
+INFO 2021-09-18 03:45:29 train.py: 82] Epoch 10, iter 2600/6416, lr 0.010000, loss 2.075372
+INFO 2021-09-18 03:48:25 train.py: 82] Epoch 10, iter 2800/6416, lr 0.010000, loss 2.048838
+INFO 2021-09-18 03:51:22 train.py: 95] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-09-18 03:51:23 train.py: 82] Epoch 10, iter 3000/6416, lr 0.010000, loss 2.040410
+INFO 2021-09-18 03:54:19 train.py: 82] Epoch 10, iter 3200/6416, lr 0.010000, loss 2.007740
+INFO 2021-09-18 03:57:15 train.py: 82] Epoch 10, iter 3400/6416, lr 0.010000, loss 1.996273
+INFO 2021-09-18 04:00:12 train.py: 82] Epoch 10, iter 3600/6416, lr 0.010000, loss 2.004067
+INFO 2021-09-18 04:03:08 train.py: 82] Epoch 10, iter 3800/6416, lr 0.010000, loss 1.950489
+INFO 2021-09-18 04:06:04 train.py: 82] Epoch 10, iter 4000/6416, lr 0.010000, loss 1.953415
+INFO 2021-09-18 04:09:01 train.py: 82] Epoch 10, iter 4200/6416, lr 0.010000, loss 1.935262
+INFO 2021-09-18 04:11:58 train.py: 82] Epoch 10, iter 4400/6416, lr 0.010000, loss 1.931762
+INFO 2021-09-18 04:14:55 train.py: 82] Epoch 10, iter 4600/6416, lr 0.010000, loss 1.914031
+INFO 2021-09-18 04:17:52 train.py: 82] Epoch 10, iter 4800/6416, lr 0.010000, loss 1.875355
+INFO 2021-09-18 04:20:48 train.py: 82] Epoch 10, iter 5000/6416, lr 0.010000, loss 1.883015
+INFO 2021-09-18 04:23:45 train.py: 82] Epoch 10, iter 5200/6416, lr 0.010000, loss 1.867335
+INFO 2021-09-18 04:26:42 train.py: 82] Epoch 10, iter 5400/6416, lr 0.010000, loss 1.841941
+INFO 2021-09-18 04:29:39 train.py: 82] Epoch 10, iter 5600/6416, lr 0.010000, loss 1.868534
+INFO 2021-09-18 04:32:35 train.py: 82] Epoch 10, iter 5800/6416, lr 0.010000, loss 1.830383
+INFO 2021-09-18 04:35:34 train.py: 95] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-09-18 04:35:34 train.py: 82] Epoch 10, iter 6000/6416, lr 0.010000, loss 1.836244
+INFO 2021-09-18 04:38:31 train.py: 82] Epoch 10, iter 6200/6416, lr 0.010000, loss 1.808811
+INFO 2021-09-18 04:41:27 train.py: 82] Epoch 10, iter 6400/6416, lr 0.010000, loss 1.782962
+INFO 2021-09-18 04:41:43 train.py: 100] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-09-18 04:41:45 train.py: 82] Epoch 11, iter 0/6416, lr 0.010000, loss 1.777685
+INFO 2021-09-18 04:44:42 train.py: 82] Epoch 11, iter 200/6416, lr 0.010000, loss 1.524485
+INFO 2021-09-18 04:47:38 train.py: 82] Epoch 11, iter 400/6416, lr 0.010000, loss 1.528034
+INFO 2021-09-18 04:50:34 train.py: 82] Epoch 11, iter 600/6416, lr 0.010000, loss 1.504700
+INFO 2021-09-18 04:53:30 train.py: 82] Epoch 11, iter 800/6416, lr 0.010000, loss 1.499111
+INFO 2021-09-18 04:56:26 train.py: 82] Epoch 11, iter 1000/6416, lr 0.010000, loss 1.524296
+INFO 2021-09-18 04:59:22 train.py: 82] Epoch 11, iter 1200/6416, lr 0.010000, loss 1.507689
+INFO 2021-09-18 05:02:17 train.py: 82] Epoch 11, iter 1400/6416, lr 0.010000, loss 1.503928
+INFO 2021-09-18 05:05:13 train.py: 82] Epoch 11, iter 1600/6416, lr 0.010000, loss 1.502943
+INFO 2021-09-18 05:08:09 train.py: 82] Epoch 11, iter 1800/6416, lr 0.010000, loss 1.525630
+INFO 2021-09-18 05:11:05 train.py: 82] Epoch 11, iter 2000/6416, lr 0.010000, loss 1.495522
+INFO 2021-09-18 05:14:01 train.py: 82] Epoch 11, iter 2200/6416, lr 0.010000, loss 1.504253
+INFO 2021-09-18 05:16:57 train.py: 82] Epoch 11, iter 2400/6416, lr 0.010000, loss 1.499877
+INFO 2021-09-18 05:19:53 train.py: 82] Epoch 11, iter 2600/6416, lr 0.010000, loss 1.518339
+INFO 2021-09-18 05:22:49 train.py: 82] Epoch 11, iter 2800/6416, lr 0.010000, loss 1.499026
+INFO 2021-09-18 05:25:47 train.py: 95] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-09-18 05:25:48 train.py: 82] Epoch 11, iter 3000/6416, lr 0.010000, loss 1.512318
+INFO 2021-09-18 05:28:44 train.py: 82] Epoch 11, iter 3200/6416, lr 0.010000, loss 1.502435
+INFO 2021-09-18 05:31:41 train.py: 82] Epoch 11, iter 3400/6416, lr 0.010000, loss 1.497613
+INFO 2021-09-18 05:34:37 train.py: 82] Epoch 11, iter 3600/6416, lr 0.010000, loss 1.498958
+INFO 2021-09-18 05:37:33 train.py: 82] Epoch 11, iter 3800/6416, lr 0.010000, loss 1.491821
+INFO 2021-09-18 05:40:29 train.py: 82] Epoch 11, iter 4000/6416, lr 0.010000, loss 1.493014
+INFO 2021-09-18 05:43:26 train.py: 82] Epoch 11, iter 4200/6416, lr 0.010000, loss 1.475280
+INFO 2021-09-18 05:46:22 train.py: 82] Epoch 11, iter 4400/6416, lr 0.010000, loss 1.490949
+INFO 2021-09-18 05:49:19 train.py: 82] Epoch 11, iter 4600/6416, lr 0.010000, loss 1.487287
+INFO 2021-09-18 05:52:15 train.py: 82] Epoch 11, iter 4800/6416, lr 0.010000, loss 1.496570
+INFO 2021-09-18 05:55:12 train.py: 82] Epoch 11, iter 5000/6416, lr 0.010000, loss 1.483571
+INFO 2021-09-18 05:58:08 train.py: 82] Epoch 11, iter 5200/6416, lr 0.010000, loss 1.488689
+INFO 2021-09-18 06:01:05 train.py: 82] Epoch 11, iter 5400/6416, lr 0.010000, loss 1.480259
+INFO 2021-09-18 06:04:01 train.py: 82] Epoch 11, iter 5600/6416, lr 0.010000, loss 1.481583
+INFO 2021-09-18 06:06:58 train.py: 82] Epoch 11, iter 5800/6416, lr 0.010000, loss 1.482504
+INFO 2021-09-18 06:09:56 train.py: 95] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-09-18 06:09:57 train.py: 82] Epoch 11, iter 6000/6416, lr 0.010000, loss 1.474071
+INFO 2021-09-18 06:12:54 train.py: 82] Epoch 11, iter 6200/6416, lr 0.010000, loss 1.477157
+INFO 2021-09-18 06:15:51 train.py: 82] Epoch 11, iter 6400/6416, lr 0.010000, loss 1.473180
+INFO 2021-09-18 06:16:06 train.py: 100] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-09-18 06:16:09 train.py: 82] Epoch 12, iter 0/6416, lr 0.010000, loss 1.470346
+INFO 2021-09-18 06:19:05 train.py: 82] Epoch 12, iter 200/6416, lr 0.010000, loss 1.234924
+INFO 2021-09-18 06:22:02 train.py: 82] Epoch 12, iter 400/6416, lr 0.010000, loss 1.211312
+INFO 2021-09-18 06:24:58 train.py: 82] Epoch 12, iter 600/6416, lr 0.010000, loss 1.212760
+INFO 2021-09-18 06:27:54 train.py: 82] Epoch 12, iter 800/6416, lr 0.010000, loss 1.234937
+INFO 2021-09-18 06:30:50 train.py: 82] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.235776
+INFO 2021-09-18 06:33:46 train.py: 82] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.224307
+INFO 2021-09-18 06:36:42 train.py: 82] Epoch 12, iter 1400/6416, lr 0.010000, loss 1.244601
+INFO 2021-09-18 06:39:38 train.py: 82] Epoch 12, iter 1600/6416, lr 0.010000, loss 1.250141
+INFO 2021-09-18 06:42:34 train.py: 82] Epoch 12, iter 1800/6416, lr 0.010000, loss 1.232014
+INFO 2021-09-18 06:45:30 train.py: 82] Epoch 12, iter 2000/6416, lr 0.010000, loss 1.257584
+INFO 2021-09-18 06:48:26 train.py: 82] Epoch 12, iter 2200/6416, lr 0.010000, loss 1.261288
+INFO 2021-09-18 06:51:22 train.py: 82] Epoch 12, iter 2400/6416, lr 0.010000, loss 1.254686
+INFO 2021-09-18 06:54:18 train.py: 82] Epoch 12, iter 2600/6416, lr 0.010000, loss 1.261828
+INFO 2021-09-18 06:57:14 train.py: 82] Epoch 12, iter 2800/6416, lr 0.010000, loss 1.264168
+INFO 2021-09-18 07:00:12 train.py: 95] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-09-18 07:00:13 train.py: 82] Epoch 12, iter 3000/6416, lr 0.010000, loss 1.256683
+INFO 2021-09-18 07:03:10 train.py: 82] Epoch 12, iter 3200/6416, lr 0.010000, loss 1.273424
+INFO 2021-09-18 07:06:06 train.py: 82] Epoch 12, iter 3400/6416, lr 0.010000, loss 1.269503
+INFO 2021-09-18 07:09:03 train.py: 82] Epoch 12, iter 3600/6416, lr 0.010000, loss 1.268754
+INFO 2021-09-18 07:11:59 train.py: 82] Epoch 12, iter 3800/6416, lr 0.010000, loss 1.277676
+INFO 2021-09-18 07:14:56 train.py: 82] Epoch 12, iter 4000/6416, lr 0.010000, loss 1.285016
+INFO 2021-09-18 07:17:52 train.py: 82] Epoch 12, iter 4200/6416, lr 0.010000, loss 1.278501
+INFO 2021-09-18 07:20:49 train.py: 82] Epoch 12, iter 4400/6416, lr 0.010000, loss 1.287451
+INFO 2021-09-18 07:23:45 train.py: 82] Epoch 12, iter 4600/6416, lr 0.010000, loss 1.294658
+INFO 2021-09-18 07:26:42 train.py: 82] Epoch 12, iter 4800/6416, lr 0.010000, loss 1.305582
+INFO 2021-09-18 07:29:38 train.py: 82] Epoch 12, iter 5000/6416, lr 0.010000, loss 1.291133
+INFO 2021-09-18 07:32:35 train.py: 82] Epoch 12, iter 5200/6416, lr 0.010000, loss 1.294039
+INFO 2021-09-18 07:35:31 train.py: 82] Epoch 12, iter 5400/6416, lr 0.010000, loss 1.302877
+INFO 2021-09-18 07:38:28 train.py: 82] Epoch 12, iter 5600/6416, lr 0.010000, loss 1.310678
+INFO 2021-09-18 07:41:24 train.py: 82] Epoch 12, iter 5800/6416, lr 0.010000, loss 1.314489
+INFO 2021-09-18 07:44:23 train.py: 95] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-09-18 07:44:23 train.py: 82] Epoch 12, iter 6000/6416, lr 0.010000, loss 1.309482
+INFO 2021-09-18 07:47:20 train.py: 82] Epoch 12, iter 6200/6416, lr 0.010000, loss 1.311983
+INFO 2021-09-18 07:50:16 train.py: 82] Epoch 12, iter 6400/6416, lr 0.010000, loss 1.313067
+INFO 2021-09-18 07:50:31 train.py: 100] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-09-18 07:50:34 train.py: 82] Epoch 13, iter 0/6416, lr 0.001000, loss 1.257330
+INFO 2021-09-18 07:53:30 train.py: 82] Epoch 13, iter 200/6416, lr 0.001000, loss 1.028950
+INFO 2021-09-18 07:56:27 train.py: 82] Epoch 13, iter 400/6416, lr 0.001000, loss 1.000358
+INFO 2021-09-18 07:59:23 train.py: 82] Epoch 13, iter 600/6416, lr 0.001000, loss 1.003336
+INFO 2021-09-18 08:02:19 train.py: 82] Epoch 13, iter 800/6416, lr 0.001000, loss 1.005784
+INFO 2021-09-18 08:05:14 train.py: 82] Epoch 13, iter 1000/6416, lr 0.001000, loss 0.987267
+INFO 2021-09-18 08:08:10 train.py: 82] Epoch 13, iter 1200/6416, lr 0.001000, loss 0.983041
+INFO 2021-09-18 08:11:06 train.py: 82] Epoch 13, iter 1400/6416, lr 0.001000, loss 0.977014
+INFO 2021-09-18 08:14:02 train.py: 82] Epoch 13, iter 1600/6416, lr 0.001000, loss 0.991434
+INFO 2021-09-18 08:16:58 train.py: 82] Epoch 13, iter 1800/6416, lr 0.001000, loss 0.971608
+INFO 2021-09-18 08:19:54 train.py: 82] Epoch 13, iter 2000/6416, lr 0.001000, loss 0.988868
+INFO 2021-09-18 08:22:50 train.py: 82] Epoch 13, iter 2200/6416, lr 0.001000, loss 0.993155
+INFO 2021-09-18 08:25:46 train.py: 82] Epoch 13, iter 2400/6416, lr 0.001000, loss 0.968953
+INFO 2021-09-18 08:28:42 train.py: 82] Epoch 13, iter 2600/6416, lr 0.001000, loss 0.986511
+INFO 2021-09-18 08:31:38 train.py: 82] Epoch 13, iter 2800/6416, lr 0.001000, loss 0.980281
+INFO 2021-09-18 08:34:36 train.py: 95] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-09-18 08:34:37 train.py: 82] Epoch 13, iter 3000/6416, lr 0.001000, loss 0.983388
+INFO 2021-09-18 08:37:33 train.py: 82] Epoch 13, iter 3200/6416, lr 0.001000, loss 0.988416
+INFO 2021-09-18 08:40:30 train.py: 82] Epoch 13, iter 3400/6416, lr 0.001000, loss 0.973327
+INFO 2021-09-18 08:43:26 train.py: 82] Epoch 13, iter 3600/6416, lr 0.001000, loss 0.982067
+INFO 2021-09-18 08:46:22 train.py: 82] Epoch 13, iter 3800/6416, lr 0.001000, loss 0.976763
+INFO 2021-09-18 08:49:18 train.py: 82] Epoch 13, iter 4000/6416, lr 0.001000, loss 0.973787
+INFO 2021-09-18 08:52:15 train.py: 82] Epoch 13, iter 4200/6416, lr 0.001000, loss 0.982412
+INFO 2021-09-18 08:55:11 train.py: 82] Epoch 13, iter 4400/6416, lr 0.001000, loss 0.981923
+INFO 2021-09-18 08:58:08 train.py: 82] Epoch 13, iter 4600/6416, lr 0.001000, loss 0.980630
+INFO 2021-09-18 09:01:04 train.py: 82] Epoch 13, iter 4800/6416, lr 0.001000, loss 0.975949
+INFO 2021-09-18 09:04:01 train.py: 82] Epoch 13, iter 5000/6416, lr 0.001000, loss 0.979984
+INFO 2021-09-18 09:06:57 train.py: 82] Epoch 13, iter 5200/6416, lr 0.001000, loss 0.983873
+INFO 2021-09-18 09:09:53 train.py: 82] Epoch 13, iter 5400/6416, lr 0.001000, loss 0.972562
+INFO 2021-09-18 09:12:50 train.py: 82] Epoch 13, iter 5600/6416, lr 0.001000, loss 0.981552
+INFO 2021-09-18 09:15:46 train.py: 82] Epoch 13, iter 5800/6416, lr 0.001000, loss 0.968218
+INFO 2021-09-18 09:18:44 train.py: 95] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-09-18 09:18:45 train.py: 82] Epoch 13, iter 6000/6416, lr 0.001000, loss 0.977834
+INFO 2021-09-18 09:21:41 train.py: 82] Epoch 13, iter 6200/6416, lr 0.001000, loss 0.978507
+INFO 2021-09-18 09:24:38 train.py: 82] Epoch 13, iter 6400/6416, lr 0.001000, loss 0.976699
+INFO 2021-09-18 09:24:54 train.py: 100] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-09-18 09:24:56 train.py: 82] Epoch 14, iter 0/6416, lr 0.001000, loss 0.966510
+INFO 2021-09-18 09:27:52 train.py: 82] Epoch 14, iter 200/6416, lr 0.001000, loss 0.956110
+INFO 2021-09-18 09:30:49 train.py: 82] Epoch 14, iter 400/6416, lr 0.001000, loss 0.946157
+INFO 2021-09-18 09:33:45 train.py: 82] Epoch 14, iter 600/6416, lr 0.001000, loss 0.938465
+INFO 2021-09-18 09:36:41 train.py: 82] Epoch 14, iter 800/6416, lr 0.001000, loss 0.943507
+INFO 2021-09-18 09:39:37 train.py: 82] Epoch 14, iter 1000/6416, lr 0.001000, loss 0.942004
+INFO 2021-09-18 09:42:33 train.py: 82] Epoch 14, iter 1200/6416, lr 0.001000, loss 0.943810
+INFO 2021-09-18 09:45:29 train.py: 82] Epoch 14, iter 1400/6416, lr 0.001000, loss 0.941365
+INFO 2021-09-18 09:48:25 train.py: 82] Epoch 14, iter 1600/6416, lr 0.001000, loss 0.953532
+INFO 2021-09-18 09:51:21 train.py: 82] Epoch 14, iter 1800/6416, lr 0.001000, loss 0.942934
+INFO 2021-09-18 09:54:17 train.py: 82] Epoch 14, iter 2000/6416, lr 0.001000, loss 0.948742
+INFO 2021-09-18 09:57:13 train.py: 82] Epoch 14, iter 2200/6416, lr 0.001000, loss 0.948447
+INFO 2021-09-18 10:00:10 train.py: 82] Epoch 14, iter 2400/6416, lr 0.001000, loss 0.946558
+INFO 2021-09-18 10:03:06 train.py: 82] Epoch 14, iter 2600/6416, lr 0.001000, loss 0.953501
+INFO 2021-09-18 10:06:02 train.py: 82] Epoch 14, iter 2800/6416, lr 0.001000, loss 0.945308
+INFO 2021-09-18 10:09:00 train.py: 95] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-09-18 10:09:01 train.py: 82] Epoch 14, iter 3000/6416, lr 0.001000, loss 0.949884
+INFO 2021-09-18 10:11:57 train.py: 82] Epoch 14, iter 3200/6416, lr 0.001000, loss 0.950832
+INFO 2021-09-18 10:14:53 train.py: 82] Epoch 14, iter 3400/6416, lr 0.001000, loss 0.945578
+INFO 2021-09-18 10:17:50 train.py: 82] Epoch 14, iter 3600/6416, lr 0.001000, loss 0.949999
+INFO 2021-09-18 10:20:46 train.py: 82] Epoch 14, iter 3800/6416, lr 0.001000, loss 0.939212
+INFO 2021-09-18 10:23:42 train.py: 82] Epoch 14, iter 4000/6416, lr 0.001000, loss 0.958105
+INFO 2021-09-18 10:26:39 train.py: 82] Epoch 14, iter 4200/6416, lr 0.001000, loss 0.955731
+INFO 2021-09-18 10:29:35 train.py: 82] Epoch 14, iter 4400/6416, lr 0.001000, loss 0.941838
+INFO 2021-09-18 10:32:31 train.py: 82] Epoch 14, iter 4600/6416, lr 0.001000, loss 0.952432
+INFO 2021-09-18 10:35:28 train.py: 82] Epoch 14, iter 4800/6416, lr 0.001000, loss 0.955755
+INFO 2021-09-18 10:38:24 train.py: 82] Epoch 14, iter 5000/6416, lr 0.001000, loss 0.942293
+INFO 2021-09-18 10:41:20 train.py: 82] Epoch 14, iter 5200/6416, lr 0.001000, loss 0.956310
+INFO 2021-09-18 10:44:16 train.py: 82] Epoch 14, iter 5400/6416, lr 0.001000, loss 0.957801
+INFO 2021-09-18 10:47:12 train.py: 82] Epoch 14, iter 5600/6416, lr 0.001000, loss 0.952182
+INFO 2021-09-18 10:50:09 train.py: 82] Epoch 14, iter 5800/6416, lr 0.001000, loss 0.951778
+INFO 2021-09-18 10:53:07 train.py: 95] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-09-18 10:53:08 train.py: 82] Epoch 14, iter 6000/6416, lr 0.001000, loss 0.949939
+INFO 2021-09-18 10:56:04 train.py: 82] Epoch 14, iter 6200/6416, lr 0.001000, loss 0.950231
+INFO 2021-09-18 10:59:00 train.py: 82] Epoch 14, iter 6400/6416, lr 0.001000, loss 0.955365
+INFO 2021-09-18 10:59:16 train.py: 100] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-09-18 10:59:18 train.py: 82] Epoch 15, iter 0/6416, lr 0.001000, loss 0.917994
+INFO 2021-09-18 11:02:14 train.py: 82] Epoch 15, iter 200/6416, lr 0.001000, loss 0.913979
+INFO 2021-09-18 11:05:11 train.py: 82] Epoch 15, iter 400/6416, lr 0.001000, loss 0.920250
+INFO 2021-09-18 11:08:07 train.py: 82] Epoch 15, iter 600/6416, lr 0.001000, loss 0.911822
+INFO 2021-09-18 11:11:04 train.py: 82] Epoch 15, iter 800/6416, lr 0.001000, loss 0.915877
+INFO 2021-09-18 11:14:00 train.py: 82] Epoch 15, iter 1000/6416, lr 0.001000, loss 0.917582
+INFO 2021-09-18 11:16:56 train.py: 82] Epoch 15, iter 1200/6416, lr 0.001000, loss 0.933239
+INFO 2021-09-18 11:19:52 train.py: 82] Epoch 15, iter 1400/6416, lr 0.001000, loss 0.913189
+INFO 2021-09-18 11:22:47 train.py: 82] Epoch 15, iter 1600/6416, lr 0.001000, loss 0.923403
+INFO 2021-09-18 11:25:43 train.py: 82] Epoch 15, iter 1800/6416, lr 0.001000, loss 0.923825
+INFO 2021-09-18 11:28:39 train.py: 82] Epoch 15, iter 2000/6416, lr 0.001000, loss 0.916485
+INFO 2021-09-18 11:31:35 train.py: 82] Epoch 15, iter 2200/6416, lr 0.001000, loss 0.914219
+INFO 2021-09-18 11:34:31 train.py: 82] Epoch 15, iter 2400/6416, lr 0.001000, loss 0.915130
+INFO 2021-09-18 11:37:27 train.py: 82] Epoch 15, iter 2600/6416, lr 0.001000, loss 0.914441
+INFO 2021-09-18 11:40:23 train.py: 82] Epoch 15, iter 2800/6416, lr 0.001000, loss 0.902331
+INFO 2021-09-18 11:43:21 train.py: 95] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-09-18 11:43:22 train.py: 82] Epoch 15, iter 3000/6416, lr 0.001000, loss 0.911609
+INFO 2021-09-18 11:46:18 train.py: 82] Epoch 15, iter 3200/6416, lr 0.001000, loss 0.916117
+INFO 2021-09-18 11:49:14 train.py: 82] Epoch 15, iter 3400/6416, lr 0.001000, loss 0.905673
+INFO 2021-09-18 11:52:10 train.py: 82] Epoch 15, iter 3600/6416, lr 0.001000, loss 0.929923
+INFO 2021-09-18 11:55:07 train.py: 82] Epoch 15, iter 3800/6416, lr 0.001000, loss 0.925836
+INFO 2021-09-18 11:58:03 train.py: 82] Epoch 15, iter 4000/6416, lr 0.001000, loss 0.928902
+INFO 2021-09-18 12:00:59 train.py: 82] Epoch 15, iter 4200/6416, lr 0.001000, loss 0.923974
+INFO 2021-09-18 12:03:55 train.py: 82] Epoch 15, iter 4400/6416, lr 0.001000, loss 0.926927
+INFO 2021-09-18 12:06:52 train.py: 82] Epoch 15, iter 4600/6416, lr 0.001000, loss 0.932579
+INFO 2021-09-18 12:09:48 train.py: 82] Epoch 15, iter 4800/6416, lr 0.001000, loss 0.931709
+INFO 2021-09-18 12:12:44 train.py: 82] Epoch 15, iter 5000/6416, lr 0.001000, loss 0.939475
+INFO 2021-09-18 12:15:40 train.py: 82] Epoch 15, iter 5200/6416, lr 0.001000, loss 0.937369
+INFO 2021-09-18 12:18:37 train.py: 82] Epoch 15, iter 5400/6416, lr 0.001000, loss 0.939057
+INFO 2021-09-18 12:21:33 train.py: 82] Epoch 15, iter 5600/6416, lr 0.001000, loss 0.917767
+INFO 2021-09-18 12:24:29 train.py: 82] Epoch 15, iter 5800/6416, lr 0.001000, loss 0.929093
+INFO 2021-09-18 12:27:27 train.py: 95] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-09-18 12:27:28 train.py: 82] Epoch 15, iter 6000/6416, lr 0.001000, loss 0.921136
+INFO 2021-09-18 12:30:24 train.py: 82] Epoch 15, iter 6200/6416, lr 0.001000, loss 0.925819
+INFO 2021-09-18 12:33:20 train.py: 82] Epoch 15, iter 6400/6416, lr 0.001000, loss 0.922700
+INFO 2021-09-18 12:33:36 train.py: 100] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-09-18 12:33:38 train.py: 82] Epoch 16, iter 0/6416, lr 0.000100, loss 0.926071
+INFO 2021-09-18 12:36:35 train.py: 82] Epoch 16, iter 200/6416, lr 0.000100, loss 0.881997
+INFO 2021-09-18 12:39:31 train.py: 82] Epoch 16, iter 400/6416, lr 0.000100, loss 0.877790
+INFO 2021-09-18 12:42:27 train.py: 82] Epoch 16, iter 600/6416, lr 0.000100, loss 0.889376
+INFO 2021-09-18 12:45:23 train.py: 82] Epoch 16, iter 800/6416, lr 0.000100, loss 0.888077
+INFO 2021-09-18 12:48:20 train.py: 82] Epoch 16, iter 1000/6416, lr 0.000100, loss 0.883947
+INFO 2021-09-18 12:51:16 train.py: 82] Epoch 16, iter 1200/6416, lr 0.000100, loss 0.889155
+INFO 2021-09-18 12:54:12 train.py: 82] Epoch 16, iter 1400/6416, lr 0.000100, loss 0.886565
+INFO 2021-09-18 12:57:09 train.py: 82] Epoch 16, iter 1600/6416, lr 0.000100, loss 0.881857
+INFO 2021-09-18 13:00:05 train.py: 82] Epoch 16, iter 1800/6416, lr 0.000100, loss 0.891921
+INFO 2021-09-18 13:03:01 train.py: 82] Epoch 16, iter 2000/6416, lr 0.000100, loss 0.893977
+INFO 2021-09-18 13:05:58 train.py: 82] Epoch 16, iter 2200/6416, lr 0.000100, loss 0.891510
+INFO 2021-09-18 13:08:54 train.py: 82] Epoch 16, iter 2400/6416, lr 0.000100, loss 0.885443
+INFO 2021-09-18 13:11:50 train.py: 82] Epoch 16, iter 2600/6416, lr 0.000100, loss 0.885629
+INFO 2021-09-18 13:14:47 train.py: 82] Epoch 16, iter 2800/6416, lr 0.000100, loss 0.887539
+INFO 2021-09-18 13:17:45 train.py: 95] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-09-18 13:17:46 train.py: 82] Epoch 16, iter 3000/6416, lr 0.000100, loss 0.894704
+INFO 2021-09-18 13:20:42 train.py: 82] Epoch 16, iter 3200/6416, lr 0.000100, loss 0.878147
+INFO 2021-09-18 13:23:39 train.py: 82] Epoch 16, iter 3400/6416, lr 0.000100, loss 0.887074
+INFO 2021-09-18 13:26:35 train.py: 82] Epoch 16, iter 3600/6416, lr 0.000100, loss 0.893055
+INFO 2021-09-18 13:29:32 train.py: 82] Epoch 16, iter 3800/6416, lr 0.000100, loss 0.887524
+INFO 2021-09-18 13:32:28 train.py: 82] Epoch 16, iter 4000/6416, lr 0.000100, loss 0.889063
+INFO 2021-09-18 13:35:25 train.py: 82] Epoch 16, iter 4200/6416, lr 0.000100, loss 0.882367
+INFO 2021-09-18 13:38:21 train.py: 82] Epoch 16, iter 4400/6416, lr 0.000100, loss 0.889608
+INFO 2021-09-18 13:41:17 train.py: 82] Epoch 16, iter 4600/6416, lr 0.000100, loss 0.890188
+INFO 2021-09-18 13:44:14 train.py: 82] Epoch 16, iter 4800/6416, lr 0.000100, loss 0.883644
+INFO 2021-09-18 13:47:14 train.py: 82] Epoch 16, iter 5000/6416, lr 0.000100, loss 0.898921
+INFO 2021-09-18 13:50:20 train.py: 82] Epoch 16, iter 5200/6416, lr 0.000100, loss 0.887187
+INFO 2021-09-18 13:53:22 train.py: 82] Epoch 16, iter 5400/6416, lr 0.000100, loss 0.891199
+INFO 2021-09-18 13:56:23 train.py: 82] Epoch 16, iter 5600/6416, lr 0.000100, loss 0.889662
+INFO 2021-09-18 13:59:27 train.py: 82] Epoch 16, iter 5800/6416, lr 0.000100, loss 0.898723
+INFO 2021-09-18 14:02:35 train.py: 95] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-09-18 14:02:35 train.py: 82] Epoch 16, iter 6000/6416, lr 0.000100, loss 0.892579
+INFO 2021-09-18 14:05:48 train.py: 82] Epoch 16, iter 6200/6416, lr 0.000100, loss 0.898475
+INFO 2021-09-18 14:09:00 train.py: 82] Epoch 16, iter 6400/6416, lr 0.000100, loss 0.885162
+INFO 2021-09-18 14:09:16 train.py: 100] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-09-18 14:09:18 train.py: 82] Epoch 17, iter 0/6416, lr 0.000100, loss 0.890773
+INFO 2021-09-18 14:12:16 train.py: 82] Epoch 17, iter 200/6416, lr 0.000100, loss 0.879062
+INFO 2021-09-18 14:15:12 train.py: 82] Epoch 17, iter 400/6416, lr 0.000100, loss 0.881150
+INFO 2021-09-18 14:18:09 train.py: 82] Epoch 17, iter 600/6416, lr 0.000100, loss 0.883788
+INFO 2021-09-18 14:21:05 train.py: 82] Epoch 17, iter 800/6416, lr 0.000100, loss 0.886351
+INFO 2021-09-18 14:24:01 train.py: 82] Epoch 17, iter 1000/6416, lr 0.000100, loss 0.888901
+INFO 2021-09-18 14:26:57 train.py: 82] Epoch 17, iter 1200/6416, lr 0.000100, loss 0.891962
+INFO 2021-09-18 14:29:52 train.py: 82] Epoch 17, iter 1400/6416, lr 0.000100, loss 0.878181
+INFO 2021-09-18 14:32:48 train.py: 82] Epoch 17, iter 1600/6416, lr 0.000100, loss 0.880641
+INFO 2021-09-18 14:35:44 train.py: 82] Epoch 17, iter 1800/6416, lr 0.000100, loss 0.884812
+INFO 2021-09-18 14:38:40 train.py: 82] Epoch 17, iter 2000/6416, lr 0.000100, loss 0.892285
+INFO 2021-09-18 14:41:36 train.py: 82] Epoch 17, iter 2200/6416, lr 0.000100, loss 0.883714
+INFO 2021-09-18 14:44:32 train.py: 82] Epoch 17, iter 2400/6416, lr 0.000100, loss 0.885584
+INFO 2021-09-18 14:47:28 train.py: 82] Epoch 17, iter 2600/6416, lr 0.000100, loss 0.892373
+INFO 2021-09-18 14:50:24 train.py: 82] Epoch 17, iter 2800/6416, lr 0.000100, loss 0.888797
+INFO 2021-09-18 14:53:22 train.py: 95] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-09-18 14:53:22 train.py: 82] Epoch 17, iter 3000/6416, lr 0.000100, loss 0.894852
+INFO 2021-09-18 14:56:20 train.py: 82] Epoch 17, iter 3200/6416, lr 0.000100, loss 0.878462
+INFO 2021-09-18 14:59:17 train.py: 82] Epoch 17, iter 3400/6416, lr 0.000100, loss 0.877219
+INFO 2021-09-18 15:02:18 train.py: 82] Epoch 17, iter 3600/6416, lr 0.000100, loss 0.878370
+INFO 2021-09-18 15:05:23 train.py: 82] Epoch 17, iter 3800/6416, lr 0.000100, loss 0.887341
+INFO 2021-09-18 15:08:35 train.py: 82] Epoch 17, iter 4000/6416, lr 0.000100, loss 0.888016
+INFO 2021-09-18 15:11:47 train.py: 82] Epoch 17, iter 4200/6416, lr 0.000100, loss 0.892289
+INFO 2021-09-18 15:15:17 train.py: 82] Epoch 17, iter 4400/6416, lr 0.000100, loss 0.876705
+INFO 2021-09-18 15:18:27 train.py: 82] Epoch 17, iter 4600/6416, lr 0.000100, loss 0.902008
+INFO 2021-09-18 15:21:41 train.py: 82] Epoch 17, iter 4800/6416, lr 0.000100, loss 0.878303
+INFO 2021-09-18 15:24:59 train.py: 82] Epoch 17, iter 5000/6416, lr 0.000100, loss 0.886773
+INFO 2021-09-18 15:28:24 train.py: 82] Epoch 17, iter 5200/6416, lr 0.000100, loss 0.879008
+INFO 2021-09-18 15:31:54 train.py: 82] Epoch 17, iter 5400/6416, lr 0.000100, loss 0.886785
+INFO 2021-09-18 15:35:20 train.py: 82] Epoch 17, iter 5600/6416, lr 0.000100, loss 0.885537
+INFO 2021-09-18 15:38:48 train.py: 82] Epoch 17, iter 5800/6416, lr 0.000100, loss 0.891631
+INFO 2021-09-18 15:42:29 train.py: 95] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-09-18 15:42:30 train.py: 82] Epoch 17, iter 6000/6416, lr 0.000100, loss 0.874014
+INFO 2021-09-18 15:46:05 train.py: 82] Epoch 17, iter 6200/6416, lr 0.000100, loss 0.877176
+INFO 2021-09-18 15:49:32 train.py: 82] Epoch 17, iter 6400/6416, lr 0.000100, loss 0.884876
+INFO 2021-09-18 15:49:52 train.py: 100] Save checkpoint Epoch_17.pt to disk...
+INFO 2021-09-18 15:49:52 train.py: 183] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/ResNeSt50/.gitkeep b/bob/bio/facexzoo/models/backbones/ResNeSt50/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_agedb30.txt b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_agedb30.txt
new file mode 100644
index 0000000000000000000000000000000000000000..2ea7922b4124ebf5ead002b3d00edfb0173fd5fc
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_agedb30.txt
@@ -0,0 +1,34 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.9798333333333333 | 0.002269633135090605 |
+| Epoch_13_batch_5999.pt | 0.9796666666666667 | 0.0024062675364119645 |
+| Epoch_17_batch_5999.pt | 0.9796666666666667 | 0.002380476142847607 |
+| Epoch_14_batch_5999.pt | 0.9795 | 0.0023966282900136616 |
+| Epoch_17_batch_2999.pt | 0.9793333333333333 | 0.0022525705480792506 |
+| Epoch_16.pt | 0.9791666666666666 | 0.002450119674577765 |
+| Epoch_16_batch_5999.pt | 0.9791666666666666 | 0.002239516041194037 |
+| Epoch_13_batch_2999.pt | 0.9791666666666666 | 0.002560984571470238 |
+| Epoch_11_batch_5999.pt | 0.9789999999999999 | 0.002333333333333326 |
+| Epoch_15_batch_5999.pt | 0.9788333333333334 | 0.002166666666666659 |
+| Epoch_15_batch_2999.pt | 0.9788333333333332 | 0.0025706078447242835 |
+| Epoch_10.pt | 0.9785000000000001 | 0.002349809553617398 |
+| Epoch_15.pt | 0.9783333333333333 | 0.0025215123817578225 |
+| Epoch_16_batch_2999.pt | 0.9781666666666666 | 0.0025873624493766684 |
+| Epoch_17.pt | 0.9778333333333334 | 0.002582586510926545 |
+| Epoch_12.pt | 0.9778333333333334 | 0.0025706078447242787 |
+| Epoch_10_batch_2999.pt | 0.9776666666666667 | 0.0021970799925872387 |
+| Epoch_12_batch_2999.pt | 0.9776666666666667 | 0.0023067266102251897 |
+| Epoch_14.pt | 0.9776666666666666 | 0.002560381915956203 |
+| Epoch_13.pt | 0.9771666666666666 | 0.0028765763257776913 |
+| Epoch_10_batch_5999.pt | 0.9770000000000001 | 0.0024570382652773313 |
+| Epoch_11_batch_2999.pt | 0.9766666666666668 | 0.002222222222222222 |
+| Epoch_11.pt | 0.9763333333333334 | 0.0024570382652773304 |
+| Epoch_12_batch_5999.pt | 0.9755 | 0.002509857110683663 |
+| Epoch_8_batch_5999.pt | 0.9705 | 0.0017924739783224148 |
+| Epoch_9_batch_5999.pt | 0.9684999999999999 | 0.002575405996999836 |
+| Epoch_9_batch_2999.pt | 0.9684999999999999 | 0.00181642030269687 |
+| Epoch_8_batch_2999.pt | 0.9664999999999999 | 0.0027938424357067037 |
+| Epoch_9.pt | 0.9646666666666667 | 0.0024190601174530328 |
+| Epoch_8.pt | 0.9630000000000001 | 0.003733399470313655 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..2fd9515f67d211f1ffe1793a9843194a8840309d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_calfw.txt
@@ -0,0 +1,34 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9555000000000001 | 0.0037437190197403256 |
+| Epoch_17_batch_2999.pt | 0.9551666666666667 | 0.003722222222222224 |
+| Epoch_12.pt | 0.9546666666666667 | 0.00391104797928845 |
+| Epoch_10.pt | 0.9546666666666667 | 0.0039503086299180435 |
+| Epoch_17.pt | 0.9545000000000001 | 0.0035490391674854334 |
+| Epoch_14_batch_2999.pt | 0.9545000000000001 | 0.003583656316193371 |
+| Epoch_13.pt | 0.9543333333333335 | 0.0038506052113696522 |
+| Epoch_10_batch_2999.pt | 0.9543333333333335 | 0.0034444444444444483 |
+| Epoch_13_batch_2999.pt | 0.9543333333333335 | 0.0036951753662924232 |
+| Epoch_15_batch_2999.pt | 0.9543333333333333 | 0.003703518513888659 |
+| Epoch_11_batch_2999.pt | 0.9543333333333333 | 0.0037614024177357215 |
+| Epoch_17_batch_5999.pt | 0.9541666666666668 | 0.0037700083505163434 |
+| Epoch_14_batch_5999.pt | 0.9541666666666668 | 0.003593976442141304 |
+| Epoch_11_batch_5999.pt | 0.9541666666666668 | 0.0037122586382862498 |
+| Epoch_16_batch_2999.pt | 0.9541666666666668 | 0.003653596232768307 |
+| Epoch_13_batch_5999.pt | 0.9540000000000001 | 0.0036106837353937623 |
+| Epoch_10_batch_5999.pt | 0.954 | 0.00365317382728302 |
+| Epoch_14.pt | 0.9538333333333334 | 0.0034876324701598946 |
+| Epoch_16_batch_5999.pt | 0.9538333333333334 | 0.0037683706404692493 |
+| Epoch_12_batch_2999.pt | 0.9536666666666667 | 0.0032848323331321023 |
+| Epoch_16.pt | 0.9535 | 0.0035438174297371255 |
+| Epoch_12_batch_5999.pt | 0.9533333333333335 | 0.0037184890068181174 |
+| Epoch_11.pt | 0.9530000000000001 | 0.003572874486847453 |
+| Epoch_15_batch_5999.pt | 0.9530000000000001 | 0.003749897117930266 |
+| Epoch_9_batch_5999.pt | 0.9475 | 0.003728849821011072 |
+| Epoch_9.pt | 0.9473333333333332 | 0.003842581436842354 |
+| Epoch_8_batch_5999.pt | 0.9471666666666667 | 0.0030128326362125 |
+| Epoch_8_batch_2999.pt | 0.9451666666666666 | 0.003224615969095875 |
+| Epoch_9_batch_2999.pt | 0.9445 | 0.0038813418832685676 |
+| Epoch_8.pt | 0.9431666666666667 | 0.0035350973933184486 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f1f56183cb4d2b9a89981a3c0faf8c84029b27c9
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_cplfw.txt
@@ -0,0 +1,34 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.8998333333333333 | 0.004130958098893619 |
+| Epoch_16_batch_5999.pt | 0.8995 | 0.003936611941790411 |
+| Epoch_17_batch_2999.pt | 0.899 | 0.004173697771331771 |
+| Epoch_17_batch_5999.pt | 0.8985000000000001 | 0.004093430448841925 |
+| Epoch_16_batch_2999.pt | 0.8976666666666666 | 0.0038425814368423473 |
+| Epoch_15_batch_5999.pt | 0.8975000000000002 | 0.004562663830983816 |
+| Epoch_16.pt | 0.8975 | 0.004276334532872405 |
+| Epoch_14.pt | 0.8971666666666668 | 0.003735465660892049 |
+| Epoch_14_batch_2999.pt | 0.8969999999999999 | 0.003887301263230197 |
+| Epoch_13_batch_2999.pt | 0.8968333333333334 | 0.004219662967057385 |
+| Epoch_13.pt | 0.8968333333333331 | 0.004651098370211355 |
+| Epoch_14_batch_5999.pt | 0.8966666666666668 | 0.004360314860077465 |
+| Epoch_13_batch_5999.pt | 0.8963333333333333 | 0.004732342097565968 |
+| Epoch_15.pt | 0.8963333333333333 | 0.004178132372658477 |
+| Epoch_17.pt | 0.8950000000000001 | 0.004374448818895448 |
+| Epoch_12_batch_5999.pt | 0.8948333333333334 | 0.0040326598764354425 |
+| Epoch_11_batch_2999.pt | 0.8933333333333333 | 0.004667989230577778 |
+| Epoch_11.pt | 0.8923333333333334 | 0.005584369721390699 |
+| Epoch_12.pt | 0.8908333333333334 | 0.004715681745091333 |
+| Epoch_12_batch_2999.pt | 0.8908333333333334 | 0.004844813951249548 |
+| Epoch_10.pt | 0.8901666666666666 | 0.004703885969055433 |
+| Epoch_10_batch_2999.pt | 0.8883333333333333 | 0.004962824763139404 |
+| Epoch_10_batch_5999.pt | 0.8876666666666665 | 0.0049888765156985895 |
+| Epoch_11_batch_5999.pt | 0.8861666666666667 | 0.005488765180792174 |
+| Epoch_8_batch_5999.pt | 0.8633333333333335 | 0.005725188012439223 |
+| Epoch_9_batch_2999.pt | 0.8633333333333333 | 0.005061351988413492 |
+| Epoch_8_batch_2999.pt | 0.8591666666666666 | 0.006723140433157004 |
+| Epoch_9_batch_5999.pt | 0.8563333333333333 | 0.006951329919829802 |
+| Epoch_8.pt | 0.8543333333333335 | 0.005753152651555383 |
+| Epoch_9.pt | 0.8446666666666667 | 0.007608474807008882 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..97b0a119408a8f9bf2d5e42d5f51eac4f8e6a6c5
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_lfw.txt
@@ -0,0 +1,34 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.998 | 0.000544331053951814 |
+| Epoch_14.pt | 0.998 | 0.0006478835438716985 |
+| Epoch_12_batch_5999.pt | 0.998 | 0.0006478835438716985 |
+| Epoch_15.pt | 0.998 | 0.0006478835438716985 |
+| Epoch_13.pt | 0.998 | 0.0006478835438716985 |
+| Epoch_15_batch_2999.pt | 0.998 | 0.0006478835438716985 |
+| Epoch_13_batch_5999.pt | 0.9978333333333333 | 0.0007049209744694171 |
+| Epoch_16.pt | 0.9976666666666667 | 0.0007934920476158739 |
+| Epoch_14_batch_2999.pt | 0.9976666666666667 | 0.0007934920476158739 |
+| Epoch_17_batch_5999.pt | 0.9976666666666667 | 0.0007934920476158739 |
+| Epoch_11.pt | 0.9976666666666667 | 0.0007114582486036506 |
+| Epoch_13_batch_2999.pt | 0.9976666666666667 | 0.0007934920476158739 |
+| Epoch_10_batch_5999.pt | 0.9974999999999999 | 0.0007556372504853036 |
+| Epoch_14_batch_5999.pt | 0.9974999999999999 | 0.0008333333333333352 |
+| Epoch_15_batch_5999.pt | 0.9974999999999999 | 0.0008333333333333352 |
+| Epoch_12_batch_2999.pt | 0.9974999999999999 | 0.0007954345035153545 |
+| Epoch_8_batch_2999.pt | 0.9973333333333333 | 0.0009362388636862596 |
+| Epoch_10_batch_2999.pt | 0.9973333333333333 | 0.0008678055195451802 |
+| Epoch_16_batch_5999.pt | 0.9971666666666665 | 0.0007876359377087692 |
+| Epoch_8_batch_5999.pt | 0.9971666666666665 | 0.00086245414979222 |
+| Epoch_11_batch_5999.pt | 0.9971666666666665 | 0.0007474235581707618 |
+| Epoch_16_batch_2999.pt | 0.9971666666666665 | 0.0007876359377087692 |
+| Epoch_11_batch_2999.pt | 0.9971666666666665 | 0.0008624541497922263 |
+| Epoch_17_batch_2999.pt | 0.9971666666666665 | 0.0007876359377087692 |
+| Epoch_9_batch_5999.pt | 0.9970000000000001 | 0.0008534606386520699 |
+| Epoch_12.pt | 0.9969999999999999 | 0.0008888888888888863 |
+| Epoch_9.pt | 0.9969999999999999 | 0.0006478835438716991 |
+| Epoch_10.pt | 0.9969999999999999 | 0.0007370277311900897 |
+| Epoch_9_batch_2999.pt | 0.9966666666666667 | 0.0007856742013183886 |
+| Epoch_8.pt | 0.9959999999999999 | 0.0011439589045541157 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..804f6f69279c268e1bc9455591bc24ab22d3dca5
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNeSt50/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.9708072431883568 |
+| 2 | 0.979757085020243 |
+| 3 | 0.982731686996368 |
+| 4 | 0.9846518348803 |
+| 5 | 0.985803923610659 |
+| 6 | 0.9866045276436205 |
+| 7 | 0.9872554252313941 |
+| 8 | 0.9878412330603903 |
+| 9 | 0.9882382805889321 |
+| 10 | 0.9886808909486181 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNeSt50/log.log b/bob/bio/facexzoo/models/backbones/ResNeSt50/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..09ce751b93e338c2ecbe5443dd0717abdf30f426
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNeSt50/log.log
@@ -0,0 +1,655 @@
+INFO 2021-02-24 17:04:52 train.py: 176] Start optimization.
+INFO 2021-02-24 17:04:52 train.py: 177] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='ResNeSt', batch_size=512, data_root='/home/wangjun492/wj_data/face_database/facex-zoo/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-resnest', train_file='/home/wangjun492/wj_data/face_database/facex-zoo/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f2966c4dda0>)
+backbone param:
+{'depth': 50, 'drop_ratio': 0.4, 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2021-02-24 17:05:17 train.py: 78] Epoch 0, iter 0/6416, lr 0.100000, loss 16.313322
+INFO 2021-02-24 17:10:50 train.py: 78] Epoch 0, iter 200/6416, lr 0.100000, loss 15.620893
+INFO 2021-02-24 17:16:23 train.py: 78] Epoch 0, iter 400/6416, lr 0.100000, loss 15.359382
+INFO 2021-02-24 17:21:56 train.py: 78] Epoch 0, iter 600/6416, lr 0.100000, loss 15.326296
+INFO 2021-02-24 17:27:29 train.py: 78] Epoch 0, iter 800/6416, lr 0.100000, loss 15.290645
+INFO 2021-02-24 17:33:02 train.py: 78] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.263804
+INFO 2021-02-24 17:38:35 train.py: 78] Epoch 0, iter 1200/6416, lr 0.100000, loss 15.189003
+INFO 2021-02-24 17:44:08 train.py: 78] Epoch 0, iter 1400/6416, lr 0.100000, loss 15.069101
+INFO 2021-02-24 17:49:41 train.py: 78] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.929240
+INFO 2021-02-24 17:55:15 train.py: 78] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.731436
+INFO 2021-02-24 18:00:48 train.py: 78] Epoch 0, iter 2000/6416, lr 0.100000, loss 14.503152
+INFO 2021-02-24 18:06:21 train.py: 78] Epoch 0, iter 2200/6416, lr 0.100000, loss 14.254705
+INFO 2021-02-24 18:11:55 train.py: 78] Epoch 0, iter 2400/6416, lr 0.100000, loss 13.988201
+INFO 2021-02-24 18:17:28 train.py: 78] Epoch 0, iter 2600/6416, lr 0.100000, loss 13.714421
+INFO 2021-02-24 18:23:02 train.py: 78] Epoch 0, iter 2800/6416, lr 0.100000, loss 13.426085
+INFO 2021-02-24 18:28:34 train.py: 91] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-02-24 18:28:36 train.py: 78] Epoch 0, iter 3000/6416, lr 0.100000, loss 13.129318
+INFO 2021-02-24 18:34:09 train.py: 78] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.801013
+INFO 2021-02-24 18:39:43 train.py: 78] Epoch 0, iter 3400/6416, lr 0.100000, loss 12.485551
+INFO 2021-02-24 18:45:16 train.py: 78] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.217961
+INFO 2021-02-24 18:50:50 train.py: 78] Epoch 0, iter 3800/6416, lr 0.100000, loss 11.977710
+INFO 2021-02-24 18:56:23 train.py: 78] Epoch 0, iter 4000/6416, lr 0.100000, loss 11.816155
+INFO 2021-02-24 19:01:55 train.py: 78] Epoch 0, iter 4200/6416, lr 0.100000, loss 11.793206
+INFO 2021-02-24 19:07:27 train.py: 78] Epoch 0, iter 4400/6416, lr 0.100000, loss 11.890083
+INFO 2021-02-24 19:12:59 train.py: 78] Epoch 0, iter 4600/6416, lr 0.100000, loss 12.079075
+INFO 2021-02-24 19:18:31 train.py: 78] Epoch 0, iter 4800/6416, lr 0.100000, loss 12.349208
+INFO 2021-02-24 19:24:01 train.py: 78] Epoch 0, iter 5000/6416, lr 0.100000, loss 12.663674
+INFO 2021-02-24 19:29:32 train.py: 78] Epoch 0, iter 5200/6416, lr 0.100000, loss 12.983118
+INFO 2021-02-24 19:35:02 train.py: 78] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.257001
+INFO 2021-02-24 19:40:32 train.py: 78] Epoch 0, iter 5600/6416, lr 0.100000, loss 13.512850
+INFO 2021-02-24 19:46:02 train.py: 78] Epoch 0, iter 5800/6416, lr 0.100000, loss 13.716488
+INFO 2021-02-24 19:51:30 train.py: 91] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-02-24 19:51:32 train.py: 78] Epoch 0, iter 6000/6416, lr 0.100000, loss 13.848291
+INFO 2021-02-24 19:57:01 train.py: 78] Epoch 0, iter 6200/6416, lr 0.100000, loss 13.889544
+INFO 2021-02-24 20:02:30 train.py: 78] Epoch 0, iter 6400/6416, lr 0.100000, loss 13.888664
+INFO 2021-02-24 20:02:54 train.py: 96] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-02-24 20:02:56 train.py: 78] Epoch 1, iter 0/6416, lr 0.100000, loss 13.808955
+INFO 2021-02-24 20:08:25 train.py: 78] Epoch 1, iter 200/6416, lr 0.100000, loss 13.758435
+INFO 2021-02-24 20:13:53 train.py: 78] Epoch 1, iter 400/6416, lr 0.100000, loss 13.620849
+INFO 2021-02-24 20:19:21 train.py: 78] Epoch 1, iter 600/6416, lr 0.100000, loss 13.442502
+INFO 2021-02-24 20:24:49 train.py: 78] Epoch 1, iter 800/6416, lr 0.100000, loss 13.226353
+INFO 2021-02-24 20:30:16 train.py: 78] Epoch 1, iter 1000/6416, lr 0.100000, loss 12.996416
+INFO 2021-02-24 20:35:44 train.py: 78] Epoch 1, iter 1200/6416, lr 0.100000, loss 12.742379
+INFO 2021-02-24 20:41:11 train.py: 78] Epoch 1, iter 1400/6416, lr 0.100000, loss 12.504958
+INFO 2021-02-24 20:46:39 train.py: 78] Epoch 1, iter 1600/6416, lr 0.100000, loss 12.231348
+INFO 2021-02-24 20:52:07 train.py: 78] Epoch 1, iter 1800/6416, lr 0.100000, loss 11.963732
+INFO 2021-02-24 20:57:34 train.py: 78] Epoch 1, iter 2000/6416, lr 0.100000, loss 11.657655
+INFO 2021-02-24 21:03:02 train.py: 78] Epoch 1, iter 2200/6416, lr 0.100000, loss 11.404816
+INFO 2021-02-24 21:08:30 train.py: 78] Epoch 1, iter 2400/6416, lr 0.100000, loss 11.153934
+INFO 2021-02-24 21:13:57 train.py: 78] Epoch 1, iter 2600/6416, lr 0.100000, loss 10.886155
+INFO 2021-02-24 21:19:25 train.py: 78] Epoch 1, iter 2800/6416, lr 0.100000, loss 10.667871
+INFO 2021-02-24 21:24:52 train.py: 91] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-02-24 21:24:53 train.py: 78] Epoch 1, iter 3000/6416, lr 0.100000, loss 10.421691
+INFO 2021-02-24 21:30:21 train.py: 78] Epoch 1, iter 3200/6416, lr 0.100000, loss 10.206822
+INFO 2021-02-24 21:35:49 train.py: 78] Epoch 1, iter 3400/6416, lr 0.100000, loss 9.992142
+INFO 2021-02-24 21:41:16 train.py: 78] Epoch 1, iter 3600/6416, lr 0.100000, loss 9.818099
+INFO 2021-02-24 21:46:44 train.py: 78] Epoch 1, iter 3800/6416, lr 0.100000, loss 9.645288
+INFO 2021-02-24 21:52:12 train.py: 78] Epoch 1, iter 4000/6416, lr 0.100000, loss 9.474552
+INFO 2021-02-24 21:57:39 train.py: 78] Epoch 1, iter 4200/6416, lr 0.100000, loss 9.332040
+INFO 2021-02-24 22:03:07 train.py: 78] Epoch 1, iter 4400/6416, lr 0.100000, loss 9.165296
+INFO 2021-02-24 22:08:35 train.py: 78] Epoch 1, iter 4600/6416, lr 0.100000, loss 9.037102
+INFO 2021-02-24 22:14:03 train.py: 78] Epoch 1, iter 4800/6416, lr 0.100000, loss 8.872455
+INFO 2021-02-24 22:19:31 train.py: 78] Epoch 1, iter 5000/6416, lr 0.100000, loss 8.755043
+INFO 2021-02-24 22:24:59 train.py: 78] Epoch 1, iter 5200/6416, lr 0.100000, loss 8.623330
+INFO 2021-02-24 22:30:26 train.py: 78] Epoch 1, iter 5400/6416, lr 0.100000, loss 8.477971
+INFO 2021-02-24 22:35:54 train.py: 78] Epoch 1, iter 5600/6416, lr 0.100000, loss 8.392763
+INFO 2021-02-24 22:41:22 train.py: 78] Epoch 1, iter 5800/6416, lr 0.100000, loss 8.297782
+INFO 2021-02-24 22:46:49 train.py: 91] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-02-24 22:46:50 train.py: 78] Epoch 1, iter 6000/6416, lr 0.100000, loss 8.201840
+INFO 2021-02-24 22:52:18 train.py: 78] Epoch 1, iter 6200/6416, lr 0.100000, loss 8.099176
+INFO 2021-02-24 22:57:46 train.py: 78] Epoch 1, iter 6400/6416, lr 0.100000, loss 7.983228
+INFO 2021-02-24 22:58:10 train.py: 96] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-02-24 22:58:13 train.py: 78] Epoch 2, iter 0/6416, lr 0.100000, loss 7.863767
+INFO 2021-02-24 23:03:40 train.py: 78] Epoch 2, iter 200/6416, lr 0.100000, loss 7.333290
+INFO 2021-02-24 23:09:07 train.py: 78] Epoch 2, iter 400/6416, lr 0.100000, loss 7.314851
+INFO 2021-02-24 23:14:34 train.py: 78] Epoch 2, iter 600/6416, lr 0.100000, loss 7.328369
+INFO 2021-02-24 23:20:01 train.py: 78] Epoch 2, iter 800/6416, lr 0.100000, loss 7.328868
+INFO 2021-02-24 23:25:27 train.py: 78] Epoch 2, iter 1000/6416, lr 0.100000, loss 7.342762
+INFO 2021-02-24 23:30:54 train.py: 78] Epoch 2, iter 1200/6416, lr 0.100000, loss 7.317062
+INFO 2021-02-24 23:36:21 train.py: 78] Epoch 2, iter 1400/6416, lr 0.100000, loss 7.302503
+INFO 2021-02-24 23:41:48 train.py: 78] Epoch 2, iter 1600/6416, lr 0.100000, loss 7.279334
+INFO 2021-02-24 23:47:15 train.py: 78] Epoch 2, iter 1800/6416, lr 0.100000, loss 7.227969
+INFO 2021-02-24 23:52:43 train.py: 78] Epoch 2, iter 2000/6416, lr 0.100000, loss 7.190882
+INFO 2021-02-24 23:58:10 train.py: 78] Epoch 2, iter 2200/6416, lr 0.100000, loss 7.155687
+INFO 2021-02-25 00:03:37 train.py: 78] Epoch 2, iter 2400/6416, lr 0.100000, loss 7.101299
+INFO 2021-02-25 00:09:05 train.py: 78] Epoch 2, iter 2600/6416, lr 0.100000, loss 7.038573
+INFO 2021-02-25 00:14:32 train.py: 78] Epoch 2, iter 2800/6416, lr 0.100000, loss 6.990323
+INFO 2021-02-25 00:19:58 train.py: 91] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-02-25 00:20:00 train.py: 78] Epoch 2, iter 3000/6416, lr 0.100000, loss 6.954301
+INFO 2021-02-25 00:25:28 train.py: 78] Epoch 2, iter 3200/6416, lr 0.100000, loss 6.909274
+INFO 2021-02-25 00:30:55 train.py: 78] Epoch 2, iter 3400/6416, lr 0.100000, loss 6.849148
+INFO 2021-02-25 00:36:23 train.py: 78] Epoch 2, iter 3600/6416, lr 0.100000, loss 6.829742
+INFO 2021-02-25 00:41:51 train.py: 78] Epoch 2, iter 3800/6416, lr 0.100000, loss 6.808652
+INFO 2021-02-25 00:47:18 train.py: 78] Epoch 2, iter 4000/6416, lr 0.100000, loss 6.728413
+INFO 2021-02-25 00:52:46 train.py: 78] Epoch 2, iter 4200/6416, lr 0.100000, loss 6.722016
+INFO 2021-02-25 00:58:14 train.py: 78] Epoch 2, iter 4400/6416, lr 0.100000, loss 6.668323
+INFO 2021-02-25 01:03:41 train.py: 78] Epoch 2, iter 4600/6416, lr 0.100000, loss 6.618915
+INFO 2021-02-25 01:09:09 train.py: 78] Epoch 2, iter 4800/6416, lr 0.100000, loss 6.585573
+INFO 2021-02-25 01:14:36 train.py: 78] Epoch 2, iter 5000/6416, lr 0.100000, loss 6.551705
+INFO 2021-02-25 01:20:04 train.py: 78] Epoch 2, iter 5200/6416, lr 0.100000, loss 6.519191
+INFO 2021-02-25 01:25:32 train.py: 78] Epoch 2, iter 5400/6416, lr 0.100000, loss 6.463274
+INFO 2021-02-25 01:30:59 train.py: 78] Epoch 2, iter 5600/6416, lr 0.100000, loss 6.418169
+INFO 2021-02-25 01:36:27 train.py: 78] Epoch 2, iter 5800/6416, lr 0.100000, loss 6.364018
+INFO 2021-02-25 01:41:54 train.py: 91] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-02-25 01:41:55 train.py: 78] Epoch 2, iter 6000/6416, lr 0.100000, loss 6.359388
+INFO 2021-02-25 01:47:23 train.py: 78] Epoch 2, iter 6200/6416, lr 0.100000, loss 6.282158
+INFO 2021-02-25 01:52:51 train.py: 78] Epoch 2, iter 6400/6416, lr 0.100000, loss 6.291159
+INFO 2021-02-25 01:53:15 train.py: 96] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-02-25 01:53:18 train.py: 78] Epoch 3, iter 0/6416, lr 0.100000, loss 6.288440
+INFO 2021-02-25 01:58:45 train.py: 78] Epoch 3, iter 200/6416, lr 0.100000, loss 5.738819
+INFO 2021-02-25 02:04:11 train.py: 78] Epoch 3, iter 400/6416, lr 0.100000, loss 5.688125
+INFO 2021-02-25 02:09:38 train.py: 78] Epoch 3, iter 600/6416, lr 0.100000, loss 5.752232
+INFO 2021-02-25 02:15:05 train.py: 78] Epoch 3, iter 800/6416, lr 0.100000, loss 5.803626
+INFO 2021-02-25 02:20:32 train.py: 78] Epoch 3, iter 1000/6416, lr 0.100000, loss 5.864418
+INFO 2021-02-25 02:25:58 train.py: 78] Epoch 3, iter 1200/6416, lr 0.100000, loss 5.874163
+INFO 2021-02-25 02:31:25 train.py: 78] Epoch 3, iter 1400/6416, lr 0.100000, loss 5.866870
+INFO 2021-02-25 02:36:52 train.py: 78] Epoch 3, iter 1600/6416, lr 0.100000, loss 5.891690
+INFO 2021-02-25 02:42:19 train.py: 78] Epoch 3, iter 1800/6416, lr 0.100000, loss 5.888111
+INFO 2021-02-25 02:47:46 train.py: 78] Epoch 3, iter 2000/6416, lr 0.100000, loss 5.898130
+INFO 2021-02-25 02:53:14 train.py: 78] Epoch 3, iter 2200/6416, lr 0.100000, loss 5.867494
+INFO 2021-02-25 02:58:41 train.py: 78] Epoch 3, iter 2400/6416, lr 0.100000, loss 5.864499
+INFO 2021-02-25 03:04:08 train.py: 78] Epoch 3, iter 2600/6416, lr 0.100000, loss 5.830463
+INFO 2021-02-25 03:09:35 train.py: 78] Epoch 3, iter 2800/6416, lr 0.100000, loss 5.830445
+INFO 2021-02-25 03:15:01 train.py: 91] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-02-25 03:15:03 train.py: 78] Epoch 3, iter 3000/6416, lr 0.100000, loss 5.800021
+INFO 2021-02-25 03:20:31 train.py: 78] Epoch 3, iter 3200/6416, lr 0.100000, loss 5.813142
+INFO 2021-02-25 03:25:58 train.py: 78] Epoch 3, iter 3400/6416, lr 0.100000, loss 5.759068
+INFO 2021-02-25 03:31:25 train.py: 78] Epoch 3, iter 3600/6416, lr 0.100000, loss 5.764986
+INFO 2021-02-25 03:36:53 train.py: 78] Epoch 3, iter 3800/6416, lr 0.100000, loss 5.711055
+INFO 2021-02-25 03:42:21 train.py: 78] Epoch 3, iter 4000/6416, lr 0.100000, loss 5.737976
+INFO 2021-02-25 03:47:48 train.py: 78] Epoch 3, iter 4200/6416, lr 0.100000, loss 5.713059
+INFO 2021-02-25 03:53:16 train.py: 78] Epoch 3, iter 4400/6416, lr 0.100000, loss 5.680029
+INFO 2021-02-25 03:58:43 train.py: 78] Epoch 3, iter 4600/6416, lr 0.100000, loss 5.661024
+INFO 2021-02-25 04:04:11 train.py: 78] Epoch 3, iter 4800/6416, lr 0.100000, loss 5.640687
+INFO 2021-02-25 04:09:39 train.py: 78] Epoch 3, iter 5000/6416, lr 0.100000, loss 5.629790
+INFO 2021-02-25 04:15:06 train.py: 78] Epoch 3, iter 5200/6416, lr 0.100000, loss 5.582526
+INFO 2021-02-25 04:20:34 train.py: 78] Epoch 3, iter 5400/6416, lr 0.100000, loss 5.605536
+INFO 2021-02-25 04:26:01 train.py: 78] Epoch 3, iter 5600/6416, lr 0.100000, loss 5.551062
+INFO 2021-02-25 04:31:29 train.py: 78] Epoch 3, iter 5800/6416, lr 0.100000, loss 5.555254
+INFO 2021-02-25 04:36:56 train.py: 91] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-02-25 04:36:57 train.py: 78] Epoch 3, iter 6000/6416, lr 0.100000, loss 5.545827
+INFO 2021-02-25 04:42:25 train.py: 78] Epoch 3, iter 6200/6416, lr 0.100000, loss 5.550858
+INFO 2021-02-25 04:47:53 train.py: 78] Epoch 3, iter 6400/6416, lr 0.100000, loss 5.506866
+INFO 2021-02-25 04:48:17 train.py: 96] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-02-25 04:48:19 train.py: 78] Epoch 4, iter 0/6416, lr 0.100000, loss 5.477970
+INFO 2021-02-25 04:53:47 train.py: 78] Epoch 4, iter 200/6416, lr 0.100000, loss 4.973440
+INFO 2021-02-25 04:59:13 train.py: 78] Epoch 4, iter 400/6416, lr 0.100000, loss 4.942978
+INFO 2021-02-25 05:04:40 train.py: 78] Epoch 4, iter 600/6416, lr 0.100000, loss 5.004058
+INFO 2021-02-25 05:10:07 train.py: 78] Epoch 4, iter 800/6416, lr 0.100000, loss 5.066273
+INFO 2021-02-25 05:15:34 train.py: 78] Epoch 4, iter 1000/6416, lr 0.100000, loss 5.140277
+INFO 2021-02-25 05:21:01 train.py: 78] Epoch 4, iter 1200/6416, lr 0.100000, loss 5.154495
+INFO 2021-02-25 05:26:27 train.py: 78] Epoch 4, iter 1400/6416, lr 0.100000, loss 5.166434
+INFO 2021-02-25 05:31:54 train.py: 78] Epoch 4, iter 1600/6416, lr 0.100000, loss 5.229045
+INFO 2021-02-25 05:37:22 train.py: 78] Epoch 4, iter 1800/6416, lr 0.100000, loss 5.174682
+INFO 2021-02-25 05:42:49 train.py: 78] Epoch 4, iter 2000/6416, lr 0.100000, loss 5.207808
+INFO 2021-02-25 05:48:16 train.py: 78] Epoch 4, iter 2200/6416, lr 0.100000, loss 5.174642
+INFO 2021-02-25 05:53:43 train.py: 78] Epoch 4, iter 2400/6416, lr 0.100000, loss 5.227486
+INFO 2021-02-25 05:59:10 train.py: 78] Epoch 4, iter 2600/6416, lr 0.100000, loss 5.223038
+INFO 2021-02-25 06:04:37 train.py: 78] Epoch 4, iter 2800/6416, lr 0.100000, loss 5.245311
+INFO 2021-02-25 06:10:04 train.py: 91] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-02-25 06:10:05 train.py: 78] Epoch 4, iter 3000/6416, lr 0.100000, loss 5.222831
+INFO 2021-02-25 06:15:33 train.py: 78] Epoch 4, iter 3200/6416, lr 0.100000, loss 5.221403
+INFO 2021-02-25 06:21:00 train.py: 78] Epoch 4, iter 3400/6416, lr 0.100000, loss 5.200797
+INFO 2021-02-25 06:26:28 train.py: 78] Epoch 4, iter 3600/6416, lr 0.100000, loss 5.169596
+INFO 2021-02-25 06:31:55 train.py: 78] Epoch 4, iter 3800/6416, lr 0.100000, loss 5.160110
+INFO 2021-02-25 06:37:23 train.py: 78] Epoch 4, iter 4000/6416, lr 0.100000, loss 5.199481
+INFO 2021-02-25 06:42:51 train.py: 78] Epoch 4, iter 4200/6416, lr 0.100000, loss 5.213140
+INFO 2021-02-25 06:48:18 train.py: 78] Epoch 4, iter 4400/6416, lr 0.100000, loss 5.141887
+INFO 2021-02-25 06:53:46 train.py: 78] Epoch 4, iter 4600/6416, lr 0.100000, loss 5.106738
+INFO 2021-02-25 06:59:13 train.py: 78] Epoch 4, iter 4800/6416, lr 0.100000, loss 5.107464
+INFO 2021-02-25 07:04:41 train.py: 78] Epoch 4, iter 5000/6416, lr 0.100000, loss 5.093096
+INFO 2021-02-25 07:10:08 train.py: 78] Epoch 4, iter 5200/6416, lr 0.100000, loss 5.106639
+INFO 2021-02-25 07:15:36 train.py: 78] Epoch 4, iter 5400/6416, lr 0.100000, loss 5.129879
+INFO 2021-02-25 07:21:04 train.py: 78] Epoch 4, iter 5600/6416, lr 0.100000, loss 5.071923
+INFO 2021-02-25 07:26:31 train.py: 78] Epoch 4, iter 5800/6416, lr 0.100000, loss 5.059435
+INFO 2021-02-25 07:31:58 train.py: 91] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-02-25 07:31:59 train.py: 78] Epoch 4, iter 6000/6416, lr 0.100000, loss 5.066874
+INFO 2021-02-25 07:37:27 train.py: 78] Epoch 4, iter 6200/6416, lr 0.100000, loss 5.043874
+INFO 2021-02-25 07:42:55 train.py: 78] Epoch 4, iter 6400/6416, lr 0.100000, loss 5.045064
+INFO 2021-02-25 07:43:19 train.py: 96] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-02-25 07:43:21 train.py: 78] Epoch 5, iter 0/6416, lr 0.100000, loss 5.080495
+INFO 2021-02-25 07:48:48 train.py: 78] Epoch 5, iter 200/6416, lr 0.100000, loss 4.551879
+INFO 2021-02-25 07:54:15 train.py: 78] Epoch 5, iter 400/6416, lr 0.100000, loss 4.528073
+INFO 2021-02-25 07:59:42 train.py: 78] Epoch 5, iter 600/6416, lr 0.100000, loss 4.571969
+INFO 2021-02-25 08:05:09 train.py: 78] Epoch 5, iter 800/6416, lr 0.100000, loss 4.628106
+INFO 2021-02-25 08:10:35 train.py: 78] Epoch 5, iter 1000/6416, lr 0.100000, loss 4.701971
+INFO 2021-02-25 08:16:02 train.py: 78] Epoch 5, iter 1200/6416, lr 0.100000, loss 4.717321
+INFO 2021-02-25 08:21:29 train.py: 78] Epoch 5, iter 1400/6416, lr 0.100000, loss 4.721839
+INFO 2021-02-25 08:26:56 train.py: 78] Epoch 5, iter 1600/6416, lr 0.100000, loss 4.806374
+INFO 2021-02-25 08:32:23 train.py: 78] Epoch 5, iter 1800/6416, lr 0.100000, loss 4.805767
+INFO 2021-02-25 08:37:50 train.py: 78] Epoch 5, iter 2000/6416, lr 0.100000, loss 4.847204
+INFO 2021-02-25 08:43:17 train.py: 78] Epoch 5, iter 2200/6416, lr 0.100000, loss 4.826196
+INFO 2021-02-25 08:48:44 train.py: 78] Epoch 5, iter 2400/6416, lr 0.100000, loss 4.859384
+INFO 2021-02-25 08:54:11 train.py: 78] Epoch 5, iter 2600/6416, lr 0.100000, loss 4.812125
+INFO 2021-02-25 08:59:39 train.py: 78] Epoch 5, iter 2800/6416, lr 0.100000, loss 4.866214
+INFO 2021-02-25 09:05:05 train.py: 91] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-02-25 09:05:06 train.py: 78] Epoch 5, iter 3000/6416, lr 0.100000, loss 4.841835
+INFO 2021-02-25 09:10:34 train.py: 78] Epoch 5, iter 3200/6416, lr 0.100000, loss 4.802345
+INFO 2021-02-25 09:16:01 train.py: 78] Epoch 5, iter 3400/6416, lr 0.100000, loss 4.830935
+INFO 2021-02-25 09:21:29 train.py: 78] Epoch 5, iter 3600/6416, lr 0.100000, loss 4.829578
+INFO 2021-02-25 09:26:56 train.py: 78] Epoch 5, iter 3800/6416, lr 0.100000, loss 4.852716
+INFO 2021-02-25 09:32:24 train.py: 78] Epoch 5, iter 4000/6416, lr 0.100000, loss 4.806939
+INFO 2021-02-25 09:37:51 train.py: 78] Epoch 5, iter 4200/6416, lr 0.100000, loss 4.786034
+INFO 2021-02-25 09:43:19 train.py: 78] Epoch 5, iter 4400/6416, lr 0.100000, loss 4.837186
+INFO 2021-02-25 09:48:46 train.py: 78] Epoch 5, iter 4600/6416, lr 0.100000, loss 4.794805
+INFO 2021-02-25 09:54:14 train.py: 78] Epoch 5, iter 4800/6416, lr 0.100000, loss 4.805823
+INFO 2021-02-25 09:59:42 train.py: 78] Epoch 5, iter 5000/6416, lr 0.100000, loss 4.787804
+INFO 2021-02-25 10:05:09 train.py: 78] Epoch 5, iter 5200/6416, lr 0.100000, loss 4.806678
+INFO 2021-02-25 10:10:37 train.py: 78] Epoch 5, iter 5400/6416, lr 0.100000, loss 4.790378
+INFO 2021-02-25 10:16:04 train.py: 78] Epoch 5, iter 5600/6416, lr 0.100000, loss 4.740289
+INFO 2021-02-25 10:21:32 train.py: 78] Epoch 5, iter 5800/6416, lr 0.100000, loss 4.763237
+INFO 2021-02-25 10:26:59 train.py: 91] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-02-25 10:27:00 train.py: 78] Epoch 5, iter 6000/6416, lr 0.100000, loss 4.789376
+INFO 2021-02-25 10:32:28 train.py: 78] Epoch 5, iter 6200/6416, lr 0.100000, loss 4.766370
+INFO 2021-02-25 10:37:56 train.py: 78] Epoch 5, iter 6400/6416, lr 0.100000, loss 4.724789
+INFO 2021-02-25 10:38:20 train.py: 96] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-02-25 10:38:23 train.py: 78] Epoch 6, iter 0/6416, lr 0.100000, loss 4.732681
+INFO 2021-02-25 10:43:50 train.py: 78] Epoch 6, iter 200/6416, lr 0.100000, loss 4.254043
+INFO 2021-02-25 10:49:16 train.py: 78] Epoch 6, iter 400/6416, lr 0.100000, loss 4.230677
+INFO 2021-02-25 10:54:43 train.py: 78] Epoch 6, iter 600/6416, lr 0.100000, loss 4.279142
+INFO 2021-02-25 11:00:09 train.py: 78] Epoch 6, iter 800/6416, lr 0.100000, loss 4.355736
+INFO 2021-02-25 11:05:36 train.py: 78] Epoch 6, iter 1000/6416, lr 0.100000, loss 4.426550
+INFO 2021-02-25 11:11:03 train.py: 78] Epoch 6, iter 1200/6416, lr 0.100000, loss 4.426828
+INFO 2021-02-25 11:16:30 train.py: 78] Epoch 6, iter 1400/6416, lr 0.100000, loss 4.444814
+INFO 2021-02-25 11:21:57 train.py: 78] Epoch 6, iter 1600/6416, lr 0.100000, loss 4.523520
+INFO 2021-02-25 11:27:24 train.py: 78] Epoch 6, iter 1800/6416, lr 0.100000, loss 4.514382
+INFO 2021-02-25 11:32:51 train.py: 78] Epoch 6, iter 2000/6416, lr 0.100000, loss 4.556363
+INFO 2021-02-25 11:38:18 train.py: 78] Epoch 6, iter 2200/6416, lr 0.100000, loss 4.552444
+INFO 2021-02-25 11:43:45 train.py: 78] Epoch 6, iter 2400/6416, lr 0.100000, loss 4.560605
+INFO 2021-02-25 11:49:13 train.py: 78] Epoch 6, iter 2600/6416, lr 0.100000, loss 4.588173
+INFO 2021-02-25 11:54:40 train.py: 78] Epoch 6, iter 2800/6416, lr 0.100000, loss 4.595977
+INFO 2021-02-25 12:00:06 train.py: 91] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-02-25 12:00:08 train.py: 78] Epoch 6, iter 3000/6416, lr 0.100000, loss 4.560906
+INFO 2021-02-25 12:05:35 train.py: 78] Epoch 6, iter 3200/6416, lr 0.100000, loss 4.562175
+INFO 2021-02-25 12:11:03 train.py: 78] Epoch 6, iter 3400/6416, lr 0.100000, loss 4.587999
+INFO 2021-02-25 12:16:30 train.py: 78] Epoch 6, iter 3600/6416, lr 0.100000, loss 4.587679
+INFO 2021-02-25 12:21:58 train.py: 78] Epoch 6, iter 3800/6416, lr 0.100000, loss 4.566212
+INFO 2021-02-25 12:27:25 train.py: 78] Epoch 6, iter 4000/6416, lr 0.100000, loss 4.623884
+INFO 2021-02-25 12:32:53 train.py: 78] Epoch 6, iter 4200/6416, lr 0.100000, loss 4.578886
+INFO 2021-02-25 12:38:20 train.py: 78] Epoch 6, iter 4400/6416, lr 0.100000, loss 4.588638
+INFO 2021-02-25 12:43:48 train.py: 78] Epoch 6, iter 4600/6416, lr 0.100000, loss 4.579136
+INFO 2021-02-25 12:49:15 train.py: 78] Epoch 6, iter 4800/6416, lr 0.100000, loss 4.584031
+INFO 2021-02-25 12:54:43 train.py: 78] Epoch 6, iter 5000/6416, lr 0.100000, loss 4.551115
+INFO 2021-02-25 13:00:10 train.py: 78] Epoch 6, iter 5200/6416, lr 0.100000, loss 4.582152
+INFO 2021-02-25 13:05:38 train.py: 78] Epoch 6, iter 5400/6416, lr 0.100000, loss 4.565033
+INFO 2021-02-25 13:11:06 train.py: 78] Epoch 6, iter 5600/6416, lr 0.100000, loss 4.562807
+INFO 2021-02-25 13:16:33 train.py: 78] Epoch 6, iter 5800/6416, lr 0.100000, loss 4.568973
+INFO 2021-02-25 13:22:00 train.py: 91] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-02-25 13:22:02 train.py: 78] Epoch 6, iter 6000/6416, lr 0.100000, loss 4.570001
+INFO 2021-02-25 13:27:29 train.py: 78] Epoch 6, iter 6200/6416, lr 0.100000, loss 4.504335
+INFO 2021-02-25 13:32:57 train.py: 78] Epoch 6, iter 6400/6416, lr 0.100000, loss 4.541264
+INFO 2021-02-25 13:33:21 train.py: 96] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-02-25 13:33:23 train.py: 78] Epoch 7, iter 0/6416, lr 0.100000, loss 4.483522
+INFO 2021-02-25 13:38:50 train.py: 78] Epoch 7, iter 200/6416, lr 0.100000, loss 4.046358
+INFO 2021-02-25 13:44:17 train.py: 78] Epoch 7, iter 400/6416, lr 0.100000, loss 3.995566
+INFO 2021-02-25 13:49:44 train.py: 78] Epoch 7, iter 600/6416, lr 0.100000, loss 4.119862
+INFO 2021-02-25 13:55:10 train.py: 78] Epoch 7, iter 800/6416, lr 0.100000, loss 4.127078
+INFO 2021-02-25 14:00:37 train.py: 78] Epoch 7, iter 1000/6416, lr 0.100000, loss 4.214201
+INFO 2021-02-25 14:06:04 train.py: 78] Epoch 7, iter 1200/6416, lr 0.100000, loss 4.245915
+INFO 2021-02-25 14:11:30 train.py: 78] Epoch 7, iter 1400/6416, lr 0.100000, loss 4.296237
+INFO 2021-02-25 14:16:57 train.py: 78] Epoch 7, iter 1600/6416, lr 0.100000, loss 4.285793
+INFO 2021-02-25 14:22:24 train.py: 78] Epoch 7, iter 1800/6416, lr 0.100000, loss 4.356745
+INFO 2021-02-25 14:27:51 train.py: 78] Epoch 7, iter 2000/6416, lr 0.100000, loss 4.336755
+INFO 2021-02-25 14:33:18 train.py: 78] Epoch 7, iter 2200/6416, lr 0.100000, loss 4.378328
+INFO 2021-02-25 14:38:46 train.py: 78] Epoch 7, iter 2400/6416, lr 0.100000, loss 4.395332
+INFO 2021-02-25 14:44:13 train.py: 78] Epoch 7, iter 2600/6416, lr 0.100000, loss 4.380206
+INFO 2021-02-25 14:49:40 train.py: 78] Epoch 7, iter 2800/6416, lr 0.100000, loss 4.390752
+INFO 2021-02-25 14:55:06 train.py: 91] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-02-25 14:55:08 train.py: 78] Epoch 7, iter 3000/6416, lr 0.100000, loss 4.419343
+INFO 2021-02-25 15:00:35 train.py: 78] Epoch 7, iter 3200/6416, lr 0.100000, loss 4.423469
+INFO 2021-02-25 15:06:03 train.py: 78] Epoch 7, iter 3400/6416, lr 0.100000, loss 4.410455
+INFO 2021-02-25 15:11:30 train.py: 78] Epoch 7, iter 3600/6416, lr 0.100000, loss 4.400200
+INFO 2021-02-25 15:16:58 train.py: 78] Epoch 7, iter 3800/6416, lr 0.100000, loss 4.373593
+INFO 2021-02-25 15:22:25 train.py: 78] Epoch 7, iter 4000/6416, lr 0.100000, loss 4.406134
+INFO 2021-02-25 15:27:53 train.py: 78] Epoch 7, iter 4200/6416, lr 0.100000, loss 4.391583
+INFO 2021-02-25 15:33:20 train.py: 78] Epoch 7, iter 4400/6416, lr 0.100000, loss 4.379079
+INFO 2021-02-25 15:38:48 train.py: 78] Epoch 7, iter 4600/6416, lr 0.100000, loss 4.405875
+INFO 2021-02-25 15:44:16 train.py: 78] Epoch 7, iter 4800/6416, lr 0.100000, loss 4.433525
+INFO 2021-02-25 15:49:43 train.py: 78] Epoch 7, iter 5000/6416, lr 0.100000, loss 4.402329
+INFO 2021-02-25 15:55:11 train.py: 78] Epoch 7, iter 5200/6416, lr 0.100000, loss 4.397792
+INFO 2021-02-25 16:00:38 train.py: 78] Epoch 7, iter 5400/6416, lr 0.100000, loss 4.399441
+INFO 2021-02-25 16:06:06 train.py: 78] Epoch 7, iter 5600/6416, lr 0.100000, loss 4.375682
+INFO 2021-02-25 16:11:33 train.py: 78] Epoch 7, iter 5800/6416, lr 0.100000, loss 4.394716
+INFO 2021-02-25 16:17:00 train.py: 91] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-02-25 16:17:01 train.py: 78] Epoch 7, iter 6000/6416, lr 0.100000, loss 4.397713
+INFO 2021-02-25 16:22:29 train.py: 78] Epoch 7, iter 6200/6416, lr 0.100000, loss 4.375446
+INFO 2021-02-25 16:27:57 train.py: 78] Epoch 7, iter 6400/6416, lr 0.100000, loss 4.380827
+INFO 2021-02-25 16:28:21 train.py: 96] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-02-25 16:28:23 train.py: 78] Epoch 8, iter 0/6416, lr 0.100000, loss 4.354991
+INFO 2021-02-25 16:33:50 train.py: 78] Epoch 8, iter 200/6416, lr 0.100000, loss 3.899804
+INFO 2021-02-25 16:39:17 train.py: 78] Epoch 8, iter 400/6416, lr 0.100000, loss 3.872152
+INFO 2021-02-25 16:44:44 train.py: 78] Epoch 8, iter 600/6416, lr 0.100000, loss 3.933592
+INFO 2021-02-25 16:50:10 train.py: 78] Epoch 8, iter 800/6416, lr 0.100000, loss 3.973056
+INFO 2021-02-25 16:55:37 train.py: 78] Epoch 8, iter 1000/6416, lr 0.100000, loss 4.077531
+INFO 2021-02-25 17:01:04 train.py: 78] Epoch 8, iter 1200/6416, lr 0.100000, loss 4.107879
+INFO 2021-02-25 17:06:30 train.py: 78] Epoch 8, iter 1400/6416, lr 0.100000, loss 4.137148
+INFO 2021-02-25 17:11:57 train.py: 78] Epoch 8, iter 1600/6416, lr 0.100000, loss 4.176589
+INFO 2021-02-25 17:17:24 train.py: 78] Epoch 8, iter 1800/6416, lr 0.100000, loss 4.210612
+INFO 2021-02-25 17:22:51 train.py: 78] Epoch 8, iter 2000/6416, lr 0.100000, loss 4.188962
+INFO 2021-02-25 17:28:18 train.py: 78] Epoch 8, iter 2200/6416, lr 0.100000, loss 4.257363
+INFO 2021-02-25 17:33:45 train.py: 78] Epoch 8, iter 2400/6416, lr 0.100000, loss 4.228372
+INFO 2021-02-25 17:39:13 train.py: 78] Epoch 8, iter 2600/6416, lr 0.100000, loss 4.261803
+INFO 2021-02-25 17:44:40 train.py: 78] Epoch 8, iter 2800/6416, lr 0.100000, loss 4.272180
+INFO 2021-02-25 17:50:06 train.py: 91] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-02-25 17:50:08 train.py: 78] Epoch 8, iter 3000/6416, lr 0.100000, loss 4.275108
+INFO 2021-02-25 17:55:35 train.py: 78] Epoch 8, iter 3200/6416, lr 0.100000, loss 4.239371
+INFO 2021-02-25 18:01:03 train.py: 78] Epoch 8, iter 3400/6416, lr 0.100000, loss 4.262291
+INFO 2021-02-25 18:06:30 train.py: 78] Epoch 8, iter 3600/6416, lr 0.100000, loss 4.235420
+INFO 2021-02-25 18:11:57 train.py: 78] Epoch 8, iter 3800/6416, lr 0.100000, loss 4.271760
+INFO 2021-02-25 18:17:25 train.py: 78] Epoch 8, iter 4000/6416, lr 0.100000, loss 4.280106
+INFO 2021-02-25 18:22:52 train.py: 78] Epoch 8, iter 4200/6416, lr 0.100000, loss 4.258138
+INFO 2021-02-25 18:28:20 train.py: 78] Epoch 8, iter 4400/6416, lr 0.100000, loss 4.259981
+INFO 2021-02-25 18:33:47 train.py: 78] Epoch 8, iter 4600/6416, lr 0.100000, loss 4.248852
+INFO 2021-02-25 18:39:15 train.py: 78] Epoch 8, iter 4800/6416, lr 0.100000, loss 4.266390
+INFO 2021-02-25 18:44:42 train.py: 78] Epoch 8, iter 5000/6416, lr 0.100000, loss 4.267292
+INFO 2021-02-25 18:50:10 train.py: 78] Epoch 8, iter 5200/6416, lr 0.100000, loss 4.244896
+INFO 2021-02-25 18:55:37 train.py: 78] Epoch 8, iter 5400/6416, lr 0.100000, loss 4.239629
+INFO 2021-02-25 19:01:05 train.py: 78] Epoch 8, iter 5600/6416, lr 0.100000, loss 4.265155
+INFO 2021-02-25 19:06:32 train.py: 78] Epoch 8, iter 5800/6416, lr 0.100000, loss 4.271514
+INFO 2021-02-25 19:11:59 train.py: 91] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-02-25 19:12:01 train.py: 78] Epoch 8, iter 6000/6416, lr 0.100000, loss 4.259593
+INFO 2021-02-25 19:17:28 train.py: 78] Epoch 8, iter 6200/6416, lr 0.100000, loss 4.270696
+INFO 2021-02-25 19:22:56 train.py: 78] Epoch 8, iter 6400/6416, lr 0.100000, loss 4.234374
+INFO 2021-02-25 19:23:20 train.py: 96] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-02-25 19:23:23 train.py: 78] Epoch 9, iter 0/6416, lr 0.100000, loss 4.238643
+INFO 2021-02-25 19:28:50 train.py: 78] Epoch 9, iter 200/6416, lr 0.100000, loss 3.798645
+INFO 2021-02-25 19:34:16 train.py: 78] Epoch 9, iter 400/6416, lr 0.100000, loss 3.755003
+INFO 2021-02-25 19:39:43 train.py: 78] Epoch 9, iter 600/6416, lr 0.100000, loss 3.782248
+INFO 2021-02-25 19:45:10 train.py: 78] Epoch 9, iter 800/6416, lr 0.100000, loss 3.853883
+INFO 2021-02-25 19:50:36 train.py: 78] Epoch 9, iter 1000/6416, lr 0.100000, loss 3.912669
+INFO 2021-02-25 19:56:03 train.py: 78] Epoch 9, iter 1200/6416, lr 0.100000, loss 3.974457
+INFO 2021-02-25 20:01:30 train.py: 78] Epoch 9, iter 1400/6416, lr 0.100000, loss 4.023855
+INFO 2021-02-25 20:06:57 train.py: 78] Epoch 9, iter 1600/6416, lr 0.100000, loss 4.063848
+INFO 2021-02-25 20:12:24 train.py: 78] Epoch 9, iter 1800/6416, lr 0.100000, loss 4.115235
+INFO 2021-02-25 20:17:51 train.py: 78] Epoch 9, iter 2000/6416, lr 0.100000, loss 4.103210
+INFO 2021-02-25 20:23:18 train.py: 78] Epoch 9, iter 2200/6416, lr 0.100000, loss 4.101963
+INFO 2021-02-25 20:28:45 train.py: 78] Epoch 9, iter 2400/6416, lr 0.100000, loss 4.115857
+INFO 2021-02-25 20:34:12 train.py: 78] Epoch 9, iter 2600/6416, lr 0.100000, loss 4.122640
+INFO 2021-02-25 20:39:39 train.py: 78] Epoch 9, iter 2800/6416, lr 0.100000, loss 4.126902
+INFO 2021-02-25 20:45:06 train.py: 91] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-02-25 20:45:07 train.py: 78] Epoch 9, iter 3000/6416, lr 0.100000, loss 4.129934
+INFO 2021-02-25 20:50:35 train.py: 78] Epoch 9, iter 3200/6416, lr 0.100000, loss 4.138831
+INFO 2021-02-25 20:56:02 train.py: 78] Epoch 9, iter 3400/6416, lr 0.100000, loss 4.154145
+INFO 2021-02-25 21:01:29 train.py: 78] Epoch 9, iter 3600/6416, lr 0.100000, loss 4.180038
+INFO 2021-02-25 21:06:57 train.py: 78] Epoch 9, iter 3800/6416, lr 0.100000, loss 4.174316
+INFO 2021-02-25 21:12:24 train.py: 78] Epoch 9, iter 4000/6416, lr 0.100000, loss 4.173374
+INFO 2021-02-25 21:17:51 train.py: 78] Epoch 9, iter 4200/6416, lr 0.100000, loss 4.168531
+INFO 2021-02-25 21:23:19 train.py: 78] Epoch 9, iter 4400/6416, lr 0.100000, loss 4.179376
+INFO 2021-02-25 21:28:46 train.py: 78] Epoch 9, iter 4600/6416, lr 0.100000, loss 4.158136
+INFO 2021-02-25 21:34:14 train.py: 78] Epoch 9, iter 4800/6416, lr 0.100000, loss 4.158712
+INFO 2021-02-25 21:39:42 train.py: 78] Epoch 9, iter 5000/6416, lr 0.100000, loss 4.154439
+INFO 2021-02-25 21:45:09 train.py: 78] Epoch 9, iter 5200/6416, lr 0.100000, loss 4.156886
+INFO 2021-02-25 21:50:37 train.py: 78] Epoch 9, iter 5400/6416, lr 0.100000, loss 4.141296
+INFO 2021-02-25 21:56:04 train.py: 78] Epoch 9, iter 5600/6416, lr 0.100000, loss 4.186354
+INFO 2021-02-25 22:01:32 train.py: 78] Epoch 9, iter 5800/6416, lr 0.100000, loss 4.128818
+INFO 2021-02-25 22:06:58 train.py: 91] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-02-25 22:07:00 train.py: 78] Epoch 9, iter 6000/6416, lr 0.100000, loss 4.139489
+INFO 2021-02-25 22:12:27 train.py: 78] Epoch 9, iter 6200/6416, lr 0.100000, loss 4.139924
+INFO 2021-02-25 22:17:55 train.py: 78] Epoch 9, iter 6400/6416, lr 0.100000, loss 4.156645
+INFO 2021-02-25 22:18:19 train.py: 96] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-02-25 22:18:22 train.py: 78] Epoch 10, iter 0/6416, lr 0.010000, loss 4.117951
+INFO 2021-02-25 22:23:49 train.py: 78] Epoch 10, iter 200/6416, lr 0.010000, loss 3.038555
+INFO 2021-02-25 22:29:15 train.py: 78] Epoch 10, iter 400/6416, lr 0.010000, loss 2.809822
+INFO 2021-02-25 22:34:42 train.py: 78] Epoch 10, iter 600/6416, lr 0.010000, loss 2.720037
+INFO 2021-02-25 22:40:08 train.py: 78] Epoch 10, iter 800/6416, lr 0.010000, loss 2.629070
+INFO 2021-02-25 22:45:35 train.py: 78] Epoch 10, iter 1000/6416, lr 0.010000, loss 2.594028
+INFO 2021-02-25 22:51:02 train.py: 78] Epoch 10, iter 1200/6416, lr 0.010000, loss 2.536712
+INFO 2021-02-25 22:56:28 train.py: 78] Epoch 10, iter 1400/6416, lr 0.010000, loss 2.492130
+INFO 2021-02-25 23:01:55 train.py: 78] Epoch 10, iter 1600/6416, lr 0.010000, loss 2.450587
+INFO 2021-02-25 23:07:22 train.py: 78] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.433430
+INFO 2021-02-25 23:12:49 train.py: 78] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.401746
+INFO 2021-02-25 23:18:16 train.py: 78] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.348785
+INFO 2021-02-25 23:23:43 train.py: 78] Epoch 10, iter 2400/6416, lr 0.010000, loss 2.343040
+INFO 2021-02-25 23:29:10 train.py: 78] Epoch 10, iter 2600/6416, lr 0.010000, loss 2.335307
+INFO 2021-02-25 23:34:38 train.py: 78] Epoch 10, iter 2800/6416, lr 0.010000, loss 2.312495
+INFO 2021-02-25 23:40:04 train.py: 91] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-02-25 23:40:05 train.py: 78] Epoch 10, iter 3000/6416, lr 0.010000, loss 2.290390
+INFO 2021-02-25 23:45:33 train.py: 78] Epoch 10, iter 3200/6416, lr 0.010000, loss 2.272417
+INFO 2021-02-25 23:51:00 train.py: 78] Epoch 10, iter 3400/6416, lr 0.010000, loss 2.255538
+INFO 2021-02-25 23:56:27 train.py: 78] Epoch 10, iter 3600/6416, lr 0.010000, loss 2.240760
+INFO 2021-02-26 00:01:55 train.py: 78] Epoch 10, iter 3800/6416, lr 0.010000, loss 2.198393
+INFO 2021-02-26 00:07:22 train.py: 78] Epoch 10, iter 4000/6416, lr 0.010000, loss 2.194825
+INFO 2021-02-26 00:12:50 train.py: 78] Epoch 10, iter 4200/6416, lr 0.010000, loss 2.175206
+INFO 2021-02-26 00:18:17 train.py: 78] Epoch 10, iter 4400/6416, lr 0.010000, loss 2.189447
+INFO 2021-02-26 00:23:45 train.py: 78] Epoch 10, iter 4600/6416, lr 0.010000, loss 2.169700
+INFO 2021-02-26 00:29:12 train.py: 78] Epoch 10, iter 4800/6416, lr 0.010000, loss 2.133254
+INFO 2021-02-26 00:34:40 train.py: 78] Epoch 10, iter 5000/6416, lr 0.010000, loss 2.138729
+INFO 2021-02-26 00:40:07 train.py: 78] Epoch 10, iter 5200/6416, lr 0.010000, loss 2.114797
+INFO 2021-02-26 00:45:35 train.py: 78] Epoch 10, iter 5400/6416, lr 0.010000, loss 2.117085
+INFO 2021-02-26 00:51:02 train.py: 78] Epoch 10, iter 5600/6416, lr 0.010000, loss 2.105980
+INFO 2021-02-26 00:56:30 train.py: 78] Epoch 10, iter 5800/6416, lr 0.010000, loss 2.089937
+INFO 2021-02-26 01:01:56 train.py: 91] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-02-26 01:01:58 train.py: 78] Epoch 10, iter 6000/6416, lr 0.010000, loss 2.066743
+INFO 2021-02-26 01:07:25 train.py: 78] Epoch 10, iter 6200/6416, lr 0.010000, loss 2.068470
+INFO 2021-02-26 01:12:53 train.py: 78] Epoch 10, iter 6400/6416, lr 0.010000, loss 2.055863
+INFO 2021-02-26 01:13:17 train.py: 96] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-02-26 01:13:20 train.py: 78] Epoch 11, iter 0/6416, lr 0.010000, loss 1.989165
+INFO 2021-02-26 01:18:47 train.py: 78] Epoch 11, iter 200/6416, lr 0.010000, loss 1.758327
+INFO 2021-02-26 01:24:13 train.py: 78] Epoch 11, iter 400/6416, lr 0.010000, loss 1.759515
+INFO 2021-02-26 01:29:40 train.py: 78] Epoch 11, iter 600/6416, lr 0.010000, loss 1.751019
+INFO 2021-02-26 01:35:06 train.py: 78] Epoch 11, iter 800/6416, lr 0.010000, loss 1.763562
+INFO 2021-02-26 01:40:33 train.py: 78] Epoch 11, iter 1000/6416, lr 0.010000, loss 1.762703
+INFO 2021-02-26 01:45:59 train.py: 78] Epoch 11, iter 1200/6416, lr 0.010000, loss 1.751586
+INFO 2021-02-26 01:51:26 train.py: 78] Epoch 11, iter 1400/6416, lr 0.010000, loss 1.742817
+INFO 2021-02-26 01:56:53 train.py: 78] Epoch 11, iter 1600/6416, lr 0.010000, loss 1.752074
+INFO 2021-02-26 02:02:20 train.py: 78] Epoch 11, iter 1800/6416, lr 0.010000, loss 1.753101
+INFO 2021-02-26 02:07:47 train.py: 78] Epoch 11, iter 2000/6416, lr 0.010000, loss 1.726724
+INFO 2021-02-26 02:13:14 train.py: 78] Epoch 11, iter 2200/6416, lr 0.010000, loss 1.742941
+INFO 2021-02-26 02:18:41 train.py: 78] Epoch 11, iter 2400/6416, lr 0.010000, loss 1.757681
+INFO 2021-02-26 02:24:08 train.py: 78] Epoch 11, iter 2600/6416, lr 0.010000, loss 1.760140
+INFO 2021-02-26 02:29:35 train.py: 78] Epoch 11, iter 2800/6416, lr 0.010000, loss 1.748676
+INFO 2021-02-26 02:35:01 train.py: 91] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-02-26 02:35:03 train.py: 78] Epoch 11, iter 3000/6416, lr 0.010000, loss 1.750115
+INFO 2021-02-26 02:40:30 train.py: 78] Epoch 11, iter 3200/6416, lr 0.010000, loss 1.736650
+INFO 2021-02-26 02:45:57 train.py: 78] Epoch 11, iter 3400/6416, lr 0.010000, loss 1.769724
+INFO 2021-02-26 02:51:25 train.py: 78] Epoch 11, iter 3600/6416, lr 0.010000, loss 1.763623
+INFO 2021-02-26 02:56:52 train.py: 78] Epoch 11, iter 3800/6416, lr 0.010000, loss 1.746507
+INFO 2021-02-26 03:02:20 train.py: 78] Epoch 11, iter 4000/6416, lr 0.010000, loss 1.756656
+INFO 2021-02-26 03:07:47 train.py: 78] Epoch 11, iter 4200/6416, lr 0.010000, loss 1.716344
+INFO 2021-02-26 03:13:15 train.py: 78] Epoch 11, iter 4400/6416, lr 0.010000, loss 1.754668
+INFO 2021-02-26 03:18:42 train.py: 78] Epoch 11, iter 4600/6416, lr 0.010000, loss 1.736323
+INFO 2021-02-26 03:24:10 train.py: 78] Epoch 11, iter 4800/6416, lr 0.010000, loss 1.745712
+INFO 2021-02-26 03:29:37 train.py: 78] Epoch 11, iter 5000/6416, lr 0.010000, loss 1.739246
+INFO 2021-02-26 03:35:05 train.py: 78] Epoch 11, iter 5200/6416, lr 0.010000, loss 1.738633
+INFO 2021-02-26 03:40:32 train.py: 78] Epoch 11, iter 5400/6416, lr 0.010000, loss 1.740162
+INFO 2021-02-26 03:46:00 train.py: 78] Epoch 11, iter 5600/6416, lr 0.010000, loss 1.720936
+INFO 2021-02-26 03:51:27 train.py: 78] Epoch 11, iter 5800/6416, lr 0.010000, loss 1.733776
+INFO 2021-02-26 03:56:54 train.py: 91] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-02-26 03:56:55 train.py: 78] Epoch 11, iter 6000/6416, lr 0.010000, loss 1.732810
+INFO 2021-02-26 04:02:23 train.py: 78] Epoch 11, iter 6200/6416, lr 0.010000, loss 1.732687
+INFO 2021-02-26 04:07:51 train.py: 78] Epoch 11, iter 6400/6416, lr 0.010000, loss 1.751937
+INFO 2021-02-26 04:08:15 train.py: 96] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-02-26 04:08:17 train.py: 78] Epoch 12, iter 0/6416, lr 0.010000, loss 1.712993
+INFO 2021-02-26 04:13:44 train.py: 78] Epoch 12, iter 200/6416, lr 0.010000, loss 1.466138
+INFO 2021-02-26 04:19:11 train.py: 78] Epoch 12, iter 400/6416, lr 0.010000, loss 1.469204
+INFO 2021-02-26 04:24:38 train.py: 78] Epoch 12, iter 600/6416, lr 0.010000, loss 1.479212
+INFO 2021-02-26 04:30:04 train.py: 78] Epoch 12, iter 800/6416, lr 0.010000, loss 1.474471
+INFO 2021-02-26 04:35:31 train.py: 78] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.480727
+INFO 2021-02-26 04:40:58 train.py: 78] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.484029
+INFO 2021-02-26 04:46:25 train.py: 78] Epoch 12, iter 1400/6416, lr 0.010000, loss 1.480404
+INFO 2021-02-26 04:51:52 train.py: 78] Epoch 12, iter 1600/6416, lr 0.010000, loss 1.479380
+INFO 2021-02-26 04:57:19 train.py: 78] Epoch 12, iter 1800/6416, lr 0.010000, loss 1.493175
+INFO 2021-02-26 05:02:46 train.py: 78] Epoch 12, iter 2000/6416, lr 0.010000, loss 1.504001
+INFO 2021-02-26 05:08:13 train.py: 78] Epoch 12, iter 2200/6416, lr 0.010000, loss 1.519298
+INFO 2021-02-26 05:13:40 train.py: 78] Epoch 12, iter 2400/6416, lr 0.010000, loss 1.532432
+INFO 2021-02-26 05:19:07 train.py: 78] Epoch 12, iter 2600/6416, lr 0.010000, loss 1.519235
+INFO 2021-02-26 05:24:35 train.py: 78] Epoch 12, iter 2800/6416, lr 0.010000, loss 1.517812
+INFO 2021-02-26 05:30:01 train.py: 91] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-02-26 05:30:03 train.py: 78] Epoch 12, iter 3000/6416, lr 0.010000, loss 1.518538
+INFO 2021-02-26 05:35:30 train.py: 78] Epoch 12, iter 3200/6416, lr 0.010000, loss 1.524167
+INFO 2021-02-26 05:40:57 train.py: 78] Epoch 12, iter 3400/6416, lr 0.010000, loss 1.525306
+INFO 2021-02-26 05:46:25 train.py: 78] Epoch 12, iter 3600/6416, lr 0.010000, loss 1.546135
+INFO 2021-02-26 05:51:52 train.py: 78] Epoch 12, iter 3800/6416, lr 0.010000, loss 1.546478
+INFO 2021-02-26 05:57:20 train.py: 78] Epoch 12, iter 4000/6416, lr 0.010000, loss 1.545730
+INFO 2021-02-26 06:02:47 train.py: 78] Epoch 12, iter 4200/6416, lr 0.010000, loss 1.560959
+INFO 2021-02-26 06:08:15 train.py: 78] Epoch 12, iter 4400/6416, lr 0.010000, loss 1.541709
+INFO 2021-02-26 06:13:43 train.py: 78] Epoch 12, iter 4600/6416, lr 0.010000, loss 1.551759
+INFO 2021-02-26 06:19:10 train.py: 78] Epoch 12, iter 4800/6416, lr 0.010000, loss 1.542566
+INFO 2021-02-26 06:24:37 train.py: 78] Epoch 12, iter 5000/6416, lr 0.010000, loss 1.554449
+INFO 2021-02-26 06:30:05 train.py: 78] Epoch 12, iter 5200/6416, lr 0.010000, loss 1.571143
+INFO 2021-02-26 06:35:32 train.py: 78] Epoch 12, iter 5400/6416, lr 0.010000, loss 1.577824
+INFO 2021-02-26 06:41:00 train.py: 78] Epoch 12, iter 5600/6416, lr 0.010000, loss 1.581001
+INFO 2021-02-26 06:46:27 train.py: 78] Epoch 12, iter 5800/6416, lr 0.010000, loss 1.564784
+INFO 2021-02-26 06:51:54 train.py: 91] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-02-26 06:51:56 train.py: 78] Epoch 12, iter 6000/6416, lr 0.010000, loss 1.577631
+INFO 2021-02-26 06:57:23 train.py: 78] Epoch 12, iter 6200/6416, lr 0.010000, loss 1.584215
+INFO 2021-02-26 07:02:51 train.py: 78] Epoch 12, iter 6400/6416, lr 0.010000, loss 1.562081
+INFO 2021-02-26 07:03:15 train.py: 96] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-02-26 07:03:17 train.py: 78] Epoch 13, iter 0/6416, lr 0.001000, loss 1.562756
+INFO 2021-02-26 07:08:44 train.py: 78] Epoch 13, iter 200/6416, lr 0.001000, loss 1.277383
+INFO 2021-02-26 07:14:11 train.py: 78] Epoch 13, iter 400/6416, lr 0.001000, loss 1.236589
+INFO 2021-02-26 07:19:38 train.py: 78] Epoch 13, iter 600/6416, lr 0.001000, loss 1.251469
+INFO 2021-02-26 07:25:04 train.py: 78] Epoch 13, iter 800/6416, lr 0.001000, loss 1.230098
+INFO 2021-02-26 07:30:31 train.py: 78] Epoch 13, iter 1000/6416, lr 0.001000, loss 1.233839
+INFO 2021-02-26 07:35:58 train.py: 78] Epoch 13, iter 1200/6416, lr 0.001000, loss 1.245220
+INFO 2021-02-26 07:41:25 train.py: 78] Epoch 13, iter 1400/6416, lr 0.001000, loss 1.226836
+INFO 2021-02-26 07:46:51 train.py: 78] Epoch 13, iter 1600/6416, lr 0.001000, loss 1.228899
+INFO 2021-02-26 07:52:18 train.py: 78] Epoch 13, iter 1800/6416, lr 0.001000, loss 1.219979
+INFO 2021-02-26 07:57:45 train.py: 78] Epoch 13, iter 2000/6416, lr 0.001000, loss 1.227166
+INFO 2021-02-26 08:03:12 train.py: 78] Epoch 13, iter 2200/6416, lr 0.001000, loss 1.213129
+INFO 2021-02-26 08:08:40 train.py: 78] Epoch 13, iter 2400/6416, lr 0.001000, loss 1.225888
+INFO 2021-02-26 08:14:07 train.py: 78] Epoch 13, iter 2600/6416, lr 0.001000, loss 1.224813
+INFO 2021-02-26 08:19:34 train.py: 78] Epoch 13, iter 2800/6416, lr 0.001000, loss 1.217198
+INFO 2021-02-26 08:25:00 train.py: 91] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-02-26 08:25:02 train.py: 78] Epoch 13, iter 3000/6416, lr 0.001000, loss 1.222988
+INFO 2021-02-26 08:30:29 train.py: 78] Epoch 13, iter 3200/6416, lr 0.001000, loss 1.213856
+INFO 2021-02-26 08:35:57 train.py: 78] Epoch 13, iter 3400/6416, lr 0.001000, loss 1.216411
+INFO 2021-02-26 08:41:24 train.py: 78] Epoch 13, iter 3600/6416, lr 0.001000, loss 1.210488
+INFO 2021-02-26 08:46:52 train.py: 78] Epoch 13, iter 3800/6416, lr 0.001000, loss 1.210495
+INFO 2021-02-26 08:52:19 train.py: 78] Epoch 13, iter 4000/6416, lr 0.001000, loss 1.221563
+INFO 2021-02-26 08:57:47 train.py: 78] Epoch 13, iter 4200/6416, lr 0.001000, loss 1.220773
+INFO 2021-02-26 09:03:14 train.py: 78] Epoch 13, iter 4400/6416, lr 0.001000, loss 1.220611
+INFO 2021-02-26 09:08:42 train.py: 78] Epoch 13, iter 4600/6416, lr 0.001000, loss 1.225989
+INFO 2021-02-26 09:14:09 train.py: 78] Epoch 13, iter 4800/6416, lr 0.001000, loss 1.212040
+INFO 2021-02-26 09:19:37 train.py: 78] Epoch 13, iter 5000/6416, lr 0.001000, loss 1.217591
+INFO 2021-02-26 09:25:04 train.py: 78] Epoch 13, iter 5200/6416, lr 0.001000, loss 1.217382
+INFO 2021-02-26 09:30:32 train.py: 78] Epoch 13, iter 5400/6416, lr 0.001000, loss 1.212948
+INFO 2021-02-26 09:35:59 train.py: 78] Epoch 13, iter 5600/6416, lr 0.001000, loss 1.223484
+INFO 2021-02-26 09:41:27 train.py: 78] Epoch 13, iter 5800/6416, lr 0.001000, loss 1.219447
+INFO 2021-02-26 09:46:53 train.py: 91] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-02-26 09:46:55 train.py: 78] Epoch 13, iter 6000/6416, lr 0.001000, loss 1.212093
+INFO 2021-02-26 09:52:23 train.py: 78] Epoch 13, iter 6200/6416, lr 0.001000, loss 1.229453
+INFO 2021-02-26 09:57:50 train.py: 78] Epoch 13, iter 6400/6416, lr 0.001000, loss 1.230379
+INFO 2021-02-26 09:58:14 train.py: 96] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-02-26 09:58:17 train.py: 78] Epoch 14, iter 0/6416, lr 0.001000, loss 1.201706
+INFO 2021-02-26 10:03:44 train.py: 78] Epoch 14, iter 200/6416, lr 0.001000, loss 1.182069
+INFO 2021-02-26 10:09:11 train.py: 78] Epoch 14, iter 400/6416, lr 0.001000, loss 1.181349
+INFO 2021-02-26 10:14:37 train.py: 78] Epoch 14, iter 600/6416, lr 0.001000, loss 1.177843
+INFO 2021-02-26 10:20:04 train.py: 78] Epoch 14, iter 800/6416, lr 0.001000, loss 1.169752
+INFO 2021-02-26 10:25:30 train.py: 78] Epoch 14, iter 1000/6416, lr 0.001000, loss 1.178197
+INFO 2021-02-26 10:30:57 train.py: 78] Epoch 14, iter 1200/6416, lr 0.001000, loss 1.171137
+INFO 2021-02-26 10:36:24 train.py: 78] Epoch 14, iter 1400/6416, lr 0.001000, loss 1.177152
+INFO 2021-02-26 10:41:51 train.py: 78] Epoch 14, iter 1600/6416, lr 0.001000, loss 1.175818
+INFO 2021-02-26 10:47:17 train.py: 78] Epoch 14, iter 1800/6416, lr 0.001000, loss 1.201867
+INFO 2021-02-26 10:52:44 train.py: 78] Epoch 14, iter 2000/6416, lr 0.001000, loss 1.177558
+INFO 2021-02-26 10:58:11 train.py: 78] Epoch 14, iter 2200/6416, lr 0.001000, loss 1.184976
+INFO 2021-02-26 11:03:38 train.py: 78] Epoch 14, iter 2400/6416, lr 0.001000, loss 1.176128
+INFO 2021-02-26 11:09:05 train.py: 78] Epoch 14, iter 2600/6416, lr 0.001000, loss 1.182863
+INFO 2021-02-26 11:14:33 train.py: 78] Epoch 14, iter 2800/6416, lr 0.001000, loss 1.199468
+INFO 2021-02-26 11:19:59 train.py: 91] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-02-26 11:20:00 train.py: 78] Epoch 14, iter 3000/6416, lr 0.001000, loss 1.187565
+INFO 2021-02-26 11:25:28 train.py: 78] Epoch 14, iter 3200/6416, lr 0.001000, loss 1.174894
+INFO 2021-02-26 11:30:55 train.py: 78] Epoch 14, iter 3400/6416, lr 0.001000, loss 1.182255
+INFO 2021-02-26 11:36:23 train.py: 78] Epoch 14, iter 3600/6416, lr 0.001000, loss 1.182152
+INFO 2021-02-26 11:41:50 train.py: 78] Epoch 14, iter 3800/6416, lr 0.001000, loss 1.202862
+INFO 2021-02-26 11:47:17 train.py: 78] Epoch 14, iter 4000/6416, lr 0.001000, loss 1.190542
+INFO 2021-02-26 11:52:45 train.py: 78] Epoch 14, iter 4200/6416, lr 0.001000, loss 1.185105
+INFO 2021-02-26 11:58:12 train.py: 78] Epoch 14, iter 4400/6416, lr 0.001000, loss 1.184710
+INFO 2021-02-26 12:03:40 train.py: 78] Epoch 14, iter 4600/6416, lr 0.001000, loss 1.183482
+INFO 2021-02-26 12:09:07 train.py: 78] Epoch 14, iter 4800/6416, lr 0.001000, loss 1.187509
+INFO 2021-02-26 12:14:35 train.py: 78] Epoch 14, iter 5000/6416, lr 0.001000, loss 1.189766
+INFO 2021-02-26 12:20:02 train.py: 78] Epoch 14, iter 5200/6416, lr 0.001000, loss 1.188259
+INFO 2021-02-26 12:25:30 train.py: 78] Epoch 14, iter 5400/6416, lr 0.001000, loss 1.192548
+INFO 2021-02-26 12:30:57 train.py: 78] Epoch 14, iter 5600/6416, lr 0.001000, loss 1.185499
+INFO 2021-02-26 12:36:25 train.py: 78] Epoch 14, iter 5800/6416, lr 0.001000, loss 1.188568
+INFO 2021-02-26 12:41:51 train.py: 91] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-02-26 12:41:53 train.py: 78] Epoch 14, iter 6000/6416, lr 0.001000, loss 1.189552
+INFO 2021-02-26 12:47:21 train.py: 78] Epoch 14, iter 6200/6416, lr 0.001000, loss 1.181539
+INFO 2021-02-26 12:52:48 train.py: 78] Epoch 14, iter 6400/6416, lr 0.001000, loss 1.185364
+INFO 2021-02-26 12:53:12 train.py: 96] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-02-26 12:53:15 train.py: 78] Epoch 15, iter 0/6416, lr 0.001000, loss 1.213076
+INFO 2021-02-26 12:58:42 train.py: 78] Epoch 15, iter 200/6416, lr 0.001000, loss 1.142559
+INFO 2021-02-26 13:04:09 train.py: 78] Epoch 15, iter 400/6416, lr 0.001000, loss 1.171237
+INFO 2021-02-26 13:09:35 train.py: 78] Epoch 15, iter 600/6416, lr 0.001000, loss 1.154295
+INFO 2021-02-26 13:15:02 train.py: 78] Epoch 15, iter 800/6416, lr 0.001000, loss 1.154924
+INFO 2021-02-26 13:20:28 train.py: 78] Epoch 15, iter 1000/6416, lr 0.001000, loss 1.146455
+INFO 2021-02-26 13:25:55 train.py: 78] Epoch 15, iter 1200/6416, lr 0.001000, loss 1.150984
+INFO 2021-02-26 13:31:22 train.py: 78] Epoch 15, iter 1400/6416, lr 0.001000, loss 1.164542
+INFO 2021-02-26 13:36:48 train.py: 78] Epoch 15, iter 1600/6416, lr 0.001000, loss 1.146822
+INFO 2021-02-26 13:42:15 train.py: 78] Epoch 15, iter 1800/6416, lr 0.001000, loss 1.166525
+INFO 2021-02-26 13:47:42 train.py: 78] Epoch 15, iter 2000/6416, lr 0.001000, loss 1.159171
+INFO 2021-02-26 13:53:09 train.py: 78] Epoch 15, iter 2200/6416, lr 0.001000, loss 1.159972
+INFO 2021-02-26 13:58:36 train.py: 78] Epoch 15, iter 2400/6416, lr 0.001000, loss 1.156786
+INFO 2021-02-26 14:04:03 train.py: 78] Epoch 15, iter 2600/6416, lr 0.001000, loss 1.147948
+INFO 2021-02-26 14:09:31 train.py: 78] Epoch 15, iter 2800/6416, lr 0.001000, loss 1.150009
+INFO 2021-02-26 14:14:57 train.py: 91] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-02-26 14:14:58 train.py: 78] Epoch 15, iter 3000/6416, lr 0.001000, loss 1.157991
+INFO 2021-02-26 14:20:26 train.py: 78] Epoch 15, iter 3200/6416, lr 0.001000, loss 1.153339
+INFO 2021-02-26 14:25:53 train.py: 78] Epoch 15, iter 3400/6416, lr 0.001000, loss 1.164965
+INFO 2021-02-26 14:31:21 train.py: 78] Epoch 15, iter 3600/6416, lr 0.001000, loss 1.152254
+INFO 2021-02-26 14:36:48 train.py: 78] Epoch 15, iter 3800/6416, lr 0.001000, loss 1.155359
+INFO 2021-02-26 14:42:16 train.py: 78] Epoch 15, iter 4000/6416, lr 0.001000, loss 1.165722
+INFO 2021-02-26 14:47:43 train.py: 78] Epoch 15, iter 4200/6416, lr 0.001000, loss 1.168520
+INFO 2021-02-26 14:53:11 train.py: 78] Epoch 15, iter 4400/6416, lr 0.001000, loss 1.150805
+INFO 2021-02-26 14:58:39 train.py: 78] Epoch 15, iter 4600/6416, lr 0.001000, loss 1.170032
+INFO 2021-02-26 15:04:06 train.py: 78] Epoch 15, iter 4800/6416, lr 0.001000, loss 1.157200
+INFO 2021-02-26 15:09:34 train.py: 78] Epoch 15, iter 5000/6416, lr 0.001000, loss 1.170168
+INFO 2021-02-26 15:15:01 train.py: 78] Epoch 15, iter 5200/6416, lr 0.001000, loss 1.169286
+INFO 2021-02-26 15:20:29 train.py: 78] Epoch 15, iter 5400/6416, lr 0.001000, loss 1.155026
+INFO 2021-02-26 15:25:56 train.py: 78] Epoch 15, iter 5600/6416, lr 0.001000, loss 1.164509
+INFO 2021-02-26 15:31:24 train.py: 78] Epoch 15, iter 5800/6416, lr 0.001000, loss 1.159596
+INFO 2021-02-26 15:36:50 train.py: 91] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-02-26 15:36:52 train.py: 78] Epoch 15, iter 6000/6416, lr 0.001000, loss 1.173726
+INFO 2021-02-26 15:42:19 train.py: 78] Epoch 15, iter 6200/6416, lr 0.001000, loss 1.176526
+INFO 2021-02-26 15:47:47 train.py: 78] Epoch 15, iter 6400/6416, lr 0.001000, loss 1.163946
+INFO 2021-02-26 15:48:11 train.py: 96] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-02-26 15:48:14 train.py: 78] Epoch 16, iter 0/6416, lr 0.000100, loss 1.158254
+INFO 2021-02-26 15:53:41 train.py: 78] Epoch 16, iter 200/6416, lr 0.000100, loss 1.120436
+INFO 2021-02-26 15:59:07 train.py: 78] Epoch 16, iter 400/6416, lr 0.000100, loss 1.124684
+INFO 2021-02-26 16:04:34 train.py: 78] Epoch 16, iter 600/6416, lr 0.000100, loss 1.140950
+INFO 2021-02-26 16:10:01 train.py: 78] Epoch 16, iter 800/6416, lr 0.000100, loss 1.122044
+INFO 2021-02-26 16:15:27 train.py: 78] Epoch 16, iter 1000/6416, lr 0.000100, loss 1.135761
+INFO 2021-02-26 16:20:54 train.py: 78] Epoch 16, iter 1200/6416, lr 0.000100, loss 1.118623
+INFO 2021-02-26 16:26:21 train.py: 78] Epoch 16, iter 1400/6416, lr 0.000100, loss 1.137746
+INFO 2021-02-26 16:31:48 train.py: 78] Epoch 16, iter 1600/6416, lr 0.000100, loss 1.117278
+INFO 2021-02-26 16:37:14 train.py: 78] Epoch 16, iter 1800/6416, lr 0.000100, loss 1.122779
+INFO 2021-02-26 16:42:41 train.py: 78] Epoch 16, iter 2000/6416, lr 0.000100, loss 1.112222
+INFO 2021-02-26 16:48:08 train.py: 78] Epoch 16, iter 2200/6416, lr 0.000100, loss 1.114890
+INFO 2021-02-26 16:53:36 train.py: 78] Epoch 16, iter 2400/6416, lr 0.000100, loss 1.127713
+INFO 2021-02-26 16:59:03 train.py: 78] Epoch 16, iter 2600/6416, lr 0.000100, loss 1.123855
+INFO 2021-02-26 17:04:30 train.py: 78] Epoch 16, iter 2800/6416, lr 0.000100, loss 1.130603
+INFO 2021-02-26 17:09:56 train.py: 91] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-02-26 17:09:58 train.py: 78] Epoch 16, iter 3000/6416, lr 0.000100, loss 1.123724
+INFO 2021-02-26 17:15:25 train.py: 78] Epoch 16, iter 3200/6416, lr 0.000100, loss 1.118672
+INFO 2021-02-26 17:20:53 train.py: 78] Epoch 16, iter 3400/6416, lr 0.000100, loss 1.126539
+INFO 2021-02-26 17:26:20 train.py: 78] Epoch 16, iter 3600/6416, lr 0.000100, loss 1.126895
+INFO 2021-02-26 17:31:47 train.py: 78] Epoch 16, iter 3800/6416, lr 0.000100, loss 1.132515
+INFO 2021-02-26 17:37:15 train.py: 78] Epoch 16, iter 4000/6416, lr 0.000100, loss 1.122580
+INFO 2021-02-26 17:42:42 train.py: 78] Epoch 16, iter 4200/6416, lr 0.000100, loss 1.127590
+INFO 2021-02-26 17:48:10 train.py: 78] Epoch 16, iter 4400/6416, lr 0.000100, loss 1.119929
+INFO 2021-02-26 17:53:37 train.py: 78] Epoch 16, iter 4600/6416, lr 0.000100, loss 1.117680
+INFO 2021-02-26 17:59:05 train.py: 78] Epoch 16, iter 4800/6416, lr 0.000100, loss 1.123241
+INFO 2021-02-26 18:04:32 train.py: 78] Epoch 16, iter 5000/6416, lr 0.000100, loss 1.125865
+INFO 2021-02-26 18:10:00 train.py: 78] Epoch 16, iter 5200/6416, lr 0.000100, loss 1.121079
+INFO 2021-02-26 18:15:27 train.py: 78] Epoch 16, iter 5400/6416, lr 0.000100, loss 1.118097
+INFO 2021-02-26 18:20:55 train.py: 78] Epoch 16, iter 5600/6416, lr 0.000100, loss 1.114844
+INFO 2021-02-26 18:26:22 train.py: 78] Epoch 16, iter 5800/6416, lr 0.000100, loss 1.125234
+INFO 2021-02-26 18:31:49 train.py: 91] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-02-26 18:31:50 train.py: 78] Epoch 16, iter 6000/6416, lr 0.000100, loss 1.118507
+INFO 2021-02-26 18:37:18 train.py: 78] Epoch 16, iter 6200/6416, lr 0.000100, loss 1.116235
+INFO 2021-02-26 18:42:46 train.py: 78] Epoch 16, iter 6400/6416, lr 0.000100, loss 1.107636
+INFO 2021-02-26 18:43:10 train.py: 96] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-02-26 18:43:12 train.py: 78] Epoch 17, iter 0/6416, lr 0.000100, loss 1.136899
+INFO 2021-02-26 18:48:39 train.py: 78] Epoch 17, iter 200/6416, lr 0.000100, loss 1.116615
+INFO 2021-02-26 18:54:06 train.py: 78] Epoch 17, iter 400/6416, lr 0.000100, loss 1.118499
+INFO 2021-02-26 18:59:33 train.py: 78] Epoch 17, iter 600/6416, lr 0.000100, loss 1.118284
+INFO 2021-02-26 19:04:59 train.py: 78] Epoch 17, iter 800/6416, lr 0.000100, loss 1.121926
+INFO 2021-02-26 19:10:26 train.py: 78] Epoch 17, iter 1000/6416, lr 0.000100, loss 1.118190
+INFO 2021-02-26 19:15:53 train.py: 78] Epoch 17, iter 1200/6416, lr 0.000100, loss 1.123708
+INFO 2021-02-26 19:21:20 train.py: 78] Epoch 17, iter 1400/6416, lr 0.000100, loss 1.118808
+INFO 2021-02-26 19:26:47 train.py: 78] Epoch 17, iter 1600/6416, lr 0.000100, loss 1.123949
+INFO 2021-02-26 19:32:14 train.py: 78] Epoch 17, iter 1800/6416, lr 0.000100, loss 1.122720
+INFO 2021-02-26 19:37:41 train.py: 78] Epoch 17, iter 2000/6416, lr 0.000100, loss 1.119571
+INFO 2021-02-26 19:43:08 train.py: 78] Epoch 17, iter 2200/6416, lr 0.000100, loss 1.112629
+INFO 2021-02-26 19:48:35 train.py: 78] Epoch 17, iter 2400/6416, lr 0.000100, loss 1.120738
+INFO 2021-02-26 19:54:02 train.py: 78] Epoch 17, iter 2600/6416, lr 0.000100, loss 1.121630
+INFO 2021-02-26 19:59:29 train.py: 78] Epoch 17, iter 2800/6416, lr 0.000100, loss 1.128484
+INFO 2021-02-26 20:04:55 train.py: 91] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-02-26 20:04:57 train.py: 78] Epoch 17, iter 3000/6416, lr 0.000100, loss 1.116445
+INFO 2021-02-26 20:10:25 train.py: 78] Epoch 17, iter 3200/6416, lr 0.000100, loss 1.121110
+INFO 2021-02-26 20:15:52 train.py: 78] Epoch 17, iter 3400/6416, lr 0.000100, loss 1.128289
+INFO 2021-02-26 20:21:19 train.py: 78] Epoch 17, iter 3600/6416, lr 0.000100, loss 1.110137
+INFO 2021-02-26 20:26:47 train.py: 78] Epoch 17, iter 3800/6416, lr 0.000100, loss 1.122043
+INFO 2021-02-26 20:32:14 train.py: 78] Epoch 17, iter 4000/6416, lr 0.000100, loss 1.120359
+INFO 2021-02-26 20:37:42 train.py: 78] Epoch 17, iter 4200/6416, lr 0.000100, loss 1.109291
+INFO 2021-02-26 20:43:09 train.py: 78] Epoch 17, iter 4400/6416, lr 0.000100, loss 1.123513
+INFO 2021-02-26 20:48:37 train.py: 78] Epoch 17, iter 4600/6416, lr 0.000100, loss 1.116357
+INFO 2021-02-26 20:54:04 train.py: 78] Epoch 17, iter 4800/6416, lr 0.000100, loss 1.134884
+INFO 2021-02-26 20:59:31 train.py: 78] Epoch 17, iter 5000/6416, lr 0.000100, loss 1.121153
+INFO 2021-02-26 21:04:59 train.py: 78] Epoch 17, iter 5200/6416, lr 0.000100, loss 1.126199
+INFO 2021-02-26 21:10:26 train.py: 78] Epoch 17, iter 5400/6416, lr 0.000100, loss 1.130628
+INFO 2021-02-26 21:15:54 train.py: 78] Epoch 17, iter 5600/6416, lr 0.000100, loss 1.118594
+INFO 2021-02-26 21:21:22 train.py: 78] Epoch 17, iter 5800/6416, lr 0.000100, loss 1.104147
+INFO 2021-02-26 21:26:48 train.py: 91] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-02-26 21:26:50 train.py: 78] Epoch 17, iter 6000/6416, lr 0.000100, loss 1.119507
+INFO 2021-02-26 21:32:17 train.py: 78] Epoch 17, iter 6200/6416, lr 0.000100, loss 1.104828
+INFO 2021-02-26 21:37:45 train.py: 78] Epoch 17, iter 6400/6416, lr 0.000100, loss 1.123017
+INFO 2021-02-26 21:38:09 train.py: 96] Save checkpoint Epoch_17.pt to disk...
+INFO 2021-02-26 21:38:10 train.py: 179] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/.gitkeep b/bob/bio/facexzoo/models/backbones/ResNet152_irse/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3d21c6d25dc492bacc8123c474b9d2ed709978f8
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9813333333333334 | 0.0025555555555555535 |
+| Epoch_17_batch_5999.pt | 0.9810000000000001 | 0.0026434171674156273 |
+| Epoch_16.pt | 0.9803333333333335 | 0.002696591355447021 |
+| Epoch_15_batch_5999.pt | 0.9803333333333333 | 0.003010270485365345 |
+| Epoch_12_batch_5999.pt | 0.9801666666666667 | 0.002514771177279177 |
+| Epoch_13_batch_2999.pt | 0.9798333333333333 | 0.0025873624493766723 |
+| Epoch_15_batch_2999.pt | 0.9798333333333333 | 0.0029757248312982424 |
+| Epoch_15.pt | 0.9798333333333333 | 0.0026579719364234807 |
+| Epoch_16_batch_5999.pt | 0.9798333333333333 | 0.0029549080326604898 |
+| Epoch_14.pt | 0.9796666666666667 | 0.0025915341754867926 |
+| Epoch_11_batch_5999.pt | 0.9795 | 0.001988392240908136 |
+| Epoch_13_batch_5999.pt | 0.9793333333333335 | 0.0028240588949197407 |
+| Epoch_14_batch_2999.pt | 0.9790000000000001 | 0.0027577052546646327 |
+| Epoch_14_batch_5999.pt | 0.9790000000000001 | 0.0027577052546646327 |
+| Epoch_17.pt | 0.9789999999999999 | 0.00293131243517176 |
+| Epoch_12_batch_2999.pt | 0.9788333333333334 | 0.002166666666666665 |
+| Epoch_17_batch_2999.pt | 0.9788333333333334 | 0.002756025945513878 |
+| Epoch_13.pt | 0.9786666666666669 | 0.0029585615457098555 |
+| Epoch_10_batch_2999.pt | 0.9781666666666666 | 0.002692582403567246 |
+| Epoch_11.pt | 0.9781666666666666 | 0.0024273391416614986 |
+| Epoch_11_batch_2999.pt | 0.9780000000000001 | 0.0023544022333796717 |
+| Epoch_12.pt | 0.9778333333333332 | 0.002897955856832214 |
+| Epoch_10_batch_5999.pt | 0.9768333333333334 | 0.002055555555555559 |
+| Epoch_10.pt | 0.9765 | 0.0023888888888888883 |
+| Epoch_7_batch_5999.pt | 0.9729999999999999 | 0.002469567863432539 |
+| Epoch_9_batch_5999.pt | 0.9726666666666667 | 0.0014740554623801753 |
+| Epoch_9_batch_2999.pt | 0.9713333333333333 | 0.002567604446286962 |
+| Epoch_6_batch_2999.pt | 0.9708333333333332 | 0.0020971762320196536 |
+| Epoch_8_batch_5999.pt | 0.9706666666666667 | 0.0030852096393144002 |
+| Epoch_8_batch_2999.pt | 0.9701666666666668 | 0.002622904807580643 |
+| Epoch_5.pt | 0.9698333333333334 | 0.0016187558093703899 |
+| Epoch_7.pt | 0.9696666666666667 | 0.002045772515502442 |
+| Epoch_6_batch_5999.pt | 0.9691666666666668 | 0.0022532555322970246 |
+| Epoch_3_batch_5999.pt | 0.9686666666666668 | 0.002719386277893431 |
+| Epoch_8.pt | 0.9685 | 0.0027267535981795677 |
+| Epoch_5_batch_2999.pt | 0.9681666666666665 | 0.0024526377838493454 |
+| Epoch_5_batch_5999.pt | 0.968 | 0.0030510067150546594 |
+| Epoch_4_batch_5999.pt | 0.9678333333333333 | 0.002422247706287962 |
+| Epoch_7_batch_2999.pt | 0.9676666666666668 | 0.0023985592383247655 |
+| Epoch_9.pt | 0.9675 | 0.0031254629286744965 |
+| Epoch_4_batch_2999.pt | 0.9648333333333333 | 0.0028158501994387177 |
+| Epoch_4.pt | 0.9643333333333335 | 0.002372684056006958 |
+| Epoch_6.pt | 0.9636666666666667 | 0.0024444444444444414 |
+| Epoch_3_batch_2999.pt | 0.9628333333333334 | 0.002629955639676581 |
+| Epoch_3.pt | 0.9628333333333334 | 0.0027108606441679727 |
+| Epoch_2_batch_5999.pt | 0.9613333333333334 | 0.0031210159789307017 |
+| Epoch_2.pt | 0.9565000000000001 | 0.0029963970133692875 |
+| Epoch_2_batch_2999.pt | 0.9561666666666667 | 0.003575033454043592 |
+| Epoch_1_batch_5999.pt | 0.9514999999999999 | 0.0028808649249920386 |
+| Epoch_1.pt | 0.9503333333333334 | 0.0038232556742411657 |
+| Epoch_1_batch_2999.pt | 0.9316666666666666 | 0.004849589520621147 |
+| Epoch_0.pt | 0.8753333333333334 | 0.005873144571734449 |
+| Epoch_0_batch_5999.pt | 0.8705 | 0.007762087348130007 |
+| Epoch_0_batch_2999.pt | 0.6849999999999999 | 0.00524698620138559 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6529b50e5ded48f011658789480708378876fbda
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.9555000000000001 | 0.0035228530805528255 |
+| Epoch_12_batch_5999.pt | 0.9553333333333333 | 0.0033035708327374532 |
+| Epoch_12_batch_2999.pt | 0.9551666666666667 | 0.0036552853665768868 |
+| Epoch_13_batch_2999.pt | 0.9550000000000001 | 0.003415650255319869 |
+| Epoch_15_batch_5999.pt | 0.9550000000000001 | 0.0035048467323527824 |
+| Epoch_15.pt | 0.9550000000000001 | 0.003469443332443559 |
+| Epoch_17_batch_2999.pt | 0.9550000000000001 | 0.003451605459335352 |
+| Epoch_13.pt | 0.9548333333333334 | 0.003482318654269964 |
+| Epoch_16_batch_2999.pt | 0.9548333333333334 | 0.0033834199952411386 |
+| Epoch_11_batch_5999.pt | 0.9546666666666669 | 0.0036666666666666636 |
+| Epoch_13_batch_5999.pt | 0.9546666666666667 | 0.0035642255405212153 |
+| Epoch_14_batch_2999.pt | 0.9546666666666667 | 0.003782676562344275 |
+| Epoch_14_batch_5999.pt | 0.9546666666666667 | 0.0036158089100606665 |
+| Epoch_12.pt | 0.9545 | 0.003478771600989619 |
+| Epoch_15_batch_2999.pt | 0.9545 | 0.003592258480072717 |
+| Epoch_16_batch_5999.pt | 0.9545 | 0.0035663897434897846 |
+| Epoch_17_batch_5999.pt | 0.9545 | 0.0038333333333333388 |
+| Epoch_16.pt | 0.9543333333333335 | 0.0033351846710674804 |
+| Epoch_11_batch_2999.pt | 0.9541666666666668 | 0.0036955929710139096 |
+| Epoch_11.pt | 0.9540000000000001 | 0.0035676876351116303 |
+| Epoch_10.pt | 0.9538333333333334 | 0.003668770440244244 |
+| Epoch_17.pt | 0.9538333333333334 | 0.0032683480177961646 |
+| Epoch_10_batch_2999.pt | 0.9536666666666667 | 0.003581502546952496 |
+| Epoch_10_batch_5999.pt | 0.9531666666666666 | 0.0036213529972738425 |
+| Epoch_9_batch_5999.pt | 0.9511666666666667 | 0.0033152286106301497 |
+| Epoch_8_batch_5999.pt | 0.9503333333333334 | 0.004065816547451553 |
+| Epoch_7_batch_2999.pt | 0.9491666666666667 | 0.0038268866887321102 |
+| Epoch_8_batch_2999.pt | 0.9490000000000001 | 0.003567687635111631 |
+| Epoch_8.pt | 0.9481666666666667 | 0.0037634532343199896 |
+| Epoch_5_batch_2999.pt | 0.9481666666666666 | 0.003672133971035721 |
+| Epoch_9_batch_2999.pt | 0.9473333333333335 | 0.00271256791460748 |
+| Epoch_6_batch_5999.pt | 0.9470000000000001 | 0.003130889511912304 |
+| Epoch_6.pt | 0.9469999999999998 | 0.0037333994703136505 |
+| Epoch_5.pt | 0.9465 | 0.004070747800382749 |
+| Epoch_6_batch_2999.pt | 0.9463333333333332 | 0.0032087842395985876 |
+| Epoch_7_batch_5999.pt | 0.9454999999999998 | 0.0036687704402442313 |
+| Epoch_3.pt | 0.945 | 0.00405973909032113 |
+| Epoch_4_batch_5999.pt | 0.945 | 0.003676753801727619 |
+| Epoch_7.pt | 0.945 | 0.003181738014061412 |
+| Epoch_4_batch_2999.pt | 0.9448333333333332 | 0.0029963970133692844 |
+| Epoch_4.pt | 0.9443333333333334 | 0.004319064827905955 |
+| Epoch_5_batch_5999.pt | 0.9441666666666666 | 0.0033170900530062612 |
+| Epoch_9.pt | 0.9434999999999999 | 0.0036298658275376963 |
+| Epoch_3_batch_5999.pt | 0.9433333333333334 | 0.003617515688022157 |
+| Epoch_3_batch_2999.pt | 0.9425000000000001 | 0.004446180216603749 |
+| Epoch_2_batch_5999.pt | 0.9404999999999999 | 0.0037271940261598873 |
+| Epoch_2.pt | 0.9393333333333335 | 0.0030449310235654884 |
+| Epoch_2_batch_2999.pt | 0.9381666666666666 | 0.0037222222222222244 |
+| Epoch_1_batch_5999.pt | 0.9325000000000001 | 0.003524604872347089 |
+| Epoch_1.pt | 0.9285 | 0.004788424829281446 |
+| Epoch_1_batch_2999.pt | 0.9228333333333334 | 0.004444791653104356 |
+| Epoch_0.pt | 0.883 | 0.005004935835357867 |
+| Epoch_0_batch_5999.pt | 0.8643333333333333 | 0.003584947956644817 |
+| Epoch_0_batch_2999.pt | 0.6809999999999999 | 0.0062321349610578464 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..41e91bd0b279754471107992e360f33c427b54f4
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.8971666666666666 | 0.00495691311748764 |
+| Epoch_13.pt | 0.8968333333333334 | 0.004877828397611301 |
+| Epoch_15_batch_2999.pt | 0.8965 | 0.005640735877043695 |
+| Epoch_13_batch_5999.pt | 0.8963333333333333 | 0.0050844716393989116 |
+| Epoch_11_batch_5999.pt | 0.8961666666666668 | 0.004944444444444445 |
+| Epoch_16.pt | 0.8961666666666668 | 0.0055890656035680695 |
+| Epoch_15_batch_5999.pt | 0.8958333333333334 | 0.005512331854046509 |
+| Epoch_14.pt | 0.8956666666666667 | 0.005573304979548648 |
+| Epoch_14_batch_5999.pt | 0.8955 | 0.005863940865078349 |
+| Epoch_12.pt | 0.8953333333333335 | 0.00519852780679537 |
+| Epoch_14_batch_2999.pt | 0.8953333333333335 | 0.005379430416404531 |
+| Epoch_17_batch_5999.pt | 0.8953333333333333 | 0.00587839736284947 |
+| Epoch_16_batch_2999.pt | 0.8951666666666668 | 0.005668028158881761 |
+| Epoch_17.pt | 0.8950000000000001 | 0.005561108336107648 |
+| Epoch_12_batch_2999.pt | 0.8946666666666667 | 0.006145358581788812 |
+| Epoch_16_batch_5999.pt | 0.8946666666666667 | 0.00571331561430331 |
+| Epoch_17_batch_2999.pt | 0.8946666666666665 | 0.005783117190965826 |
+| Epoch_10.pt | 0.8945000000000001 | 0.005550275268448907 |
+| Epoch_12_batch_5999.pt | 0.8939999999999999 | 0.006030785220049735 |
+| Epoch_11.pt | 0.8934999999999998 | 0.005290627632024459 |
+| Epoch_15.pt | 0.8931666666666667 | 0.0056734708604996124 |
+| Epoch_11_batch_2999.pt | 0.8925000000000001 | 0.005484264808871577 |
+| Epoch_10_batch_5999.pt | 0.8903333333333334 | 0.005447844743169452 |
+| Epoch_10_batch_2999.pt | 0.8871666666666667 | 0.0050494468588397744 |
+| Epoch_9_batch_2999.pt | 0.8666666666666668 | 0.006295206877429018 |
+| Epoch_6_batch_2999.pt | 0.8665 | 0.0059340511425975155 |
+| Epoch_4_batch_5999.pt | 0.8651666666666668 | 0.005743756674924583 |
+| Epoch_9_batch_5999.pt | 0.8644999999999999 | 0.006587727085129446 |
+| Epoch_7_batch_2999.pt | 0.8640000000000001 | 0.005358735800794444 |
+| Epoch_5_batch_2999.pt | 0.8636666666666667 | 0.004477653706579981 |
+| Epoch_6_batch_5999.pt | 0.8633333333333333 | 0.005091750772173148 |
+| Epoch_5.pt | 0.8629999999999999 | 0.006434858771364556 |
+| Epoch_7.pt | 0.8621666666666667 | 0.005957928631684579 |
+| Epoch_3_batch_5999.pt | 0.8619999999999999 | 0.00703957069398096 |
+| Epoch_8_batch_5999.pt | 0.8616666666666667 | 0.006662035428409452 |
+| Epoch_2_batch_5999.pt | 0.8601666666666666 | 0.006032064527932027 |
+| Epoch_8_batch_2999.pt | 0.8596666666666668 | 0.0048673769141798886 |
+| Epoch_8.pt | 0.8583333333333334 | 0.006196374886043859 |
+| Epoch_7_batch_5999.pt | 0.8578333333333334 | 0.0051523098020495985 |
+| Epoch_5_batch_5999.pt | 0.857 | 0.005576626707257601 |
+| Epoch_4_batch_2999.pt | 0.8551666666666666 | 0.006976371054238714 |
+| Epoch_6.pt | 0.8546666666666667 | 0.007083278866892988 |
+| Epoch_3.pt | 0.8545 | 0.006440851488835504 |
+| Epoch_4.pt | 0.8543333333333335 | 0.005806552280201743 |
+| Epoch_9.pt | 0.8526666666666667 | 0.006667592528301102 |
+| Epoch_3_batch_2999.pt | 0.8526666666666666 | 0.005693834655697309 |
+| Epoch_2_batch_2999.pt | 0.8458333333333334 | 0.003745367509040703 |
+| Epoch_1_batch_5999.pt | 0.8423333333333334 | 0.005595412583117408 |
+| Epoch_2.pt | 0.8401666666666665 | 0.006188649506641124 |
+| Epoch_1.pt | 0.8381666666666666 | 0.007198465199927541 |
+| Epoch_1_batch_2999.pt | 0.8153333333333332 | 0.006357653111519986 |
+| Epoch_0.pt | 0.7558333333333334 | 0.006827911127890455 |
+| Epoch_0_batch_5999.pt | 0.7536666666666667 | 0.007118052168020876 |
+| Epoch_0_batch_2999.pt | 0.6213333333333333 | 0.007229910437776639 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..65e2c0bed80184c1d3493bb2df6501b53e3bea7c
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11_batch_2999.pt | 0.9984999999999999 | 0.000524110062892033 |
+| Epoch_13_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_13_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_13.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_14_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_15_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_15.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_16_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_16_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_16.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_17_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_17_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_17.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_11.pt | 0.9983333333333333 | 0.000608580619450185 |
+| Epoch_10_batch_5999.pt | 0.9981666666666668 | 0.0005800170282728054 |
+| Epoch_10.pt | 0.9981666666666668 | 0.0005800170282728054 |
+| Epoch_12_batch_2999.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_12.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_10_batch_2999.pt | 0.998 | 0.000544331053951814 |
+| Epoch_11_batch_5999.pt | 0.998 | 0.0007370277311900908 |
+| Epoch_12_batch_5999.pt | 0.998 | 0.0005983516452371637 |
+| Epoch_14_batch_2999.pt | 0.9978333333333333 | 0.0008258927081843653 |
+| Epoch_15_batch_5999.pt | 0.9978333333333333 | 0.0007049209744694192 |
+| Epoch_7.pt | 0.9976666666666667 | 0.0007535922203472518 |
+| Epoch_14.pt | 0.9976666666666667 | 0.0008678055195451877 |
+| Epoch_8_batch_5999.pt | 0.9973333333333333 | 0.0009686442096757044 |
+| Epoch_9_batch_2999.pt | 0.9973333333333333 | 0.000753592220347253 |
+| Epoch_8_batch_2999.pt | 0.9971666666666668 | 0.0009638528651609697 |
+| Epoch_7_batch_2999.pt | 0.9971666666666665 | 0.0009953596037316056 |
+| Epoch_4.pt | 0.9969999999999999 | 0.0007370277311900913 |
+| Epoch_3.pt | 0.9968333333333333 | 0.0009111788592698171 |
+| Epoch_6_batch_2999.pt | 0.9968333333333333 | 0.0008766518798921921 |
+| Epoch_6.pt | 0.9968333333333333 | 0.0008031573497111664 |
+| Epoch_9_batch_5999.pt | 0.9968333333333333 | 0.0008407081083567489 |
+| Epoch_2_batch_2999.pt | 0.9964999999999999 | 0.0007637626158259766 |
+| Epoch_3_batch_2999.pt | 0.9964999999999999 | 0.0009765775461803867 |
+| Epoch_6_batch_5999.pt | 0.9964999999999999 | 0.0012533904636309393 |
+| Epoch_8.pt | 0.9964999999999999 | 0.0006309898162000248 |
+| Epoch_2_batch_5999.pt | 0.9963333333333333 | 0.0010183501544346332 |
+| Epoch_4_batch_2999.pt | 0.9963333333333333 | 0.0012372809695177837 |
+| Epoch_9.pt | 0.9961666666666666 | 0.0009953596037316098 |
+| Epoch_1_batch_5999.pt | 0.9959999999999999 | 0.0009686442096757069 |
+| Epoch_3_batch_5999.pt | 0.9959999999999999 | 0.0009026709338484409 |
+| Epoch_7_batch_5999.pt | 0.9959999999999999 | 0.0011439589045541155 |
+| Epoch_5_batch_2999.pt | 0.9958333333333332 | 0.0012729376930432825 |
+| Epoch_1.pt | 0.9956666666666667 | 0.0009026709338484367 |
+| Epoch_2.pt | 0.9956666666666667 | 0.0009026709338484435 |
+| Epoch_5.pt | 0.9956666666666667 | 0.0009362388636862605 |
+| Epoch_4_batch_5999.pt | 0.9953333333333333 | 0.0011600340565456116 |
+| Epoch_5_batch_5999.pt | 0.9953333333333333 | 0.0011331154474650662 |
+| Epoch_1_batch_2999.pt | 0.9945 | 0.0010258991840344132 |
+| Epoch_0.pt | 0.9848333333333332 | 0.0019317042945237446 |
+| Epoch_0_batch_5999.pt | 0.984 | 0.002701165729142936 |
+| Epoch_0_batch_2999.pt | 0.9145000000000001 | 0.005533567598789261 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..90d6de2396abfcd7db92afb43ab88cb160a96c90
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_12_batch_5999.pt | 0.9585000000000001 | 0.002114762923408262 |
+| Epoch_15_batch_5999.pt | 0.9581666666666667 | 0.002870131411797666 |
+| Epoch_17_batch_2999.pt | 0.9576666666666668 | 0.002701165729142934 |
+| Epoch_16.pt | 0.9573333333333334 | 0.002824058894919745 |
+| Epoch_11.pt | 0.9570000000000001 | 0.0031991511219751087 |
+| Epoch_16_batch_2999.pt | 0.9568333333333335 | 0.0028158501994387094 |
+| Epoch_16_batch_5999.pt | 0.9568333333333335 | 0.0026463345252921264 |
+| Epoch_14_batch_5999.pt | 0.9566666666666667 | 0.0030832082056692542 |
+| Epoch_12_batch_2999.pt | 0.9566666666666664 | 0.0030731814857643033 |
+| Epoch_15.pt | 0.9565000000000001 | 0.0028915585954829154 |
+| Epoch_15_batch_2999.pt | 0.9560000000000001 | 0.0025962936545662106 |
+| Epoch_17_batch_5999.pt | 0.9558333333333333 | 0.002548904388212106 |
+| Epoch_14_batch_2999.pt | 0.9556666666666669 | 0.002867441755680878 |
+| Epoch_10_batch_5999.pt | 0.9556666666666667 | 0.0028349668493717955 |
+| Epoch_14.pt | 0.9555 | 0.002756025945513882 |
+| Epoch_17.pt | 0.9555 | 0.0028114624286399987 |
+| Epoch_11_batch_2999.pt | 0.9551666666666666 | 0.0027267535981795677 |
+| Epoch_10.pt | 0.9550000000000001 | 0.00262936879248872 |
+| Epoch_13_batch_5999.pt | 0.9550000000000001 | 0.003113094605804871 |
+| Epoch_13_batch_2999.pt | 0.9548333333333332 | 0.002539198862672544 |
+| Epoch_13.pt | 0.9546666666666667 | 0.0030510067150546646 |
+| Epoch_11_batch_5999.pt | 0.9545 | 0.0022089883724523405 |
+| Epoch_10_batch_2999.pt | 0.9520000000000002 | 0.0036158089100606656 |
+| Epoch_12.pt | 0.9506666666666665 | 0.0023200681130912263 |
+| Epoch_8_batch_2999.pt | 0.9353333333333333 | 0.003020506048681821 |
+| Epoch_9_batch_2999.pt | 0.9336666666666668 | 0.0031111111111111096 |
+| Epoch_8_batch_5999.pt | 0.9321666666666667 | 0.004105476619298289 |
+| Epoch_7_batch_5999.pt | 0.9285 | 0.0050396575427451495 |
+| Epoch_9_batch_5999.pt | 0.9283333333333333 | 0.004059739090321137 |
+| Epoch_6.pt | 0.9261666666666667 | 0.0033888888888888914 |
+| Epoch_5.pt | 0.9260000000000002 | 0.005410323921751044 |
+| Epoch_7_batch_2999.pt | 0.9251666666666665 | 0.003318950451387522 |
+| Epoch_4_batch_5999.pt | 0.9245000000000001 | 0.004142895152778256 |
+| Epoch_4_batch_2999.pt | 0.9235 | 0.0033002992756175904 |
+| Epoch_6_batch_5999.pt | 0.9231666666666667 | 0.0034016154477425845 |
+| Epoch_5_batch_2999.pt | 0.9226666666666666 | 0.003818408950542123 |
+| Epoch_6_batch_2999.pt | 0.9219999999999999 | 0.004170738750916325 |
+| Epoch_8.pt | 0.9198333333333334 | 0.003047464033756107 |
+| Epoch_7.pt | 0.9191666666666667 | 0.004383259399460977 |
+| Epoch_3_batch_5999.pt | 0.9176666666666666 | 0.0041514537093932016 |
+| Epoch_9.pt | 0.9163333333333334 | 0.003887301263230207 |
+| Epoch_3_batch_2999.pt | 0.9159999999999998 | 0.004007708621526988 |
+| Epoch_3.pt | 0.9153333333333332 | 0.0033774853674601504 |
+| Epoch_5_batch_5999.pt | 0.9136666666666665 | 0.0052340292059172835 |
+| Epoch_2.pt | 0.9121666666666666 | 0.0045409657970932075 |
+| Epoch_4.pt | 0.9105000000000001 | 0.005140315117400181 |
+| Epoch_2_batch_5999.pt | 0.9091666666666667 | 0.003703935177951839 |
+| Epoch_1_batch_5999.pt | 0.8968333333333334 | 0.0029128281619818504 |
+| Epoch_2_batch_2999.pt | 0.8956666666666667 | 0.005271633850789535 |
+| Epoch_1.pt | 0.8773333333333333 | 0.004114113018127469 |
+| Epoch_1_batch_2999.pt | 0.8480000000000001 | 0.004660048216780169 |
+| Epoch_0_batch_5999.pt | 0.766 | 0.005562218227015034 |
+| Epoch_0.pt | 0.7656666666666666 | 0.007180219742846001 |
+| Epoch_0_batch_2999.pt | 0.588 | 0.010020965676341762 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..88bdd36f6fe13a56dbc3d6a2f469333eb429747c
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9543333333333335 | 0.002712567914607495 |
+| Epoch_16_batch_2999.pt | 0.9540000000000001 | 0.0028566578071516557 |
+| Epoch_14_batch_2999.pt | 0.954 | 0.0029418227321941636 |
+| Epoch_13_batch_5999.pt | 0.9536666666666666 | 0.0027532248207475327 |
+| Epoch_13.pt | 0.9535 | 0.0026810952248919277 |
+| Epoch_16.pt | 0.9535 | 0.002715410979966658 |
+| Epoch_12_batch_5999.pt | 0.9531666666666666 | 0.0022005891467042583 |
+| Epoch_14_batch_5999.pt | 0.9531666666666666 | 0.0030776975521032367 |
+| Epoch_14.pt | 0.953 | 0.0029270977494043372 |
+| Epoch_15_batch_2999.pt | 0.9528333333333334 | 0.002833333333333332 |
+| Epoch_17_batch_5999.pt | 0.9528333333333332 | 0.0030332519321597125 |
+| Epoch_15_batch_5999.pt | 0.9525 | 0.0031841621957571357 |
+| Epoch_15.pt | 0.9525 | 0.0025609845714702514 |
+| Epoch_17_batch_2999.pt | 0.9523333333333334 | 0.003055050463303896 |
+| Epoch_16_batch_5999.pt | 0.9518333333333333 | 0.0029339435392246424 |
+| Epoch_17.pt | 0.9515 | 0.0028048679025254124 |
+| Epoch_11.pt | 0.9503333333333333 | 0.002494438257849298 |
+| Epoch_12_batch_2999.pt | 0.9501666666666667 | 0.002527014536787585 |
+| Epoch_11_batch_5999.pt | 0.9498333333333331 | 0.0021293075440818763 |
+| Epoch_12.pt | 0.9496666666666667 | 0.0025312857221987824 |
+| Epoch_10_batch_5999.pt | 0.9490000000000001 | 0.002436856911051256 |
+| Epoch_10.pt | 0.9481666666666666 | 0.0027716599296147572 |
+| Epoch_10_batch_2999.pt | 0.9456666666666667 | 0.0015355861067872512 |
+| Epoch_11_batch_2999.pt | 0.9441666666666666 | 0.0026323017201071385 |
+| Epoch_9_batch_5999.pt | 0.9259999999999999 | 0.0031348302177035244 |
+| Epoch_8_batch_5999.pt | 0.9231666666666667 | 0.003097689302100148 |
+| Epoch_7_batch_5999.pt | 0.923 | 0.002863133050383362 |
+| Epoch_8_batch_2999.pt | 0.9221666666666668 | 0.0036094013046179493 |
+| Epoch_5_batch_2999.pt | 0.922 | 0.0025555555555555596 |
+| Epoch_5.pt | 0.9215 | 0.0038204291659898297 |
+| Epoch_6_batch_2999.pt | 0.9209999999999999 | 0.00347122207818074 |
+| Epoch_7_batch_2999.pt | 0.9208333333333334 | 0.004254627110787955 |
+| Epoch_9_batch_2999.pt | 0.9199999999999999 | 0.003998456492321463 |
+| Epoch_4_batch_5999.pt | 0.9175000000000001 | 0.0031451510788262677 |
+| Epoch_6_batch_5999.pt | 0.917 | 0.0028846122190549287 |
+| Epoch_8.pt | 0.917 | 0.004058218303944369 |
+| Epoch_3_batch_5999.pt | 0.9163333333333332 | 0.003511884584284245 |
+| Epoch_3_batch_2999.pt | 0.9146666666666666 | 0.003349958540373629 |
+| Epoch_4_batch_2999.pt | 0.9138333333333334 | 0.003913020367320081 |
+| Epoch_7.pt | 0.9125 | 0.003515837184954837 |
+| Epoch_6.pt | 0.9118333333333333 | 0.003629865827537696 |
+| Epoch_5_batch_5999.pt | 0.9101666666666667 | 0.0025754059969998306 |
+| Epoch_2_batch_5999.pt | 0.9096666666666666 | 0.004784233364802442 |
+| Epoch_3.pt | 0.9011666666666667 | 0.004805154125815265 |
+| Epoch_2.pt | 0.8965 | 0.0029860788111948223 |
+| Epoch_9.pt | 0.8965 | 0.0039003798485484713 |
+| Epoch_4.pt | 0.8944999999999999 | 0.004238637680047214 |
+| Epoch_1_batch_5999.pt | 0.8873333333333333 | 0.003471222078180738 |
+| Epoch_2_batch_2999.pt | 0.8833333333333332 | 0.0037515428924742534 |
+| Epoch_1.pt | 0.8821666666666665 | 0.005122270426382004 |
+| Epoch_1_batch_2999.pt | 0.8546666666666667 | 0.004936635531449569 |
+| Epoch_0.pt | 0.7931666666666667 | 0.006052496679006336 |
+| Epoch_0_batch_5999.pt | 0.7801666666666667 | 0.004405734198423739 |
+| Epoch_0_batch_2999.pt | 0.6598333333333334 | 0.007241214344651393 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..5bac63d14fbd36cda12e00c6f789523ba91336be
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.9908333333333333 | 0.0008695819912499194 |
+| Epoch_13_batch_2999.pt | 0.9906666666666666 | 0.0007535922203472496 |
+| Epoch_13.pt | 0.9906666666666666 | 0.0009362388636862635 |
+| Epoch_14_batch_5999.pt | 0.9906666666666666 | 0.0008678055195451845 |
+| Epoch_14.pt | 0.9904999999999999 | 0.0008624541497922234 |
+| Epoch_15_batch_2999.pt | 0.9903333333333334 | 0.0009229582069908943 |
+| Epoch_16_batch_5999.pt | 0.9903333333333333 | 0.0010772621905369623 |
+| Epoch_15_batch_5999.pt | 0.9901666666666665 | 0.0009444444444444424 |
+| Epoch_17_batch_2999.pt | 0.9901666666666665 | 0.0010076865081787242 |
+| Epoch_11.pt | 0.9899999999999999 | 0.0010540925533894627 |
+| Epoch_15.pt | 0.9898333333333333 | 0.0011772011166898408 |
+| Epoch_12_batch_2999.pt | 0.9898333333333331 | 0.0009765775461803873 |
+| Epoch_14_batch_2999.pt | 0.9894999999999999 | 0.000825892708184358 |
+| Epoch_17_batch_5999.pt | 0.9894999999999999 | 0.000963852865160968 |
+| Epoch_10.pt | 0.9893333333333333 | 0.0010304020550550815 |
+| Epoch_12.pt | 0.9893333333333333 | 0.0007535922203472558 |
+| Epoch_10_batch_5999.pt | 0.9891666666666665 | 0.0011180339887499004 |
+| Epoch_16_batch_2999.pt | 0.9891666666666665 | 0.0009043789220055396 |
+| Epoch_11_batch_5999.pt | 0.9889999999999999 | 0.0010886621079036394 |
+| Epoch_17.pt | 0.9888333333333332 | 0.0009312808119022304 |
+| Epoch_12_batch_5999.pt | 0.9884999999999998 | 0.0009444444444444462 |
+| Epoch_16.pt | 0.9883333333333333 | 0.0011111111111111087 |
+| Epoch_10_batch_2999.pt | 0.9881666666666666 | 0.0011235415786753741 |
+| Epoch_11_batch_2999.pt | 0.9881666666666666 | 0.001177201116689838 |
+| Epoch_9_batch_5999.pt | 0.9790000000000001 | 0.0013877773329774217 |
+| Epoch_9_batch_2999.pt | 0.976 | 0.0014315665251916796 |
+| Epoch_8_batch_2999.pt | 0.9753333333333334 | 0.0026270200927859784 |
+| Epoch_6_batch_2999.pt | 0.9736666666666667 | 0.0014010578014353875 |
+| Epoch_7_batch_5999.pt | 0.9734999999999999 | 0.001458267194267411 |
+| Epoch_4_batch_5999.pt | 0.9730000000000001 | 0.001567415108851769 |
+| Epoch_6_batch_5999.pt | 0.9726666666666667 | 0.0023465235646603203 |
+| Epoch_8_batch_5999.pt | 0.9721666666666667 | 0.0018600743380870254 |
+| Epoch_7_batch_2999.pt | 0.9718333333333333 | 0.0025633937766798513 |
+| Epoch_7.pt | 0.9701666666666664 | 0.002269633135090615 |
+| Epoch_5.pt | 0.97 | 0.0030530292586742366 |
+| Epoch_5_batch_2999.pt | 0.9693333333333334 | 0.0012472191289246478 |
+| Epoch_8.pt | 0.9683333333333334 | 0.0028109134757052343 |
+| Epoch_5_batch_5999.pt | 0.9678333333333334 | 0.002222916558193617 |
+| Epoch_3_batch_2999.pt | 0.9676666666666668 | 0.0023200681130912397 |
+| Epoch_3.pt | 0.9666666666666666 | 0.0018921540406584946 |
+| Epoch_4_batch_2999.pt | 0.9665000000000001 | 0.001915659961062966 |
+| Epoch_3_batch_5999.pt | 0.9663333333333334 | 0.001855921454276676 |
+| Epoch_9.pt | 0.9646666666666668 | 0.002650413431528127 |
+| Epoch_4.pt | 0.9646666666666665 | 0.0024317854031377004 |
+| Epoch_2_batch_5999.pt | 0.9644999999999999 | 0.002208988372452337 |
+| Epoch_6.pt | 0.9639999999999999 | 0.0018790593916986366 |
+| Epoch_2.pt | 0.9628333333333334 | 0.0026063786901644255 |
+| Epoch_2_batch_2999.pt | 0.9559999999999998 | 0.0029627314724385272 |
+| Epoch_1_batch_5999.pt | 0.9478333333333333 | 0.002629955639676583 |
+| Epoch_1.pt | 0.9470000000000001 | 0.003081205471969342 |
+| Epoch_1_batch_2999.pt | 0.9291666666666666 | 0.003035286306588386 |
+| Epoch_0.pt | 0.8735000000000002 | 0.0022966696231771534 |
+| Epoch_0_batch_5999.pt | 0.8663333333333334 | 0.005498596902959382 |
+| Epoch_0_batch_2999.pt | 0.7356666666666667 | 0.005074749887408547 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..79947719bf218f1d0fb4d7dd7acf14e0d1a1a5b1
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,40 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9626666666666667 | 0.0017777777777777796 |
+| Epoch_15_batch_5999.pt | 0.9626666666666667 | 0.0020214894887400333 |
+| Epoch_16.pt | 0.9625 | 0.001829963232444516 |
+| Epoch_17_batch_5999.pt | 0.9623333333333333 | 0.0022662308949301267 |
+| Epoch_14_batch_5999.pt | 0.9623333333333332 | 0.002051798368068823 |
+| Epoch_12_batch_2999.pt | 0.9621666666666666 | 0.002277777777777779 |
+| Epoch_16_batch_2999.pt | 0.9620000000000001 | 0.0017177360926378088 |
+| Epoch_17.pt | 0.962 | 0.0019212907184211804 |
+| Epoch_13.pt | 0.9618333333333334 | 0.0018994801108087572 |
+| Epoch_16_batch_5999.pt | 0.9618333333333332 | 0.0019945914523351368 |
+| Epoch_17_batch_2999.pt | 0.9615 | 0.0017293758240303758 |
+| Epoch_13_batch_5999.pt | 0.9613333333333334 | 0.0021631024815479787 |
+| Epoch_14_batch_2999.pt | 0.961 | 0.0022662308949301267 |
+| Epoch_14.pt | 0.9601666666666666 | 0.0020705161281760255 |
+| Epoch_15.pt | 0.9601666666666666 | 0.0017471316881684884 |
+| Epoch_13_batch_2999.pt | 0.96 | 0.002263505420829262 |
+| Epoch_11.pt | 0.9599999999999997 | 0.0022222222222222244 |
+| Epoch_12_batch_5999.pt | 0.9594999999999999 | 0.002331348362039706 |
+| Epoch_10_batch_5999.pt | 0.959 | 0.0024993826398226654 |
+| Epoch_11_batch_5999.pt | 0.9586666666666666 | 0.0018392161508052006 |
+| Epoch_11_batch_2999.pt | 0.9573333333333334 | 0.002140151142695358 |
+| Epoch_12.pt | 0.9571666666666667 | 0.0019092044752710689 |
+| Epoch_10_batch_2999.pt | 0.9564999999999999 | 0.0018666997351568324 |
+| Epoch_10.pt | 0.9563333333333333 | 0.0021914536581462258 |
+| Epoch_8_batch_2999.pt | 0.9448333333333332 | 0.0022696331350906137 |
+| Epoch_9_batch_5999.pt | 0.9443333333333334 | 0.002655067365633006 |
+| Epoch_6_batch_2999.pt | 0.9421666666666667 | 0.0025343321617682267 |
+| Epoch_7_batch_5999.pt | 0.9421666666666667 | 0.002865826750339763 |
+| Epoch_7.pt | 0.942 | 0.0027866524897743198 |
+| Epoch_5.pt | 0.9403333333333335 | 0.002775554665954841 |
+| Epoch_7_batch_2999.pt | 0.9395 | 0.0019883922409081414 |
+| Epoch_3_batch_5999.pt | 0.9393333333333332 | 0.0031249691356500485 |
+| Epoch_9_batch_2999.pt | 0.9383333333333332 | 0.0028867513459481286 |
+| Epoch_8_batch_5999.pt | 0.9378333333333332 | 0.002618193703988476 |
+| Epoch_5_batch_2999.pt | 0.9375 | 0.0028246052930858104 |
+| Epoch_4_batch_5999.pt | 0.9368333333333334 | 0.0025633937766798456 |
+| Epoch_4_batch_2999.pt | 0.9356666666666665 | 0.0032979604621457353 |
diff --git a/bob/bio/facexzoo/models/backbones/ResNet152_irse/log.log b/bob/bio/facexzoo/models/backbones/ResNet152_irse/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..71ec8397a0b012137ab35ef285551943759ca0e4
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet152_irse/log.log
@@ -0,0 +1,655 @@
+INFO 2020-12-02 16:32:30 train.py: 177] Start optimization.
+INFO 2020-12-02 16:32:30 train.py: 178] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='ResNet', batch_size=512, data_root='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-resnet152', train_file='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f24172304a8>)
+backbone param:
+{'depth': 152, 'drop_ratio': 0.4, 'net_mode': 'ir_se', 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-12-02 16:32:52 train.py: 79] Epoch 0, iter 0/6416, lr 0.100000, loss 16.178095
+INFO 2020-12-02 16:38:23 train.py: 79] Epoch 0, iter 200/6416, lr 0.100000, loss 15.722174
+INFO 2020-12-02 16:43:54 train.py: 79] Epoch 0, iter 400/6416, lr 0.100000, loss 15.293269
+INFO 2020-12-02 16:49:25 train.py: 79] Epoch 0, iter 600/6416, lr 0.100000, loss 15.186101
+INFO 2020-12-02 16:54:56 train.py: 79] Epoch 0, iter 800/6416, lr 0.100000, loss 15.095860
+INFO 2020-12-02 17:00:28 train.py: 79] Epoch 0, iter 1000/6416, lr 0.100000, loss 14.994113
+INFO 2020-12-02 17:06:00 train.py: 79] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.869156
+INFO 2020-12-02 17:11:32 train.py: 79] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.682795
+INFO 2020-12-02 17:17:04 train.py: 79] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.429536
+INFO 2020-12-02 17:22:36 train.py: 79] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.161485
+INFO 2020-12-02 17:28:08 train.py: 79] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.890918
+INFO 2020-12-02 17:33:40 train.py: 79] Epoch 0, iter 2200/6416, lr 0.100000, loss 13.619366
+INFO 2020-12-02 17:39:12 train.py: 79] Epoch 0, iter 2400/6416, lr 0.100000, loss 13.314948
+INFO 2020-12-02 17:44:43 train.py: 79] Epoch 0, iter 2600/6416, lr 0.100000, loss 13.002598
+INFO 2020-12-02 17:50:15 train.py: 79] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.655731
+INFO 2020-12-02 17:55:45 train.py: 92] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-12-02 17:55:47 train.py: 79] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.355666
+INFO 2020-12-02 18:01:19 train.py: 79] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.111276
+INFO 2020-12-02 18:06:50 train.py: 79] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.970935
+INFO 2020-12-02 18:12:21 train.py: 79] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.004759
+INFO 2020-12-02 18:17:51 train.py: 79] Epoch 0, iter 3800/6416, lr 0.100000, loss 12.173104
+INFO 2020-12-02 18:23:22 train.py: 79] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.465508
+INFO 2020-12-02 18:28:51 train.py: 79] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.848047
+INFO 2020-12-02 18:34:21 train.py: 79] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.270324
+INFO 2020-12-02 18:39:50 train.py: 79] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.661355
+INFO 2020-12-02 18:45:18 train.py: 79] Epoch 0, iter 4800/6416, lr 0.100000, loss 13.982073
+INFO 2020-12-02 18:50:47 train.py: 79] Epoch 0, iter 5000/6416, lr 0.100000, loss 14.266946
+INFO 2020-12-02 18:56:15 train.py: 79] Epoch 0, iter 5200/6416, lr 0.100000, loss 14.419807
+INFO 2020-12-02 19:01:42 train.py: 79] Epoch 0, iter 5400/6416, lr 0.100000, loss 14.488699
+INFO 2020-12-02 19:07:10 train.py: 79] Epoch 0, iter 5600/6416, lr 0.100000, loss 14.469730
+INFO 2020-12-02 19:12:37 train.py: 79] Epoch 0, iter 5800/6416, lr 0.100000, loss 14.368183
+INFO 2020-12-02 19:18:03 train.py: 92] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-12-02 19:18:05 train.py: 79] Epoch 0, iter 6000/6416, lr 0.100000, loss 14.185414
+INFO 2020-12-02 19:23:32 train.py: 79] Epoch 0, iter 6200/6416, lr 0.100000, loss 13.941561
+INFO 2020-12-02 19:28:58 train.py: 79] Epoch 0, iter 6400/6416, lr 0.100000, loss 13.675249
+INFO 2020-12-02 19:29:22 train.py: 97] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-12-02 19:29:25 train.py: 79] Epoch 1, iter 0/6416, lr 0.100000, loss 13.437862
+INFO 2020-12-02 19:34:52 train.py: 79] Epoch 1, iter 200/6416, lr 0.100000, loss 13.193642
+INFO 2020-12-02 19:40:18 train.py: 79] Epoch 1, iter 400/6416, lr 0.100000, loss 12.860866
+INFO 2020-12-02 19:45:45 train.py: 79] Epoch 1, iter 600/6416, lr 0.100000, loss 12.522454
+INFO 2020-12-02 19:51:11 train.py: 79] Epoch 1, iter 800/6416, lr 0.100000, loss 12.201575
+INFO 2020-12-02 19:56:37 train.py: 79] Epoch 1, iter 1000/6416, lr 0.100000, loss 11.883126
+INFO 2020-12-02 20:02:03 train.py: 79] Epoch 1, iter 1200/6416, lr 0.100000, loss 11.625596
+INFO 2020-12-02 20:07:29 train.py: 79] Epoch 1, iter 1400/6416, lr 0.100000, loss 11.293950
+INFO 2020-12-02 20:12:56 train.py: 79] Epoch 1, iter 1600/6416, lr 0.100000, loss 11.037240
+INFO 2020-12-02 20:18:22 train.py: 79] Epoch 1, iter 1800/6416, lr 0.100000, loss 10.740246
+INFO 2020-12-02 20:23:48 train.py: 79] Epoch 1, iter 2000/6416, lr 0.100000, loss 10.469370
+INFO 2020-12-02 20:29:14 train.py: 79] Epoch 1, iter 2200/6416, lr 0.100000, loss 10.215419
+INFO 2020-12-02 20:34:41 train.py: 79] Epoch 1, iter 2400/6416, lr 0.100000, loss 9.969037
+INFO 2020-12-02 20:40:07 train.py: 79] Epoch 1, iter 2600/6416, lr 0.100000, loss 9.750741
+INFO 2020-12-02 20:45:33 train.py: 79] Epoch 1, iter 2800/6416, lr 0.100000, loss 9.539316
+INFO 2020-12-02 20:50:58 train.py: 92] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-12-02 20:51:00 train.py: 79] Epoch 1, iter 3000/6416, lr 0.100000, loss 9.350437
+INFO 2020-12-02 20:56:26 train.py: 79] Epoch 1, iter 3200/6416, lr 0.100000, loss 9.149724
+INFO 2020-12-02 21:01:52 train.py: 79] Epoch 1, iter 3400/6416, lr 0.100000, loss 9.004838
+INFO 2020-12-02 21:07:17 train.py: 79] Epoch 1, iter 3600/6416, lr 0.100000, loss 8.810904
+INFO 2020-12-02 21:12:43 train.py: 79] Epoch 1, iter 3800/6416, lr 0.100000, loss 8.683816
+INFO 2020-12-02 21:18:09 train.py: 79] Epoch 1, iter 4000/6416, lr 0.100000, loss 8.540958
+INFO 2020-12-02 21:23:36 train.py: 79] Epoch 1, iter 4200/6416, lr 0.100000, loss 8.400602
+INFO 2020-12-02 21:29:02 train.py: 79] Epoch 1, iter 4400/6416, lr 0.100000, loss 8.271969
+INFO 2020-12-02 21:34:28 train.py: 79] Epoch 1, iter 4600/6416, lr 0.100000, loss 8.164128
+INFO 2020-12-02 21:39:54 train.py: 79] Epoch 1, iter 4800/6416, lr 0.100000, loss 8.013629
+INFO 2020-12-02 21:45:20 train.py: 79] Epoch 1, iter 5000/6416, lr 0.100000, loss 7.945607
+INFO 2020-12-02 21:50:45 train.py: 79] Epoch 1, iter 5200/6416, lr 0.100000, loss 7.818117
+INFO 2020-12-02 21:56:11 train.py: 79] Epoch 1, iter 5400/6416, lr 0.100000, loss 7.735166
+INFO 2020-12-02 22:01:37 train.py: 79] Epoch 1, iter 5600/6416, lr 0.100000, loss 7.675649
+INFO 2020-12-02 22:07:03 train.py: 79] Epoch 1, iter 5800/6416, lr 0.100000, loss 7.530980
+INFO 2020-12-02 22:12:28 train.py: 92] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-12-02 22:12:30 train.py: 79] Epoch 1, iter 6000/6416, lr 0.100000, loss 7.466132
+INFO 2020-12-02 22:17:56 train.py: 79] Epoch 1, iter 6200/6416, lr 0.100000, loss 7.415678
+INFO 2020-12-02 22:23:21 train.py: 79] Epoch 1, iter 6400/6416, lr 0.100000, loss 7.315983
+INFO 2020-12-02 22:23:45 train.py: 97] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-12-02 22:23:48 train.py: 79] Epoch 2, iter 0/6416, lr 0.100000, loss 7.279906
+INFO 2020-12-02 22:29:14 train.py: 79] Epoch 2, iter 200/6416, lr 0.100000, loss 6.741915
+INFO 2020-12-02 22:34:40 train.py: 79] Epoch 2, iter 400/6416, lr 0.100000, loss 6.704193
+INFO 2020-12-02 22:40:06 train.py: 79] Epoch 2, iter 600/6416, lr 0.100000, loss 6.755705
+INFO 2020-12-02 22:45:31 train.py: 79] Epoch 2, iter 800/6416, lr 0.100000, loss 6.749712
+INFO 2020-12-02 22:50:58 train.py: 79] Epoch 2, iter 1000/6416, lr 0.100000, loss 6.715226
+INFO 2020-12-02 22:56:23 train.py: 79] Epoch 2, iter 1200/6416, lr 0.100000, loss 6.709964
+INFO 2020-12-02 23:01:49 train.py: 79] Epoch 2, iter 1400/6416, lr 0.100000, loss 6.712732
+INFO 2020-12-02 23:07:15 train.py: 79] Epoch 2, iter 1600/6416, lr 0.100000, loss 6.657225
+INFO 2020-12-02 23:12:41 train.py: 79] Epoch 2, iter 1800/6416, lr 0.100000, loss 6.636164
+INFO 2020-12-02 23:18:06 train.py: 79] Epoch 2, iter 2000/6416, lr 0.100000, loss 6.612790
+INFO 2020-12-02 23:23:32 train.py: 79] Epoch 2, iter 2200/6416, lr 0.100000, loss 6.584645
+INFO 2020-12-02 23:28:58 train.py: 79] Epoch 2, iter 2400/6416, lr 0.100000, loss 6.568062
+INFO 2020-12-02 23:34:24 train.py: 79] Epoch 2, iter 2600/6416, lr 0.100000, loss 6.504225
+INFO 2020-12-02 23:39:50 train.py: 79] Epoch 2, iter 2800/6416, lr 0.100000, loss 6.490968
+INFO 2020-12-02 23:45:15 train.py: 92] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-12-02 23:45:17 train.py: 79] Epoch 2, iter 3000/6416, lr 0.100000, loss 6.462917
+INFO 2020-12-02 23:50:43 train.py: 79] Epoch 2, iter 3200/6416, lr 0.100000, loss 6.406101
+INFO 2020-12-02 23:56:09 train.py: 79] Epoch 2, iter 3400/6416, lr 0.100000, loss 6.382000
+INFO 2020-12-03 00:01:34 train.py: 79] Epoch 2, iter 3600/6416, lr 0.100000, loss 6.336562
+INFO 2020-12-03 00:07:01 train.py: 79] Epoch 2, iter 3800/6416, lr 0.100000, loss 6.274786
+INFO 2020-12-03 00:12:27 train.py: 79] Epoch 2, iter 4000/6416, lr 0.100000, loss 6.282790
+INFO 2020-12-03 00:17:53 train.py: 79] Epoch 2, iter 4200/6416, lr 0.100000, loss 6.250135
+INFO 2020-12-03 00:23:18 train.py: 79] Epoch 2, iter 4400/6416, lr 0.100000, loss 6.208534
+INFO 2020-12-03 00:28:45 train.py: 79] Epoch 2, iter 4600/6416, lr 0.100000, loss 6.142354
+INFO 2020-12-03 00:34:11 train.py: 79] Epoch 2, iter 4800/6416, lr 0.100000, loss 6.137732
+INFO 2020-12-03 00:39:37 train.py: 79] Epoch 2, iter 5000/6416, lr 0.100000, loss 6.095399
+INFO 2020-12-03 00:45:03 train.py: 79] Epoch 2, iter 5200/6416, lr 0.100000, loss 6.076461
+INFO 2020-12-03 00:50:29 train.py: 79] Epoch 2, iter 5400/6416, lr 0.100000, loss 6.044541
+INFO 2020-12-03 00:55:54 train.py: 79] Epoch 2, iter 5600/6416, lr 0.100000, loss 6.026438
+INFO 2020-12-03 01:01:20 train.py: 79] Epoch 2, iter 5800/6416, lr 0.100000, loss 5.990023
+INFO 2020-12-03 01:06:45 train.py: 92] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-12-03 01:06:47 train.py: 79] Epoch 2, iter 6000/6416, lr 0.100000, loss 5.941019
+INFO 2020-12-03 01:12:13 train.py: 79] Epoch 2, iter 6200/6416, lr 0.100000, loss 5.895717
+INFO 2020-12-03 01:17:39 train.py: 79] Epoch 2, iter 6400/6416, lr 0.100000, loss 5.911855
+INFO 2020-12-03 01:18:02 train.py: 97] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-12-03 01:18:05 train.py: 79] Epoch 3, iter 0/6416, lr 0.100000, loss 5.816688
+INFO 2020-12-03 01:23:31 train.py: 79] Epoch 3, iter 200/6416, lr 0.100000, loss 5.340988
+INFO 2020-12-03 01:28:57 train.py: 79] Epoch 3, iter 400/6416, lr 0.100000, loss 5.316447
+INFO 2020-12-03 01:34:23 train.py: 79] Epoch 3, iter 600/6416, lr 0.100000, loss 5.390077
+INFO 2020-12-03 01:39:48 train.py: 79] Epoch 3, iter 800/6416, lr 0.100000, loss 5.470877
+INFO 2020-12-03 01:45:14 train.py: 79] Epoch 3, iter 1000/6416, lr 0.100000, loss 5.502773
+INFO 2020-12-03 01:50:40 train.py: 79] Epoch 3, iter 1200/6416, lr 0.100000, loss 5.506761
+INFO 2020-12-03 01:56:05 train.py: 79] Epoch 3, iter 1400/6416, lr 0.100000, loss 5.522129
+INFO 2020-12-03 02:01:31 train.py: 79] Epoch 3, iter 1600/6416, lr 0.100000, loss 5.550251
+INFO 2020-12-03 02:06:57 train.py: 79] Epoch 3, iter 1800/6416, lr 0.100000, loss 5.540093
+INFO 2020-12-03 02:12:22 train.py: 79] Epoch 3, iter 2000/6416, lr 0.100000, loss 5.519903
+INFO 2020-12-03 02:17:48 train.py: 79] Epoch 3, iter 2200/6416, lr 0.100000, loss 5.533010
+INFO 2020-12-03 02:23:14 train.py: 79] Epoch 3, iter 2400/6416, lr 0.100000, loss 5.524664
+INFO 2020-12-03 02:28:39 train.py: 79] Epoch 3, iter 2600/6416, lr 0.100000, loss 5.506803
+INFO 2020-12-03 02:34:05 train.py: 79] Epoch 3, iter 2800/6416, lr 0.100000, loss 5.503340
+INFO 2020-12-03 02:39:30 train.py: 92] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-12-03 02:39:31 train.py: 79] Epoch 3, iter 3000/6416, lr 0.100000, loss 5.485541
+INFO 2020-12-03 02:44:57 train.py: 79] Epoch 3, iter 3200/6416, lr 0.100000, loss 5.475738
+INFO 2020-12-03 02:50:22 train.py: 79] Epoch 3, iter 3400/6416, lr 0.100000, loss 5.460465
+INFO 2020-12-03 02:55:48 train.py: 79] Epoch 3, iter 3600/6416, lr 0.100000, loss 5.451768
+INFO 2020-12-03 03:01:14 train.py: 79] Epoch 3, iter 3800/6416, lr 0.100000, loss 5.438681
+INFO 2020-12-03 03:06:39 train.py: 79] Epoch 3, iter 4000/6416, lr 0.100000, loss 5.395285
+INFO 2020-12-03 03:12:05 train.py: 79] Epoch 3, iter 4200/6416, lr 0.100000, loss 5.405740
+INFO 2020-12-03 03:17:30 train.py: 79] Epoch 3, iter 4400/6416, lr 0.100000, loss 5.387588
+INFO 2020-12-03 03:22:56 train.py: 79] Epoch 3, iter 4600/6416, lr 0.100000, loss 5.382697
+INFO 2020-12-03 03:28:21 train.py: 79] Epoch 3, iter 4800/6416, lr 0.100000, loss 5.338904
+INFO 2020-12-03 03:33:47 train.py: 79] Epoch 3, iter 5000/6416, lr 0.100000, loss 5.325464
+INFO 2020-12-03 03:39:12 train.py: 79] Epoch 3, iter 5200/6416, lr 0.100000, loss 5.323090
+INFO 2020-12-03 03:44:38 train.py: 79] Epoch 3, iter 5400/6416, lr 0.100000, loss 5.322617
+INFO 2020-12-03 03:50:04 train.py: 79] Epoch 3, iter 5600/6416, lr 0.100000, loss 5.282720
+INFO 2020-12-03 03:55:29 train.py: 79] Epoch 3, iter 5800/6416, lr 0.100000, loss 5.246655
+INFO 2020-12-03 04:00:54 train.py: 92] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-12-03 04:00:55 train.py: 79] Epoch 3, iter 6000/6416, lr 0.100000, loss 5.282314
+INFO 2020-12-03 04:06:21 train.py: 79] Epoch 3, iter 6200/6416, lr 0.100000, loss 5.250553
+INFO 2020-12-03 04:11:47 train.py: 79] Epoch 3, iter 6400/6416, lr 0.100000, loss 5.219490
+INFO 2020-12-03 04:12:11 train.py: 97] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-12-03 04:12:14 train.py: 79] Epoch 4, iter 0/6416, lr 0.100000, loss 5.258139
+INFO 2020-12-03 04:17:39 train.py: 79] Epoch 4, iter 200/6416, lr 0.100000, loss 4.713236
+INFO 2020-12-03 04:23:05 train.py: 79] Epoch 4, iter 400/6416, lr 0.100000, loss 4.693161
+INFO 2020-12-03 04:28:31 train.py: 79] Epoch 4, iter 600/6416, lr 0.100000, loss 4.773143
+INFO 2020-12-03 04:33:57 train.py: 79] Epoch 4, iter 800/6416, lr 0.100000, loss 4.850218
+INFO 2020-12-03 04:39:22 train.py: 79] Epoch 4, iter 1000/6416, lr 0.100000, loss 4.858812
+INFO 2020-12-03 04:44:48 train.py: 79] Epoch 4, iter 1200/6416, lr 0.100000, loss 4.919182
+INFO 2020-12-03 04:50:14 train.py: 79] Epoch 4, iter 1400/6416, lr 0.100000, loss 4.929719
+INFO 2020-12-03 04:55:39 train.py: 79] Epoch 4, iter 1600/6416, lr 0.100000, loss 4.975892
+INFO 2020-12-03 05:01:05 train.py: 79] Epoch 4, iter 1800/6416, lr 0.100000, loss 4.952683
+INFO 2020-12-03 05:06:30 train.py: 79] Epoch 4, iter 2000/6416, lr 0.100000, loss 4.979694
+INFO 2020-12-03 05:11:55 train.py: 79] Epoch 4, iter 2200/6416, lr 0.100000, loss 5.015690
+INFO 2020-12-03 05:17:21 train.py: 79] Epoch 4, iter 2400/6416, lr 0.100000, loss 5.009107
+INFO 2020-12-03 05:22:46 train.py: 79] Epoch 4, iter 2600/6416, lr 0.100000, loss 4.987703
+INFO 2020-12-03 05:28:12 train.py: 79] Epoch 4, iter 2800/6416, lr 0.100000, loss 4.990482
+INFO 2020-12-03 05:33:36 train.py: 92] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-12-03 05:33:37 train.py: 79] Epoch 4, iter 3000/6416, lr 0.100000, loss 4.973592
+INFO 2020-12-03 05:39:02 train.py: 79] Epoch 4, iter 3200/6416, lr 0.100000, loss 4.944998
+INFO 2020-12-03 05:44:28 train.py: 79] Epoch 4, iter 3400/6416, lr 0.100000, loss 4.968152
+INFO 2020-12-03 05:49:53 train.py: 79] Epoch 4, iter 3600/6416, lr 0.100000, loss 4.961744
+INFO 2020-12-03 05:55:19 train.py: 79] Epoch 4, iter 3800/6416, lr 0.100000, loss 4.953355
+INFO 2020-12-03 06:00:44 train.py: 79] Epoch 4, iter 4000/6416, lr 0.100000, loss 4.941136
+INFO 2020-12-03 06:06:09 train.py: 79] Epoch 4, iter 4200/6416, lr 0.100000, loss 4.954097
+INFO 2020-12-03 06:11:34 train.py: 79] Epoch 4, iter 4400/6416, lr 0.100000, loss 4.926616
+INFO 2020-12-03 06:16:59 train.py: 79] Epoch 4, iter 4600/6416, lr 0.100000, loss 4.929985
+INFO 2020-12-03 06:22:25 train.py: 79] Epoch 4, iter 4800/6416, lr 0.100000, loss 4.916453
+INFO 2020-12-03 06:27:50 train.py: 79] Epoch 4, iter 5000/6416, lr 0.100000, loss 4.874083
+INFO 2020-12-03 06:33:15 train.py: 79] Epoch 4, iter 5200/6416, lr 0.100000, loss 4.899871
+INFO 2020-12-03 06:38:40 train.py: 79] Epoch 4, iter 5400/6416, lr 0.100000, loss 4.897398
+INFO 2020-12-03 06:44:05 train.py: 79] Epoch 4, iter 5600/6416, lr 0.100000, loss 4.879323
+INFO 2020-12-03 06:49:31 train.py: 79] Epoch 4, iter 5800/6416, lr 0.100000, loss 4.858275
+INFO 2020-12-03 06:54:55 train.py: 92] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-12-03 06:54:56 train.py: 79] Epoch 4, iter 6000/6416, lr 0.100000, loss 4.884626
+INFO 2020-12-03 07:00:22 train.py: 79] Epoch 4, iter 6200/6416, lr 0.100000, loss 4.869890
+INFO 2020-12-03 07:05:47 train.py: 79] Epoch 4, iter 6400/6416, lr 0.100000, loss 4.850183
+INFO 2020-12-03 07:06:11 train.py: 97] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-12-03 07:06:14 train.py: 79] Epoch 5, iter 0/6416, lr 0.100000, loss 4.767923
+INFO 2020-12-03 07:11:40 train.py: 79] Epoch 5, iter 200/6416, lr 0.100000, loss 4.354355
+INFO 2020-12-03 07:17:05 train.py: 79] Epoch 5, iter 400/6416, lr 0.100000, loss 4.337549
+INFO 2020-12-03 07:22:31 train.py: 79] Epoch 5, iter 600/6416, lr 0.100000, loss 4.402054
+INFO 2020-12-03 07:27:57 train.py: 79] Epoch 5, iter 800/6416, lr 0.100000, loss 4.495166
+INFO 2020-12-03 07:33:22 train.py: 79] Epoch 5, iter 1000/6416, lr 0.100000, loss 4.532663
+INFO 2020-12-03 07:38:48 train.py: 79] Epoch 5, iter 1200/6416, lr 0.100000, loss 4.553198
+INFO 2020-12-03 07:44:13 train.py: 79] Epoch 5, iter 1400/6416, lr 0.100000, loss 4.594005
+INFO 2020-12-03 07:49:38 train.py: 79] Epoch 5, iter 1600/6416, lr 0.100000, loss 4.625065
+INFO 2020-12-03 07:55:03 train.py: 79] Epoch 5, iter 1800/6416, lr 0.100000, loss 4.628173
+INFO 2020-12-03 08:00:28 train.py: 79] Epoch 5, iter 2000/6416, lr 0.100000, loss 4.648198
+INFO 2020-12-03 08:05:54 train.py: 79] Epoch 5, iter 2200/6416, lr 0.100000, loss 4.649925
+INFO 2020-12-03 08:11:19 train.py: 79] Epoch 5, iter 2400/6416, lr 0.100000, loss 4.667757
+INFO 2020-12-03 08:16:44 train.py: 79] Epoch 5, iter 2600/6416, lr 0.100000, loss 4.626384
+INFO 2020-12-03 08:22:09 train.py: 79] Epoch 5, iter 2800/6416, lr 0.100000, loss 4.668897
+INFO 2020-12-03 08:27:33 train.py: 92] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-12-03 08:27:35 train.py: 79] Epoch 5, iter 3000/6416, lr 0.100000, loss 4.646401
+INFO 2020-12-03 08:33:00 train.py: 79] Epoch 5, iter 3200/6416, lr 0.100000, loss 4.678051
+INFO 2020-12-03 08:38:26 train.py: 79] Epoch 5, iter 3400/6416, lr 0.100000, loss 4.677237
+INFO 2020-12-03 08:43:51 train.py: 79] Epoch 5, iter 3600/6416, lr 0.100000, loss 4.660041
+INFO 2020-12-03 08:49:16 train.py: 79] Epoch 5, iter 3800/6416, lr 0.100000, loss 4.656218
+INFO 2020-12-03 08:54:41 train.py: 79] Epoch 5, iter 4000/6416, lr 0.100000, loss 4.652777
+INFO 2020-12-03 09:00:07 train.py: 79] Epoch 5, iter 4200/6416, lr 0.100000, loss 4.661610
+INFO 2020-12-03 09:05:32 train.py: 79] Epoch 5, iter 4400/6416, lr 0.100000, loss 4.635444
+INFO 2020-12-03 09:10:57 train.py: 79] Epoch 5, iter 4600/6416, lr 0.100000, loss 4.652587
+INFO 2020-12-03 09:16:22 train.py: 79] Epoch 5, iter 4800/6416, lr 0.100000, loss 4.638926
+INFO 2020-12-03 09:21:47 train.py: 79] Epoch 5, iter 5000/6416, lr 0.100000, loss 4.647194
+INFO 2020-12-03 09:27:13 train.py: 79] Epoch 5, iter 5200/6416, lr 0.100000, loss 4.620812
+INFO 2020-12-03 09:32:38 train.py: 79] Epoch 5, iter 5400/6416, lr 0.100000, loss 4.613374
+INFO 2020-12-03 09:38:03 train.py: 79] Epoch 5, iter 5600/6416, lr 0.100000, loss 4.584471
+INFO 2020-12-03 09:43:28 train.py: 79] Epoch 5, iter 5800/6416, lr 0.100000, loss 4.587631
+INFO 2020-12-03 09:48:53 train.py: 92] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-12-03 09:48:55 train.py: 79] Epoch 5, iter 6000/6416, lr 0.100000, loss 4.581571
+INFO 2020-12-03 09:54:20 train.py: 79] Epoch 5, iter 6200/6416, lr 0.100000, loss 4.601777
+INFO 2020-12-03 09:59:45 train.py: 79] Epoch 5, iter 6400/6416, lr 0.100000, loss 4.588366
+INFO 2020-12-03 10:00:09 train.py: 97] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-12-03 10:00:12 train.py: 79] Epoch 6, iter 0/6416, lr 0.100000, loss 4.574776
+INFO 2020-12-03 10:05:38 train.py: 79] Epoch 6, iter 200/6416, lr 0.100000, loss 4.083318
+INFO 2020-12-03 10:11:03 train.py: 79] Epoch 6, iter 400/6416, lr 0.100000, loss 4.079606
+INFO 2020-12-03 10:16:29 train.py: 79] Epoch 6, iter 600/6416, lr 0.100000, loss 4.155832
+INFO 2020-12-03 10:21:54 train.py: 79] Epoch 6, iter 800/6416, lr 0.100000, loss 4.191371
+INFO 2020-12-03 10:27:19 train.py: 79] Epoch 6, iter 1000/6416, lr 0.100000, loss 4.276350
+INFO 2020-12-03 10:32:44 train.py: 79] Epoch 6, iter 1200/6416, lr 0.100000, loss 4.286451
+INFO 2020-12-03 10:38:10 train.py: 79] Epoch 6, iter 1400/6416, lr 0.100000, loss 4.369414
+INFO 2020-12-03 10:43:35 train.py: 79] Epoch 6, iter 1600/6416, lr 0.100000, loss 4.371112
+INFO 2020-12-03 10:49:00 train.py: 79] Epoch 6, iter 1800/6416, lr 0.100000, loss 4.390935
+INFO 2020-12-03 10:54:25 train.py: 79] Epoch 6, iter 2000/6416, lr 0.100000, loss 4.412332
+INFO 2020-12-03 10:59:51 train.py: 79] Epoch 6, iter 2200/6416, lr 0.100000, loss 4.432450
+INFO 2020-12-03 11:05:16 train.py: 79] Epoch 6, iter 2400/6416, lr 0.100000, loss 4.451447
+INFO 2020-12-03 11:10:41 train.py: 79] Epoch 6, iter 2600/6416, lr 0.100000, loss 4.434381
+INFO 2020-12-03 11:16:07 train.py: 79] Epoch 6, iter 2800/6416, lr 0.100000, loss 4.440942
+INFO 2020-12-03 11:21:30 train.py: 92] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-12-03 11:21:32 train.py: 79] Epoch 6, iter 3000/6416, lr 0.100000, loss 4.422063
+INFO 2020-12-03 11:26:57 train.py: 79] Epoch 6, iter 3200/6416, lr 0.100000, loss 4.434628
+INFO 2020-12-03 11:32:22 train.py: 79] Epoch 6, iter 3400/6416, lr 0.100000, loss 4.463880
+INFO 2020-12-03 11:37:48 train.py: 79] Epoch 6, iter 3600/6416, lr 0.100000, loss 4.454169
+INFO 2020-12-03 11:43:13 train.py: 79] Epoch 6, iter 3800/6416, lr 0.100000, loss 4.428558
+INFO 2020-12-03 11:48:38 train.py: 79] Epoch 6, iter 4000/6416, lr 0.100000, loss 4.453789
+INFO 2020-12-03 11:54:04 train.py: 79] Epoch 6, iter 4200/6416, lr 0.100000, loss 4.443664
+INFO 2020-12-03 11:59:29 train.py: 79] Epoch 6, iter 4400/6416, lr 0.100000, loss 4.467308
+INFO 2020-12-03 12:04:54 train.py: 79] Epoch 6, iter 4600/6416, lr 0.100000, loss 4.413052
+INFO 2020-12-03 12:10:19 train.py: 79] Epoch 6, iter 4800/6416, lr 0.100000, loss 4.444799
+INFO 2020-12-03 12:15:45 train.py: 79] Epoch 6, iter 5000/6416, lr 0.100000, loss 4.390673
+INFO 2020-12-03 12:21:10 train.py: 79] Epoch 6, iter 5200/6416, lr 0.100000, loss 4.460484
+INFO 2020-12-03 12:26:35 train.py: 79] Epoch 6, iter 5400/6416, lr 0.100000, loss 4.422217
+INFO 2020-12-03 12:32:00 train.py: 79] Epoch 6, iter 5600/6416, lr 0.100000, loss 4.445330
+INFO 2020-12-03 12:37:26 train.py: 79] Epoch 6, iter 5800/6416, lr 0.100000, loss 4.444832
+INFO 2020-12-03 12:42:50 train.py: 92] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-12-03 12:42:52 train.py: 79] Epoch 6, iter 6000/6416, lr 0.100000, loss 4.398113
+INFO 2020-12-03 12:48:17 train.py: 79] Epoch 6, iter 6200/6416, lr 0.100000, loss 4.422164
+INFO 2020-12-03 12:53:42 train.py: 79] Epoch 6, iter 6400/6416, lr 0.100000, loss 4.419172
+INFO 2020-12-03 12:54:06 train.py: 97] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-12-03 12:54:09 train.py: 79] Epoch 7, iter 0/6416, lr 0.100000, loss 4.392308
+INFO 2020-12-03 12:59:35 train.py: 79] Epoch 7, iter 200/6416, lr 0.100000, loss 3.921637
+INFO 2020-12-03 13:05:01 train.py: 79] Epoch 7, iter 400/6416, lr 0.100000, loss 3.914008
+INFO 2020-12-03 13:10:26 train.py: 79] Epoch 7, iter 600/6416, lr 0.100000, loss 3.993223
+INFO 2020-12-03 13:15:52 train.py: 79] Epoch 7, iter 800/6416, lr 0.100000, loss 4.034022
+INFO 2020-12-03 13:21:17 train.py: 79] Epoch 7, iter 1000/6416, lr 0.100000, loss 4.083853
+INFO 2020-12-03 13:26:43 train.py: 79] Epoch 7, iter 1200/6416, lr 0.100000, loss 4.128044
+INFO 2020-12-03 13:32:08 train.py: 79] Epoch 7, iter 1400/6416, lr 0.100000, loss 4.176273
+INFO 2020-12-03 13:37:33 train.py: 79] Epoch 7, iter 1600/6416, lr 0.100000, loss 4.194730
+INFO 2020-12-03 13:42:58 train.py: 79] Epoch 7, iter 1800/6416, lr 0.100000, loss 4.216888
+INFO 2020-12-03 13:48:24 train.py: 79] Epoch 7, iter 2000/6416, lr 0.100000, loss 4.232902
+INFO 2020-12-03 13:53:49 train.py: 79] Epoch 7, iter 2200/6416, lr 0.100000, loss 4.280568
+INFO 2020-12-03 13:59:14 train.py: 79] Epoch 7, iter 2400/6416, lr 0.100000, loss 4.276624
+INFO 2020-12-03 14:04:39 train.py: 79] Epoch 7, iter 2600/6416, lr 0.100000, loss 4.268450
+INFO 2020-12-03 14:10:04 train.py: 79] Epoch 7, iter 2800/6416, lr 0.100000, loss 4.270746
+INFO 2020-12-03 14:15:28 train.py: 92] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-12-03 14:15:30 train.py: 79] Epoch 7, iter 3000/6416, lr 0.100000, loss 4.261979
+INFO 2020-12-03 14:20:55 train.py: 79] Epoch 7, iter 3200/6416, lr 0.100000, loss 4.285834
+INFO 2020-12-03 14:26:20 train.py: 79] Epoch 7, iter 3400/6416, lr 0.100000, loss 4.295862
+INFO 2020-12-03 14:31:46 train.py: 79] Epoch 7, iter 3600/6416, lr 0.100000, loss 4.283829
+INFO 2020-12-03 14:37:11 train.py: 79] Epoch 7, iter 3800/6416, lr 0.100000, loss 4.259717
+INFO 2020-12-03 14:42:36 train.py: 79] Epoch 7, iter 4000/6416, lr 0.100000, loss 4.292408
+INFO 2020-12-03 14:48:01 train.py: 79] Epoch 7, iter 4200/6416, lr 0.100000, loss 4.278243
+INFO 2020-12-03 14:53:27 train.py: 79] Epoch 7, iter 4400/6416, lr 0.100000, loss 4.305061
+INFO 2020-12-03 14:58:52 train.py: 79] Epoch 7, iter 4600/6416, lr 0.100000, loss 4.289528
+INFO 2020-12-03 15:04:17 train.py: 79] Epoch 7, iter 4800/6416, lr 0.100000, loss 4.306961
+INFO 2020-12-03 15:09:42 train.py: 79] Epoch 7, iter 5000/6416, lr 0.100000, loss 4.303247
+INFO 2020-12-03 15:15:07 train.py: 79] Epoch 7, iter 5200/6416, lr 0.100000, loss 4.255533
+INFO 2020-12-03 15:20:33 train.py: 79] Epoch 7, iter 5400/6416, lr 0.100000, loss 4.290896
+INFO 2020-12-03 15:25:58 train.py: 79] Epoch 7, iter 5600/6416, lr 0.100000, loss 4.242422
+INFO 2020-12-03 15:31:23 train.py: 79] Epoch 7, iter 5800/6416, lr 0.100000, loss 4.237866
+INFO 2020-12-03 15:36:47 train.py: 92] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-12-03 15:36:49 train.py: 79] Epoch 7, iter 6000/6416, lr 0.100000, loss 4.258237
+INFO 2020-12-03 15:42:14 train.py: 79] Epoch 7, iter 6200/6416, lr 0.100000, loss 4.272571
+INFO 2020-12-03 15:47:39 train.py: 79] Epoch 7, iter 6400/6416, lr 0.100000, loss 4.270667
+INFO 2020-12-03 15:48:03 train.py: 97] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-12-03 15:48:06 train.py: 79] Epoch 8, iter 0/6416, lr 0.100000, loss 4.248033
+INFO 2020-12-03 15:53:32 train.py: 79] Epoch 8, iter 200/6416, lr 0.100000, loss 3.778519
+INFO 2020-12-03 15:58:57 train.py: 79] Epoch 8, iter 400/6416, lr 0.100000, loss 3.747538
+INFO 2020-12-03 16:04:22 train.py: 79] Epoch 8, iter 600/6416, lr 0.100000, loss 3.809143
+INFO 2020-12-03 16:09:48 train.py: 79] Epoch 8, iter 800/6416, lr 0.100000, loss 3.911213
+INFO 2020-12-03 16:15:12 train.py: 79] Epoch 8, iter 1000/6416, lr 0.100000, loss 3.968898
+INFO 2020-12-03 16:20:37 train.py: 79] Epoch 8, iter 1200/6416, lr 0.100000, loss 4.009813
+INFO 2020-12-03 16:26:02 train.py: 79] Epoch 8, iter 1400/6416, lr 0.100000, loss 4.036713
+INFO 2020-12-03 16:31:28 train.py: 79] Epoch 8, iter 1600/6416, lr 0.100000, loss 4.058011
+INFO 2020-12-03 16:36:53 train.py: 79] Epoch 8, iter 1800/6416, lr 0.100000, loss 4.090811
+INFO 2020-12-03 16:42:18 train.py: 79] Epoch 8, iter 2000/6416, lr 0.100000, loss 4.086257
+INFO 2020-12-03 16:47:43 train.py: 79] Epoch 8, iter 2200/6416, lr 0.100000, loss 4.135278
+INFO 2020-12-03 16:53:08 train.py: 79] Epoch 8, iter 2400/6416, lr 0.100000, loss 4.140927
+INFO 2020-12-03 16:58:33 train.py: 79] Epoch 8, iter 2600/6416, lr 0.100000, loss 4.124890
+INFO 2020-12-03 17:03:58 train.py: 79] Epoch 8, iter 2800/6416, lr 0.100000, loss 4.173703
+INFO 2020-12-03 17:09:22 train.py: 92] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-12-03 17:09:24 train.py: 79] Epoch 8, iter 3000/6416, lr 0.100000, loss 4.184023
+INFO 2020-12-03 17:14:49 train.py: 79] Epoch 8, iter 3200/6416, lr 0.100000, loss 4.165257
+INFO 2020-12-03 17:20:14 train.py: 79] Epoch 8, iter 3400/6416, lr 0.100000, loss 4.146530
+INFO 2020-12-03 17:25:40 train.py: 79] Epoch 8, iter 3600/6416, lr 0.100000, loss 4.154909
+INFO 2020-12-03 17:31:05 train.py: 79] Epoch 8, iter 3800/6416, lr 0.100000, loss 4.171395
+INFO 2020-12-03 17:36:30 train.py: 79] Epoch 8, iter 4000/6416, lr 0.100000, loss 4.131770
+INFO 2020-12-03 17:41:55 train.py: 79] Epoch 8, iter 4200/6416, lr 0.100000, loss 4.159292
+INFO 2020-12-03 17:47:20 train.py: 79] Epoch 8, iter 4400/6416, lr 0.100000, loss 4.159086
+INFO 2020-12-03 17:52:45 train.py: 79] Epoch 8, iter 4600/6416, lr 0.100000, loss 4.161169
+INFO 2020-12-03 17:58:11 train.py: 79] Epoch 8, iter 4800/6416, lr 0.100000, loss 4.176258
+INFO 2020-12-03 18:03:36 train.py: 79] Epoch 8, iter 5000/6416, lr 0.100000, loss 4.174733
+INFO 2020-12-03 18:09:01 train.py: 79] Epoch 8, iter 5200/6416, lr 0.100000, loss 4.162446
+INFO 2020-12-03 18:14:26 train.py: 79] Epoch 8, iter 5400/6416, lr 0.100000, loss 4.163522
+INFO 2020-12-03 18:19:51 train.py: 79] Epoch 8, iter 5600/6416, lr 0.100000, loss 4.156853
+INFO 2020-12-03 18:25:16 train.py: 79] Epoch 8, iter 5800/6416, lr 0.100000, loss 4.143952
+INFO 2020-12-03 18:30:40 train.py: 92] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-12-03 18:30:42 train.py: 79] Epoch 8, iter 6000/6416, lr 0.100000, loss 4.175537
+INFO 2020-12-03 18:36:07 train.py: 79] Epoch 8, iter 6200/6416, lr 0.100000, loss 4.148403
+INFO 2020-12-03 18:41:32 train.py: 79] Epoch 8, iter 6400/6416, lr 0.100000, loss 4.140968
+INFO 2020-12-03 18:41:56 train.py: 97] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-12-03 18:41:59 train.py: 79] Epoch 9, iter 0/6416, lr 0.100000, loss 4.117626
+INFO 2020-12-03 18:47:25 train.py: 79] Epoch 9, iter 200/6416, lr 0.100000, loss 3.670153
+INFO 2020-12-03 18:52:50 train.py: 79] Epoch 9, iter 400/6416, lr 0.100000, loss 3.637191
+INFO 2020-12-03 18:58:16 train.py: 79] Epoch 9, iter 600/6416, lr 0.100000, loss 3.723079
+INFO 2020-12-03 19:03:41 train.py: 79] Epoch 9, iter 800/6416, lr 0.100000, loss 3.782626
+INFO 2020-12-03 19:09:06 train.py: 79] Epoch 9, iter 1000/6416, lr 0.100000, loss 3.825913
+INFO 2020-12-03 19:14:32 train.py: 79] Epoch 9, iter 1200/6416, lr 0.100000, loss 3.869075
+INFO 2020-12-03 19:19:57 train.py: 79] Epoch 9, iter 1400/6416, lr 0.100000, loss 3.910492
+INFO 2020-12-03 19:25:22 train.py: 79] Epoch 9, iter 1600/6416, lr 0.100000, loss 3.942527
+INFO 2020-12-03 19:30:48 train.py: 79] Epoch 9, iter 1800/6416, lr 0.100000, loss 3.973456
+INFO 2020-12-03 19:36:13 train.py: 79] Epoch 9, iter 2000/6416, lr 0.100000, loss 4.017191
+INFO 2020-12-03 19:41:38 train.py: 79] Epoch 9, iter 2200/6416, lr 0.100000, loss 4.030699
+INFO 2020-12-03 19:47:03 train.py: 79] Epoch 9, iter 2400/6416, lr 0.100000, loss 4.042986
+INFO 2020-12-03 19:52:29 train.py: 79] Epoch 9, iter 2600/6416, lr 0.100000, loss 4.053050
+INFO 2020-12-03 19:57:54 train.py: 79] Epoch 9, iter 2800/6416, lr 0.100000, loss 4.034845
+INFO 2020-12-03 20:03:19 train.py: 92] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-12-03 20:03:20 train.py: 79] Epoch 9, iter 3000/6416, lr 0.100000, loss 4.062791
+INFO 2020-12-03 20:08:46 train.py: 79] Epoch 9, iter 3200/6416, lr 0.100000, loss 4.053587
+INFO 2020-12-03 20:14:12 train.py: 79] Epoch 9, iter 3400/6416, lr 0.100000, loss 4.057787
+INFO 2020-12-03 20:19:37 train.py: 79] Epoch 9, iter 3600/6416, lr 0.100000, loss 4.053538
+INFO 2020-12-03 20:25:03 train.py: 79] Epoch 9, iter 3800/6416, lr 0.100000, loss 4.068157
+INFO 2020-12-03 20:30:28 train.py: 79] Epoch 9, iter 4000/6416, lr 0.100000, loss 4.116380
+INFO 2020-12-03 20:35:53 train.py: 79] Epoch 9, iter 4200/6416, lr 0.100000, loss 4.055538
+INFO 2020-12-03 20:41:18 train.py: 79] Epoch 9, iter 4400/6416, lr 0.100000, loss 4.080986
+INFO 2020-12-03 20:46:43 train.py: 79] Epoch 9, iter 4600/6416, lr 0.100000, loss 4.066792
+INFO 2020-12-03 20:52:09 train.py: 79] Epoch 9, iter 4800/6416, lr 0.100000, loss 4.083648
+INFO 2020-12-03 20:57:34 train.py: 79] Epoch 9, iter 5000/6416, lr 0.100000, loss 4.043272
+INFO 2020-12-03 21:02:59 train.py: 79] Epoch 9, iter 5200/6416, lr 0.100000, loss 4.077645
+INFO 2020-12-03 21:08:25 train.py: 79] Epoch 9, iter 5400/6416, lr 0.100000, loss 4.080447
+INFO 2020-12-03 21:13:51 train.py: 79] Epoch 9, iter 5600/6416, lr 0.100000, loss 4.072473
+INFO 2020-12-03 21:19:16 train.py: 79] Epoch 9, iter 5800/6416, lr 0.100000, loss 4.057376
+INFO 2020-12-03 21:24:40 train.py: 92] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-12-03 21:24:42 train.py: 79] Epoch 9, iter 6000/6416, lr 0.100000, loss 4.046439
+INFO 2020-12-03 21:30:07 train.py: 79] Epoch 9, iter 6200/6416, lr 0.100000, loss 4.045076
+INFO 2020-12-03 21:35:33 train.py: 79] Epoch 9, iter 6400/6416, lr 0.100000, loss 4.044658
+INFO 2020-12-03 21:35:57 train.py: 97] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-12-03 21:36:00 train.py: 79] Epoch 10, iter 0/6416, lr 0.010000, loss 4.014374
+INFO 2020-12-03 21:41:25 train.py: 79] Epoch 10, iter 200/6416, lr 0.010000, loss 3.045075
+INFO 2020-12-03 21:46:51 train.py: 79] Epoch 10, iter 400/6416, lr 0.010000, loss 2.772279
+INFO 2020-12-03 21:52:16 train.py: 79] Epoch 10, iter 600/6416, lr 0.010000, loss 2.661138
+INFO 2020-12-03 21:57:41 train.py: 79] Epoch 10, iter 800/6416, lr 0.010000, loss 2.591469
+INFO 2020-12-03 22:03:06 train.py: 79] Epoch 10, iter 1000/6416, lr 0.010000, loss 2.528380
+INFO 2020-12-03 22:08:32 train.py: 79] Epoch 10, iter 1200/6416, lr 0.010000, loss 2.488472
+INFO 2020-12-03 22:13:57 train.py: 79] Epoch 10, iter 1400/6416, lr 0.010000, loss 2.453629
+INFO 2020-12-03 22:19:22 train.py: 79] Epoch 10, iter 1600/6416, lr 0.010000, loss 2.428982
+INFO 2020-12-03 22:24:47 train.py: 79] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.373689
+INFO 2020-12-03 22:30:12 train.py: 79] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.351944
+INFO 2020-12-03 22:35:37 train.py: 79] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.322801
+INFO 2020-12-03 22:41:02 train.py: 79] Epoch 10, iter 2400/6416, lr 0.010000, loss 2.277094
+INFO 2020-12-03 22:46:27 train.py: 79] Epoch 10, iter 2600/6416, lr 0.010000, loss 2.283659
+INFO 2020-12-03 22:51:52 train.py: 79] Epoch 10, iter 2800/6416, lr 0.010000, loss 2.256703
+INFO 2020-12-03 22:57:15 train.py: 92] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-12-03 22:57:17 train.py: 79] Epoch 10, iter 3000/6416, lr 0.010000, loss 2.237671
+INFO 2020-12-03 23:02:42 train.py: 79] Epoch 10, iter 3200/6416, lr 0.010000, loss 2.204354
+INFO 2020-12-03 23:08:07 train.py: 79] Epoch 10, iter 3400/6416, lr 0.010000, loss 2.196482
+INFO 2020-12-03 23:13:32 train.py: 79] Epoch 10, iter 3600/6416, lr 0.010000, loss 2.171479
+INFO 2020-12-03 23:18:57 train.py: 79] Epoch 10, iter 3800/6416, lr 0.010000, loss 2.157480
+INFO 2020-12-03 23:24:22 train.py: 79] Epoch 10, iter 4000/6416, lr 0.010000, loss 2.142582
+INFO 2020-12-03 23:29:47 train.py: 79] Epoch 10, iter 4200/6416, lr 0.010000, loss 2.118018
+INFO 2020-12-03 23:35:13 train.py: 79] Epoch 10, iter 4400/6416, lr 0.010000, loss 2.128270
+INFO 2020-12-03 23:40:38 train.py: 79] Epoch 10, iter 4600/6416, lr 0.010000, loss 2.094376
+INFO 2020-12-03 23:46:03 train.py: 79] Epoch 10, iter 4800/6416, lr 0.010000, loss 2.072259
+INFO 2020-12-03 23:51:28 train.py: 79] Epoch 10, iter 5000/6416, lr 0.010000, loss 2.081495
+INFO 2020-12-03 23:56:53 train.py: 79] Epoch 10, iter 5200/6416, lr 0.010000, loss 2.045563
+INFO 2020-12-04 00:02:18 train.py: 79] Epoch 10, iter 5400/6416, lr 0.010000, loss 2.044969
+INFO 2020-12-04 00:07:44 train.py: 79] Epoch 10, iter 5600/6416, lr 0.010000, loss 2.032908
+INFO 2020-12-04 00:13:09 train.py: 79] Epoch 10, iter 5800/6416, lr 0.010000, loss 2.007562
+INFO 2020-12-04 00:18:33 train.py: 92] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-12-04 00:18:35 train.py: 79] Epoch 10, iter 6000/6416, lr 0.010000, loss 1.989081
+INFO 2020-12-04 00:24:00 train.py: 79] Epoch 10, iter 6200/6416, lr 0.010000, loss 2.007318
+INFO 2020-12-04 00:29:25 train.py: 79] Epoch 10, iter 6400/6416, lr 0.010000, loss 1.985579
+INFO 2020-12-04 00:29:50 train.py: 97] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-12-04 00:29:53 train.py: 79] Epoch 11, iter 0/6416, lr 0.010000, loss 1.945808
+INFO 2020-12-04 00:35:18 train.py: 79] Epoch 11, iter 200/6416, lr 0.010000, loss 1.689300
+INFO 2020-12-04 00:40:43 train.py: 79] Epoch 11, iter 400/6416, lr 0.010000, loss 1.669825
+INFO 2020-12-04 00:46:08 train.py: 79] Epoch 11, iter 600/6416, lr 0.010000, loss 1.684342
+INFO 2020-12-04 00:51:33 train.py: 79] Epoch 11, iter 800/6416, lr 0.010000, loss 1.675694
+INFO 2020-12-04 00:56:58 train.py: 79] Epoch 11, iter 1000/6416, lr 0.010000, loss 1.672571
+INFO 2020-12-04 01:02:23 train.py: 79] Epoch 11, iter 1200/6416, lr 0.010000, loss 1.673412
+INFO 2020-12-04 01:07:48 train.py: 79] Epoch 11, iter 1400/6416, lr 0.010000, loss 1.673887
+INFO 2020-12-04 01:13:13 train.py: 79] Epoch 11, iter 1600/6416, lr 0.010000, loss 1.668809
+INFO 2020-12-04 01:18:38 train.py: 79] Epoch 11, iter 1800/6416, lr 0.010000, loss 1.665705
+INFO 2020-12-04 01:24:03 train.py: 79] Epoch 11, iter 2000/6416, lr 0.010000, loss 1.669471
+INFO 2020-12-04 01:29:28 train.py: 79] Epoch 11, iter 2200/6416, lr 0.010000, loss 1.661018
+INFO 2020-12-04 01:34:53 train.py: 79] Epoch 11, iter 2400/6416, lr 0.010000, loss 1.677115
+INFO 2020-12-04 01:40:18 train.py: 79] Epoch 11, iter 2600/6416, lr 0.010000, loss 1.666816
+INFO 2020-12-04 01:45:43 train.py: 79] Epoch 11, iter 2800/6416, lr 0.010000, loss 1.664066
+INFO 2020-12-04 01:51:07 train.py: 92] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-12-04 01:51:08 train.py: 79] Epoch 11, iter 3000/6416, lr 0.010000, loss 1.676867
+INFO 2020-12-04 01:56:33 train.py: 79] Epoch 11, iter 3200/6416, lr 0.010000, loss 1.662786
+INFO 2020-12-04 02:01:58 train.py: 79] Epoch 11, iter 3400/6416, lr 0.010000, loss 1.664614
+INFO 2020-12-04 02:07:23 train.py: 79] Epoch 11, iter 3600/6416, lr 0.010000, loss 1.664018
+INFO 2020-12-04 02:12:48 train.py: 79] Epoch 11, iter 3800/6416, lr 0.010000, loss 1.661976
+INFO 2020-12-04 02:18:13 train.py: 79] Epoch 11, iter 4000/6416, lr 0.010000, loss 1.641390
+INFO 2020-12-04 02:23:38 train.py: 79] Epoch 11, iter 4200/6416, lr 0.010000, loss 1.648104
+INFO 2020-12-04 02:29:03 train.py: 79] Epoch 11, iter 4400/6416, lr 0.010000, loss 1.665530
+INFO 2020-12-04 02:34:28 train.py: 79] Epoch 11, iter 4600/6416, lr 0.010000, loss 1.624464
+INFO 2020-12-04 02:39:53 train.py: 79] Epoch 11, iter 4800/6416, lr 0.010000, loss 1.646179
+INFO 2020-12-04 02:45:18 train.py: 79] Epoch 11, iter 5000/6416, lr 0.010000, loss 1.651610
+INFO 2020-12-04 02:50:43 train.py: 79] Epoch 11, iter 5200/6416, lr 0.010000, loss 1.644582
+INFO 2020-12-04 02:56:08 train.py: 79] Epoch 11, iter 5400/6416, lr 0.010000, loss 1.655777
+INFO 2020-12-04 03:01:33 train.py: 79] Epoch 11, iter 5600/6416, lr 0.010000, loss 1.652178
+INFO 2020-12-04 03:06:58 train.py: 79] Epoch 11, iter 5800/6416, lr 0.010000, loss 1.646212
+INFO 2020-12-04 03:12:22 train.py: 92] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-12-04 03:12:23 train.py: 79] Epoch 11, iter 6000/6416, lr 0.010000, loss 1.635665
+INFO 2020-12-04 03:17:48 train.py: 79] Epoch 11, iter 6200/6416, lr 0.010000, loss 1.637378
+INFO 2020-12-04 03:23:14 train.py: 79] Epoch 11, iter 6400/6416, lr 0.010000, loss 1.637290
+INFO 2020-12-04 03:23:38 train.py: 97] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-12-04 03:23:41 train.py: 79] Epoch 12, iter 0/6416, lr 0.010000, loss 1.586589
+INFO 2020-12-04 03:29:06 train.py: 79] Epoch 12, iter 200/6416, lr 0.010000, loss 1.378230
+INFO 2020-12-04 03:34:31 train.py: 79] Epoch 12, iter 400/6416, lr 0.010000, loss 1.359621
+INFO 2020-12-04 03:39:56 train.py: 79] Epoch 12, iter 600/6416, lr 0.010000, loss 1.374209
+INFO 2020-12-04 03:45:22 train.py: 79] Epoch 12, iter 800/6416, lr 0.010000, loss 1.361567
+INFO 2020-12-04 03:50:47 train.py: 79] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.370249
+INFO 2020-12-04 03:56:12 train.py: 79] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.384933
+INFO 2020-12-04 04:01:37 train.py: 79] Epoch 12, iter 1400/6416, lr 0.010000, loss 1.401534
+INFO 2020-12-04 04:07:02 train.py: 79] Epoch 12, iter 1600/6416, lr 0.010000, loss 1.404983
+INFO 2020-12-04 04:12:27 train.py: 79] Epoch 12, iter 1800/6416, lr 0.010000, loss 1.388584
+INFO 2020-12-04 04:17:52 train.py: 79] Epoch 12, iter 2000/6416, lr 0.010000, loss 1.404521
+INFO 2020-12-04 04:23:17 train.py: 79] Epoch 12, iter 2200/6416, lr 0.010000, loss 1.395233
+INFO 2020-12-04 04:28:42 train.py: 79] Epoch 12, iter 2400/6416, lr 0.010000, loss 1.399625
+INFO 2020-12-04 04:34:08 train.py: 79] Epoch 12, iter 2600/6416, lr 0.010000, loss 1.407339
+INFO 2020-12-04 04:39:33 train.py: 79] Epoch 12, iter 2800/6416, lr 0.010000, loss 1.416177
+INFO 2020-12-04 04:44:57 train.py: 92] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-12-04 04:44:59 train.py: 79] Epoch 12, iter 3000/6416, lr 0.010000, loss 1.404588
+INFO 2020-12-04 04:50:24 train.py: 79] Epoch 12, iter 3200/6416, lr 0.010000, loss 1.425275
+INFO 2020-12-04 04:55:49 train.py: 79] Epoch 12, iter 3400/6416, lr 0.010000, loss 1.436283
+INFO 2020-12-04 05:01:15 train.py: 79] Epoch 12, iter 3600/6416, lr 0.010000, loss 1.432694
+INFO 2020-12-04 05:06:40 train.py: 79] Epoch 12, iter 3800/6416, lr 0.010000, loss 1.446021
+INFO 2020-12-04 05:12:05 train.py: 79] Epoch 12, iter 4000/6416, lr 0.010000, loss 1.451524
+INFO 2020-12-04 05:17:30 train.py: 79] Epoch 12, iter 4200/6416, lr 0.010000, loss 1.435503
+INFO 2020-12-04 05:22:55 train.py: 79] Epoch 12, iter 4400/6416, lr 0.010000, loss 1.445245
+INFO 2020-12-04 05:28:20 train.py: 79] Epoch 12, iter 4600/6416, lr 0.010000, loss 1.455817
+INFO 2020-12-04 05:33:46 train.py: 79] Epoch 12, iter 4800/6416, lr 0.010000, loss 1.450810
+INFO 2020-12-04 05:39:11 train.py: 79] Epoch 12, iter 5000/6416, lr 0.010000, loss 1.447359
+INFO 2020-12-04 05:44:36 train.py: 79] Epoch 12, iter 5200/6416, lr 0.010000, loss 1.457142
+INFO 2020-12-04 05:50:01 train.py: 79] Epoch 12, iter 5400/6416, lr 0.010000, loss 1.455452
+INFO 2020-12-04 05:55:26 train.py: 79] Epoch 12, iter 5600/6416, lr 0.010000, loss 1.463986
+INFO 2020-12-04 06:00:51 train.py: 79] Epoch 12, iter 5800/6416, lr 0.010000, loss 1.465392
+INFO 2020-12-04 06:06:15 train.py: 92] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-12-04 06:06:16 train.py: 79] Epoch 12, iter 6000/6416, lr 0.010000, loss 1.475555
+INFO 2020-12-04 06:11:41 train.py: 79] Epoch 12, iter 6200/6416, lr 0.010000, loss 1.478011
+INFO 2020-12-04 06:17:06 train.py: 79] Epoch 12, iter 6400/6416, lr 0.010000, loss 1.482399
+INFO 2020-12-04 06:17:30 train.py: 97] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-12-04 06:17:33 train.py: 79] Epoch 13, iter 0/6416, lr 0.001000, loss 1.496635
+INFO 2020-12-04 06:22:58 train.py: 79] Epoch 13, iter 200/6416, lr 0.001000, loss 1.166110
+INFO 2020-12-04 06:28:23 train.py: 79] Epoch 13, iter 400/6416, lr 0.001000, loss 1.137528
+INFO 2020-12-04 06:33:49 train.py: 79] Epoch 13, iter 600/6416, lr 0.001000, loss 1.144449
+INFO 2020-12-04 06:39:14 train.py: 79] Epoch 13, iter 800/6416, lr 0.001000, loss 1.136726
+INFO 2020-12-04 06:44:39 train.py: 79] Epoch 13, iter 1000/6416, lr 0.001000, loss 1.123902
+INFO 2020-12-04 06:50:04 train.py: 79] Epoch 13, iter 1200/6416, lr 0.001000, loss 1.135582
+INFO 2020-12-04 06:55:29 train.py: 79] Epoch 13, iter 1400/6416, lr 0.001000, loss 1.114972
+INFO 2020-12-04 07:00:55 train.py: 79] Epoch 13, iter 1600/6416, lr 0.001000, loss 1.126917
+INFO 2020-12-04 07:06:20 train.py: 79] Epoch 13, iter 1800/6416, lr 0.001000, loss 1.113411
+INFO 2020-12-04 07:11:45 train.py: 79] Epoch 13, iter 2000/6416, lr 0.001000, loss 1.121289
+INFO 2020-12-04 07:17:10 train.py: 79] Epoch 13, iter 2200/6416, lr 0.001000, loss 1.122110
+INFO 2020-12-04 07:22:35 train.py: 79] Epoch 13, iter 2400/6416, lr 0.001000, loss 1.119020
+INFO 2020-12-04 07:28:00 train.py: 79] Epoch 13, iter 2600/6416, lr 0.001000, loss 1.113377
+INFO 2020-12-04 07:33:25 train.py: 79] Epoch 13, iter 2800/6416, lr 0.001000, loss 1.116419
+INFO 2020-12-04 07:38:49 train.py: 92] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-12-04 07:38:51 train.py: 79] Epoch 13, iter 3000/6416, lr 0.001000, loss 1.117171
+INFO 2020-12-04 07:44:16 train.py: 79] Epoch 13, iter 3200/6416, lr 0.001000, loss 1.109928
+INFO 2020-12-04 07:49:41 train.py: 79] Epoch 13, iter 3400/6416, lr 0.001000, loss 1.110337
+INFO 2020-12-04 07:55:06 train.py: 79] Epoch 13, iter 3600/6416, lr 0.001000, loss 1.106861
+INFO 2020-12-04 08:00:31 train.py: 79] Epoch 13, iter 3800/6416, lr 0.001000, loss 1.119624
+INFO 2020-12-04 08:05:56 train.py: 79] Epoch 13, iter 4000/6416, lr 0.001000, loss 1.117525
+INFO 2020-12-04 08:11:21 train.py: 79] Epoch 13, iter 4200/6416, lr 0.001000, loss 1.108125
+INFO 2020-12-04 08:16:46 train.py: 79] Epoch 13, iter 4400/6416, lr 0.001000, loss 1.112367
+INFO 2020-12-04 08:22:11 train.py: 79] Epoch 13, iter 4600/6416, lr 0.001000, loss 1.112227
+INFO 2020-12-04 08:27:36 train.py: 79] Epoch 13, iter 4800/6416, lr 0.001000, loss 1.110203
+INFO 2020-12-04 08:33:02 train.py: 79] Epoch 13, iter 5000/6416, lr 0.001000, loss 1.106777
+INFO 2020-12-04 08:38:27 train.py: 79] Epoch 13, iter 5200/6416, lr 0.001000, loss 1.111004
+INFO 2020-12-04 08:43:52 train.py: 79] Epoch 13, iter 5400/6416, lr 0.001000, loss 1.104399
+INFO 2020-12-04 08:49:17 train.py: 79] Epoch 13, iter 5600/6416, lr 0.001000, loss 1.113798
+INFO 2020-12-04 08:54:42 train.py: 79] Epoch 13, iter 5800/6416, lr 0.001000, loss 1.113383
+INFO 2020-12-04 09:00:06 train.py: 92] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-12-04 09:00:08 train.py: 79] Epoch 13, iter 6000/6416, lr 0.001000, loss 1.111207
+INFO 2020-12-04 09:05:33 train.py: 79] Epoch 13, iter 6200/6416, lr 0.001000, loss 1.116616
+INFO 2020-12-04 09:10:58 train.py: 79] Epoch 13, iter 6400/6416, lr 0.001000, loss 1.118895
+INFO 2020-12-04 09:11:22 train.py: 97] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-12-04 09:11:24 train.py: 79] Epoch 14, iter 0/6416, lr 0.001000, loss 1.123141
+INFO 2020-12-04 09:16:50 train.py: 79] Epoch 14, iter 200/6416, lr 0.001000, loss 1.068512
+INFO 2020-12-04 09:22:15 train.py: 79] Epoch 14, iter 400/6416, lr 0.001000, loss 1.069994
+INFO 2020-12-04 09:27:40 train.py: 79] Epoch 14, iter 600/6416, lr 0.001000, loss 1.073890
+INFO 2020-12-04 09:33:05 train.py: 79] Epoch 14, iter 800/6416, lr 0.001000, loss 1.077538
+INFO 2020-12-04 09:38:31 train.py: 79] Epoch 14, iter 1000/6416, lr 0.001000, loss 1.068196
+INFO 2020-12-04 09:43:56 train.py: 79] Epoch 14, iter 1200/6416, lr 0.001000, loss 1.070067
+INFO 2020-12-04 09:49:21 train.py: 79] Epoch 14, iter 1400/6416, lr 0.001000, loss 1.079795
+INFO 2020-12-04 09:54:46 train.py: 79] Epoch 14, iter 1600/6416, lr 0.001000, loss 1.072936
+INFO 2020-12-04 10:00:11 train.py: 79] Epoch 14, iter 1800/6416, lr 0.001000, loss 1.078108
+INFO 2020-12-04 10:05:35 train.py: 79] Epoch 14, iter 2000/6416, lr 0.001000, loss 1.064167
+INFO 2020-12-04 10:11:00 train.py: 79] Epoch 14, iter 2200/6416, lr 0.001000, loss 1.068662
+INFO 2020-12-04 10:16:25 train.py: 79] Epoch 14, iter 2400/6416, lr 0.001000, loss 1.074293
+INFO 2020-12-04 10:21:50 train.py: 79] Epoch 14, iter 2600/6416, lr 0.001000, loss 1.075261
+INFO 2020-12-04 10:27:15 train.py: 79] Epoch 14, iter 2800/6416, lr 0.001000, loss 1.078322
+INFO 2020-12-04 10:32:39 train.py: 92] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-12-04 10:32:41 train.py: 79] Epoch 14, iter 3000/6416, lr 0.001000, loss 1.058165
+INFO 2020-12-04 10:38:06 train.py: 79] Epoch 14, iter 3200/6416, lr 0.001000, loss 1.084067
+INFO 2020-12-04 10:43:31 train.py: 79] Epoch 14, iter 3400/6416, lr 0.001000, loss 1.093464
+INFO 2020-12-04 10:48:56 train.py: 79] Epoch 14, iter 3600/6416, lr 0.001000, loss 1.072578
+INFO 2020-12-04 10:54:21 train.py: 79] Epoch 14, iter 3800/6416, lr 0.001000, loss 1.069096
+INFO 2020-12-04 10:59:46 train.py: 79] Epoch 14, iter 4000/6416, lr 0.001000, loss 1.077155
+INFO 2020-12-04 11:05:11 train.py: 79] Epoch 14, iter 4200/6416, lr 0.001000, loss 1.077070
+INFO 2020-12-04 11:10:36 train.py: 79] Epoch 14, iter 4400/6416, lr 0.001000, loss 1.075185
+INFO 2020-12-04 11:16:01 train.py: 79] Epoch 14, iter 4600/6416, lr 0.001000, loss 1.078912
+INFO 2020-12-04 11:21:26 train.py: 79] Epoch 14, iter 4800/6416, lr 0.001000, loss 1.079929
+INFO 2020-12-04 11:26:51 train.py: 79] Epoch 14, iter 5000/6416, lr 0.001000, loss 1.070700
+INFO 2020-12-04 11:32:16 train.py: 79] Epoch 14, iter 5200/6416, lr 0.001000, loss 1.069580
+INFO 2020-12-04 11:37:41 train.py: 79] Epoch 14, iter 5400/6416, lr 0.001000, loss 1.073560
+INFO 2020-12-04 11:43:07 train.py: 79] Epoch 14, iter 5600/6416, lr 0.001000, loss 1.066351
+INFO 2020-12-04 11:48:32 train.py: 79] Epoch 14, iter 5800/6416, lr 0.001000, loss 1.082068
+INFO 2020-12-04 11:53:56 train.py: 92] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-12-04 11:53:58 train.py: 79] Epoch 14, iter 6000/6416, lr 0.001000, loss 1.076696
+INFO 2020-12-04 11:59:23 train.py: 79] Epoch 14, iter 6200/6416, lr 0.001000, loss 1.072486
+INFO 2020-12-04 12:04:48 train.py: 79] Epoch 14, iter 6400/6416, lr 0.001000, loss 1.075122
+INFO 2020-12-04 12:05:12 train.py: 97] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-12-04 12:05:15 train.py: 79] Epoch 15, iter 0/6416, lr 0.001000, loss 1.093127
+INFO 2020-12-04 12:10:40 train.py: 79] Epoch 15, iter 200/6416, lr 0.001000, loss 1.038921
+INFO 2020-12-04 12:16:05 train.py: 79] Epoch 15, iter 400/6416, lr 0.001000, loss 1.021710
+INFO 2020-12-04 12:21:30 train.py: 79] Epoch 15, iter 600/6416, lr 0.001000, loss 1.036300
+INFO 2020-12-04 12:26:55 train.py: 79] Epoch 15, iter 800/6416, lr 0.001000, loss 1.036481
+INFO 2020-12-04 12:32:21 train.py: 79] Epoch 15, iter 1000/6416, lr 0.001000, loss 1.035608
+INFO 2020-12-04 12:37:46 train.py: 79] Epoch 15, iter 1200/6416, lr 0.001000, loss 1.034641
+INFO 2020-12-04 12:43:11 train.py: 79] Epoch 15, iter 1400/6416, lr 0.001000, loss 1.038440
+INFO 2020-12-04 12:48:36 train.py: 79] Epoch 15, iter 1600/6416, lr 0.001000, loss 1.043738
+INFO 2020-12-04 12:54:01 train.py: 79] Epoch 15, iter 1800/6416, lr 0.001000, loss 1.039400
+INFO 2020-12-04 12:59:26 train.py: 79] Epoch 15, iter 2000/6416, lr 0.001000, loss 1.043403
+INFO 2020-12-04 13:04:51 train.py: 79] Epoch 15, iter 2200/6416, lr 0.001000, loss 1.052029
+INFO 2020-12-04 13:10:16 train.py: 79] Epoch 15, iter 2400/6416, lr 0.001000, loss 1.031198
+INFO 2020-12-04 13:15:41 train.py: 79] Epoch 15, iter 2600/6416, lr 0.001000, loss 1.038019
+INFO 2020-12-04 13:21:05 train.py: 79] Epoch 15, iter 2800/6416, lr 0.001000, loss 1.044535
+INFO 2020-12-04 13:26:29 train.py: 92] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-12-04 13:26:31 train.py: 79] Epoch 15, iter 3000/6416, lr 0.001000, loss 1.046154
+INFO 2020-12-04 13:31:56 train.py: 79] Epoch 15, iter 3200/6416, lr 0.001000, loss 1.046215
+INFO 2020-12-04 13:37:21 train.py: 79] Epoch 15, iter 3400/6416, lr 0.001000, loss 1.036994
+INFO 2020-12-04 13:42:46 train.py: 79] Epoch 15, iter 3600/6416, lr 0.001000, loss 1.042916
+INFO 2020-12-04 13:48:11 train.py: 79] Epoch 15, iter 3800/6416, lr 0.001000, loss 1.053086
+INFO 2020-12-04 13:53:36 train.py: 79] Epoch 15, iter 4000/6416, lr 0.001000, loss 1.046281
+INFO 2020-12-04 13:59:01 train.py: 79] Epoch 15, iter 4200/6416, lr 0.001000, loss 1.046119
+INFO 2020-12-04 14:04:26 train.py: 79] Epoch 15, iter 4400/6416, lr 0.001000, loss 1.068466
+INFO 2020-12-04 14:09:51 train.py: 79] Epoch 15, iter 4600/6416, lr 0.001000, loss 1.052074
+INFO 2020-12-04 14:15:16 train.py: 79] Epoch 15, iter 4800/6416, lr 0.001000, loss 1.048229
+INFO 2020-12-04 14:20:42 train.py: 79] Epoch 15, iter 5000/6416, lr 0.001000, loss 1.043999
+INFO 2020-12-04 14:26:07 train.py: 79] Epoch 15, iter 5200/6416, lr 0.001000, loss 1.038350
+INFO 2020-12-04 14:31:31 train.py: 79] Epoch 15, iter 5400/6416, lr 0.001000, loss 1.055580
+INFO 2020-12-04 14:36:56 train.py: 79] Epoch 15, iter 5600/6416, lr 0.001000, loss 1.042690
+INFO 2020-12-04 14:42:21 train.py: 79] Epoch 15, iter 5800/6416, lr 0.001000, loss 1.051873
+INFO 2020-12-04 14:47:45 train.py: 92] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-12-04 14:47:47 train.py: 79] Epoch 15, iter 6000/6416, lr 0.001000, loss 1.061339
+INFO 2020-12-04 14:53:12 train.py: 79] Epoch 15, iter 6200/6416, lr 0.001000, loss 1.056667
+INFO 2020-12-04 14:58:37 train.py: 79] Epoch 15, iter 6400/6416, lr 0.001000, loss 1.052175
+INFO 2020-12-04 14:59:01 train.py: 97] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-12-04 14:59:04 train.py: 79] Epoch 16, iter 0/6416, lr 0.000100, loss 1.035008
+INFO 2020-12-04 15:04:29 train.py: 79] Epoch 16, iter 200/6416, lr 0.000100, loss 0.999609
+INFO 2020-12-04 15:09:55 train.py: 79] Epoch 16, iter 400/6416, lr 0.000100, loss 1.001416
+INFO 2020-12-04 15:15:20 train.py: 79] Epoch 16, iter 600/6416, lr 0.000100, loss 0.997739
+INFO 2020-12-04 15:20:45 train.py: 79] Epoch 16, iter 800/6416, lr 0.000100, loss 1.007479
+INFO 2020-12-04 15:26:10 train.py: 79] Epoch 16, iter 1000/6416, lr 0.000100, loss 1.007760
+INFO 2020-12-04 15:31:35 train.py: 79] Epoch 16, iter 1200/6416, lr 0.000100, loss 1.008566
+INFO 2020-12-04 15:37:00 train.py: 79] Epoch 16, iter 1400/6416, lr 0.000100, loss 1.005205
+INFO 2020-12-04 15:42:25 train.py: 79] Epoch 16, iter 1600/6416, lr 0.000100, loss 1.015564
+INFO 2020-12-04 15:47:50 train.py: 79] Epoch 16, iter 1800/6416, lr 0.000100, loss 1.006663
+INFO 2020-12-04 15:53:14 train.py: 79] Epoch 16, iter 2000/6416, lr 0.000100, loss 0.998463
+INFO 2020-12-04 15:58:39 train.py: 79] Epoch 16, iter 2200/6416, lr 0.000100, loss 1.000941
+INFO 2020-12-04 16:04:04 train.py: 79] Epoch 16, iter 2400/6416, lr 0.000100, loss 1.019119
+INFO 2020-12-04 16:09:29 train.py: 79] Epoch 16, iter 2600/6416, lr 0.000100, loss 0.996453
+INFO 2020-12-04 16:14:54 train.py: 79] Epoch 16, iter 2800/6416, lr 0.000100, loss 1.002589
+INFO 2020-12-04 16:20:18 train.py: 92] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-12-04 16:20:19 train.py: 79] Epoch 16, iter 3000/6416, lr 0.000100, loss 1.004591
+INFO 2020-12-04 16:25:45 train.py: 79] Epoch 16, iter 3200/6416, lr 0.000100, loss 1.006190
+INFO 2020-12-04 16:31:10 train.py: 79] Epoch 16, iter 3400/6416, lr 0.000100, loss 0.997385
+INFO 2020-12-04 16:36:35 train.py: 79] Epoch 16, iter 3600/6416, lr 0.000100, loss 0.998832
+INFO 2020-12-04 16:42:00 train.py: 79] Epoch 16, iter 3800/6416, lr 0.000100, loss 1.007833
+INFO 2020-12-04 16:47:25 train.py: 79] Epoch 16, iter 4000/6416, lr 0.000100, loss 0.999878
+INFO 2020-12-04 16:52:50 train.py: 79] Epoch 16, iter 4200/6416, lr 0.000100, loss 1.002242
+INFO 2020-12-04 16:58:15 train.py: 79] Epoch 16, iter 4400/6416, lr 0.000100, loss 1.008763
+INFO 2020-12-04 17:03:40 train.py: 79] Epoch 16, iter 4600/6416, lr 0.000100, loss 1.017141
+INFO 2020-12-04 17:09:05 train.py: 79] Epoch 16, iter 4800/6416, lr 0.000100, loss 0.994520
+INFO 2020-12-04 17:14:31 train.py: 79] Epoch 16, iter 5000/6416, lr 0.000100, loss 0.995979
+INFO 2020-12-04 17:19:56 train.py: 79] Epoch 16, iter 5200/6416, lr 0.000100, loss 1.001580
+INFO 2020-12-04 17:25:21 train.py: 79] Epoch 16, iter 5400/6416, lr 0.000100, loss 1.002718
+INFO 2020-12-04 17:30:46 train.py: 79] Epoch 16, iter 5600/6416, lr 0.000100, loss 1.018163
+INFO 2020-12-04 17:36:11 train.py: 79] Epoch 16, iter 5800/6416, lr 0.000100, loss 1.001848
+INFO 2020-12-04 17:41:34 train.py: 92] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-12-04 17:41:36 train.py: 79] Epoch 16, iter 6000/6416, lr 0.000100, loss 1.019192
+INFO 2020-12-04 17:47:01 train.py: 79] Epoch 16, iter 6200/6416, lr 0.000100, loss 1.004056
+INFO 2020-12-04 17:52:26 train.py: 79] Epoch 16, iter 6400/6416, lr 0.000100, loss 1.012894
+INFO 2020-12-04 17:52:50 train.py: 97] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-12-04 17:52:52 train.py: 79] Epoch 17, iter 0/6416, lr 0.000100, loss 1.016697
+INFO 2020-12-04 17:58:18 train.py: 79] Epoch 17, iter 200/6416, lr 0.000100, loss 1.003436
+INFO 2020-12-04 18:03:43 train.py: 79] Epoch 17, iter 400/6416, lr 0.000100, loss 1.001899
+INFO 2020-12-04 18:09:09 train.py: 79] Epoch 17, iter 600/6416, lr 0.000100, loss 0.997006
+INFO 2020-12-04 18:14:34 train.py: 79] Epoch 17, iter 800/6416, lr 0.000100, loss 1.003413
+INFO 2020-12-04 18:19:58 train.py: 79] Epoch 17, iter 1000/6416, lr 0.000100, loss 1.013359
+INFO 2020-12-04 18:25:23 train.py: 79] Epoch 17, iter 1200/6416, lr 0.000100, loss 1.005548
+INFO 2020-12-04 18:30:48 train.py: 79] Epoch 17, iter 1400/6416, lr 0.000100, loss 0.993645
+INFO 2020-12-04 18:36:13 train.py: 79] Epoch 17, iter 1600/6416, lr 0.000100, loss 1.007651
+INFO 2020-12-04 18:41:38 train.py: 79] Epoch 17, iter 1800/6416, lr 0.000100, loss 0.993668
+INFO 2020-12-04 18:47:03 train.py: 79] Epoch 17, iter 2000/6416, lr 0.000100, loss 1.001400
+INFO 2020-12-04 18:52:28 train.py: 79] Epoch 17, iter 2200/6416, lr 0.000100, loss 0.999513
+INFO 2020-12-04 18:57:53 train.py: 79] Epoch 17, iter 2400/6416, lr 0.000100, loss 1.002893
+INFO 2020-12-04 19:03:18 train.py: 79] Epoch 17, iter 2600/6416, lr 0.000100, loss 0.996499
+INFO 2020-12-04 19:08:43 train.py: 79] Epoch 17, iter 2800/6416, lr 0.000100, loss 0.999717
+INFO 2020-12-04 19:14:06 train.py: 92] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-12-04 19:14:08 train.py: 79] Epoch 17, iter 3000/6416, lr 0.000100, loss 0.987270
+INFO 2020-12-04 19:19:33 train.py: 79] Epoch 17, iter 3200/6416, lr 0.000100, loss 0.994398
+INFO 2020-12-04 19:24:58 train.py: 79] Epoch 17, iter 3400/6416, lr 0.000100, loss 1.003996
+INFO 2020-12-04 19:30:23 train.py: 79] Epoch 17, iter 3600/6416, lr 0.000100, loss 1.006844
+INFO 2020-12-04 19:35:48 train.py: 79] Epoch 17, iter 3800/6416, lr 0.000100, loss 0.998328
+INFO 2020-12-04 19:41:13 train.py: 79] Epoch 17, iter 4000/6416, lr 0.000100, loss 1.009933
+INFO 2020-12-04 19:46:38 train.py: 79] Epoch 17, iter 4200/6416, lr 0.000100, loss 0.999123
+INFO 2020-12-04 19:52:03 train.py: 79] Epoch 17, iter 4400/6416, lr 0.000100, loss 1.008997
+INFO 2020-12-04 19:57:28 train.py: 79] Epoch 17, iter 4600/6416, lr 0.000100, loss 1.004571
+INFO 2020-12-04 20:02:53 train.py: 79] Epoch 17, iter 4800/6416, lr 0.000100, loss 0.997180
+INFO 2020-12-04 20:08:18 train.py: 79] Epoch 17, iter 5000/6416, lr 0.000100, loss 0.998194
+INFO 2020-12-04 20:13:43 train.py: 79] Epoch 17, iter 5200/6416, lr 0.000100, loss 0.999965
+INFO 2020-12-04 20:19:08 train.py: 79] Epoch 17, iter 5400/6416, lr 0.000100, loss 0.998162
+INFO 2020-12-04 20:24:33 train.py: 79] Epoch 17, iter 5600/6416, lr 0.000100, loss 1.007894
+INFO 2020-12-04 20:29:58 train.py: 79] Epoch 17, iter 5800/6416, lr 0.000100, loss 1.006676
+INFO 2020-12-04 20:35:22 train.py: 92] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-12-04 20:35:23 train.py: 79] Epoch 17, iter 6000/6416, lr 0.000100, loss 1.004134
+INFO 2020-12-04 20:40:48 train.py: 79] Epoch 17, iter 6200/6416, lr 0.000100, loss 1.004713
+INFO 2020-12-04 20:46:13 train.py: 79] Epoch 17, iter 6400/6416, lr 0.000100, loss 0.991361
+INFO 2020-12-04 20:46:37 train.py: 97] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-12-04 20:46:38 train.py: 180] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/.gitkeep b/bob/bio/facexzoo/models/backbones/ResNet50_ir/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0b6d7cbc25c28e40f70c9ec7827fe96722080c49
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.9776666666666667 | 0.0028458329944145975 |
+| Epoch_15.pt | 0.977 | 0.002819683897877668 |
+| Epoch_14.pt | 0.9766666666666668 | 0.00275546594836383 |
+| Epoch_17_batch_2999.pt | 0.9763333333333334 | 0.0027419917065006766 |
+| Epoch_16_batch_5999.pt | 0.9761666666666666 | 0.0028114624286399953 |
+| Epoch_14_batch_2999.pt | 0.976 | 0.0027352296944646985 |
+| Epoch_12_batch_2999.pt | 0.976 | 0.0030651364942519315 |
+| Epoch_16.pt | 0.976 | 0.002487003253955483 |
+| Epoch_11_batch_2999.pt | 0.9758333333333333 | 0.0029943362173804177 |
+| Epoch_13.pt | 0.9758333333333333 | 0.00290008514136404 |
+| Epoch_17_batch_5999.pt | 0.9756666666666666 | 0.00263171539607266 |
+| Epoch_11_batch_5999.pt | 0.9755 | 0.0034609818060383313 |
+| Epoch_11.pt | 0.9754999999999999 | 0.0029297326385411587 |
+| Epoch_17.pt | 0.9753333333333332 | 0.002730712383876551 |
+| Epoch_15_batch_2999.pt | 0.975 | 0.0026988795114424634 |
+| Epoch_16_batch_2999.pt | 0.9749999999999999 | 0.002744242007828548 |
+| Epoch_10_batch_2999.pt | 0.9748333333333333 | 0.002954908032660492 |
+| Epoch_13_batch_2999.pt | 0.9746666666666668 | 0.002916534388534813 |
+| Epoch_14_batch_5999.pt | 0.9746666666666666 | 0.002968975381308303 |
+| Epoch_15_batch_5999.pt | 0.9741666666666667 | 0.0023992025424099048 |
+| Epoch_10.pt | 0.9741666666666667 | 0.0027916321169780232 |
+| Epoch_12.pt | 0.9739999999999999 | 0.0025843785221362193 |
+| Epoch_10_batch_5999.pt | 0.9736666666666668 | 0.0030307070437746303 |
+| Epoch_12_batch_5999.pt | 0.9731666666666667 | 0.0029339435392246407 |
+| Epoch_9_batch_5999.pt | 0.969 | 0.002779999111821524 |
+| Epoch_7_batch_5999.pt | 0.9686666666666666 | 0.003431876713662339 |
+| Epoch_8_batch_2999.pt | 0.9683333333333332 | 0.002810913475705229 |
+| Epoch_5_batch_5999.pt | 0.9671666666666665 | 0.002291287847477926 |
+| Epoch_5.pt | 0.9666666666666666 | 0.002865288212939474 |
+| Epoch_7_batch_2999.pt | 0.9661666666666667 | 0.003152991916369494 |
+| Epoch_8_batch_5999.pt | 0.9658333333333335 | 0.003745367509040708 |
+| Epoch_6_batch_5999.pt | 0.9654999999999999 | 0.0027448042948968127 |
+| Epoch_7.pt | 0.9648333333333333 | 0.0025391988626725457 |
+| Epoch_4_batch_2999.pt | 0.9648333333333332 | 0.0025754059969998384 |
+| Epoch_8.pt | 0.9638333333333333 | 0.0025098571106836665 |
+| Epoch_9.pt | 0.9636666666666669 | 0.003256522419945082 |
+| Epoch_6_batch_2999.pt | 0.9635 | 0.0029234049148717527 |
+| Epoch_9_batch_2999.pt | 0.9635 | 0.004241549332583639 |
+| Epoch_3_batch_5999.pt | 0.9628333333333332 | 0.0032871804872193367 |
+| Epoch_5_batch_2999.pt | 0.961 | 0.003124969135650052 |
+| Epoch_4_batch_5999.pt | 0.9606666666666666 | 0.0033259176771323943 |
+| Epoch_6.pt | 0.9603333333333334 | 0.003733399470313658 |
+| Epoch_3_batch_2999.pt | 0.9594999999999999 | 0.0027672020900916207 |
+| Epoch_4.pt | 0.9570000000000001 | 0.003208784239598593 |
+| Epoch_2_batch_5999.pt | 0.9564999999999999 | 0.00204048529691189 |
+| Epoch_3.pt | 0.9538333333333332 | 0.003881341883268571 |
+| Epoch_2_batch_2999.pt | 0.9505000000000001 | 0.0039366119417904195 |
+| Epoch_2.pt | 0.9483333333333335 | 0.0036599264648890296 |
+| Epoch_1_batch_5999.pt | 0.9398333333333332 | 0.004489012649559063 |
+| Epoch_1.pt | 0.9380000000000001 | 0.005504207145111479 |
+| Epoch_1_batch_2999.pt | 0.922 | 0.005017254179944729 |
+| Epoch_0_batch_5999.pt | 0.8675 | 0.007475267863403268 |
+| Epoch_0.pt | 0.8576666666666666 | 0.00704132423011326 |
+| Epoch_0_batch_2999.pt | 0.6758333333333334 | 0.004225510415512534 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..78899a887edd07a16b7d8a78ef63801fe90b295d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_12_batch_2999.pt | 0.9546666666666669 | 0.0031991511219751022 |
+| Epoch_14.pt | 0.9546666666666667 | 0.003312900336861243 |
+| Epoch_13.pt | 0.9543333333333335 | 0.0033536418383970207 |
+| Epoch_11_batch_5999.pt | 0.9543333333333335 | 0.0036868133384526888 |
+| Epoch_15_batch_2999.pt | 0.9541666666666668 | 0.0032131097202047253 |
+| Epoch_17_batch_2999.pt | 0.9541666666666668 | 0.0032417987690878153 |
+| Epoch_13_batch_2999.pt | 0.9541666666666668 | 0.0033448873829978647 |
+| Epoch_14_batch_2999.pt | 0.9540000000000001 | 0.0032885885748775018 |
+| Epoch_10_batch_2999.pt | 0.9538333333333334 | 0.003258890972458331 |
+| Epoch_15.pt | 0.9536666666666667 | 0.003256522419945083 |
+| Epoch_12.pt | 0.9535 | 0.0038204291659898184 |
+| Epoch_17.pt | 0.9531666666666666 | 0.003224615969095881 |
+| Epoch_10_batch_5999.pt | 0.9530000000000001 | 0.0033499585403736335 |
+| Epoch_14_batch_5999.pt | 0.9528333333333334 | 0.0034251250315107465 |
+| Epoch_16.pt | 0.9528333333333334 | 0.003531603350069519 |
+| Epoch_12_batch_5999.pt | 0.9526666666666668 | 0.0033536418383970194 |
+| Epoch_15_batch_5999.pt | 0.9523333333333334 | 0.0033444259873983135 |
+| Epoch_16_batch_2999.pt | 0.9523333333333334 | 0.0034174569998288227 |
+| Epoch_16_batch_5999.pt | 0.9521666666666666 | 0.003333796264150556 |
+| Epoch_11_batch_2999.pt | 0.952 | 0.0034587516480607556 |
+| Epoch_13_batch_5999.pt | 0.952 | 0.003395712619985805 |
+| Epoch_17_batch_5999.pt | 0.952 | 0.003349958540373629 |
+| Epoch_10.pt | 0.9518333333333334 | 0.0034466838598324672 |
+| Epoch_11.pt | 0.9518333333333334 | 0.0036468318738391774 |
+| Epoch_7_batch_5999.pt | 0.9486666666666668 | 0.003520662115056637 |
+| Epoch_7_batch_2999.pt | 0.9484999999999999 | 0.0028593576072128367 |
+| Epoch_6_batch_2999.pt | 0.9471666666666667 | 0.00223675801546638 |
+| Epoch_9_batch_5999.pt | 0.9470000000000001 | 0.003958114029012637 |
+| Epoch_8_batch_5999.pt | 0.9463333333333332 | 0.0034318767136623336 |
+| Epoch_8.pt | 0.9461666666666666 | 0.003776552088746075 |
+| Epoch_7.pt | 0.9458333333333334 | 0.0032322640461752952 |
+| Epoch_4.pt | 0.9456666666666665 | 0.003515398226568084 |
+| Epoch_4_batch_5999.pt | 0.9446666666666665 | 0.0031700761372638006 |
+| Epoch_8_batch_2999.pt | 0.9445 | 0.0035140810224152117 |
+| Epoch_5_batch_2999.pt | 0.9443333333333334 | 0.0030041124076879435 |
+| Epoch_5.pt | 0.9441666666666666 | 0.0032322640461753017 |
+| Epoch_5_batch_5999.pt | 0.944 | 0.0028995529668221945 |
+| Epoch_3_batch_5999.pt | 0.9431666666666665 | 0.0033742854753467133 |
+| Epoch_9.pt | 0.9418333333333333 | 0.0042488197344441965 |
+| Epoch_6_batch_5999.pt | 0.9410000000000001 | 0.004562325592810626 |
+| Epoch_4_batch_2999.pt | 0.9403333333333335 | 0.0039969123885788705 |
+| Epoch_9_batch_2999.pt | 0.9400000000000001 | 0.00352241499647719 |
+| Epoch_3_batch_2999.pt | 0.9386666666666666 | 0.003020506048681815 |
+| Epoch_6.pt | 0.9386666666666665 | 0.003973678831610592 |
+| Epoch_3.pt | 0.9371666666666666 | 0.002897955856832215 |
+| Epoch_2_batch_5999.pt | 0.933 | 0.00424409538995206 |
+| Epoch_2_batch_2999.pt | 0.9321666666666667 | 0.0034251250315107465 |
+| Epoch_2.pt | 0.93 | 0.002641080960889938 |
+| Epoch_1_batch_5999.pt | 0.9268333333333334 | 0.0032150302880511773 |
+| Epoch_1.pt | 0.9178333333333335 | 0.002152374514201173 |
+| Epoch_1_batch_2999.pt | 0.9146666666666666 | 0.004686465935325139 |
+| Epoch_0_batch_5999.pt | 0.8631666666666666 | 0.005541370780182171 |
+| Epoch_0.pt | 0.8554999999999999 | 0.004135438531988101 |
+| Epoch_0_batch_2999.pt | 0.6685 | 0.007578600476683353 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ca43adc87ad14ab2b7c81f239f2585fc792f63a5
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.882 | 0.005675374572943168 |
+| Epoch_17.pt | 0.882 | 0.006230153675056075 |
+| Epoch_14_batch_5999.pt | 0.8816666666666666 | 0.005947299418254502 |
+| Epoch_15_batch_2999.pt | 0.8805 | 0.00566039958455144 |
+| Epoch_16_batch_2999.pt | 0.8801666666666668 | 0.006645568466799031 |
+| Epoch_13_batch_5999.pt | 0.8801666666666665 | 0.005907987896708517 |
+| Epoch_14_batch_2999.pt | 0.8800000000000001 | 0.006191391873668908 |
+| Epoch_13_batch_2999.pt | 0.8793333333333335 | 0.00636929362131845 |
+| Epoch_17_batch_2999.pt | 0.8793333333333333 | 0.006446359868604571 |
+| Epoch_12.pt | 0.8791666666666667 | 0.0065886640396335665 |
+| Epoch_16.pt | 0.8788333333333334 | 0.00644085148883551 |
+| Epoch_15.pt | 0.8786666666666667 | 0.00617041927862881 |
+| Epoch_12_batch_2999.pt | 0.8779999999999999 | 0.005745637099487944 |
+| Epoch_11_batch_5999.pt | 0.8779999999999999 | 0.006235105709599201 |
+| Epoch_16_batch_5999.pt | 0.8776666666666666 | 0.006662961933584421 |
+| Epoch_13.pt | 0.8773333333333333 | 0.006403124237432855 |
+| Epoch_11.pt | 0.8771666666666667 | 0.006620441595058467 |
+| Epoch_17_batch_5999.pt | 0.8771666666666667 | 0.006597090641962685 |
+| Epoch_10.pt | 0.8771666666666664 | 0.006161659628924347 |
+| Epoch_14.pt | 0.8766666666666667 | 0.006662035428409453 |
+| Epoch_12_batch_5999.pt | 0.8756666666666668 | 0.006475023237481537 |
+| Epoch_10_batch_2999.pt | 0.8745 | 0.005555833326389239 |
+| Epoch_11_batch_2999.pt | 0.8728333333333333 | 0.0067222222222222275 |
+| Epoch_10_batch_5999.pt | 0.8728333333333333 | 0.006161659628924348 |
+| Epoch_5_batch_2999.pt | 0.8539999999999999 | 0.007036939570475114 |
+| Epoch_9_batch_2999.pt | 0.8526666666666667 | 0.005600925849390468 |
+| Epoch_5_batch_5999.pt | 0.8511666666666666 | 0.005931970297037756 |
+| Epoch_7_batch_5999.pt | 0.8508333333333334 | 0.006316009444945595 |
+| Epoch_8_batch_2999.pt | 0.8504999999999999 | 0.005176215787285458 |
+| Epoch_9_batch_5999.pt | 0.8488333333333331 | 0.006772538958958988 |
+| Epoch_6_batch_5999.pt | 0.8476666666666667 | 0.006795968303721347 |
+| Epoch_8_batch_5999.pt | 0.8473333333333335 | 0.007714824594151085 |
+| Epoch_7_batch_2999.pt | 0.8466666666666667 | 0.007888106377466154 |
+| Epoch_3_batch_5999.pt | 0.8460000000000001 | 0.005720873627491551 |
+| Epoch_7.pt | 0.8451666666666668 | 0.006173669697259828 |
+| Epoch_6_batch_2999.pt | 0.8451666666666666 | 0.007434695108612373 |
+| Epoch_6.pt | 0.8441666666666666 | 0.006311120892011203 |
+| Epoch_8.pt | 0.844 | 0.007573508083534208 |
+| Epoch_4_batch_2999.pt | 0.8438333333333334 | 0.006979025012382233 |
+| Epoch_4_batch_5999.pt | 0.8435 | 0.005695189653135811 |
+| Epoch_3_batch_2999.pt | 0.8426666666666666 | 0.00791622805802528 |
+| Epoch_5.pt | 0.8408333333333333 | 0.00634526167143739 |
+| Epoch_9.pt | 0.8391666666666666 | 0.007214739713655818 |
+| Epoch_2_batch_5999.pt | 0.8376666666666667 | 0.0063059838327389865 |
+| Epoch_4.pt | 0.836 | 0.006667592528301113 |
+| Epoch_3.pt | 0.8343333333333334 | 0.007978365809340552 |
+| Epoch_2.pt | 0.8236666666666667 | 0.006986759965449988 |
+| Epoch_2_batch_2999.pt | 0.8206666666666667 | 0.006939776208723194 |
+| Epoch_1_batch_5999.pt | 0.8201666666666666 | 0.007566372975210781 |
+| Epoch_1.pt | 0.8175000000000001 | 0.007594060382063062 |
+| Epoch_1_batch_2999.pt | 0.8056666666666666 | 0.008931149539603814 |
+| Epoch_0_batch_5999.pt | 0.7378333333333333 | 0.010030970559739987 |
+| Epoch_0.pt | 0.7271666666666666 | 0.008039447496525035 |
+| Epoch_0_batch_2999.pt | 0.6039999999999999 | 0.007396195218063386 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d6ad8b7d8ae4616045f191b2fafda5030e2445cf
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_10.pt | 0.9978333333333333 | 0.0008258927081843653 |
+| Epoch_11_batch_2999.pt | 0.9976666666666667 | 0.0009362388636862609 |
+| Epoch_12_batch_5999.pt | 0.9974999999999999 | 0.0009043789220055372 |
+| Epoch_16.pt | 0.9974999999999999 | 0.0010015420209622187 |
+| Epoch_17.pt | 0.9973333333333333 | 0.0009686442096757043 |
+| Epoch_10_batch_2999.pt | 0.9973333333333333 | 0.0010599324460188297 |
+| Epoch_14_batch_5999.pt | 0.9973333333333333 | 0.0009686442096757043 |
+| Epoch_16_batch_5999.pt | 0.9973333333333333 | 0.0009686442096757043 |
+| Epoch_9.pt | 0.9971666666666665 | 0.0008624541497922207 |
+| Epoch_10_batch_5999.pt | 0.9971666666666665 | 0.0009312808119022322 |
+| Epoch_14.pt | 0.9971666666666665 | 0.0009312808119022322 |
+| Epoch_11.pt | 0.9970000000000001 | 0.0010772621905369634 |
+| Epoch_15_batch_2999.pt | 0.9970000000000001 | 0.0010482201257840677 |
+| Epoch_13_batch_2999.pt | 0.9970000000000001 | 0.0009875771574795094 |
+| Epoch_11_batch_5999.pt | 0.9970000000000001 | 0.0010772621905369634 |
+| Epoch_15.pt | 0.9969999999999999 | 0.0008888888888888863 |
+| Epoch_14_batch_2999.pt | 0.9968333333333333 | 0.0008766518798921921 |
+| Epoch_6_batch_2999.pt | 0.9968333333333333 | 0.0007637626158259699 |
+| Epoch_13.pt | 0.9968333333333333 | 0.001007686508178725 |
+| Epoch_8_batch_2999.pt | 0.9968333333333333 | 0.001007686508178725 |
+| Epoch_16_batch_2999.pt | 0.9968333333333333 | 0.001007686508178725 |
+| Epoch_15_batch_5999.pt | 0.9966666666666667 | 0.0009938079899999067 |
+| Epoch_13_batch_5999.pt | 0.9966666666666667 | 0.0008958064164776142 |
+| Epoch_4_batch_5999.pt | 0.9966666666666667 | 0.0012668615834434834 |
+| Epoch_12.pt | 0.9966666666666667 | 0.0008240220541217368 |
+| Epoch_17_batch_2999.pt | 0.9966666666666667 | 0.0009938079899999067 |
+| Epoch_6.pt | 0.9966666666666667 | 0.0009296222517045275 |
+| Epoch_17_batch_5999.pt | 0.9966666666666667 | 0.0009938079899999067 |
+| Epoch_7_batch_5999.pt | 0.9964999999999999 | 0.0011235415786753746 |
+| Epoch_5_batch_2999.pt | 0.9964999999999999 | 0.0010378634273483015 |
+| Epoch_12_batch_2999.pt | 0.9964999999999999 | 0.0008407081083567495 |
+| Epoch_9_batch_5999.pt | 0.9963333333333333 | 0.001160034056545621 |
+| Epoch_4_batch_2999.pt | 0.9963333333333333 | 0.0011055415967851372 |
+| Epoch_2_batch_5999.pt | 0.9963333333333333 | 0.0011331154474650631 |
+| Epoch_5_batch_5999.pt | 0.9963333333333333 | 0.0012120791238484144 |
+| Epoch_8_batch_5999.pt | 0.9961666666666666 | 0.0008258927081843622 |
+| Epoch_3.pt | 0.9959999999999999 | 0.0010304020550550763 |
+| Epoch_4.pt | 0.9959999999999999 | 0.0009026709338484372 |
+| Epoch_5.pt | 0.9956666666666667 | 0.0010886621079036305 |
+| Epoch_6_batch_5999.pt | 0.9956666666666665 | 0.0009026709338484367 |
+| Epoch_7.pt | 0.9956666666666665 | 0.0010599324460188332 |
+| Epoch_8.pt | 0.9955 | 0.0010844011831079527 |
+| Epoch_7_batch_2999.pt | 0.9954999999999998 | 0.0010844011831079527 |
+| Epoch_2.pt | 0.9953333333333333 | 0.0009229582069908945 |
+| Epoch_3_batch_2999.pt | 0.9953333333333332 | 0.0008164965809277296 |
+| Epoch_3_batch_5999.pt | 0.9951666666666666 | 0.0010957268290731133 |
+| Epoch_9_batch_2999.pt | 0.9948333333333335 | 0.0011772011166898413 |
+| Epoch_1.pt | 0.9946666666666667 | 0.0009229582069908971 |
+| Epoch_1_batch_5999.pt | 0.9941666666666666 | 0.0009043789220055396 |
+| Epoch_2_batch_2999.pt | 0.9941666666666664 | 0.0011453071182271246 |
+| Epoch_1_batch_2999.pt | 0.9911666666666668 | 0.0014497764834110972 |
+| Epoch_0_batch_5999.pt | 0.9789999999999999 | 0.0019436506316150984 |
+| Epoch_0.pt | 0.9748333333333333 | 0.0026925824035672567 |
+| Epoch_0_batch_2999.pt | 0.9081666666666666 | 0.00527309733985212 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..1227308f139401ce190099b032c307d19b16cd59
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9524999999999999 | 0.0022939803135625047 |
+| Epoch_16_batch_5999.pt | 0.9521666666666666 | 0.00234454976066769 |
+| Epoch_17_batch_2999.pt | 0.9516666666666665 | 0.002675909906398299 |
+| Epoch_16.pt | 0.9511666666666667 | 0.0023445497606676847 |
+| Epoch_14_batch_5999.pt | 0.9508333333333333 | 0.0024876236863597945 |
+| Epoch_11_batch_2999.pt | 0.9506666666666668 | 0.002834966849371794 |
+| Epoch_17.pt | 0.9505000000000001 | 0.0024349563334234606 |
+| Epoch_15_batch_2999.pt | 0.9501666666666667 | 0.002634645711417646 |
+| Epoch_15_batch_5999.pt | 0.95 | 0.002557969874049184 |
+| Epoch_12.pt | 0.95 | 0.0018425693279752213 |
+| Epoch_14.pt | 0.9493333333333333 | 0.002655067365633006 |
+| Epoch_13_batch_2999.pt | 0.9491666666666665 | 0.002052550355767911 |
+| Epoch_17_batch_5999.pt | 0.9489999999999998 | 0.002643417167415624 |
+| Epoch_13.pt | 0.9488333333333335 | 0.0025221243250702625 |
+| Epoch_12_batch_5999.pt | 0.9488333333333333 | 0.0031135902820449007 |
+| Epoch_13_batch_5999.pt | 0.9484999999999999 | 0.002527014536787587 |
+| Epoch_14_batch_2999.pt | 0.9476666666666667 | 0.002499382639822663 |
+| Epoch_11_batch_5999.pt | 0.9471666666666667 | 0.0031627656283311797 |
+| Epoch_15.pt | 0.9471666666666666 | 0.002618193703988486 |
+| Epoch_11.pt | 0.9456666666666665 | 0.003567687635111629 |
+| Epoch_10_batch_5999.pt | 0.9455 | 0.0026994512473902986 |
+| Epoch_12_batch_2999.pt | 0.9453333333333334 | 0.0030205060486818173 |
+| Epoch_10.pt | 0.9451666666666666 | 0.002575405996999833 |
+| Epoch_10_batch_2999.pt | 0.9416666666666667 | 0.0029397236789606516 |
+| Epoch_7_batch_2999.pt | 0.9174999999999999 | 0.0033541019662496844 |
+| Epoch_9_batch_5999.pt | 0.9173333333333332 | 0.0029627314724385355 |
+| Epoch_6_batch_5999.pt | 0.9168333333333333 | 0.003290934048804416 |
+| Epoch_8_batch_2999.pt | 0.9133333333333334 | 0.0046080980785172775 |
+| Epoch_9_batch_2999.pt | 0.9119999999999999 | 0.004525647078759181 |
+| Epoch_7_batch_5999.pt | 0.9118333333333333 | 0.0034645470728153108 |
+| Epoch_7.pt | 0.9108333333333334 | 0.003222701113838484 |
+| Epoch_8_batch_5999.pt | 0.9048333333333334 | 0.003986474044646721 |
+| Epoch_3_batch_5999.pt | 0.9046666666666667 | 0.003911047979288454 |
+| Epoch_6.pt | 0.9031666666666667 | 0.0051487143307100665 |
+| Epoch_4_batch_2999.pt | 0.9016666666666666 | 0.004540625942600565 |
+| Epoch_5_batch_2999.pt | 0.9015000000000001 | 0.002880864924992031 |
+| Epoch_5_batch_5999.pt | 0.9013333333333332 | 0.003807075933113734 |
+| Epoch_6_batch_2999.pt | 0.8963333333333334 | 0.005526591162420401 |
+| Epoch_4.pt | 0.8961666666666666 | 0.0047143725607948425 |
+| Epoch_3.pt | 0.8955 | 0.004931944248881376 |
+| Epoch_9.pt | 0.8950000000000001 | 0.004746668747398624 |
+| Epoch_2.pt | 0.8941666666666667 | 0.004663027681720735 |
+| Epoch_8.pt | 0.8926666666666667 | 0.005277485372016852 |
+| Epoch_5.pt | 0.8924999999999998 | 0.005026166101414908 |
+| Epoch_4_batch_5999.pt | 0.8888333333333334 | 0.0048879418274609164 |
+| Epoch_3_batch_2999.pt | 0.8861666666666667 | 0.004655078204114267 |
+| Epoch_2_batch_2999.pt | 0.8860000000000001 | 0.004964068423041923 |
+| Epoch_2_batch_5999.pt | 0.8795 | 0.0044724810139063756 |
+| Epoch_1_batch_5999.pt | 0.8723333333333333 | 0.004857220665420977 |
+| Epoch_1.pt | 0.8636666666666667 | 0.0031797973380564867 |
+| Epoch_1_batch_2999.pt | 0.8298333333333334 | 0.00554137078018218 |
+| Epoch_0_batch_5999.pt | 0.7493333333333333 | 0.003961231882216864 |
+| Epoch_0.pt | 0.733 | 0.005459163776239215 |
+| Epoch_0_batch_2999.pt | 0.5816666666666668 | 0.008465616732800198 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ceb19d29b909d653b375be0bc344a82bd8848440
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_12_batch_5999.pt | 0.9395 | 0.003211188003692987 |
+| Epoch_17_batch_5999.pt | 0.9395 | 0.003760171390892827 |
+| Epoch_13_batch_5999.pt | 0.9383333333333335 | 0.003478327964999673 |
+| Epoch_13_batch_2999.pt | 0.9380000000000001 | 0.003377485367460146 |
+| Epoch_14_batch_2999.pt | 0.9376666666666666 | 0.0031446603773521917 |
+| Epoch_14_batch_5999.pt | 0.9376666666666666 | 0.003203007845644349 |
+| Epoch_12_batch_2999.pt | 0.9376666666666665 | 0.003670032125536394 |
+| Epoch_15_batch_2999.pt | 0.9375 | 0.003426926780731359 |
+| Epoch_17_batch_2999.pt | 0.9375 | 0.003335647344950685 |
+| Epoch_16.pt | 0.9373333333333334 | 0.0030951973949298046 |
+| Epoch_15_batch_5999.pt | 0.937 | 0.003449816599168893 |
+| Epoch_16_batch_2999.pt | 0.9365 | 0.0030976893021001438 |
+| Epoch_12.pt | 0.9363333333333334 | 0.0026736020923368753 |
+| Epoch_13.pt | 0.9359999999999999 | 0.003686813338452678 |
+| Epoch_10_batch_5999.pt | 0.9353333333333333 | 0.0036582394740342934 |
+| Epoch_14.pt | 0.9353333333333333 | 0.003590109871423003 |
+| Epoch_11_batch_5999.pt | 0.9349999999999999 | 0.003042903097250916 |
+| Epoch_15.pt | 0.9348333333333333 | 0.003365126160201735 |
+| Epoch_17.pt | 0.9346666666666668 | 0.003081205471969343 |
+| Epoch_10.pt | 0.9343333333333333 | 0.0036531738272830233 |
+| Epoch_16_batch_5999.pt | 0.9341666666666667 | 0.0030756912301480844 |
+| Epoch_11.pt | 0.9334999999999999 | 0.0035175924707281595 |
+| Epoch_11_batch_2999.pt | 0.9331666666666667 | 0.0037387691907536068 |
+| Epoch_10_batch_2999.pt | 0.9306666666666666 | 0.004396968652757644 |
+| Epoch_7_batch_5999.pt | 0.9091666666666667 | 0.004521894610027698 |
+| Epoch_9.pt | 0.9061666666666668 | 0.0032207851201412822 |
+| Epoch_6_batch_5999.pt | 0.9061666666666668 | 0.004938198301925088 |
+| Epoch_9_batch_5999.pt | 0.9038333333333333 | 0.0045680722357048275 |
+| Epoch_7_batch_2999.pt | 0.9021666666666667 | 0.0027672020900916133 |
+| Epoch_6_batch_2999.pt | 0.9008333333333335 | 0.004319422114983375 |
+| Epoch_9_batch_2999.pt | 0.9005000000000001 | 0.004465574769801906 |
+| Epoch_3_batch_5999.pt | 0.8988333333333334 | 0.00399111667903596 |
+| Epoch_5_batch_2999.pt | 0.8983333333333332 | 0.0034066021592790855 |
+| Epoch_8_batch_5999.pt | 0.8983333333333332 | 0.002876039801232168 |
+| Epoch_4_batch_2999.pt | 0.8981666666666668 | 0.0037305048809332286 |
+| Epoch_4_batch_5999.pt | 0.8975 | 0.003381595065229603 |
+| Epoch_5_batch_5999.pt | 0.8965000000000002 | 0.0032626770800057788 |
+| Epoch_8_batch_2999.pt | 0.8963333333333333 | 0.003981438414926426 |
+| Epoch_7.pt | 0.8963333333333333 | 0.003989182904670284 |
+| Epoch_6.pt | 0.8946666666666665 | 0.004192880503136269 |
+| Epoch_5.pt | 0.8931666666666667 | 0.004197662488858571 |
+| Epoch_8.pt | 0.892 | 0.005004935835357872 |
+| Epoch_2.pt | 0.8913333333333334 | 0.0027532248207475193 |
+| Epoch_3_batch_2999.pt | 0.8905 | 0.0038892856940415826 |
+| Epoch_3.pt | 0.8858333333333335 | 0.0030555555555555583 |
+| Epoch_2_batch_2999.pt | 0.8846666666666667 | 0.00406581654745155 |
+| Epoch_4.pt | 0.8838333333333332 | 0.004635144863025813 |
+| Epoch_2_batch_5999.pt | 0.8783333333333333 | 0.004614791034954483 |
+| Epoch_1_batch_5999.pt | 0.8683333333333334 | 0.004843221048378527 |
+| Epoch_1.pt | 0.8538333333333332 | 0.004963135707330871 |
+| Epoch_1_batch_2999.pt | 0.8418333333333333 | 0.004807722696959985 |
+| Epoch_0.pt | 0.7705000000000001 | 0.005633070342316123 |
+| Epoch_0_batch_5999.pt | 0.7628333333333334 | 0.005911121553875487 |
+| Epoch_0_batch_2999.pt | 0.6415 | 0.005844961954274267 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..45267fd5d3bdfd4667a74735b6264aec367705c3
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9856666666666666 | 0.0011439589045541066 |
+| Epoch_14_batch_2999.pt | 0.9850000000000001 | 0.001405456737852613 |
+| Epoch_13_batch_2999.pt | 0.9848333333333332 | 0.0016187558093703764 |
+| Epoch_16.pt | 0.9846666666666668 | 0.0012862041003100216 |
+| Epoch_13_batch_5999.pt | 0.9843333333333334 | 0.0014315665251916788 |
+| Epoch_17.pt | 0.9841666666666666 | 0.0017435949807194596 |
+| Epoch_16_batch_5999.pt | 0.9841666666666666 | 0.001296957503325412 |
+| Epoch_13.pt | 0.9840000000000002 | 0.0015752718754175373 |
+| Epoch_17_batch_2999.pt | 0.984 | 0.0015355861067872468 |
+| Epoch_17_batch_5999.pt | 0.9838333333333334 | 0.0014497764834110966 |
+| Epoch_15_batch_5999.pt | 0.9835 | 0.0014153043558729952 |
+| Epoch_14_batch_5999.pt | 0.9835 | 0.0016377114414426249 |
+| Epoch_14.pt | 0.9834999999999999 | 0.001599575560987542 |
+| Epoch_10.pt | 0.9833333333333334 | 0.001791612832955228 |
+| Epoch_16_batch_2999.pt | 0.9833333333333334 | 0.0018257418583505463 |
+| Epoch_15.pt | 0.983 | 0.001606314699422323 |
+| Epoch_10_batch_5999.pt | 0.9823333333333334 | 0.0023465235646603173 |
+| Epoch_11_batch_5999.pt | 0.9821666666666667 | 0.0020495407495218533 |
+| Epoch_12_batch_5999.pt | 0.982 | 0.0017356110390903657 |
+| Epoch_12_batch_2999.pt | 0.9819999999999999 | 0.0020608041101101504 |
+| Epoch_11_batch_2999.pt | 0.9810000000000001 | 0.0022662308949301297 |
+| Epoch_12.pt | 0.9806666666666667 | 0.0013425606637327238 |
+| Epoch_11.pt | 0.9801666666666666 | 0.0019476164603286774 |
+| Epoch_10_batch_2999.pt | 0.9798333333333333 | 0.002283191398670498 |
+| Epoch_7_batch_2999.pt | 0.9701666666666668 | 0.002760501833067743 |
+| Epoch_9_batch_2999.pt | 0.9654999999999999 | 0.0026879934229784405 |
+| Epoch_6_batch_5999.pt | 0.9651666666666667 | 0.0023498095536174024 |
+| Epoch_8_batch_5999.pt | 0.9651666666666667 | 0.0025992639033976866 |
+| Epoch_7_batch_5999.pt | 0.9648333333333333 | 0.001994591452335135 |
+| Epoch_9_batch_5999.pt | 0.9644999999999999 | 0.0020645449084513395 |
+| Epoch_6_batch_2999.pt | 0.9636666666666667 | 0.002591534175486804 |
+| Epoch_8_batch_2999.pt | 0.9631666666666667 | 0.0033002992756175965 |
+| Epoch_5_batch_5999.pt | 0.9631666666666667 | 0.0023100692095879834 |
+| Epoch_5_batch_2999.pt | 0.9626666666666667 | 0.002723922371584722 |
+| Epoch_4_batch_5999.pt | 0.9603333333333334 | 0.0023013683530231114 |
+| Epoch_5.pt | 0.9591666666666668 | 0.0026902888917340537 |
+| Epoch_9.pt | 0.9586666666666666 | 0.0016996731711975911 |
+| Epoch_8.pt | 0.9583333333333333 | 0.003181738014061419 |
+| Epoch_6.pt | 0.9576666666666667 | 0.0014740554623801836 |
+| Epoch_4.pt | 0.9576666666666667 | 0.0028131086447049218 |
+| Epoch_7.pt | 0.9568333333333333 | 0.0026695586170519957 |
+| Epoch_3_batch_5999.pt | 0.9563333333333333 | 0.0032942149067496872 |
+| Epoch_4_batch_2999.pt | 0.9550000000000001 | 0.0021942686286812777 |
+| Epoch_3.pt | 0.954 | 0.002746490465401836 |
+| Epoch_3_batch_2999.pt | 0.9536666666666667 | 0.00268512132746546 |
+| Epoch_2_batch_5999.pt | 0.9516666666666665 | 0.0034426518632954795 |
+| Epoch_2.pt | 0.9426666666666665 | 0.002962731472438525 |
+| Epoch_2_batch_2999.pt | 0.9410000000000001 | 0.003914203323068566 |
+| Epoch_1_batch_5999.pt | 0.932 | 0.0036497928234792984 |
+| Epoch_1.pt | 0.932 | 0.003150543750835068 |
+| Epoch_1_batch_2999.pt | 0.9201666666666668 | 0.0027716599296147624 |
+| Epoch_0_batch_5999.pt | 0.8571666666666667 | 0.0027783332777888818 |
+| Epoch_0.pt | 0.8404999999999999 | 0.0042313497831843565 |
+| Epoch_0_batch_2999.pt | 0.7184999999999999 | 0.005802564354354004 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ab56551770ef0a67a9caf8adaa938fb7125a08d3
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.958 | 0.0041335722750529555 |
+| Epoch_17_batch_5999.pt | 0.9573333333333334 | 0.004210510071443643 |
+| Epoch_14_batch_2999.pt | 0.9568333333333333 | 0.003868597857456373 |
+| Epoch_13_batch_5999.pt | 0.9568333333333332 | 0.003892458679022968 |
+| Epoch_15.pt | 0.9566666666666667 | 0.003767961101736262 |
+| Epoch_17.pt | 0.9566666666666667 | 0.0037267799624996494 |
+| Epoch_13_batch_2999.pt | 0.9565000000000001 | 0.0033189504513875204 |
+| Epoch_15_batch_2999.pt | 0.9563333333333333 | 0.0034942633763369703 |
+| Epoch_14_batch_5999.pt | 0.9561666666666666 | 0.003849402711404451 |
+| Epoch_16_batch_5999.pt | 0.9560000000000002 | 0.003874576838702822 |
+| Epoch_15_batch_5999.pt | 0.9560000000000001 | 0.0035416394334464966 |
+| Epoch_12.pt | 0.9558333333333333 | 0.003977172517576765 |
+| Epoch_16.pt | 0.9558333333333333 | 0.0039616214413349055 |
+| Epoch_14.pt | 0.9556666666666664 | 0.003850605211369658 |
+| Epoch_10.pt | 0.9555 | 0.0032965563775881708 |
+| Epoch_17_batch_2999.pt | 0.9553333333333335 | 0.0038151743807531995 |
+| Epoch_11_batch_5999.pt | 0.9553333333333333 | 0.0030408738185342303 |
+| Epoch_11.pt | 0.9551666666666667 | 0.002323391518480724 |
+| Epoch_13.pt | 0.9551666666666666 | 0.0038845213569807377 |
+| Epoch_12_batch_5999.pt | 0.9546666666666667 | 0.0035815025469524884 |
+| Epoch_11_batch_2999.pt | 0.9536666666666667 | 0.003520662115056636 |
+| Epoch_12_batch_2999.pt | 0.9535 | 0.002490103870112778 |
+| Epoch_10_batch_5999.pt | 0.9528333333333332 | 0.002855036701673241 |
+| Epoch_10_batch_2999.pt | 0.9473333333333332 | 0.0032317865716108853 |
+| Epoch_7_batch_2999.pt | 0.9323333333333332 | 0.0048381202349060525 |
+| Epoch_9_batch_2999.pt | 0.9313333333333335 | 0.0034587516480607504 |
+| Epoch_7.pt | 0.9313333333333332 | 0.00310117460821174 |
+| Epoch_9_batch_5999.pt | 0.9311666666666667 | 0.0023707320325946427 |
+| Epoch_6_batch_2999.pt | 0.931 | 0.003417456999828825 |
+| Epoch_8_batch_5999.pt | 0.9305 | 0.004317992789294008 |
+| Epoch_7_batch_5999.pt | 0.9288333333333334 | 0.00418735604138028 |
+| Epoch_8_batch_2999.pt | 0.9281666666666666 | 0.002826789829601903 |
+| Epoch_6_batch_5999.pt | 0.9259999999999999 | 0.0031739681904634897 |
+| Epoch_4_batch_5999.pt | 0.9248333333333335 | 0.0029234049148717505 |
+| Epoch_4.pt | 0.9246666666666666 | 0.0030408738185342186 |
+| Epoch_5_batch_5999.pt | 0.9245000000000001 | 0.0016301556390134705 |
+| Epoch_8.pt | 0.9235000000000001 | 0.003057575090181562 |
+| Epoch_4_batch_2999.pt | 0.9221666666666668 | 0.004014249311000032 |
+| Epoch_5.pt | 0.9221666666666668 | 0.003889285694041591 |
+| Epoch_3.pt | 0.9215 | 0.0049156464729385885 |
+| Epoch_3_batch_5999.pt | 0.9208333333333334 | 0.0029736497091272527 |
+| Epoch_5_batch_2999.pt | 0.9206666666666667 | 0.003362832433427013 |
+| Epoch_6.pt | 0.9194999999999999 | 0.002981941533404366 |
+| Epoch_9.pt | 0.917 | 0.0023412563895228336 |
+| Epoch_3_batch_2999.pt | 0.9119999999999999 | 0.005004935835357869 |
+| Epoch_2_batch_5999.pt | 0.9113333333333333 | 0.004566382797341006 |
+| Epoch_2_batch_2999.pt | 0.9066666666666666 | 0.0025939150066508366 |
+| Epoch_2.pt | 0.9033333333333333 | 0.004527010841658525 |
+| Epoch_1_batch_5999.pt | 0.8918333333333333 | 0.0035525160619440366 |
+| Epoch_1.pt | 0.8881666666666668 | 0.003947573094109001 |
+| Epoch_1_batch_2999.pt | 0.8846666666666666 | 0.004861031745383853 |
+| Epoch_0_batch_5999.pt | 0.8125 | 0.006695539329552194 |
+| Epoch_0.pt | 0.7993333333333332 | 0.004569085598878245 |
+| Epoch_0_batch_2999.pt | 0.6751666666666666 | 0.003820429165989821 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/ResNet50_ir/log.log b/bob/bio/facexzoo/models/backbones/ResNet50_ir/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..64fe918d8d4b71e3e5504ada20df22f1c35a40d0
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/ResNet50_ir/log.log
@@ -0,0 +1,657 @@
+INFO 2020-12-02 16:24:52 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2020-12-02 16:24:52 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2020-12-02 16:24:52 train.py: 177] Start optimization.
+INFO 2020-12-02 16:24:52 train.py: 178] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='ResNet', batch_size=512, data_root='/home/wangjun492/wj_data/facex-zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10,13,16', tensorboardx_logdir='mv-resnet50', train_file='/home/wangjun492/wj_data/facex-zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f0c45f14e48>)
+backbone param:
+{'depth': 50, 'drop_ratio': 0.4, 'net_mode': 'ir', 'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-12-02 16:25:20 train.py: 79] Epoch 0, iter 0/6416, lr 0.100000, loss 16.332954
+INFO 2020-12-02 16:28:32 train.py: 79] Epoch 0, iter 200/6416, lr 0.100000, loss 15.724994
+INFO 2020-12-02 16:31:45 train.py: 79] Epoch 0, iter 400/6416, lr 0.100000, loss 15.269594
+INFO 2020-12-02 16:34:58 train.py: 79] Epoch 0, iter 600/6416, lr 0.100000, loss 15.151460
+INFO 2020-12-02 16:38:11 train.py: 79] Epoch 0, iter 800/6416, lr 0.100000, loss 15.060672
+INFO 2020-12-02 16:41:25 train.py: 79] Epoch 0, iter 1000/6416, lr 0.100000, loss 14.924791
+INFO 2020-12-02 16:44:38 train.py: 79] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.762907
+INFO 2020-12-02 16:47:51 train.py: 79] Epoch 0, iter 1400/6416, lr 0.100000, loss 14.583320
+INFO 2020-12-02 16:51:04 train.py: 79] Epoch 0, iter 1600/6416, lr 0.100000, loss 14.378530
+INFO 2020-12-02 16:54:17 train.py: 79] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.124073
+INFO 2020-12-02 16:57:31 train.py: 79] Epoch 0, iter 2000/6416, lr 0.100000, loss 13.869375
+INFO 2020-12-02 17:00:44 train.py: 79] Epoch 0, iter 2200/6416, lr 0.100000, loss 13.614476
+INFO 2020-12-02 17:03:57 train.py: 79] Epoch 0, iter 2400/6416, lr 0.100000, loss 13.330109
+INFO 2020-12-02 17:07:11 train.py: 79] Epoch 0, iter 2600/6416, lr 0.100000, loss 13.028246
+INFO 2020-12-02 17:10:24 train.py: 79] Epoch 0, iter 2800/6416, lr 0.100000, loss 12.702374
+INFO 2020-12-02 17:13:36 train.py: 92] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-12-02 17:13:37 train.py: 79] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.396822
+INFO 2020-12-02 17:16:50 train.py: 79] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.155658
+INFO 2020-12-02 17:20:03 train.py: 79] Epoch 0, iter 3400/6416, lr 0.100000, loss 12.003897
+INFO 2020-12-02 17:23:15 train.py: 79] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.011105
+INFO 2020-12-02 17:26:27 train.py: 79] Epoch 0, iter 3800/6416, lr 0.100000, loss 12.183243
+INFO 2020-12-02 17:29:39 train.py: 79] Epoch 0, iter 4000/6416, lr 0.100000, loss 12.469275
+INFO 2020-12-02 17:32:50 train.py: 79] Epoch 0, iter 4200/6416, lr 0.100000, loss 12.898298
+INFO 2020-12-02 17:36:01 train.py: 79] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.308281
+INFO 2020-12-02 17:39:11 train.py: 79] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.742956
+INFO 2020-12-02 17:42:21 train.py: 79] Epoch 0, iter 4800/6416, lr 0.100000, loss 14.099098
+INFO 2020-12-02 17:45:31 train.py: 79] Epoch 0, iter 5000/6416, lr 0.100000, loss 14.386579
+INFO 2020-12-02 17:48:40 train.py: 79] Epoch 0, iter 5200/6416, lr 0.100000, loss 14.528618
+INFO 2020-12-02 17:51:49 train.py: 79] Epoch 0, iter 5400/6416, lr 0.100000, loss 14.624579
+INFO 2020-12-02 17:54:57 train.py: 79] Epoch 0, iter 5600/6416, lr 0.100000, loss 14.632182
+INFO 2020-12-02 17:58:06 train.py: 79] Epoch 0, iter 5800/6416, lr 0.100000, loss 14.546628
+INFO 2020-12-02 18:01:14 train.py: 92] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-12-02 18:01:15 train.py: 79] Epoch 0, iter 6000/6416, lr 0.100000, loss 14.422989
+INFO 2020-12-02 18:04:22 train.py: 79] Epoch 0, iter 6200/6416, lr 0.100000, loss 14.221948
+INFO 2020-12-02 18:07:29 train.py: 79] Epoch 0, iter 6400/6416, lr 0.100000, loss 13.978512
+INFO 2020-12-02 18:07:43 train.py: 97] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-12-02 18:07:45 train.py: 79] Epoch 1, iter 0/6416, lr 0.100000, loss 13.879437
+INFO 2020-12-02 18:10:52 train.py: 79] Epoch 1, iter 200/6416, lr 0.100000, loss 13.540010
+INFO 2020-12-02 18:13:59 train.py: 79] Epoch 1, iter 400/6416, lr 0.100000, loss 13.260385
+INFO 2020-12-02 18:17:06 train.py: 79] Epoch 1, iter 600/6416, lr 0.100000, loss 13.009260
+INFO 2020-12-02 18:20:13 train.py: 79] Epoch 1, iter 800/6416, lr 0.100000, loss 12.713784
+INFO 2020-12-02 18:23:20 train.py: 79] Epoch 1, iter 1000/6416, lr 0.100000, loss 12.434011
+INFO 2020-12-02 18:26:26 train.py: 79] Epoch 1, iter 1200/6416, lr 0.100000, loss 12.122822
+INFO 2020-12-02 18:29:33 train.py: 79] Epoch 1, iter 1400/6416, lr 0.100000, loss 11.863502
+INFO 2020-12-02 18:32:40 train.py: 79] Epoch 1, iter 1600/6416, lr 0.100000, loss 11.586728
+INFO 2020-12-02 18:35:46 train.py: 79] Epoch 1, iter 1800/6416, lr 0.100000, loss 11.311850
+INFO 2020-12-02 18:38:53 train.py: 79] Epoch 1, iter 2000/6416, lr 0.100000, loss 11.063287
+INFO 2020-12-02 18:41:59 train.py: 79] Epoch 1, iter 2200/6416, lr 0.100000, loss 10.864117
+INFO 2020-12-02 18:45:06 train.py: 79] Epoch 1, iter 2400/6416, lr 0.100000, loss 10.600532
+INFO 2020-12-02 18:48:13 train.py: 79] Epoch 1, iter 2600/6416, lr 0.100000, loss 10.403338
+INFO 2020-12-02 18:51:20 train.py: 79] Epoch 1, iter 2800/6416, lr 0.100000, loss 10.214808
+INFO 2020-12-02 18:54:26 train.py: 92] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-12-02 18:54:27 train.py: 79] Epoch 1, iter 3000/6416, lr 0.100000, loss 10.005656
+INFO 2020-12-02 18:57:34 train.py: 79] Epoch 1, iter 3200/6416, lr 0.100000, loss 9.837706
+INFO 2020-12-02 19:00:42 train.py: 79] Epoch 1, iter 3400/6416, lr 0.100000, loss 9.649582
+INFO 2020-12-02 19:03:49 train.py: 79] Epoch 1, iter 3600/6416, lr 0.100000, loss 9.496042
+INFO 2020-12-02 19:06:57 train.py: 79] Epoch 1, iter 3800/6416, lr 0.100000, loss 9.386625
+INFO 2020-12-02 19:10:04 train.py: 79] Epoch 1, iter 4000/6416, lr 0.100000, loss 9.236245
+INFO 2020-12-02 19:13:12 train.py: 79] Epoch 1, iter 4200/6416, lr 0.100000, loss 9.107701
+INFO 2020-12-02 19:16:19 train.py: 79] Epoch 1, iter 4400/6416, lr 0.100000, loss 8.984832
+INFO 2020-12-02 19:19:27 train.py: 79] Epoch 1, iter 4600/6416, lr 0.100000, loss 8.868542
+INFO 2020-12-02 19:22:35 train.py: 79] Epoch 1, iter 4800/6416, lr 0.100000, loss 8.713447
+INFO 2020-12-02 19:25:42 train.py: 79] Epoch 1, iter 5000/6416, lr 0.100000, loss 8.605859
+INFO 2020-12-02 19:28:50 train.py: 79] Epoch 1, iter 5200/6416, lr 0.100000, loss 8.533835
+INFO 2020-12-02 19:31:58 train.py: 79] Epoch 1, iter 5400/6416, lr 0.100000, loss 8.405970
+INFO 2020-12-02 19:35:06 train.py: 79] Epoch 1, iter 5600/6416, lr 0.100000, loss 8.326429
+INFO 2020-12-02 19:38:14 train.py: 79] Epoch 1, iter 5800/6416, lr 0.100000, loss 8.217263
+INFO 2020-12-02 19:41:21 train.py: 92] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-12-02 19:41:22 train.py: 79] Epoch 1, iter 6000/6416, lr 0.100000, loss 8.140155
+INFO 2020-12-02 19:44:30 train.py: 79] Epoch 1, iter 6200/6416, lr 0.100000, loss 8.101199
+INFO 2020-12-02 19:47:37 train.py: 79] Epoch 1, iter 6400/6416, lr 0.100000, loss 8.014781
+INFO 2020-12-02 19:47:52 train.py: 97] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-12-02 19:47:54 train.py: 79] Epoch 2, iter 0/6416, lr 0.100000, loss 7.922591
+INFO 2020-12-02 19:51:01 train.py: 79] Epoch 2, iter 200/6416, lr 0.100000, loss 7.404116
+INFO 2020-12-02 19:54:08 train.py: 79] Epoch 2, iter 400/6416, lr 0.100000, loss 7.376579
+INFO 2020-12-02 19:57:15 train.py: 79] Epoch 2, iter 600/6416, lr 0.100000, loss 7.398144
+INFO 2020-12-02 20:00:22 train.py: 79] Epoch 2, iter 800/6416, lr 0.100000, loss 7.432407
+INFO 2020-12-02 20:03:29 train.py: 79] Epoch 2, iter 1000/6416, lr 0.100000, loss 7.419043
+INFO 2020-12-02 20:06:36 train.py: 79] Epoch 2, iter 1200/6416, lr 0.100000, loss 7.417946
+INFO 2020-12-02 20:09:43 train.py: 79] Epoch 2, iter 1400/6416, lr 0.100000, loss 7.426534
+INFO 2020-12-02 20:12:50 train.py: 79] Epoch 2, iter 1600/6416, lr 0.100000, loss 7.341967
+INFO 2020-12-02 20:15:57 train.py: 79] Epoch 2, iter 1800/6416, lr 0.100000, loss 7.347022
+INFO 2020-12-02 20:19:04 train.py: 79] Epoch 2, iter 2000/6416, lr 0.100000, loss 7.309699
+INFO 2020-12-02 20:22:11 train.py: 79] Epoch 2, iter 2200/6416, lr 0.100000, loss 7.240558
+INFO 2020-12-02 20:25:18 train.py: 79] Epoch 2, iter 2400/6416, lr 0.100000, loss 7.219714
+INFO 2020-12-02 20:28:25 train.py: 79] Epoch 2, iter 2600/6416, lr 0.100000, loss 7.212546
+INFO 2020-12-02 20:31:32 train.py: 79] Epoch 2, iter 2800/6416, lr 0.100000, loss 7.145501
+INFO 2020-12-02 20:34:39 train.py: 92] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-12-02 20:34:40 train.py: 79] Epoch 2, iter 3000/6416, lr 0.100000, loss 7.139050
+INFO 2020-12-02 20:37:47 train.py: 79] Epoch 2, iter 3200/6416, lr 0.100000, loss 7.083629
+INFO 2020-12-02 20:40:55 train.py: 79] Epoch 2, iter 3400/6416, lr 0.100000, loss 7.039036
+INFO 2020-12-02 20:44:02 train.py: 79] Epoch 2, iter 3600/6416, lr 0.100000, loss 7.006894
+INFO 2020-12-02 20:47:09 train.py: 79] Epoch 2, iter 3800/6416, lr 0.100000, loss 6.950217
+INFO 2020-12-02 20:50:17 train.py: 79] Epoch 2, iter 4000/6416, lr 0.100000, loss 6.927581
+INFO 2020-12-02 20:53:24 train.py: 79] Epoch 2, iter 4200/6416, lr 0.100000, loss 6.886845
+INFO 2020-12-02 20:56:32 train.py: 79] Epoch 2, iter 4400/6416, lr 0.100000, loss 6.854423
+INFO 2020-12-02 20:59:40 train.py: 79] Epoch 2, iter 4600/6416, lr 0.100000, loss 6.819378
+INFO 2020-12-02 21:02:47 train.py: 79] Epoch 2, iter 4800/6416, lr 0.100000, loss 6.809598
+INFO 2020-12-02 21:05:55 train.py: 79] Epoch 2, iter 5000/6416, lr 0.100000, loss 6.793162
+INFO 2020-12-02 21:09:03 train.py: 79] Epoch 2, iter 5200/6416, lr 0.100000, loss 6.764558
+INFO 2020-12-02 21:12:10 train.py: 79] Epoch 2, iter 5400/6416, lr 0.100000, loss 6.706105
+INFO 2020-12-02 21:15:18 train.py: 79] Epoch 2, iter 5600/6416, lr 0.100000, loss 6.642517
+INFO 2020-12-02 21:18:26 train.py: 79] Epoch 2, iter 5800/6416, lr 0.100000, loss 6.652111
+INFO 2020-12-02 21:21:33 train.py: 92] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-12-02 21:21:34 train.py: 79] Epoch 2, iter 6000/6416, lr 0.100000, loss 6.628229
+INFO 2020-12-02 21:24:41 train.py: 79] Epoch 2, iter 6200/6416, lr 0.100000, loss 6.598334
+INFO 2020-12-02 21:27:48 train.py: 79] Epoch 2, iter 6400/6416, lr 0.100000, loss 6.545922
+INFO 2020-12-02 21:28:02 train.py: 97] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-12-02 21:28:04 train.py: 79] Epoch 3, iter 0/6416, lr 0.100000, loss 6.495420
+INFO 2020-12-02 21:31:11 train.py: 79] Epoch 3, iter 200/6416, lr 0.100000, loss 5.970732
+INFO 2020-12-02 21:34:17 train.py: 79] Epoch 3, iter 400/6416, lr 0.100000, loss 6.001810
+INFO 2020-12-02 21:37:24 train.py: 79] Epoch 3, iter 600/6416, lr 0.100000, loss 6.091019
+INFO 2020-12-02 21:40:30 train.py: 79] Epoch 3, iter 800/6416, lr 0.100000, loss 6.105414
+INFO 2020-12-02 21:43:36 train.py: 79] Epoch 3, iter 1000/6416, lr 0.100000, loss 6.161345
+INFO 2020-12-02 21:46:43 train.py: 79] Epoch 3, iter 1200/6416, lr 0.100000, loss 6.203309
+INFO 2020-12-02 21:49:49 train.py: 79] Epoch 3, iter 1400/6416, lr 0.100000, loss 6.194657
+INFO 2020-12-02 21:52:55 train.py: 79] Epoch 3, iter 1600/6416, lr 0.100000, loss 6.185114
+INFO 2020-12-02 21:56:02 train.py: 79] Epoch 3, iter 1800/6416, lr 0.100000, loss 6.222416
+INFO 2020-12-02 21:59:08 train.py: 79] Epoch 3, iter 2000/6416, lr 0.100000, loss 6.181895
+INFO 2020-12-02 22:02:14 train.py: 79] Epoch 3, iter 2200/6416, lr 0.100000, loss 6.180269
+INFO 2020-12-02 22:05:21 train.py: 79] Epoch 3, iter 2400/6416, lr 0.100000, loss 6.176522
+INFO 2020-12-02 22:08:27 train.py: 79] Epoch 3, iter 2600/6416, lr 0.100000, loss 6.180500
+INFO 2020-12-02 22:11:34 train.py: 79] Epoch 3, iter 2800/6416, lr 0.100000, loss 6.137156
+INFO 2020-12-02 22:14:40 train.py: 92] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-12-02 22:14:41 train.py: 79] Epoch 3, iter 3000/6416, lr 0.100000, loss 6.180045
+INFO 2020-12-02 22:17:48 train.py: 79] Epoch 3, iter 3200/6416, lr 0.100000, loss 6.166514
+INFO 2020-12-02 22:20:56 train.py: 79] Epoch 3, iter 3400/6416, lr 0.100000, loss 6.101875
+INFO 2020-12-02 22:24:03 train.py: 79] Epoch 3, iter 3600/6416, lr 0.100000, loss 6.092663
+INFO 2020-12-02 22:27:11 train.py: 79] Epoch 3, iter 3800/6416, lr 0.100000, loss 6.091960
+INFO 2020-12-02 22:30:18 train.py: 79] Epoch 3, iter 4000/6416, lr 0.100000, loss 6.051987
+INFO 2020-12-02 22:33:26 train.py: 79] Epoch 3, iter 4200/6416, lr 0.100000, loss 6.017490
+INFO 2020-12-02 22:36:33 train.py: 79] Epoch 3, iter 4400/6416, lr 0.100000, loss 6.035057
+INFO 2020-12-02 22:39:41 train.py: 79] Epoch 3, iter 4600/6416, lr 0.100000, loss 6.036681
+INFO 2020-12-02 22:42:49 train.py: 79] Epoch 3, iter 4800/6416, lr 0.100000, loss 6.002127
+INFO 2020-12-02 22:45:56 train.py: 79] Epoch 3, iter 5000/6416, lr 0.100000, loss 5.985600
+INFO 2020-12-02 22:49:04 train.py: 79] Epoch 3, iter 5200/6416, lr 0.100000, loss 5.980595
+INFO 2020-12-02 22:52:12 train.py: 79] Epoch 3, iter 5400/6416, lr 0.100000, loss 5.925722
+INFO 2020-12-02 22:55:20 train.py: 79] Epoch 3, iter 5600/6416, lr 0.100000, loss 5.970767
+INFO 2020-12-02 22:58:27 train.py: 79] Epoch 3, iter 5800/6416, lr 0.100000, loss 5.918377
+INFO 2020-12-02 23:01:34 train.py: 92] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-12-02 23:01:35 train.py: 79] Epoch 3, iter 6000/6416, lr 0.100000, loss 5.911398
+INFO 2020-12-02 23:04:43 train.py: 79] Epoch 3, iter 6200/6416, lr 0.100000, loss 5.876962
+INFO 2020-12-02 23:07:51 train.py: 79] Epoch 3, iter 6400/6416, lr 0.100000, loss 5.840898
+INFO 2020-12-02 23:08:05 train.py: 97] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-12-02 23:08:07 train.py: 79] Epoch 4, iter 0/6416, lr 0.100000, loss 5.867492
+INFO 2020-12-02 23:11:14 train.py: 79] Epoch 4, iter 200/6416, lr 0.100000, loss 5.366016
+INFO 2020-12-02 23:14:21 train.py: 79] Epoch 4, iter 400/6416, lr 0.100000, loss 5.328521
+INFO 2020-12-02 23:17:28 train.py: 79] Epoch 4, iter 600/6416, lr 0.100000, loss 5.403710
+INFO 2020-12-02 23:20:35 train.py: 79] Epoch 4, iter 800/6416, lr 0.100000, loss 5.461301
+INFO 2020-12-02 23:23:42 train.py: 79] Epoch 4, iter 1000/6416, lr 0.100000, loss 5.519482
+INFO 2020-12-02 23:26:49 train.py: 79] Epoch 4, iter 1200/6416, lr 0.100000, loss 5.591844
+INFO 2020-12-02 23:29:56 train.py: 79] Epoch 4, iter 1400/6416, lr 0.100000, loss 5.600097
+INFO 2020-12-02 23:33:03 train.py: 79] Epoch 4, iter 1600/6416, lr 0.100000, loss 5.605669
+INFO 2020-12-02 23:36:10 train.py: 79] Epoch 4, iter 1800/6416, lr 0.100000, loss 5.576889
+INFO 2020-12-02 23:39:17 train.py: 79] Epoch 4, iter 2000/6416, lr 0.100000, loss 5.612118
+INFO 2020-12-02 23:42:24 train.py: 79] Epoch 4, iter 2200/6416, lr 0.100000, loss 5.645781
+INFO 2020-12-02 23:45:31 train.py: 79] Epoch 4, iter 2400/6416, lr 0.100000, loss 5.601656
+INFO 2020-12-02 23:48:38 train.py: 79] Epoch 4, iter 2600/6416, lr 0.100000, loss 5.605802
+INFO 2020-12-02 23:51:45 train.py: 79] Epoch 4, iter 2800/6416, lr 0.100000, loss 5.630188
+INFO 2020-12-02 23:54:52 train.py: 92] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-12-02 23:54:53 train.py: 79] Epoch 4, iter 3000/6416, lr 0.100000, loss 5.618637
+INFO 2020-12-02 23:58:00 train.py: 79] Epoch 4, iter 3200/6416, lr 0.100000, loss 5.612096
+INFO 2020-12-03 00:01:07 train.py: 79] Epoch 4, iter 3400/6416, lr 0.100000, loss 5.594699
+INFO 2020-12-03 00:04:15 train.py: 79] Epoch 4, iter 3600/6416, lr 0.100000, loss 5.594811
+INFO 2020-12-03 00:07:22 train.py: 79] Epoch 4, iter 3800/6416, lr 0.100000, loss 5.576285
+INFO 2020-12-03 00:10:30 train.py: 79] Epoch 4, iter 4000/6416, lr 0.100000, loss 5.578566
+INFO 2020-12-03 00:13:37 train.py: 79] Epoch 4, iter 4200/6416, lr 0.100000, loss 5.579606
+INFO 2020-12-03 00:16:45 train.py: 79] Epoch 4, iter 4400/6416, lr 0.100000, loss 5.560446
+INFO 2020-12-03 00:19:53 train.py: 79] Epoch 4, iter 4600/6416, lr 0.100000, loss 5.560880
+INFO 2020-12-03 00:23:00 train.py: 79] Epoch 4, iter 4800/6416, lr 0.100000, loss 5.541731
+INFO 2020-12-03 00:26:08 train.py: 79] Epoch 4, iter 5000/6416, lr 0.100000, loss 5.540355
+INFO 2020-12-03 00:29:15 train.py: 79] Epoch 4, iter 5200/6416, lr 0.100000, loss 5.524416
+INFO 2020-12-03 00:32:23 train.py: 79] Epoch 4, iter 5400/6416, lr 0.100000, loss 5.523820
+INFO 2020-12-03 00:35:31 train.py: 79] Epoch 4, iter 5600/6416, lr 0.100000, loss 5.498184
+INFO 2020-12-03 00:38:38 train.py: 79] Epoch 4, iter 5800/6416, lr 0.100000, loss 5.482068
+INFO 2020-12-03 00:41:46 train.py: 92] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-12-03 00:41:46 train.py: 79] Epoch 4, iter 6000/6416, lr 0.100000, loss 5.487048
+INFO 2020-12-03 00:44:54 train.py: 79] Epoch 4, iter 6200/6416, lr 0.100000, loss 5.471176
+INFO 2020-12-03 00:48:01 train.py: 79] Epoch 4, iter 6400/6416, lr 0.100000, loss 5.465708
+INFO 2020-12-03 00:48:15 train.py: 97] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-12-03 00:48:17 train.py: 79] Epoch 5, iter 0/6416, lr 0.100000, loss 5.451878
+INFO 2020-12-03 00:51:23 train.py: 79] Epoch 5, iter 200/6416, lr 0.100000, loss 4.967053
+INFO 2020-12-03 00:54:30 train.py: 79] Epoch 5, iter 400/6416, lr 0.100000, loss 4.949885
+INFO 2020-12-03 00:57:36 train.py: 79] Epoch 5, iter 600/6416, lr 0.100000, loss 5.028063
+INFO 2020-12-03 01:00:42 train.py: 79] Epoch 5, iter 800/6416, lr 0.100000, loss 5.128778
+INFO 2020-12-03 01:03:48 train.py: 79] Epoch 5, iter 1000/6416, lr 0.100000, loss 5.130865
+INFO 2020-12-03 01:06:55 train.py: 79] Epoch 5, iter 1200/6416, lr 0.100000, loss 5.188409
+INFO 2020-12-03 01:10:01 train.py: 79] Epoch 5, iter 1400/6416, lr 0.100000, loss 5.182377
+INFO 2020-12-03 01:13:07 train.py: 79] Epoch 5, iter 1600/6416, lr 0.100000, loss 5.228963
+INFO 2020-12-03 01:16:14 train.py: 79] Epoch 5, iter 1800/6416, lr 0.100000, loss 5.249112
+INFO 2020-12-03 01:19:20 train.py: 79] Epoch 5, iter 2000/6416, lr 0.100000, loss 5.271594
+INFO 2020-12-03 01:22:26 train.py: 79] Epoch 5, iter 2200/6416, lr 0.100000, loss 5.273448
+INFO 2020-12-03 01:25:33 train.py: 79] Epoch 5, iter 2400/6416, lr 0.100000, loss 5.280648
+INFO 2020-12-03 01:28:39 train.py: 79] Epoch 5, iter 2600/6416, lr 0.100000, loss 5.272331
+INFO 2020-12-03 01:31:46 train.py: 79] Epoch 5, iter 2800/6416, lr 0.100000, loss 5.284181
+INFO 2020-12-03 01:34:52 train.py: 92] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-12-03 01:34:53 train.py: 79] Epoch 5, iter 3000/6416, lr 0.100000, loss 5.274477
+INFO 2020-12-03 01:38:00 train.py: 79] Epoch 5, iter 3200/6416, lr 0.100000, loss 5.293002
+INFO 2020-12-03 01:41:08 train.py: 79] Epoch 5, iter 3400/6416, lr 0.100000, loss 5.278444
+INFO 2020-12-03 01:44:15 train.py: 79] Epoch 5, iter 3600/6416, lr 0.100000, loss 5.219957
+INFO 2020-12-03 01:47:23 train.py: 79] Epoch 5, iter 3800/6416, lr 0.100000, loss 5.263780
+INFO 2020-12-03 01:50:30 train.py: 79] Epoch 5, iter 4000/6416, lr 0.100000, loss 5.216983
+INFO 2020-12-03 01:53:38 train.py: 79] Epoch 5, iter 4200/6416, lr 0.100000, loss 5.261914
+INFO 2020-12-03 01:56:45 train.py: 79] Epoch 5, iter 4400/6416, lr 0.100000, loss 5.258034
+INFO 2020-12-03 01:59:53 train.py: 79] Epoch 5, iter 4600/6416, lr 0.100000, loss 5.222422
+INFO 2020-12-03 02:03:00 train.py: 79] Epoch 5, iter 4800/6416, lr 0.100000, loss 5.210544
+INFO 2020-12-03 02:06:08 train.py: 79] Epoch 5, iter 5000/6416, lr 0.100000, loss 5.223289
+INFO 2020-12-03 02:09:16 train.py: 79] Epoch 5, iter 5200/6416, lr 0.100000, loss 5.235448
+INFO 2020-12-03 02:12:23 train.py: 79] Epoch 5, iter 5400/6416, lr 0.100000, loss 5.218714
+INFO 2020-12-03 02:15:31 train.py: 79] Epoch 5, iter 5600/6416, lr 0.100000, loss 5.199490
+INFO 2020-12-03 02:18:39 train.py: 79] Epoch 5, iter 5800/6416, lr 0.100000, loss 5.235230
+INFO 2020-12-03 02:21:46 train.py: 92] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-12-03 02:21:47 train.py: 79] Epoch 5, iter 6000/6416, lr 0.100000, loss 5.208253
+INFO 2020-12-03 02:24:55 train.py: 79] Epoch 5, iter 6200/6416, lr 0.100000, loss 5.216329
+INFO 2020-12-03 02:28:03 train.py: 79] Epoch 5, iter 6400/6416, lr 0.100000, loss 5.189704
+INFO 2020-12-03 02:28:17 train.py: 97] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-12-03 02:28:19 train.py: 79] Epoch 6, iter 0/6416, lr 0.100000, loss 5.104144
+INFO 2020-12-03 02:31:26 train.py: 79] Epoch 6, iter 200/6416, lr 0.100000, loss 4.670524
+INFO 2020-12-03 02:34:33 train.py: 79] Epoch 6, iter 400/6416, lr 0.100000, loss 4.665509
+INFO 2020-12-03 02:37:40 train.py: 79] Epoch 6, iter 600/6416, lr 0.100000, loss 4.749700
+INFO 2020-12-03 02:40:47 train.py: 79] Epoch 6, iter 800/6416, lr 0.100000, loss 4.823609
+INFO 2020-12-03 02:43:54 train.py: 79] Epoch 6, iter 1000/6416, lr 0.100000, loss 4.862002
+INFO 2020-12-03 02:47:01 train.py: 79] Epoch 6, iter 1200/6416, lr 0.100000, loss 4.931015
+INFO 2020-12-03 02:50:08 train.py: 79] Epoch 6, iter 1400/6416, lr 0.100000, loss 4.945732
+INFO 2020-12-03 02:53:15 train.py: 79] Epoch 6, iter 1600/6416, lr 0.100000, loss 4.949572
+INFO 2020-12-03 02:56:22 train.py: 79] Epoch 6, iter 1800/6416, lr 0.100000, loss 4.985425
+INFO 2020-12-03 02:59:29 train.py: 79] Epoch 6, iter 2000/6416, lr 0.100000, loss 4.988755
+INFO 2020-12-03 03:02:36 train.py: 79] Epoch 6, iter 2200/6416, lr 0.100000, loss 5.008859
+INFO 2020-12-03 03:05:43 train.py: 79] Epoch 6, iter 2400/6416, lr 0.100000, loss 5.048368
+INFO 2020-12-03 03:08:50 train.py: 79] Epoch 6, iter 2600/6416, lr 0.100000, loss 5.053820
+INFO 2020-12-03 03:11:57 train.py: 79] Epoch 6, iter 2800/6416, lr 0.100000, loss 5.033558
+INFO 2020-12-03 03:15:04 train.py: 92] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-12-03 03:15:05 train.py: 79] Epoch 6, iter 3000/6416, lr 0.100000, loss 5.036939
+INFO 2020-12-03 03:18:12 train.py: 79] Epoch 6, iter 3200/6416, lr 0.100000, loss 5.029966
+INFO 2020-12-03 03:21:20 train.py: 79] Epoch 6, iter 3400/6416, lr 0.100000, loss 5.043766
+INFO 2020-12-03 03:24:27 train.py: 79] Epoch 6, iter 3600/6416, lr 0.100000, loss 5.056108
+INFO 2020-12-03 03:27:35 train.py: 79] Epoch 6, iter 3800/6416, lr 0.100000, loss 5.014114
+INFO 2020-12-03 03:30:42 train.py: 79] Epoch 6, iter 4000/6416, lr 0.100000, loss 5.028921
+INFO 2020-12-03 03:33:50 train.py: 79] Epoch 6, iter 4200/6416, lr 0.100000, loss 5.063586
+INFO 2020-12-03 03:36:57 train.py: 79] Epoch 6, iter 4400/6416, lr 0.100000, loss 5.039221
+INFO 2020-12-03 03:40:05 train.py: 79] Epoch 6, iter 4600/6416, lr 0.100000, loss 5.033477
+INFO 2020-12-03 03:43:12 train.py: 79] Epoch 6, iter 4800/6416, lr 0.100000, loss 5.026603
+INFO 2020-12-03 03:46:20 train.py: 79] Epoch 6, iter 5000/6416, lr 0.100000, loss 5.009321
+INFO 2020-12-03 03:49:27 train.py: 79] Epoch 6, iter 5200/6416, lr 0.100000, loss 5.043872
+INFO 2020-12-03 03:52:35 train.py: 79] Epoch 6, iter 5400/6416, lr 0.100000, loss 5.000646
+INFO 2020-12-03 03:55:43 train.py: 79] Epoch 6, iter 5600/6416, lr 0.100000, loss 5.012944
+INFO 2020-12-03 03:58:50 train.py: 79] Epoch 6, iter 5800/6416, lr 0.100000, loss 5.008230
+INFO 2020-12-03 04:01:58 train.py: 92] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-12-03 04:01:59 train.py: 79] Epoch 6, iter 6000/6416, lr 0.100000, loss 5.013476
+INFO 2020-12-03 04:05:06 train.py: 79] Epoch 6, iter 6200/6416, lr 0.100000, loss 4.998242
+INFO 2020-12-03 04:08:13 train.py: 79] Epoch 6, iter 6400/6416, lr 0.100000, loss 4.975437
+INFO 2020-12-03 04:08:27 train.py: 97] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-12-03 04:08:29 train.py: 79] Epoch 7, iter 0/6416, lr 0.100000, loss 5.022930
+INFO 2020-12-03 04:11:35 train.py: 79] Epoch 7, iter 200/6416, lr 0.100000, loss 4.513452
+INFO 2020-12-03 04:14:42 train.py: 79] Epoch 7, iter 400/6416, lr 0.100000, loss 4.482976
+INFO 2020-12-03 04:17:48 train.py: 79] Epoch 7, iter 600/6416, lr 0.100000, loss 4.583296
+INFO 2020-12-03 04:20:54 train.py: 79] Epoch 7, iter 800/6416, lr 0.100000, loss 4.615423
+INFO 2020-12-03 04:24:01 train.py: 79] Epoch 7, iter 1000/6416, lr 0.100000, loss 4.688656
+INFO 2020-12-03 04:27:07 train.py: 79] Epoch 7, iter 1200/6416, lr 0.100000, loss 4.745344
+INFO 2020-12-03 04:30:13 train.py: 79] Epoch 7, iter 1400/6416, lr 0.100000, loss 4.792305
+INFO 2020-12-03 04:33:20 train.py: 79] Epoch 7, iter 1600/6416, lr 0.100000, loss 4.783717
+INFO 2020-12-03 04:36:26 train.py: 79] Epoch 7, iter 1800/6416, lr 0.100000, loss 4.811773
+INFO 2020-12-03 04:39:33 train.py: 79] Epoch 7, iter 2000/6416, lr 0.100000, loss 4.817311
+INFO 2020-12-03 04:42:39 train.py: 79] Epoch 7, iter 2200/6416, lr 0.100000, loss 4.844831
+INFO 2020-12-03 04:45:46 train.py: 79] Epoch 7, iter 2400/6416, lr 0.100000, loss 4.856340
+INFO 2020-12-03 04:48:52 train.py: 79] Epoch 7, iter 2600/6416, lr 0.100000, loss 4.874803
+INFO 2020-12-03 04:51:59 train.py: 79] Epoch 7, iter 2800/6416, lr 0.100000, loss 4.854270
+INFO 2020-12-03 04:55:05 train.py: 92] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-12-03 04:55:06 train.py: 79] Epoch 7, iter 3000/6416, lr 0.100000, loss 4.856605
+INFO 2020-12-03 04:58:13 train.py: 79] Epoch 7, iter 3200/6416, lr 0.100000, loss 4.843140
+INFO 2020-12-03 05:01:20 train.py: 79] Epoch 7, iter 3400/6416, lr 0.100000, loss 4.901404
+INFO 2020-12-03 05:04:28 train.py: 79] Epoch 7, iter 3600/6416, lr 0.100000, loss 4.851626
+INFO 2020-12-03 05:07:35 train.py: 79] Epoch 7, iter 3800/6416, lr 0.100000, loss 4.883711
+INFO 2020-12-03 05:10:43 train.py: 79] Epoch 7, iter 4000/6416, lr 0.100000, loss 4.856796
+INFO 2020-12-03 05:13:50 train.py: 79] Epoch 7, iter 4200/6416, lr 0.100000, loss 4.874653
+INFO 2020-12-03 05:16:58 train.py: 79] Epoch 7, iter 4400/6416, lr 0.100000, loss 4.836083
+INFO 2020-12-03 05:20:05 train.py: 79] Epoch 7, iter 4600/6416, lr 0.100000, loss 4.861144
+INFO 2020-12-03 05:23:13 train.py: 79] Epoch 7, iter 4800/6416, lr 0.100000, loss 4.899105
+INFO 2020-12-03 05:26:20 train.py: 79] Epoch 7, iter 5000/6416, lr 0.100000, loss 4.867728
+INFO 2020-12-03 05:29:28 train.py: 79] Epoch 7, iter 5200/6416, lr 0.100000, loss 4.843084
+INFO 2020-12-03 05:32:36 train.py: 79] Epoch 7, iter 5400/6416, lr 0.100000, loss 4.857403
+INFO 2020-12-03 05:35:43 train.py: 79] Epoch 7, iter 5600/6416, lr 0.100000, loss 4.876713
+INFO 2020-12-03 05:38:51 train.py: 79] Epoch 7, iter 5800/6416, lr 0.100000, loss 4.839580
+INFO 2020-12-03 05:41:58 train.py: 92] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-12-03 05:41:59 train.py: 79] Epoch 7, iter 6000/6416, lr 0.100000, loss 4.879442
+INFO 2020-12-03 05:45:07 train.py: 79] Epoch 7, iter 6200/6416, lr 0.100000, loss 4.835701
+INFO 2020-12-03 05:48:15 train.py: 79] Epoch 7, iter 6400/6416, lr 0.100000, loss 4.831005
+INFO 2020-12-03 05:48:29 train.py: 97] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-12-03 05:48:31 train.py: 79] Epoch 8, iter 0/6416, lr 0.100000, loss 4.855350
+INFO 2020-12-03 05:51:38 train.py: 79] Epoch 8, iter 200/6416, lr 0.100000, loss 4.334920
+INFO 2020-12-03 05:54:45 train.py: 79] Epoch 8, iter 400/6416, lr 0.100000, loss 4.316042
+INFO 2020-12-03 05:57:52 train.py: 79] Epoch 8, iter 600/6416, lr 0.100000, loss 4.422725
+INFO 2020-12-03 06:00:59 train.py: 79] Epoch 8, iter 800/6416, lr 0.100000, loss 4.480498
+INFO 2020-12-03 06:04:06 train.py: 79] Epoch 8, iter 1000/6416, lr 0.100000, loss 4.541036
+INFO 2020-12-03 06:07:13 train.py: 79] Epoch 8, iter 1200/6416, lr 0.100000, loss 4.578566
+INFO 2020-12-03 06:10:20 train.py: 79] Epoch 8, iter 1400/6416, lr 0.100000, loss 4.642581
+INFO 2020-12-03 06:13:27 train.py: 79] Epoch 8, iter 1600/6416, lr 0.100000, loss 4.655586
+INFO 2020-12-03 06:16:34 train.py: 79] Epoch 8, iter 1800/6416, lr 0.100000, loss 4.634501
+INFO 2020-12-03 06:19:41 train.py: 79] Epoch 8, iter 2000/6416, lr 0.100000, loss 4.697801
+INFO 2020-12-03 06:22:48 train.py: 79] Epoch 8, iter 2200/6416, lr 0.100000, loss 4.690633
+INFO 2020-12-03 06:25:55 train.py: 79] Epoch 8, iter 2400/6416, lr 0.100000, loss 4.763088
+INFO 2020-12-03 06:29:02 train.py: 79] Epoch 8, iter 2600/6416, lr 0.100000, loss 4.749126
+INFO 2020-12-03 06:32:09 train.py: 79] Epoch 8, iter 2800/6416, lr 0.100000, loss 4.736125
+INFO 2020-12-03 06:35:16 train.py: 92] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-12-03 06:35:17 train.py: 79] Epoch 8, iter 3000/6416, lr 0.100000, loss 4.752036
+INFO 2020-12-03 06:38:24 train.py: 79] Epoch 8, iter 3200/6416, lr 0.100000, loss 4.745572
+INFO 2020-12-03 06:41:32 train.py: 79] Epoch 8, iter 3400/6416, lr 0.100000, loss 4.758962
+INFO 2020-12-03 06:44:39 train.py: 79] Epoch 8, iter 3600/6416, lr 0.100000, loss 4.771978
+INFO 2020-12-03 06:47:47 train.py: 79] Epoch 8, iter 3800/6416, lr 0.100000, loss 4.740541
+INFO 2020-12-03 06:50:54 train.py: 79] Epoch 8, iter 4000/6416, lr 0.100000, loss 4.732087
+INFO 2020-12-03 06:54:02 train.py: 79] Epoch 8, iter 4200/6416, lr 0.100000, loss 4.743392
+INFO 2020-12-03 06:57:09 train.py: 79] Epoch 8, iter 4400/6416, lr 0.100000, loss 4.722718
+INFO 2020-12-03 07:00:17 train.py: 79] Epoch 8, iter 4600/6416, lr 0.100000, loss 4.774203
+INFO 2020-12-03 07:03:24 train.py: 79] Epoch 8, iter 4800/6416, lr 0.100000, loss 4.736959
+INFO 2020-12-03 07:06:32 train.py: 79] Epoch 8, iter 5000/6416, lr 0.100000, loss 4.745756
+INFO 2020-12-03 07:09:40 train.py: 79] Epoch 8, iter 5200/6416, lr 0.100000, loss 4.745374
+INFO 2020-12-03 07:12:47 train.py: 79] Epoch 8, iter 5400/6416, lr 0.100000, loss 4.709741
+INFO 2020-12-03 07:15:55 train.py: 79] Epoch 8, iter 5600/6416, lr 0.100000, loss 4.747218
+INFO 2020-12-03 07:19:03 train.py: 79] Epoch 8, iter 5800/6416, lr 0.100000, loss 4.730891
+INFO 2020-12-03 07:22:10 train.py: 92] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-12-03 07:22:11 train.py: 79] Epoch 8, iter 6000/6416, lr 0.100000, loss 4.718416
+INFO 2020-12-03 07:25:18 train.py: 79] Epoch 8, iter 6200/6416, lr 0.100000, loss 4.721629
+INFO 2020-12-03 07:28:25 train.py: 79] Epoch 8, iter 6400/6416, lr 0.100000, loss 4.725244
+INFO 2020-12-03 07:28:39 train.py: 97] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-12-03 07:28:41 train.py: 79] Epoch 9, iter 0/6416, lr 0.100000, loss 4.708601
+INFO 2020-12-03 07:31:49 train.py: 79] Epoch 9, iter 200/6416, lr 0.100000, loss 4.243455
+INFO 2020-12-03 07:34:56 train.py: 79] Epoch 9, iter 400/6416, lr 0.100000, loss 4.201859
+INFO 2020-12-03 07:38:02 train.py: 79] Epoch 9, iter 600/6416, lr 0.100000, loss 4.286142
+INFO 2020-12-03 07:41:09 train.py: 79] Epoch 9, iter 800/6416, lr 0.100000, loss 4.379097
+INFO 2020-12-03 07:44:16 train.py: 79] Epoch 9, iter 1000/6416, lr 0.100000, loss 4.433232
+INFO 2020-12-03 07:47:23 train.py: 79] Epoch 9, iter 1200/6416, lr 0.100000, loss 4.478835
+INFO 2020-12-03 07:50:30 train.py: 79] Epoch 9, iter 1400/6416, lr 0.100000, loss 4.501469
+INFO 2020-12-03 07:53:37 train.py: 79] Epoch 9, iter 1600/6416, lr 0.100000, loss 4.549438
+INFO 2020-12-03 07:56:44 train.py: 79] Epoch 9, iter 1800/6416, lr 0.100000, loss 4.564158
+INFO 2020-12-03 07:59:51 train.py: 79] Epoch 9, iter 2000/6416, lr 0.100000, loss 4.566252
+INFO 2020-12-03 08:02:58 train.py: 79] Epoch 9, iter 2200/6416, lr 0.100000, loss 4.587315
+INFO 2020-12-03 08:06:05 train.py: 79] Epoch 9, iter 2400/6416, lr 0.100000, loss 4.598708
+INFO 2020-12-03 08:09:12 train.py: 79] Epoch 9, iter 2600/6416, lr 0.100000, loss 4.576349
+INFO 2020-12-03 08:12:19 train.py: 79] Epoch 9, iter 2800/6416, lr 0.100000, loss 4.634216
+INFO 2020-12-03 08:15:26 train.py: 92] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-12-03 08:15:27 train.py: 79] Epoch 9, iter 3000/6416, lr 0.100000, loss 4.616995
+INFO 2020-12-03 08:18:34 train.py: 79] Epoch 9, iter 3200/6416, lr 0.100000, loss 4.627426
+INFO 2020-12-03 08:21:40 train.py: 79] Epoch 9, iter 3400/6416, lr 0.100000, loss 4.630696
+INFO 2020-12-03 08:24:47 train.py: 79] Epoch 9, iter 3600/6416, lr 0.100000, loss 4.654315
+INFO 2020-12-03 08:27:54 train.py: 79] Epoch 9, iter 3800/6416, lr 0.100000, loss 4.634898
+INFO 2020-12-03 08:31:01 train.py: 79] Epoch 9, iter 4000/6416, lr 0.100000, loss 4.635881
+INFO 2020-12-03 08:34:08 train.py: 79] Epoch 9, iter 4200/6416, lr 0.100000, loss 4.645761
+INFO 2020-12-03 08:37:15 train.py: 79] Epoch 9, iter 4400/6416, lr 0.100000, loss 4.660442
+INFO 2020-12-03 08:40:22 train.py: 79] Epoch 9, iter 4600/6416, lr 0.100000, loss 4.654919
+INFO 2020-12-03 08:43:29 train.py: 79] Epoch 9, iter 4800/6416, lr 0.100000, loss 4.661503
+INFO 2020-12-03 08:46:36 train.py: 79] Epoch 9, iter 5000/6416, lr 0.100000, loss 4.637967
+INFO 2020-12-03 08:49:43 train.py: 79] Epoch 9, iter 5200/6416, lr 0.100000, loss 4.638848
+INFO 2020-12-03 08:52:50 train.py: 79] Epoch 9, iter 5400/6416, lr 0.100000, loss 4.634365
+INFO 2020-12-03 08:55:57 train.py: 79] Epoch 9, iter 5600/6416, lr 0.100000, loss 4.619052
+INFO 2020-12-03 08:59:04 train.py: 79] Epoch 9, iter 5800/6416, lr 0.100000, loss 4.669011
+INFO 2020-12-03 09:02:10 train.py: 92] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-12-03 09:02:11 train.py: 79] Epoch 9, iter 6000/6416, lr 0.100000, loss 4.644300
+INFO 2020-12-03 09:05:19 train.py: 79] Epoch 9, iter 6200/6416, lr 0.100000, loss 4.620329
+INFO 2020-12-03 09:08:27 train.py: 79] Epoch 9, iter 6400/6416, lr 0.100000, loss 4.648147
+INFO 2020-12-03 09:08:41 train.py: 97] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-12-03 09:08:43 train.py: 79] Epoch 10, iter 0/6416, lr 0.010000, loss 4.587415
+INFO 2020-12-03 09:11:50 train.py: 79] Epoch 10, iter 200/6416, lr 0.010000, loss 3.600639
+INFO 2020-12-03 09:14:57 train.py: 79] Epoch 10, iter 400/6416, lr 0.010000, loss 3.326887
+INFO 2020-12-03 09:18:04 train.py: 79] Epoch 10, iter 600/6416, lr 0.010000, loss 3.218004
+INFO 2020-12-03 09:21:11 train.py: 79] Epoch 10, iter 800/6416, lr 0.010000, loss 3.135760
+INFO 2020-12-03 09:24:18 train.py: 79] Epoch 10, iter 1000/6416, lr 0.010000, loss 3.111210
+INFO 2020-12-03 09:27:25 train.py: 79] Epoch 10, iter 1200/6416, lr 0.010000, loss 3.050307
+INFO 2020-12-03 09:30:32 train.py: 79] Epoch 10, iter 1400/6416, lr 0.010000, loss 2.967692
+INFO 2020-12-03 09:33:39 train.py: 79] Epoch 10, iter 1600/6416, lr 0.010000, loss 2.966347
+INFO 2020-12-03 09:36:45 train.py: 79] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.902957
+INFO 2020-12-03 09:39:52 train.py: 79] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.886213
+INFO 2020-12-03 09:42:59 train.py: 79] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.868429
+INFO 2020-12-03 09:46:06 train.py: 79] Epoch 10, iter 2400/6416, lr 0.010000, loss 2.804226
+INFO 2020-12-03 09:49:14 train.py: 79] Epoch 10, iter 2600/6416, lr 0.010000, loss 2.800321
+INFO 2020-12-03 09:52:21 train.py: 79] Epoch 10, iter 2800/6416, lr 0.010000, loss 2.782938
+INFO 2020-12-03 09:55:27 train.py: 92] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-12-03 09:55:28 train.py: 79] Epoch 10, iter 3000/6416, lr 0.010000, loss 2.750414
+INFO 2020-12-03 09:58:36 train.py: 79] Epoch 10, iter 3200/6416, lr 0.010000, loss 2.738220
+INFO 2020-12-03 10:01:43 train.py: 79] Epoch 10, iter 3400/6416, lr 0.010000, loss 2.730819
+INFO 2020-12-03 10:04:50 train.py: 79] Epoch 10, iter 3600/6416, lr 0.010000, loss 2.689711
+INFO 2020-12-03 10:07:58 train.py: 79] Epoch 10, iter 3800/6416, lr 0.010000, loss 2.681723
+INFO 2020-12-03 10:11:05 train.py: 79] Epoch 10, iter 4000/6416, lr 0.010000, loss 2.657805
+INFO 2020-12-03 10:14:12 train.py: 79] Epoch 10, iter 4200/6416, lr 0.010000, loss 2.661914
+INFO 2020-12-03 10:17:20 train.py: 79] Epoch 10, iter 4400/6416, lr 0.010000, loss 2.620412
+INFO 2020-12-03 10:20:27 train.py: 79] Epoch 10, iter 4600/6416, lr 0.010000, loss 2.592250
+INFO 2020-12-03 10:23:35 train.py: 79] Epoch 10, iter 4800/6416, lr 0.010000, loss 2.569926
+INFO 2020-12-03 10:26:42 train.py: 79] Epoch 10, iter 5000/6416, lr 0.010000, loss 2.583932
+INFO 2020-12-03 10:29:50 train.py: 79] Epoch 10, iter 5200/6416, lr 0.010000, loss 2.553929
+INFO 2020-12-03 10:32:58 train.py: 79] Epoch 10, iter 5400/6416, lr 0.010000, loss 2.567436
+INFO 2020-12-03 10:36:05 train.py: 79] Epoch 10, iter 5600/6416, lr 0.010000, loss 2.529675
+INFO 2020-12-03 10:39:13 train.py: 79] Epoch 10, iter 5800/6416, lr 0.010000, loss 2.534601
+INFO 2020-12-03 10:42:20 train.py: 92] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-12-03 10:42:21 train.py: 79] Epoch 10, iter 6000/6416, lr 0.010000, loss 2.517062
+INFO 2020-12-03 10:45:28 train.py: 79] Epoch 10, iter 6200/6416, lr 0.010000, loss 2.511272
+INFO 2020-12-03 10:48:36 train.py: 79] Epoch 10, iter 6400/6416, lr 0.010000, loss 2.504296
+INFO 2020-12-03 10:48:50 train.py: 97] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-12-03 10:48:52 train.py: 79] Epoch 11, iter 0/6416, lr 0.010000, loss 2.511902
+INFO 2020-12-03 10:51:59 train.py: 79] Epoch 11, iter 200/6416, lr 0.010000, loss 2.179786
+INFO 2020-12-03 10:55:06 train.py: 79] Epoch 11, iter 400/6416, lr 0.010000, loss 2.178046
+INFO 2020-12-03 10:58:13 train.py: 79] Epoch 11, iter 600/6416, lr 0.010000, loss 2.180389
+INFO 2020-12-03 11:01:20 train.py: 79] Epoch 11, iter 800/6416, lr 0.010000, loss 2.173076
+INFO 2020-12-03 11:04:27 train.py: 79] Epoch 11, iter 1000/6416, lr 0.010000, loss 2.158977
+INFO 2020-12-03 11:07:34 train.py: 79] Epoch 11, iter 1200/6416, lr 0.010000, loss 2.160206
+INFO 2020-12-03 11:10:41 train.py: 79] Epoch 11, iter 1400/6416, lr 0.010000, loss 2.185351
+INFO 2020-12-03 11:13:48 train.py: 79] Epoch 11, iter 1600/6416, lr 0.010000, loss 2.181845
+INFO 2020-12-03 11:16:54 train.py: 79] Epoch 11, iter 1800/6416, lr 0.010000, loss 2.176098
+INFO 2020-12-03 11:20:01 train.py: 79] Epoch 11, iter 2000/6416, lr 0.010000, loss 2.155652
+INFO 2020-12-03 11:23:08 train.py: 79] Epoch 11, iter 2200/6416, lr 0.010000, loss 2.177292
+INFO 2020-12-03 11:26:15 train.py: 79] Epoch 11, iter 2400/6416, lr 0.010000, loss 2.169880
+INFO 2020-12-03 11:29:23 train.py: 79] Epoch 11, iter 2600/6416, lr 0.010000, loss 2.178358
+INFO 2020-12-03 11:32:30 train.py: 79] Epoch 11, iter 2800/6416, lr 0.010000, loss 2.159219
+INFO 2020-12-03 11:35:36 train.py: 92] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-12-03 11:35:37 train.py: 79] Epoch 11, iter 3000/6416, lr 0.010000, loss 2.171988
+INFO 2020-12-03 11:38:44 train.py: 79] Epoch 11, iter 3200/6416, lr 0.010000, loss 2.153228
+INFO 2020-12-03 11:41:50 train.py: 79] Epoch 11, iter 3400/6416, lr 0.010000, loss 2.171687
+INFO 2020-12-03 11:44:57 train.py: 79] Epoch 11, iter 3600/6416, lr 0.010000, loss 2.160353
+INFO 2020-12-03 11:48:04 train.py: 79] Epoch 11, iter 3800/6416, lr 0.010000, loss 2.179015
+INFO 2020-12-03 11:51:11 train.py: 79] Epoch 11, iter 4000/6416, lr 0.010000, loss 2.152023
+INFO 2020-12-03 11:54:18 train.py: 79] Epoch 11, iter 4200/6416, lr 0.010000, loss 2.167601
+INFO 2020-12-03 11:57:24 train.py: 79] Epoch 11, iter 4400/6416, lr 0.010000, loss 2.177003
+INFO 2020-12-03 12:00:31 train.py: 79] Epoch 11, iter 4600/6416, lr 0.010000, loss 2.156524
+INFO 2020-12-03 12:03:38 train.py: 79] Epoch 11, iter 4800/6416, lr 0.010000, loss 2.149799
+INFO 2020-12-03 12:06:45 train.py: 79] Epoch 11, iter 5000/6416, lr 0.010000, loss 2.166937
+INFO 2020-12-03 12:09:52 train.py: 79] Epoch 11, iter 5200/6416, lr 0.010000, loss 2.144454
+INFO 2020-12-03 12:12:59 train.py: 79] Epoch 11, iter 5400/6416, lr 0.010000, loss 2.136277
+INFO 2020-12-03 12:16:06 train.py: 79] Epoch 11, iter 5600/6416, lr 0.010000, loss 2.132223
+INFO 2020-12-03 12:19:13 train.py: 79] Epoch 11, iter 5800/6416, lr 0.010000, loss 2.149768
+INFO 2020-12-03 12:22:20 train.py: 92] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-12-03 12:22:21 train.py: 79] Epoch 11, iter 6000/6416, lr 0.010000, loss 2.152355
+INFO 2020-12-03 12:25:28 train.py: 79] Epoch 11, iter 6200/6416, lr 0.010000, loss 2.158323
+INFO 2020-12-03 12:28:36 train.py: 79] Epoch 11, iter 6400/6416, lr 0.010000, loss 2.155531
+INFO 2020-12-03 12:28:50 train.py: 97] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-12-03 12:28:52 train.py: 79] Epoch 12, iter 0/6416, lr 0.010000, loss 2.192659
+INFO 2020-12-03 12:32:00 train.py: 79] Epoch 12, iter 200/6416, lr 0.010000, loss 1.861888
+INFO 2020-12-03 12:35:06 train.py: 79] Epoch 12, iter 400/6416, lr 0.010000, loss 1.850325
+INFO 2020-12-03 12:38:13 train.py: 79] Epoch 12, iter 600/6416, lr 0.010000, loss 1.858519
+INFO 2020-12-03 12:41:20 train.py: 79] Epoch 12, iter 800/6416, lr 0.010000, loss 1.868805
+INFO 2020-12-03 12:44:27 train.py: 79] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.875272
+INFO 2020-12-03 12:47:34 train.py: 79] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.878553
+INFO 2020-12-03 12:50:41 train.py: 79] Epoch 12, iter 1400/6416, lr 0.010000, loss 1.884702
+INFO 2020-12-03 12:53:48 train.py: 79] Epoch 12, iter 1600/6416, lr 0.010000, loss 1.899908
+INFO 2020-12-03 12:56:55 train.py: 79] Epoch 12, iter 1800/6416, lr 0.010000, loss 1.894162
+INFO 2020-12-03 13:00:02 train.py: 79] Epoch 12, iter 2000/6416, lr 0.010000, loss 1.905304
+INFO 2020-12-03 13:03:09 train.py: 79] Epoch 12, iter 2200/6416, lr 0.010000, loss 1.923799
+INFO 2020-12-03 13:06:16 train.py: 79] Epoch 12, iter 2400/6416, lr 0.010000, loss 1.894776
+INFO 2020-12-03 13:09:23 train.py: 79] Epoch 12, iter 2600/6416, lr 0.010000, loss 1.924357
+INFO 2020-12-03 13:12:30 train.py: 79] Epoch 12, iter 2800/6416, lr 0.010000, loss 1.928320
+INFO 2020-12-03 13:15:37 train.py: 92] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-12-03 13:15:38 train.py: 79] Epoch 12, iter 3000/6416, lr 0.010000, loss 1.917324
+INFO 2020-12-03 13:18:45 train.py: 79] Epoch 12, iter 3200/6416, lr 0.010000, loss 1.932692
+INFO 2020-12-03 13:21:52 train.py: 79] Epoch 12, iter 3400/6416, lr 0.010000, loss 1.943242
+INFO 2020-12-03 13:25:00 train.py: 79] Epoch 12, iter 3600/6416, lr 0.010000, loss 1.958510
+INFO 2020-12-03 13:28:07 train.py: 79] Epoch 12, iter 3800/6416, lr 0.010000, loss 1.947445
+INFO 2020-12-03 13:31:15 train.py: 79] Epoch 12, iter 4000/6416, lr 0.010000, loss 1.955349
+INFO 2020-12-03 13:34:22 train.py: 79] Epoch 12, iter 4200/6416, lr 0.010000, loss 1.950447
+INFO 2020-12-03 13:37:30 train.py: 79] Epoch 12, iter 4400/6416, lr 0.010000, loss 1.966869
+INFO 2020-12-03 13:40:37 train.py: 79] Epoch 12, iter 4600/6416, lr 0.010000, loss 1.966571
+INFO 2020-12-03 13:43:45 train.py: 79] Epoch 12, iter 4800/6416, lr 0.010000, loss 1.961027
+INFO 2020-12-03 13:46:52 train.py: 79] Epoch 12, iter 5000/6416, lr 0.010000, loss 1.972043
+INFO 2020-12-03 13:50:00 train.py: 79] Epoch 12, iter 5200/6416, lr 0.010000, loss 1.966379
+INFO 2020-12-03 13:53:07 train.py: 79] Epoch 12, iter 5400/6416, lr 0.010000, loss 1.981322
+INFO 2020-12-03 13:56:15 train.py: 79] Epoch 12, iter 5600/6416, lr 0.010000, loss 1.978793
+INFO 2020-12-03 13:59:23 train.py: 79] Epoch 12, iter 5800/6416, lr 0.010000, loss 2.013717
+INFO 2020-12-03 14:02:30 train.py: 92] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-12-03 14:02:31 train.py: 79] Epoch 12, iter 6000/6416, lr 0.010000, loss 1.972498
+INFO 2020-12-03 14:05:38 train.py: 79] Epoch 12, iter 6200/6416, lr 0.010000, loss 2.001041
+INFO 2020-12-03 14:08:46 train.py: 79] Epoch 12, iter 6400/6416, lr 0.010000, loss 1.989047
+INFO 2020-12-03 14:09:00 train.py: 97] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-12-03 14:09:02 train.py: 79] Epoch 13, iter 0/6416, lr 0.001000, loss 1.978062
+INFO 2020-12-03 14:12:09 train.py: 79] Epoch 13, iter 200/6416, lr 0.001000, loss 1.652931
+INFO 2020-12-03 14:15:16 train.py: 79] Epoch 13, iter 400/6416, lr 0.001000, loss 1.631869
+INFO 2020-12-03 14:18:23 train.py: 79] Epoch 13, iter 600/6416, lr 0.001000, loss 1.645035
+INFO 2020-12-03 14:21:30 train.py: 79] Epoch 13, iter 800/6416, lr 0.001000, loss 1.620876
+INFO 2020-12-03 14:24:37 train.py: 79] Epoch 13, iter 1000/6416, lr 0.001000, loss 1.602443
+INFO 2020-12-03 14:27:44 train.py: 79] Epoch 13, iter 1200/6416, lr 0.001000, loss 1.621005
+INFO 2020-12-03 14:30:51 train.py: 79] Epoch 13, iter 1400/6416, lr 0.001000, loss 1.589904
+INFO 2020-12-03 14:33:58 train.py: 79] Epoch 13, iter 1600/6416, lr 0.001000, loss 1.603842
+INFO 2020-12-03 14:37:04 train.py: 79] Epoch 13, iter 1800/6416, lr 0.001000, loss 1.589838
+INFO 2020-12-03 14:40:11 train.py: 79] Epoch 13, iter 2000/6416, lr 0.001000, loss 1.607866
+INFO 2020-12-03 14:43:18 train.py: 79] Epoch 13, iter 2200/6416, lr 0.001000, loss 1.591209
+INFO 2020-12-03 14:46:25 train.py: 79] Epoch 13, iter 2400/6416, lr 0.001000, loss 1.585208
+INFO 2020-12-03 14:49:32 train.py: 79] Epoch 13, iter 2600/6416, lr 0.001000, loss 1.606543
+INFO 2020-12-03 14:52:39 train.py: 79] Epoch 13, iter 2800/6416, lr 0.001000, loss 1.567105
+INFO 2020-12-03 14:55:45 train.py: 92] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-12-03 14:55:46 train.py: 79] Epoch 13, iter 3000/6416, lr 0.001000, loss 1.606321
+INFO 2020-12-03 14:58:53 train.py: 79] Epoch 13, iter 3200/6416, lr 0.001000, loss 1.589099
+INFO 2020-12-03 15:01:59 train.py: 79] Epoch 13, iter 3400/6416, lr 0.001000, loss 1.600815
+INFO 2020-12-03 15:05:05 train.py: 79] Epoch 13, iter 3600/6416, lr 0.001000, loss 1.606442
+INFO 2020-12-03 15:08:12 train.py: 79] Epoch 13, iter 3800/6416, lr 0.001000, loss 1.589249
+INFO 2020-12-03 15:11:18 train.py: 79] Epoch 13, iter 4000/6416, lr 0.001000, loss 1.589843
+INFO 2020-12-03 15:14:25 train.py: 79] Epoch 13, iter 4200/6416, lr 0.001000, loss 1.585513
+INFO 2020-12-03 15:17:31 train.py: 79] Epoch 13, iter 4400/6416, lr 0.001000, loss 1.573967
+INFO 2020-12-03 15:20:38 train.py: 79] Epoch 13, iter 4600/6416, lr 0.001000, loss 1.598693
+INFO 2020-12-03 15:23:44 train.py: 79] Epoch 13, iter 4800/6416, lr 0.001000, loss 1.588592
+INFO 2020-12-03 15:26:51 train.py: 79] Epoch 13, iter 5000/6416, lr 0.001000, loss 1.591285
+INFO 2020-12-03 15:29:57 train.py: 79] Epoch 13, iter 5200/6416, lr 0.001000, loss 1.609549
+INFO 2020-12-03 15:33:04 train.py: 79] Epoch 13, iter 5400/6416, lr 0.001000, loss 1.593667
+INFO 2020-12-03 15:36:11 train.py: 79] Epoch 13, iter 5600/6416, lr 0.001000, loss 1.592378
+INFO 2020-12-03 15:39:18 train.py: 79] Epoch 13, iter 5800/6416, lr 0.001000, loss 1.602734
+INFO 2020-12-03 15:42:24 train.py: 92] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-12-03 15:42:25 train.py: 79] Epoch 13, iter 6000/6416, lr 0.001000, loss 1.602092
+INFO 2020-12-03 15:45:32 train.py: 79] Epoch 13, iter 6200/6416, lr 0.001000, loss 1.602258
+INFO 2020-12-03 15:48:40 train.py: 79] Epoch 13, iter 6400/6416, lr 0.001000, loss 1.602071
+INFO 2020-12-03 15:48:54 train.py: 97] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-12-03 15:48:56 train.py: 79] Epoch 14, iter 0/6416, lr 0.001000, loss 1.627704
+INFO 2020-12-03 15:52:03 train.py: 79] Epoch 14, iter 200/6416, lr 0.001000, loss 1.550562
+INFO 2020-12-03 15:55:10 train.py: 79] Epoch 14, iter 400/6416, lr 0.001000, loss 1.546345
+INFO 2020-12-03 15:58:17 train.py: 79] Epoch 14, iter 600/6416, lr 0.001000, loss 1.544102
+INFO 2020-12-03 16:01:24 train.py: 79] Epoch 14, iter 800/6416, lr 0.001000, loss 1.545626
+INFO 2020-12-03 16:04:31 train.py: 79] Epoch 14, iter 1000/6416, lr 0.001000, loss 1.560231
+INFO 2020-12-03 16:07:38 train.py: 79] Epoch 14, iter 1200/6416, lr 0.001000, loss 1.546803
+INFO 2020-12-03 16:10:45 train.py: 79] Epoch 14, iter 1400/6416, lr 0.001000, loss 1.558189
+INFO 2020-12-03 16:13:51 train.py: 79] Epoch 14, iter 1600/6416, lr 0.001000, loss 1.550768
+INFO 2020-12-03 16:16:58 train.py: 79] Epoch 14, iter 1800/6416, lr 0.001000, loss 1.547268
+INFO 2020-12-03 16:20:05 train.py: 79] Epoch 14, iter 2000/6416, lr 0.001000, loss 1.545094
+INFO 2020-12-03 16:23:12 train.py: 79] Epoch 14, iter 2200/6416, lr 0.001000, loss 1.555121
+INFO 2020-12-03 16:26:19 train.py: 79] Epoch 14, iter 2400/6416, lr 0.001000, loss 1.552616
+INFO 2020-12-03 16:29:26 train.py: 79] Epoch 14, iter 2600/6416, lr 0.001000, loss 1.560406
+INFO 2020-12-03 16:32:33 train.py: 79] Epoch 14, iter 2800/6416, lr 0.001000, loss 1.553938
+INFO 2020-12-03 16:35:39 train.py: 92] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-12-03 16:35:40 train.py: 79] Epoch 14, iter 3000/6416, lr 0.001000, loss 1.533452
+INFO 2020-12-03 16:38:47 train.py: 79] Epoch 14, iter 3200/6416, lr 0.001000, loss 1.552535
+INFO 2020-12-03 16:41:54 train.py: 79] Epoch 14, iter 3400/6416, lr 0.001000, loss 1.558537
+INFO 2020-12-03 16:45:01 train.py: 79] Epoch 14, iter 3600/6416, lr 0.001000, loss 1.553029
+INFO 2020-12-03 16:48:08 train.py: 79] Epoch 14, iter 3800/6416, lr 0.001000, loss 1.547475
+INFO 2020-12-03 16:51:15 train.py: 79] Epoch 14, iter 4000/6416, lr 0.001000, loss 1.553226
+INFO 2020-12-03 16:54:23 train.py: 79] Epoch 14, iter 4200/6416, lr 0.001000, loss 1.559306
+INFO 2020-12-03 16:57:30 train.py: 79] Epoch 14, iter 4400/6416, lr 0.001000, loss 1.568083
+INFO 2020-12-03 17:00:37 train.py: 79] Epoch 14, iter 4600/6416, lr 0.001000, loss 1.573283
+INFO 2020-12-03 17:03:44 train.py: 79] Epoch 14, iter 4800/6416, lr 0.001000, loss 1.564634
+INFO 2020-12-03 17:06:51 train.py: 79] Epoch 14, iter 5000/6416, lr 0.001000, loss 1.553898
+INFO 2020-12-03 17:09:59 train.py: 79] Epoch 14, iter 5200/6416, lr 0.001000, loss 1.559875
+INFO 2020-12-03 17:13:06 train.py: 79] Epoch 14, iter 5400/6416, lr 0.001000, loss 1.555709
+INFO 2020-12-03 17:16:13 train.py: 79] Epoch 14, iter 5600/6416, lr 0.001000, loss 1.556728
+INFO 2020-12-03 17:19:21 train.py: 79] Epoch 14, iter 5800/6416, lr 0.001000, loss 1.559506
+INFO 2020-12-03 17:22:28 train.py: 92] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-12-03 17:22:29 train.py: 79] Epoch 14, iter 6000/6416, lr 0.001000, loss 1.532555
+INFO 2020-12-03 17:25:36 train.py: 79] Epoch 14, iter 6200/6416, lr 0.001000, loss 1.559591
+INFO 2020-12-03 17:28:43 train.py: 79] Epoch 14, iter 6400/6416, lr 0.001000, loss 1.561863
+INFO 2020-12-03 17:28:58 train.py: 97] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-12-03 17:29:00 train.py: 79] Epoch 15, iter 0/6416, lr 0.001000, loss 1.585972
+INFO 2020-12-03 17:32:07 train.py: 79] Epoch 15, iter 200/6416, lr 0.001000, loss 1.504678
+INFO 2020-12-03 17:35:14 train.py: 79] Epoch 15, iter 400/6416, lr 0.001000, loss 1.511323
+INFO 2020-12-03 17:38:21 train.py: 79] Epoch 15, iter 600/6416, lr 0.001000, loss 1.500655
+INFO 2020-12-03 17:41:27 train.py: 79] Epoch 15, iter 800/6416, lr 0.001000, loss 1.511984
+INFO 2020-12-03 17:44:34 train.py: 79] Epoch 15, iter 1000/6416, lr 0.001000, loss 1.519356
+INFO 2020-12-03 17:47:41 train.py: 79] Epoch 15, iter 1200/6416, lr 0.001000, loss 1.518752
+INFO 2020-12-03 17:50:48 train.py: 79] Epoch 15, iter 1400/6416, lr 0.001000, loss 1.514031
+INFO 2020-12-03 17:53:55 train.py: 79] Epoch 15, iter 1600/6416, lr 0.001000, loss 1.516873
+INFO 2020-12-03 17:57:02 train.py: 79] Epoch 15, iter 1800/6416, lr 0.001000, loss 1.516810
+INFO 2020-12-03 18:00:09 train.py: 79] Epoch 15, iter 2000/6416, lr 0.001000, loss 1.520707
+INFO 2020-12-03 18:03:15 train.py: 79] Epoch 15, iter 2200/6416, lr 0.001000, loss 1.516459
+INFO 2020-12-03 18:06:22 train.py: 79] Epoch 15, iter 2400/6416, lr 0.001000, loss 1.508329
+INFO 2020-12-03 18:09:29 train.py: 79] Epoch 15, iter 2600/6416, lr 0.001000, loss 1.522004
+INFO 2020-12-03 18:12:36 train.py: 79] Epoch 15, iter 2800/6416, lr 0.001000, loss 1.524256
+INFO 2020-12-03 18:15:43 train.py: 92] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-12-03 18:15:44 train.py: 79] Epoch 15, iter 3000/6416, lr 0.001000, loss 1.528647
+INFO 2020-12-03 18:18:50 train.py: 79] Epoch 15, iter 3200/6416, lr 0.001000, loss 1.533105
+INFO 2020-12-03 18:21:56 train.py: 79] Epoch 15, iter 3400/6416, lr 0.001000, loss 1.531813
+INFO 2020-12-03 18:25:03 train.py: 79] Epoch 15, iter 3600/6416, lr 0.001000, loss 1.527862
+INFO 2020-12-03 18:28:09 train.py: 79] Epoch 15, iter 3800/6416, lr 0.001000, loss 1.521715
+INFO 2020-12-03 18:31:16 train.py: 79] Epoch 15, iter 4000/6416, lr 0.001000, loss 1.536820
+INFO 2020-12-03 18:34:22 train.py: 79] Epoch 15, iter 4200/6416, lr 0.001000, loss 1.528713
+INFO 2020-12-03 18:37:29 train.py: 79] Epoch 15, iter 4400/6416, lr 0.001000, loss 1.528955
+INFO 2020-12-03 18:40:36 train.py: 79] Epoch 15, iter 4600/6416, lr 0.001000, loss 1.520924
+INFO 2020-12-03 18:43:42 train.py: 79] Epoch 15, iter 4800/6416, lr 0.001000, loss 1.531768
+INFO 2020-12-03 18:46:49 train.py: 79] Epoch 15, iter 5000/6416, lr 0.001000, loss 1.539711
+INFO 2020-12-03 18:49:56 train.py: 79] Epoch 15, iter 5200/6416, lr 0.001000, loss 1.533245
+INFO 2020-12-03 18:53:02 train.py: 79] Epoch 15, iter 5400/6416, lr 0.001000, loss 1.537295
+INFO 2020-12-03 18:56:09 train.py: 79] Epoch 15, iter 5600/6416, lr 0.001000, loss 1.528401
+INFO 2020-12-03 18:59:16 train.py: 79] Epoch 15, iter 5800/6416, lr 0.001000, loss 1.543341
+INFO 2020-12-03 19:02:22 train.py: 92] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-12-03 19:02:23 train.py: 79] Epoch 15, iter 6000/6416, lr 0.001000, loss 1.547450
+INFO 2020-12-03 19:05:30 train.py: 79] Epoch 15, iter 6200/6416, lr 0.001000, loss 1.565476
+INFO 2020-12-03 19:08:38 train.py: 79] Epoch 15, iter 6400/6416, lr 0.001000, loss 1.542237
+INFO 2020-12-03 19:08:52 train.py: 97] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-12-03 19:08:54 train.py: 79] Epoch 16, iter 0/6416, lr 0.000100, loss 1.563744
+INFO 2020-12-03 19:12:01 train.py: 79] Epoch 16, iter 200/6416, lr 0.000100, loss 1.482822
+INFO 2020-12-03 19:15:08 train.py: 79] Epoch 16, iter 400/6416, lr 0.000100, loss 1.488903
+INFO 2020-12-03 19:18:15 train.py: 79] Epoch 16, iter 600/6416, lr 0.000100, loss 1.476458
+INFO 2020-12-03 19:21:22 train.py: 79] Epoch 16, iter 800/6416, lr 0.000100, loss 1.488819
+INFO 2020-12-03 19:24:29 train.py: 79] Epoch 16, iter 1000/6416, lr 0.000100, loss 1.493010
+INFO 2020-12-03 19:27:36 train.py: 79] Epoch 16, iter 1200/6416, lr 0.000100, loss 1.479292
+INFO 2020-12-03 19:30:42 train.py: 79] Epoch 16, iter 1400/6416, lr 0.000100, loss 1.487100
+INFO 2020-12-03 19:33:49 train.py: 79] Epoch 16, iter 1600/6416, lr 0.000100, loss 1.482532
+INFO 2020-12-03 19:36:56 train.py: 79] Epoch 16, iter 1800/6416, lr 0.000100, loss 1.482536
+INFO 2020-12-03 19:40:03 train.py: 79] Epoch 16, iter 2000/6416, lr 0.000100, loss 1.494960
+INFO 2020-12-03 19:43:10 train.py: 79] Epoch 16, iter 2200/6416, lr 0.000100, loss 1.490512
+INFO 2020-12-03 19:46:17 train.py: 79] Epoch 16, iter 2400/6416, lr 0.000100, loss 1.468600
+INFO 2020-12-03 19:49:24 train.py: 79] Epoch 16, iter 2600/6416, lr 0.000100, loss 1.488502
+INFO 2020-12-03 19:52:31 train.py: 79] Epoch 16, iter 2800/6416, lr 0.000100, loss 1.480113
+INFO 2020-12-03 19:55:37 train.py: 92] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-12-03 19:55:38 train.py: 79] Epoch 16, iter 3000/6416, lr 0.000100, loss 1.489426
+INFO 2020-12-03 19:58:45 train.py: 79] Epoch 16, iter 3200/6416, lr 0.000100, loss 1.484256
+INFO 2020-12-03 20:01:52 train.py: 79] Epoch 16, iter 3400/6416, lr 0.000100, loss 1.472291
+INFO 2020-12-03 20:04:59 train.py: 79] Epoch 16, iter 3600/6416, lr 0.000100, loss 1.482397
+INFO 2020-12-03 20:08:06 train.py: 79] Epoch 16, iter 3800/6416, lr 0.000100, loss 1.483921
+INFO 2020-12-03 20:11:13 train.py: 79] Epoch 16, iter 4000/6416, lr 0.000100, loss 1.485047
+INFO 2020-12-03 20:14:20 train.py: 79] Epoch 16, iter 4200/6416, lr 0.000100, loss 1.478618
+INFO 2020-12-03 20:17:27 train.py: 79] Epoch 16, iter 4400/6416, lr 0.000100, loss 1.475264
+INFO 2020-12-03 20:20:35 train.py: 79] Epoch 16, iter 4600/6416, lr 0.000100, loss 1.467394
+INFO 2020-12-03 20:23:42 train.py: 79] Epoch 16, iter 4800/6416, lr 0.000100, loss 1.483348
+INFO 2020-12-03 20:26:49 train.py: 79] Epoch 16, iter 5000/6416, lr 0.000100, loss 1.473250
+INFO 2020-12-03 20:29:56 train.py: 79] Epoch 16, iter 5200/6416, lr 0.000100, loss 1.479615
+INFO 2020-12-03 20:33:04 train.py: 79] Epoch 16, iter 5400/6416, lr 0.000100, loss 1.499258
+INFO 2020-12-03 20:36:11 train.py: 79] Epoch 16, iter 5600/6416, lr 0.000100, loss 1.486454
+INFO 2020-12-03 20:39:18 train.py: 79] Epoch 16, iter 5800/6416, lr 0.000100, loss 1.484143
+INFO 2020-12-03 20:42:25 train.py: 92] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-12-03 20:42:26 train.py: 79] Epoch 16, iter 6000/6416, lr 0.000100, loss 1.484250
+INFO 2020-12-03 20:45:33 train.py: 79] Epoch 16, iter 6200/6416, lr 0.000100, loss 1.485810
+INFO 2020-12-03 20:48:41 train.py: 79] Epoch 16, iter 6400/6416, lr 0.000100, loss 1.488005
+INFO 2020-12-03 20:48:55 train.py: 97] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-12-03 20:48:57 train.py: 79] Epoch 17, iter 0/6416, lr 0.000100, loss 1.527157
+INFO 2020-12-03 20:52:04 train.py: 79] Epoch 17, iter 200/6416, lr 0.000100, loss 1.465263
+INFO 2020-12-03 20:55:11 train.py: 79] Epoch 17, iter 400/6416, lr 0.000100, loss 1.473775
+INFO 2020-12-03 20:58:18 train.py: 79] Epoch 17, iter 600/6416, lr 0.000100, loss 1.486029
+INFO 2020-12-03 21:01:25 train.py: 79] Epoch 17, iter 800/6416, lr 0.000100, loss 1.479593
+INFO 2020-12-03 21:04:31 train.py: 79] Epoch 17, iter 1000/6416, lr 0.000100, loss 1.465834
+INFO 2020-12-03 21:07:38 train.py: 79] Epoch 17, iter 1200/6416, lr 0.000100, loss 1.476643
+INFO 2020-12-03 21:10:45 train.py: 79] Epoch 17, iter 1400/6416, lr 0.000100, loss 1.474023
+INFO 2020-12-03 21:13:52 train.py: 79] Epoch 17, iter 1600/6416, lr 0.000100, loss 1.475308
+INFO 2020-12-03 21:16:59 train.py: 79] Epoch 17, iter 1800/6416, lr 0.000100, loss 1.492811
+INFO 2020-12-03 21:20:06 train.py: 79] Epoch 17, iter 2000/6416, lr 0.000100, loss 1.484967
+INFO 2020-12-03 21:23:13 train.py: 79] Epoch 17, iter 2200/6416, lr 0.000100, loss 1.474637
+INFO 2020-12-03 21:26:19 train.py: 79] Epoch 17, iter 2400/6416, lr 0.000100, loss 1.474549
+INFO 2020-12-03 21:29:26 train.py: 79] Epoch 17, iter 2600/6416, lr 0.000100, loss 1.474134
+INFO 2020-12-03 21:32:33 train.py: 79] Epoch 17, iter 2800/6416, lr 0.000100, loss 1.480024
+INFO 2020-12-03 21:35:39 train.py: 92] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-12-03 21:35:40 train.py: 79] Epoch 17, iter 3000/6416, lr 0.000100, loss 1.482449
+INFO 2020-12-03 21:38:47 train.py: 79] Epoch 17, iter 3200/6416, lr 0.000100, loss 1.483054
+INFO 2020-12-03 21:41:53 train.py: 79] Epoch 17, iter 3400/6416, lr 0.000100, loss 1.478087
+INFO 2020-12-03 21:44:59 train.py: 79] Epoch 17, iter 3600/6416, lr 0.000100, loss 1.483813
+INFO 2020-12-03 21:48:06 train.py: 79] Epoch 17, iter 3800/6416, lr 0.000100, loss 1.489746
+INFO 2020-12-03 21:51:12 train.py: 79] Epoch 17, iter 4000/6416, lr 0.000100, loss 1.497122
+INFO 2020-12-03 21:54:19 train.py: 79] Epoch 17, iter 4200/6416, lr 0.000100, loss 1.471576
+INFO 2020-12-03 21:57:25 train.py: 79] Epoch 17, iter 4400/6416, lr 0.000100, loss 1.495841
+INFO 2020-12-03 22:00:32 train.py: 79] Epoch 17, iter 4600/6416, lr 0.000100, loss 1.474587
+INFO 2020-12-03 22:03:39 train.py: 79] Epoch 17, iter 4800/6416, lr 0.000100, loss 1.471644
+INFO 2020-12-03 22:06:45 train.py: 79] Epoch 17, iter 5000/6416, lr 0.000100, loss 1.479576
+INFO 2020-12-03 22:09:52 train.py: 79] Epoch 17, iter 5200/6416, lr 0.000100, loss 1.481074
+INFO 2020-12-03 22:12:59 train.py: 79] Epoch 17, iter 5400/6416, lr 0.000100, loss 1.467576
+INFO 2020-12-03 22:16:06 train.py: 79] Epoch 17, iter 5600/6416, lr 0.000100, loss 1.485052
+INFO 2020-12-03 22:19:12 train.py: 79] Epoch 17, iter 5800/6416, lr 0.000100, loss 1.473696
+INFO 2020-12-03 22:22:19 train.py: 92] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-12-03 22:22:19 train.py: 79] Epoch 17, iter 6000/6416, lr 0.000100, loss 1.480181
+INFO 2020-12-03 22:25:27 train.py: 79] Epoch 17, iter 6200/6416, lr 0.000100, loss 1.472788
+INFO 2020-12-03 22:28:34 train.py: 79] Epoch 17, iter 6400/6416, lr 0.000100, loss 1.485880
+INFO 2020-12-03 22:28:48 train.py: 97] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-12-03 22:28:49 train.py: 180] Optimization done!
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_S/.gitkeep b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f176c5f8baf9291b6e46f8757a4ba4b8d92e78d4
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9805000000000001 | 0.0021808651584878315 |
+| Epoch_16.pt | 0.9804999999999999 | 0.002277777777777769 |
+| Epoch_15_batch_5999.pt | 0.9801666666666667 | 0.0021865187393504864 |
+| Epoch_16_batch_5999.pt | 0.9801666666666667 | 0.0021147629234082475 |
+| Epoch_14.pt | 0.9801666666666666 | 0.001963399671896915 |
+| Epoch_17.pt | 0.9800000000000001 | 0.0021801574300387306 |
+| Epoch_15.pt | 0.9800000000000001 | 0.002222222222222217 |
+| Epoch_13.pt | 0.9800000000000001 | 0.001940472132952549 |
+| Epoch_12_batch_5999.pt | 0.9798333333333333 | 0.001994591452335138 |
+| Epoch_14_batch_2999.pt | 0.9798333333333333 | 0.0019633996718969207 |
+| Epoch_11_batch_2999.pt | 0.9798333333333333 | 0.0021147629234082466 |
+| Epoch_17_batch_2999.pt | 0.9796666666666669 | 0.0020757268546965943 |
+| Epoch_17_batch_5999.pt | 0.9795 | 0.002208988372452335 |
+| Epoch_13_batch_5999.pt | 0.9793333333333335 | 0.0021545243810739217 |
+| Epoch_15_batch_2999.pt | 0.9793333333333335 | 0.0020517983680688142 |
+| Epoch_14_batch_5999.pt | 0.9793333333333335 | 0.0020964402515681285 |
+| Epoch_11.pt | 0.9785 | 0.002129307544081864 |
+| Epoch_13_batch_2999.pt | 0.9783333333333335 | 0.0021942686286812708 |
+| Epoch_12_batch_2999.pt | 0.9783333333333333 | 0.0020786985482077404 |
+| Epoch_10_batch_2999.pt | 0.9783333333333333 | 0.0022082896571501928 |
+| Epoch_11_batch_5999.pt | 0.9778333333333334 | 0.0022641870969238704 |
+| Epoch_12.pt | 0.9776666666666667 | 0.0022249982660556373 |
+| Epoch_10.pt | 0.9776666666666667 | 0.0016329931618554526 |
+| Epoch_9_batch_5999.pt | 0.977 | 0.002444444444444445 |
+| Epoch_10_batch_5999.pt | 0.9768333333333334 | 0.0023629078131262994 |
+| Epoch_8_batch_2999.pt | 0.9765 | 0.002599263903397687 |
+| Epoch_9_batch_2999.pt | 0.9758333333333333 | 0.0020824072015545913 |
+| Epoch_8_batch_5999.pt | 0.9756666666666668 | 0.0022388268532899875 |
+| Epoch_6_batch_2999.pt | 0.9756666666666666 | 0.00260815435429011 |
+| Epoch_6.pt | 0.9754999999999999 | 0.0022367580154663757 |
+| Epoch_7_batch_5999.pt | 0.9753333333333334 | 0.002177324215807262 |
+| Epoch_6_batch_5999.pt | 0.9751666666666667 | 0.0028485430540653896 |
+| Epoch_9.pt | 0.975 | 0.0020184335693983237 |
+| Epoch_8.pt | 0.9745000000000001 | 0.001809610830544702 |
+| Epoch_7_batch_2999.pt | 0.9743333333333334 | 0.0019751543149590166 |
+| Epoch_5.pt | 0.9743333333333334 | 0.002140151142695358 |
+| Epoch_7.pt | 0.9743333333333333 | 0.0018291197370171465 |
+| Epoch_4_batch_5999.pt | 0.9731666666666665 | 0.0024651897480370928 |
+| Epoch_5_batch_2999.pt | 0.9730000000000001 | 0.0029165343885348173 |
+| Epoch_5_batch_5999.pt | 0.9726666666666667 | 0.0019751543149590157 |
+| Epoch_4.pt | 0.9725000000000001 | 0.0025968879761156085 |
+| Epoch_4_batch_2999.pt | 0.9698333333333334 | 0.0027492984514796872 |
+| Epoch_3.pt | 0.9678333333333334 | 0.001572330188676106 |
+| Epoch_3_batch_5999.pt | 0.966 | 0.0030347778408328146 |
+| Epoch_3_batch_2999.pt | 0.9628333333333334 | 0.0019253026056848255 |
+| Epoch_2_batch_5999.pt | 0.9580000000000002 | 0.003189488909868296 |
+| Epoch_2.pt | 0.9555 | 0.002844205719148935 |
+| Epoch_2_batch_2999.pt | 0.9436666666666668 | 0.0038634085890474888 |
+| Epoch_1.pt | 0.9148333333333335 | 0.005629781918018179 |
+| Epoch_1_batch_5999.pt | 0.8988333333333335 | 0.005741606864172785 |
+| Epoch_1_batch_2999.pt | 0.8113333333333334 | 0.006679617051195168 |
+| Epoch_0.pt | 0.6483333333333334 | 0.005246986201385589 |
+| Epoch_0_batch_5999.pt | 0.5900000000000001 | 0.005488484015657096 |
+| Epoch_0_batch_2999.pt | 0.4993333333333333 | 0.00563389214848321 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..cebba977d786b930af86e6f3f310995ed43cfdd0
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+---------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+---------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9591666666666667 | 0.004069231128273333 |
+| Epoch_16.pt | 0.9588333333333333 | 0.0038733817807896373 |
+| Epoch_17.pt | 0.9586666666666666 | 0.003942487777602259 |
+| Epoch_16_batch_5999.pt | 0.9585000000000001 | 0.004017323597731317 |
+| Epoch_14_batch_5999.pt | 0.9585000000000001 | 0.003963179294891511 |
+| Epoch_15.pt | 0.9583333333333334 | 0.003840974670543025 |
+| Epoch_13_batch_2999.pt | 0.9581666666666668 | 0.004130958098893623 |
+| Epoch_13_batch_5999.pt | 0.9581666666666667 | 0.003931904950937899 |
+| Epoch_17_batch_2999.pt | 0.958 | 0.0039110479792884515 |
+| Epoch_17_batch_5999.pt | 0.958 | 0.004050605807457375 |
+| Epoch_14.pt | 0.9578333333333331 | 0.004157768276015038 |
+| Epoch_14_batch_2999.pt | 0.9576666666666668 | 0.004061259307221574 |
+| Epoch_12.pt | 0.9576666666666667 | 0.004166296279833932 |
+| Epoch_13.pt | 0.9576666666666667 | 0.0038904758666692433 |
+| Epoch_12_batch_5999.pt | 0.9576666666666667 | 0.004061259307221576 |
+| Epoch_11_batch_5999.pt | 0.9571666666666667 | 0.004097951912424451 |
+| Epoch_11_batch_2999.pt | 0.9570000000000001 | 0.004058218303944377 |
+| Epoch_15_batch_2999.pt | 0.9568333333333335 | 0.00372222222222222 |
+| Epoch_15_batch_5999.pt | 0.9568333333333333 | 0.004070747800382751 |
+| Epoch_12_batch_2999.pt | 0.9566666666666667 | 0.0038570122128243995 |
+| Epoch_10_batch_5999.pt | 0.9563333333333333 | 0.004351812450570253 |
+| Epoch_10_batch_2999.pt | 0.9561666666666667 | 0.0038972133127323535 |
+| Epoch_11.pt | 0.9560000000000001 | 0.004217833976911622 |
+| Epoch_9_batch_2999.pt | 0.9556666666666669 | 0.003670032125536384 |
+| Epoch_10.pt | 0.9550000000000001 | 0.004067334492827362 |
+| Epoch_8.pt | 0.9543333333333335 | 0.004268749491621904 |
+| Epoch_7.pt | 0.9541666666666668 | 0.0036955929710139074 |
+| Epoch_8_batch_5999.pt | 0.9540000000000001 | 0.0037777777777777822 |
+| Epoch_9_batch_5999.pt | 0.954 | 0.003976784481816277 |
+| Epoch_6_batch_2999.pt | 0.9538333333333334 | 0.0036603480911175826 |
+| Epoch_9.pt | 0.9536666666666667 | 0.004027681991198194 |
+| Epoch_8_batch_2999.pt | 0.9531666666666668 | 0.003804237403504441 |
+| Epoch_7_batch_2999.pt | 0.9530000000000001 | 0.0034942633763369742 |
+| Epoch_6.pt | 0.9526666666666668 | 0.0042251451870736765 |
+| Epoch_7_batch_5999.pt | 0.9525 | 0.003914597562211731 |
+| Epoch_6_batch_5999.pt | 0.9521666666666666 | 0.00401424931100003 |
+| Epoch_5_batch_2999.pt | 0.9513333333333334 | 0.0038952329206573496 |
+| Epoch_5_batch_5999.pt | 0.9508333333333333 | 0.003435921354681382 |
+| Epoch_4_batch_5999.pt | 0.9506666666666665 | 0.004076430295076477 |
+| Epoch_4.pt | 0.9504999999999999 | 0.004067713890663442 |
+| Epoch_5.pt | 0.9501666666666667 | 0.004024999041484438 |
+| Epoch_4_batch_2999.pt | 0.9476666666666667 | 0.004361730316975678 |
+| Epoch_3.pt | 0.9465 | 0.004175546094247453 |
+| Epoch_3_batch_5999.pt | 0.9455 | 0.0038171963080670854 |
+| Epoch_3_batch_2999.pt | 0.9418333333333333 | 0.005045778090959789 |
+| Epoch_2_batch_5999.pt | 0.9349999999999999 | 0.0035832256659104667 |
+| Epoch_2.pt | 0.9343333333333332 | 0.0040612593072215765 |
+| Epoch_2_batch_2999.pt | 0.9241666666666667 | 0.0037288498210110775 |
+| Epoch_1.pt | 0.9086666666666667 | 0.0029896942326830527 |
+| Epoch_1_batch_5999.pt | 0.8951666666666667 | 0.004078322700496203 |
+| Epoch_1_batch_2999.pt | 0.8278333333333334 | 0.004150338376804396 |
+| Epoch_0.pt | 0.6188333333333335 | 0.004060119197909653 |
+| Epoch_0_batch_5999.pt | 0.5845 | 0.005369380496256557 |
+| Epoch_0_batch_2999.pt | 0.49800000000000005 | 0.002730712383876561 |
++------------------------+---------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..5f7864bd4f671e0a58710d35d0fc2b1298272a36
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9003333333333334 | 0.005350666279037561 |
+| Epoch_17.pt | 0.9001666666666667 | 0.005278947238855234 |
+| Epoch_15.pt | 0.9001666666666667 | 0.005172636932890826 |
+| Epoch_15_batch_5999.pt | 0.8998333333333333 | 0.005411749905255888 |
+| Epoch_16_batch_5999.pt | 0.8995000000000001 | 0.00520001187083165 |
+| Epoch_15_batch_2999.pt | 0.8993333333333334 | 0.005248162523830909 |
+| Epoch_17_batch_5999.pt | 0.8983333333333334 | 0.00512196914294049 |
+| Epoch_17_batch_2999.pt | 0.898 | 0.00530431922768 |
+| Epoch_16.pt | 0.898 | 0.005114733016215065 |
+| Epoch_14_batch_2999.pt | 0.8979999999999999 | 0.004809969071536302 |
+| Epoch_14.pt | 0.8976666666666668 | 0.004818944098266985 |
+| Epoch_12.pt | 0.8959999999999999 | 0.005628411174183975 |
+| Epoch_14_batch_5999.pt | 0.8959999999999999 | 0.004669311419877936 |
+| Epoch_11.pt | 0.8948333333333333 | 0.0049966037848438605 |
+| Epoch_11_batch_5999.pt | 0.8946666666666667 | 0.0056152350082192305 |
+| Epoch_12_batch_2999.pt | 0.8941666666666664 | 0.005068968775248516 |
+| Epoch_13_batch_2999.pt | 0.8939999999999999 | 0.005105068892293394 |
+| Epoch_13.pt | 0.8933333333333333 | 0.005660126945311127 |
+| Epoch_13_batch_5999.pt | 0.8928333333333335 | 0.005522401070209128 |
+| Epoch_12_batch_5999.pt | 0.8918333333333333 | 0.006218500871304899 |
+| Epoch_10.pt | 0.8916666666666666 | 0.005522121617684078 |
+| Epoch_10_batch_5999.pt | 0.8901666666666666 | 0.00510658011724355 |
+| Epoch_10_batch_2999.pt | 0.8896666666666666 | 0.005951449662596485 |
+| Epoch_9_batch_5999.pt | 0.8891666666666668 | 0.00552910369357647 |
+| Epoch_9_batch_2999.pt | 0.8891666666666668 | 0.004709132183951299 |
+| Epoch_11_batch_2999.pt | 0.889 | 0.004741464065189305 |
+| Epoch_9.pt | 0.8885 | 0.005348646996330758 |
+| Epoch_8_batch_2999.pt | 0.8883333333333333 | 0.005414885747116089 |
+| Epoch_8.pt | 0.8876666666666665 | 0.0053644923131436935 |
+| Epoch_7.pt | 0.8861666666666667 | 0.005170249653689989 |
+| Epoch_7_batch_2999.pt | 0.8859999999999999 | 0.005500841686438479 |
+| Epoch_8_batch_5999.pt | 0.884 | 0.005183068350973608 |
+| Epoch_5.pt | 0.8833333333333334 | 0.0044025806124797645 |
+| Epoch_6_batch_5999.pt | 0.8826666666666666 | 0.005289168993516405 |
+| Epoch_7_batch_5999.pt | 0.8815 | 0.00547976074092453 |
+| Epoch_6_batch_2999.pt | 0.8785000000000001 | 0.005296458168984485 |
+| Epoch_5_batch_2999.pt | 0.8768333333333335 | 0.005635261559103196 |
+| Epoch_6.pt | 0.8761666666666666 | 0.005730845713556826 |
+| Epoch_5_batch_5999.pt | 0.8753333333333334 | 0.0049553562491061725 |
+| Epoch_4_batch_5999.pt | 0.8729999999999999 | 0.005487359211051443 |
+| Epoch_4.pt | 0.8718333333333332 | 0.005354414355129774 |
+| Epoch_4_batch_2999.pt | 0.8671666666666666 | 0.0056712944066512365 |
+| Epoch_3_batch_5999.pt | 0.867 | 0.003934651379916842 |
+| Epoch_3.pt | 0.8651666666666668 | 0.0052201532544552676 |
+| Epoch_3_batch_2999.pt | 0.8563333333333333 | 0.004525647078759182 |
+| Epoch_2.pt | 0.8483333333333333 | 0.004987639041658537 |
+| Epoch_2_batch_5999.pt | 0.8468333333333333 | 0.005231964907195335 |
+| Epoch_2_batch_2999.pt | 0.8215 | 0.006316986701557505 |
+| Epoch_1.pt | 0.7905 | 0.00675428533379886 |
+| Epoch_1_batch_5999.pt | 0.7828333333333333 | 0.006713033205251915 |
+| Epoch_1_batch_2999.pt | 0.6838333333333335 | 0.010018655438237383 |
+| Epoch_0_batch_5999.pt | 0.5638333333333333 | 0.006952439841745232 |
+| Epoch_0.pt | 0.5498333333333333 | 0.006778916115700475 |
+| Epoch_0_batch_2999.pt | 0.5228333333333334 | 0.00609112894723563 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..59942c888f1705f9cb5476f8061938bddfc0e1c7
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_2999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_15_batch_5999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_16.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_14.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_17.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_15.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_8_batch_2999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_17_batch_5999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_9_batch_2999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_15_batch_2999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_16_batch_2999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_16_batch_5999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_14_batch_5999.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_7_batch_2999.pt | 0.9981666666666668 | 0.0005241100628920312 |
+| Epoch_7.pt | 0.9981666666666668 | 0.0005241100628920312 |
+| Epoch_13_batch_2999.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_14_batch_2999.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_12_batch_2999.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_10_batch_2999.pt | 0.9981666666666665 | 0.0006309898162000297 |
+| Epoch_9_batch_5999.pt | 0.998 | 0.0005983516452371659 |
+| Epoch_13.pt | 0.998 | 0.0005983516452371659 |
+| Epoch_9.pt | 0.998 | 0.000544331053951814 |
+| Epoch_8_batch_5999.pt | 0.998 | 0.0005983516452371637 |
+| Epoch_10.pt | 0.9978333333333333 | 0.000611111111111109 |
+| Epoch_11.pt | 0.9978333333333333 | 0.000611111111111109 |
+| Epoch_6.pt | 0.9978333333333333 | 0.0005583264233956013 |
+| Epoch_12_batch_5999.pt | 0.9978333333333333 | 0.0007049209744694192 |
+| Epoch_11_batch_2999.pt | 0.9978333333333333 | 0.0007049209744694192 |
+| Epoch_12.pt | 0.9976666666666667 | 0.0007934920476158739 |
+| Epoch_5_batch_5999.pt | 0.9976666666666667 | 0.000566557723732528 |
+| Epoch_10_batch_5999.pt | 0.9976666666666667 | 0.000618640484758889 |
+| Epoch_11_batch_5999.pt | 0.9976666666666667 | 0.0005665577237325287 |
+| Epoch_13_batch_5999.pt | 0.9976666666666667 | 0.0006666666666666672 |
+| Epoch_5_batch_2999.pt | 0.9976666666666667 | 0.0006666666666666653 |
+| Epoch_5.pt | 0.9974999999999999 | 0.0005692750425533078 |
+| Epoch_6_batch_5999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_8.pt | 0.9974999999999999 | 0.0007556372504853035 |
+| Epoch_6_batch_2999.pt | 0.9973333333333333 | 0.00071145824860365 |
+| Epoch_4.pt | 0.9971666666666665 | 0.000611111111111109 |
+| Epoch_7_batch_5999.pt | 0.9971666666666665 | 0.0007049209744694181 |
+| Epoch_4_batch_5999.pt | 0.9968333333333333 | 0.0005800170282728054 |
+| Epoch_4_batch_2999.pt | 0.9964999999999999 | 0.0006309898162000309 |
+| Epoch_3_batch_5999.pt | 0.9963333333333333 | 0.0006478835438717003 |
+| Epoch_2.pt | 0.9961666666666666 | 0.0008624541497922222 |
+| Epoch_3.pt | 0.9961666666666666 | 0.0007876359377087665 |
+| Epoch_2_batch_2999.pt | 0.9953333333333333 | 0.0008534606386520708 |
+| Epoch_2_batch_5999.pt | 0.9951666666666666 | 0.0006781419786518758 |
+| Epoch_3_batch_2999.pt | 0.9946666666666666 | 0.0008888888888888913 |
+| Epoch_1.pt | 0.9904999999999999 | 0.0014283289035758226 |
+| Epoch_1_batch_5999.pt | 0.9891666666666665 | 0.0017959144305827281 |
+| Epoch_1_batch_2999.pt | 0.9663333333333334 | 0.002380476142847616 |
+| Epoch_0_batch_5999.pt | 0.8395000000000001 | 0.008165154810927945 |
+| Epoch_0.pt | 0.8291666666666666 | 0.0054390562906935745 |
+| Epoch_0_batch_2999.pt | 0.5998333333333333 | 0.005662580225905394 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0fea3884dc57d07850f2907c41964049f90d7b2d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.981722795735319 |
+| 2 | 0.9867347071611753 |
+| 3 | 0.9885897652863298 |
+| 4 | 0.9893448064881472 |
+| 5 | 0.9900868297382089 |
+| 6 | 0.9906010388325501 |
+| 7 | 0.9910306312404806 |
+| 8 | 0.9913821159378783 |
+| 9 | 0.9916294570212323 |
+| 10 | 0.9918051993699312 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_S/log.log b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..ee99137bc508f0e95123d2925677e2f1ac20b19a
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_S/log.log
@@ -0,0 +1,887 @@
+Use GPU: 0 for training
+Use GPU: 1 for training
+Use GPU: 3 for training
+Use GPU: 2 for training
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 18, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.3}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 18, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.3}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 18, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.3}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 18, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.3}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+Selected optimization level O1: Insert automatic casts around Pytorch functions and Tensor methods.
+
+Defaults for this optimization level are:
+enabled : True
+opt_level : O1
+cast_model_type : None
+patch_torch_functions : True
+keep_batchnorm_fp32 : None
+master_weights : None
+loss_scale : dynamic
+Processing user overrides (additional kwargs that are not None)...
+After processing overrides, optimization options are:
+enabled : True
+opt_level : O1
+cast_model_type : None
+patch_torch_functions : True
+keep_batchnorm_fp32 : None
+master_weights : None
+loss_scale : dynamic
+INFO 2021-11-05 10:23:21 train.py: 88] Epoch 0, iter 0/6416, lr 0.000000, loss 16.274643
+INFO 2021-11-05 10:23:21 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-11-05 10:23:21 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-11-05 10:23:21 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-11-05 10:23:21 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-11-05 10:30:28 train.py: 88] Epoch 0, iter 200/6416, lr 0.000016, loss 16.222235
+INFO 2021-11-05 10:37:33 train.py: 88] Epoch 0, iter 400/6416, lr 0.000032, loss 15.818622
+INFO 2021-11-05 10:44:36 train.py: 88] Epoch 0, iter 600/6416, lr 0.000047, loss 15.370561
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-05 10:51:41 train.py: 88] Epoch 0, iter 800/6416, lr 0.000063, loss 15.145919
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+INFO 2021-11-05 10:58:50 train.py: 88] Epoch 0, iter 1000/6416, lr 0.000078, loss 15.074239
+INFO 2021-11-05 11:06:04 train.py: 88] Epoch 0, iter 1200/6416, lr 0.000094, loss 15.021375
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 2048.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 2048.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 2048.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 2048.0
+INFO 2021-11-05 11:13:19 train.py: 88] Epoch 0, iter 1400/6416, lr 0.000109, loss 15.002815
+INFO 2021-11-05 11:20:25 train.py: 88] Epoch 0, iter 1600/6416, lr 0.000125, loss 15.003175
+INFO 2021-11-05 11:27:27 train.py: 88] Epoch 0, iter 1800/6416, lr 0.000141, loss 15.014103
+INFO 2021-11-05 11:34:26 train.py: 88] Epoch 0, iter 2000/6416, lr 0.000156, loss 15.034289
+INFO 2021-11-05 11:41:31 train.py: 88] Epoch 0, iter 2200/6416, lr 0.000172, loss 15.048180
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 1024.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 1024.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 1024.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 1024.0
+INFO 2021-11-05 11:48:38 train.py: 88] Epoch 0, iter 2400/6416, lr 0.000187, loss 15.052229
+INFO 2021-11-05 11:55:50 train.py: 88] Epoch 0, iter 2600/6416, lr 0.000203, loss 15.064821
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 512.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 512.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 512.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 512.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 256.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 256.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 256.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 256.0
+INFO 2021-11-05 12:02:55 train.py: 88] Epoch 0, iter 2800/6416, lr 0.000218, loss 15.392186
+INFO 2021-11-05 12:10:10 train.py: 101] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-11-05 12:10:12 train.py: 88] Epoch 0, iter 3000/6416, lr 0.000234, loss 15.513644
+INFO 2021-11-05 12:17:18 train.py: 88] Epoch 0, iter 3200/6416, lr 0.000250, loss 15.508453
+INFO 2021-11-05 12:24:24 train.py: 88] Epoch 0, iter 3400/6416, lr 0.000265, loss 15.534392
+INFO 2021-11-05 12:31:33 train.py: 88] Epoch 0, iter 3600/6416, lr 0.000281, loss 15.556318
+INFO 2021-11-05 12:38:32 train.py: 88] Epoch 0, iter 3800/6416, lr 0.000296, loss 15.588835
+INFO 2021-11-05 12:45:37 train.py: 88] Epoch 0, iter 4000/6416, lr 0.000312, loss 15.578653
+INFO 2021-11-05 12:52:43 train.py: 88] Epoch 0, iter 4200/6416, lr 0.000327, loss 15.604295
+INFO 2021-11-05 12:59:52 train.py: 88] Epoch 0, iter 4400/6416, lr 0.000343, loss 15.616197
+INFO 2021-11-05 13:06:54 train.py: 88] Epoch 0, iter 4600/6416, lr 0.000359, loss 15.523668
+INFO 2021-11-05 13:14:02 train.py: 88] Epoch 0, iter 4800/6416, lr 0.000374, loss 15.387649
+INFO 2021-11-05 13:21:08 train.py: 88] Epoch 0, iter 5000/6416, lr 0.000390, loss 15.169683
+INFO 2021-11-05 13:28:20 train.py: 88] Epoch 0, iter 5200/6416, lr 0.000405, loss 14.932817
+INFO 2021-11-05 13:35:32 train.py: 88] Epoch 0, iter 5400/6416, lr 0.000421, loss 14.694200
+INFO 2021-11-05 13:42:45 train.py: 88] Epoch 0, iter 5600/6416, lr 0.000436, loss 14.368824
+INFO 2021-11-05 13:49:57 train.py: 88] Epoch 0, iter 5800/6416, lr 0.000452, loss 14.007963
+INFO 2021-11-05 13:57:00 train.py: 101] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-11-05 13:57:02 train.py: 88] Epoch 0, iter 6000/6416, lr 0.000468, loss 13.635418
+INFO 2021-11-05 14:04:04 train.py: 88] Epoch 0, iter 6200/6416, lr 0.000483, loss 13.300804
+INFO 2021-11-05 14:11:14 train.py: 88] Epoch 0, iter 6400/6416, lr 0.000499, loss 12.910882
+INFO 2021-11-05 14:11:51 train.py: 108] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-11-05 14:11:53 train.py: 88] Epoch 1, iter 0/6416, lr 0.000496, loss 12.714586
+INFO 2021-11-05 14:19:00 train.py: 88] Epoch 1, iter 200/6416, lr 0.000496, loss 12.648638
+INFO 2021-11-05 14:26:07 train.py: 88] Epoch 1, iter 400/6416, lr 0.000496, loss 12.275591
+INFO 2021-11-05 14:33:12 train.py: 88] Epoch 1, iter 600/6416, lr 0.000496, loss 12.020270
+INFO 2021-11-05 14:40:22 train.py: 88] Epoch 1, iter 800/6416, lr 0.000495, loss 11.813366
+INFO 2021-11-05 14:47:25 train.py: 88] Epoch 1, iter 1000/6416, lr 0.000495, loss 11.774437
+INFO 2021-11-05 14:54:34 train.py: 88] Epoch 1, iter 1200/6416, lr 0.000495, loss 11.780509
+INFO 2021-11-05 15:01:44 train.py: 88] Epoch 1, iter 1400/6416, lr 0.000494, loss 11.809932
+INFO 2021-11-05 15:08:50 train.py: 88] Epoch 1, iter 1600/6416, lr 0.000494, loss 12.056637
+INFO 2021-11-05 15:15:59 train.py: 88] Epoch 1, iter 1800/6416, lr 0.000494, loss 12.179423
+INFO 2021-11-05 15:23:04 train.py: 88] Epoch 1, iter 2000/6416, lr 0.000494, loss 12.385912
+INFO 2021-11-05 15:30:06 train.py: 88] Epoch 1, iter 2200/6416, lr 0.000493, loss 12.678529
+INFO 2021-11-05 15:37:16 train.py: 88] Epoch 1, iter 2400/6416, lr 0.000493, loss 12.902084
+INFO 2021-11-05 15:44:21 train.py: 88] Epoch 1, iter 2600/6416, lr 0.000493, loss 13.031389
+INFO 2021-11-05 15:51:23 train.py: 88] Epoch 1, iter 2800/6416, lr 0.000492, loss 13.233134
+INFO 2021-11-05 15:58:31 train.py: 101] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-11-05 15:58:33 train.py: 88] Epoch 1, iter 3000/6416, lr 0.000492, loss 13.376003
+INFO 2021-11-05 16:05:41 train.py: 88] Epoch 1, iter 3200/6416, lr 0.000492, loss 13.400894
+INFO 2021-11-05 16:12:50 train.py: 88] Epoch 1, iter 3400/6416, lr 0.000491, loss 13.399327
+INFO 2021-11-05 16:19:53 train.py: 88] Epoch 1, iter 3600/6416, lr 0.000491, loss 13.403037
+INFO 2021-11-05 16:26:55 train.py: 88] Epoch 1, iter 3800/6416, lr 0.000491, loss 13.345497
+INFO 2021-11-05 16:33:58 train.py: 88] Epoch 1, iter 4000/6416, lr 0.000490, loss 13.233253
+INFO 2021-11-05 16:41:05 train.py: 88] Epoch 1, iter 4200/6416, lr 0.000490, loss 13.084278
+INFO 2021-11-05 16:48:10 train.py: 88] Epoch 1, iter 4400/6416, lr 0.000489, loss 12.913297
+INFO 2021-11-05 16:55:14 train.py: 88] Epoch 1, iter 4600/6416, lr 0.000489, loss 12.767326
+INFO 2021-11-05 17:02:21 train.py: 88] Epoch 1, iter 4800/6416, lr 0.000489, loss 12.515195
+INFO 2021-11-05 17:09:30 train.py: 88] Epoch 1, iter 5000/6416, lr 0.000488, loss 12.256249
+INFO 2021-11-05 17:16:43 train.py: 88] Epoch 1, iter 5200/6416, lr 0.000488, loss 12.054113
+INFO 2021-11-05 17:23:51 train.py: 88] Epoch 1, iter 5400/6416, lr 0.000487, loss 11.833332
+INFO 2021-11-05 17:31:01 train.py: 88] Epoch 1, iter 5600/6416, lr 0.000487, loss 11.609206
+INFO 2021-11-05 17:38:17 train.py: 88] Epoch 1, iter 5800/6416, lr 0.000486, loss 11.389999
+INFO 2021-11-05 17:45:36 train.py: 101] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-11-05 17:45:38 train.py: 88] Epoch 1, iter 6000/6416, lr 0.000486, loss 10.981157
+INFO 2021-11-05 17:52:47 train.py: 88] Epoch 1, iter 6200/6416, lr 0.000486, loss 10.848667
+INFO 2021-11-05 17:59:58 train.py: 88] Epoch 1, iter 6400/6416, lr 0.000485, loss 10.582276
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+INFO 2021-11-05 18:00:33 train.py: 108] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-11-05 18:00:35 train.py: 88] Epoch 2, iter 0/6416, lr 0.000485, loss 10.323736
+INFO 2021-11-05 18:08:31 train.py: 88] Epoch 2, iter 200/6416, lr 0.000485, loss 10.283372
+INFO 2021-11-05 18:16:47 train.py: 88] Epoch 2, iter 400/6416, lr 0.000484, loss 10.132700
+INFO 2021-11-05 18:24:55 train.py: 88] Epoch 2, iter 600/6416, lr 0.000484, loss 9.871738
+INFO 2021-11-05 18:33:00 train.py: 88] Epoch 2, iter 800/6416, lr 0.000483, loss 9.703248
+INFO 2021-11-05 18:41:01 train.py: 88] Epoch 2, iter 1000/6416, lr 0.000483, loss 9.534842
+INFO 2021-11-05 18:48:48 train.py: 88] Epoch 2, iter 1200/6416, lr 0.000482, loss 9.295459
+INFO 2021-11-05 18:56:30 train.py: 88] Epoch 2, iter 1400/6416, lr 0.000482, loss 9.072993
+INFO 2021-11-05 19:04:16 train.py: 88] Epoch 2, iter 1600/6416, lr 0.000481, loss 8.939991
+INFO 2021-11-05 19:12:10 train.py: 88] Epoch 2, iter 1800/6416, lr 0.000481, loss 8.662765
+INFO 2021-11-05 19:19:47 train.py: 88] Epoch 2, iter 2000/6416, lr 0.000480, loss 8.524214
+INFO 2021-11-05 19:27:29 train.py: 88] Epoch 2, iter 2200/6416, lr 0.000480, loss 8.402646
+INFO 2021-11-05 19:35:19 train.py: 88] Epoch 2, iter 2400/6416, lr 0.000479, loss 8.219456
+INFO 2021-11-05 19:43:08 train.py: 88] Epoch 2, iter 2600/6416, lr 0.000479, loss 8.178787
+INFO 2021-11-05 19:51:15 train.py: 88] Epoch 2, iter 2800/6416, lr 0.000478, loss 7.999471
+INFO 2021-11-05 19:58:53 train.py: 101] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-11-05 19:58:55 train.py: 88] Epoch 2, iter 3000/6416, lr 0.000477, loss 7.870890
+INFO 2021-11-05 20:06:39 train.py: 88] Epoch 2, iter 3200/6416, lr 0.000477, loss 7.646556
+INFO 2021-11-05 20:14:20 train.py: 88] Epoch 2, iter 3400/6416, lr 0.000476, loss 7.601601
+INFO 2021-11-05 20:22:03 train.py: 88] Epoch 2, iter 3600/6416, lr 0.000476, loss 7.535017
+INFO 2021-11-05 20:29:53 train.py: 88] Epoch 2, iter 3800/6416, lr 0.000475, loss 7.368255
+INFO 2021-11-05 20:37:39 train.py: 88] Epoch 2, iter 4000/6416, lr 0.000475, loss 7.289487
+INFO 2021-11-05 20:45:30 train.py: 88] Epoch 2, iter 4200/6416, lr 0.000474, loss 7.137532
+INFO 2021-11-05 20:52:59 train.py: 88] Epoch 2, iter 4400/6416, lr 0.000473, loss 7.077837
+INFO 2021-11-05 21:00:49 train.py: 88] Epoch 2, iter 4600/6416, lr 0.000473, loss 6.949704
+INFO 2021-11-05 21:08:31 train.py: 88] Epoch 2, iter 4800/6416, lr 0.000472, loss 6.817141
+INFO 2021-11-05 21:16:18 train.py: 88] Epoch 2, iter 5000/6416, lr 0.000471, loss 6.814206
+INFO 2021-11-05 21:24:29 train.py: 88] Epoch 2, iter 5200/6416, lr 0.000471, loss 6.726458
+INFO 2021-11-05 21:32:15 train.py: 88] Epoch 2, iter 5400/6416, lr 0.000470, loss 6.606023
+INFO 2021-11-05 21:40:10 train.py: 88] Epoch 2, iter 5600/6416, lr 0.000470, loss 6.547344
+INFO 2021-11-05 21:47:58 train.py: 88] Epoch 2, iter 5800/6416, lr 0.000469, loss 6.472251
+INFO 2021-11-05 21:55:44 train.py: 101] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-11-05 21:55:46 train.py: 88] Epoch 2, iter 6000/6416, lr 0.000468, loss 6.325722
+INFO 2021-11-05 22:03:34 train.py: 88] Epoch 2, iter 6200/6416, lr 0.000468, loss 6.352963
+INFO 2021-11-05 22:11:30 train.py: 88] Epoch 2, iter 6400/6416, lr 0.000467, loss 6.214425
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-05 22:12:13 train.py: 108] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-11-05 22:12:15 train.py: 88] Epoch 3, iter 0/6416, lr 0.000467, loss 6.221016
+INFO 2021-11-05 22:19:32 train.py: 88] Epoch 3, iter 200/6416, lr 0.000466, loss 6.161653
+INFO 2021-11-05 22:26:44 train.py: 88] Epoch 3, iter 400/6416, lr 0.000465, loss 6.143535
+INFO 2021-11-05 22:33:54 train.py: 88] Epoch 3, iter 600/6416, lr 0.000465, loss 6.091019
+INFO 2021-11-05 22:41:07 train.py: 88] Epoch 3, iter 800/6416, lr 0.000464, loss 6.007750
+INFO 2021-11-05 22:48:18 train.py: 88] Epoch 3, iter 1000/6416, lr 0.000463, loss 6.018930
+INFO 2021-11-05 22:55:29 train.py: 88] Epoch 3, iter 1200/6416, lr 0.000463, loss 5.880804
+INFO 2021-11-05 23:02:38 train.py: 88] Epoch 3, iter 1400/6416, lr 0.000462, loss 5.794622
+INFO 2021-11-05 23:09:52 train.py: 88] Epoch 3, iter 1600/6416, lr 0.000461, loss 5.821633
+INFO 2021-11-05 23:17:00 train.py: 88] Epoch 3, iter 1800/6416, lr 0.000461, loss 5.645156
+INFO 2021-11-05 23:24:13 train.py: 88] Epoch 3, iter 2000/6416, lr 0.000460, loss 5.603906
+INFO 2021-11-05 23:31:25 train.py: 88] Epoch 3, iter 2200/6416, lr 0.000459, loss 5.597737
+INFO 2021-11-05 23:38:39 train.py: 88] Epoch 3, iter 2400/6416, lr 0.000458, loss 5.525972
+INFO 2021-11-05 23:45:49 train.py: 88] Epoch 3, iter 2600/6416, lr 0.000458, loss 5.574679
+INFO 2021-11-05 23:53:04 train.py: 88] Epoch 3, iter 2800/6416, lr 0.000457, loss 5.506074
+INFO 2021-11-06 00:00:15 train.py: 101] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-11-06 00:00:17 train.py: 88] Epoch 3, iter 3000/6416, lr 0.000456, loss 5.434778
+INFO 2021-11-06 00:07:26 train.py: 88] Epoch 3, iter 3200/6416, lr 0.000455, loss 5.344736
+INFO 2021-11-06 00:14:30 train.py: 88] Epoch 3, iter 3400/6416, lr 0.000454, loss 5.336427
+INFO 2021-11-06 00:21:41 train.py: 88] Epoch 3, iter 3600/6416, lr 0.000454, loss 5.319012
+INFO 2021-11-06 00:28:50 train.py: 88] Epoch 3, iter 3800/6416, lr 0.000453, loss 5.224000
+INFO 2021-11-06 00:36:00 train.py: 88] Epoch 3, iter 4000/6416, lr 0.000452, loss 5.236565
+INFO 2021-11-06 00:43:08 train.py: 88] Epoch 3, iter 4200/6416, lr 0.000451, loss 5.119843
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 00:50:17 train.py: 88] Epoch 3, iter 4400/6416, lr 0.000451, loss 5.129462
+INFO 2021-11-06 00:57:25 train.py: 88] Epoch 3, iter 4600/6416, lr 0.000450, loss 5.104172
+INFO 2021-11-06 01:04:34 train.py: 88] Epoch 3, iter 4800/6416, lr 0.000449, loss 5.008277
+INFO 2021-11-06 01:11:48 train.py: 88] Epoch 3, iter 5000/6416, lr 0.000448, loss 5.048250
+INFO 2021-11-06 01:18:59 train.py: 88] Epoch 3, iter 5200/6416, lr 0.000447, loss 5.022131
+INFO 2021-11-06 01:26:02 train.py: 88] Epoch 3, iter 5400/6416, lr 0.000446, loss 4.935270
+INFO 2021-11-06 01:33:11 train.py: 88] Epoch 3, iter 5600/6416, lr 0.000446, loss 4.925240
+INFO 2021-11-06 01:40:17 train.py: 88] Epoch 3, iter 5800/6416, lr 0.000445, loss 4.867109
+INFO 2021-11-06 01:47:26 train.py: 101] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-11-06 01:47:28 train.py: 88] Epoch 3, iter 6000/6416, lr 0.000444, loss 4.813902
+INFO 2021-11-06 01:54:31 train.py: 88] Epoch 3, iter 6200/6416, lr 0.000443, loss 4.847055
+INFO 2021-11-06 02:01:40 train.py: 88] Epoch 3, iter 6400/6416, lr 0.000442, loss 4.778060
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 02:02:16 train.py: 108] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-11-06 02:02:18 train.py: 88] Epoch 4, iter 0/6416, lr 0.000442, loss 4.727170
+INFO 2021-11-06 02:09:32 train.py: 88] Epoch 4, iter 200/6416, lr 0.000441, loss 4.738089
+INFO 2021-11-06 02:16:41 train.py: 88] Epoch 4, iter 400/6416, lr 0.000440, loss 4.739438
+INFO 2021-11-06 02:23:49 train.py: 88] Epoch 4, iter 600/6416, lr 0.000439, loss 4.730593
+INFO 2021-11-06 02:30:55 train.py: 88] Epoch 4, iter 800/6416, lr 0.000439, loss 4.658415
+INFO 2021-11-06 02:38:02 train.py: 88] Epoch 4, iter 1000/6416, lr 0.000438, loss 4.691971
+INFO 2021-11-06 02:45:13 train.py: 88] Epoch 4, iter 1200/6416, lr 0.000437, loss 4.617151
+INFO 2021-11-06 02:52:19 train.py: 88] Epoch 4, iter 1400/6416, lr 0.000436, loss 4.556927
+INFO 2021-11-06 02:59:23 train.py: 88] Epoch 4, iter 1600/6416, lr 0.000435, loss 4.608335
+INFO 2021-11-06 03:06:33 train.py: 88] Epoch 4, iter 1800/6416, lr 0.000434, loss 4.456906
+INFO 2021-11-06 03:13:34 train.py: 88] Epoch 4, iter 2000/6416, lr 0.000433, loss 4.457605
+INFO 2021-11-06 03:20:40 train.py: 88] Epoch 4, iter 2200/6416, lr 0.000432, loss 4.447750
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 03:27:52 train.py: 88] Epoch 4, iter 2400/6416, lr 0.000431, loss 4.418313
+INFO 2021-11-06 03:34:56 train.py: 88] Epoch 4, iter 2600/6416, lr 0.000430, loss 4.442612
+INFO 2021-11-06 03:42:08 train.py: 88] Epoch 4, iter 2800/6416, lr 0.000429, loss 4.422268
+INFO 2021-11-06 03:49:18 train.py: 101] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-11-06 03:49:20 train.py: 88] Epoch 4, iter 3000/6416, lr 0.000428, loss 4.391550
+INFO 2021-11-06 03:56:24 train.py: 88] Epoch 4, iter 3200/6416, lr 0.000428, loss 4.308499
+INFO 2021-11-06 04:03:31 train.py: 88] Epoch 4, iter 3400/6416, lr 0.000427, loss 4.308353
+INFO 2021-11-06 04:10:38 train.py: 88] Epoch 4, iter 3600/6416, lr 0.000426, loss 4.309207
+INFO 2021-11-06 04:17:45 train.py: 88] Epoch 4, iter 3800/6416, lr 0.000425, loss 4.238269
+INFO 2021-11-06 04:24:59 train.py: 88] Epoch 4, iter 4000/6416, lr 0.000424, loss 4.236058
+INFO 2021-11-06 04:32:06 train.py: 88] Epoch 4, iter 4200/6416, lr 0.000423, loss 4.157394
+INFO 2021-11-06 04:39:16 train.py: 88] Epoch 4, iter 4400/6416, lr 0.000422, loss 4.195104
+INFO 2021-11-06 04:46:23 train.py: 88] Epoch 4, iter 4600/6416, lr 0.000421, loss 4.181349
+INFO 2021-11-06 04:53:32 train.py: 88] Epoch 4, iter 4800/6416, lr 0.000420, loss 4.107947
+INFO 2021-11-06 05:00:39 train.py: 88] Epoch 4, iter 5000/6416, lr 0.000419, loss 4.151203
+INFO 2021-11-06 05:07:48 train.py: 88] Epoch 4, iter 5200/6416, lr 0.000418, loss 4.135777
+INFO 2021-11-06 05:15:00 train.py: 88] Epoch 4, iter 5400/6416, lr 0.000417, loss 4.061098
+INFO 2021-11-06 05:22:04 train.py: 88] Epoch 4, iter 5600/6416, lr 0.000416, loss 4.061317
+INFO 2021-11-06 05:29:07 train.py: 88] Epoch 4, iter 5800/6416, lr 0.000415, loss 4.035901
+INFO 2021-11-06 05:36:01 train.py: 101] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-11-06 05:36:02 train.py: 88] Epoch 4, iter 6000/6416, lr 0.000414, loss 4.004453
+INFO 2021-11-06 05:42:52 train.py: 88] Epoch 4, iter 6200/6416, lr 0.000413, loss 4.008494
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-11-06 05:49:47 train.py: 88] Epoch 4, iter 6400/6416, lr 0.000412, loss 3.967295
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 05:50:22 train.py: 108] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-11-06 05:50:24 train.py: 88] Epoch 5, iter 0/6416, lr 0.000412, loss 3.980309
+INFO 2021-11-06 05:57:14 train.py: 88] Epoch 5, iter 200/6416, lr 0.000411, loss 3.943311
+INFO 2021-11-06 06:04:01 train.py: 88] Epoch 5, iter 400/6416, lr 0.000410, loss 3.959548
+INFO 2021-11-06 06:10:54 train.py: 88] Epoch 5, iter 600/6416, lr 0.000408, loss 3.945531
+INFO 2021-11-06 06:17:43 train.py: 88] Epoch 5, iter 800/6416, lr 0.000407, loss 3.915572
+INFO 2021-11-06 06:24:32 train.py: 88] Epoch 5, iter 1000/6416, lr 0.000406, loss 3.952650
+INFO 2021-11-06 06:31:22 train.py: 88] Epoch 5, iter 1200/6416, lr 0.000405, loss 3.878313
+INFO 2021-11-06 06:38:09 train.py: 88] Epoch 5, iter 1400/6416, lr 0.000404, loss 3.822669
+INFO 2021-11-06 06:44:56 train.py: 88] Epoch 5, iter 1600/6416, lr 0.000403, loss 3.878812
+INFO 2021-11-06 06:51:45 train.py: 88] Epoch 5, iter 1800/6416, lr 0.000402, loss 3.750864
+INFO 2021-11-06 06:58:32 train.py: 88] Epoch 5, iter 2000/6416, lr 0.000401, loss 3.758655
+INFO 2021-11-06 07:05:26 train.py: 88] Epoch 5, iter 2200/6416, lr 0.000400, loss 3.770638
+INFO 2021-11-06 07:12:12 train.py: 88] Epoch 5, iter 2400/6416, lr 0.000399, loss 3.718840
+INFO 2021-11-06 07:19:04 train.py: 88] Epoch 5, iter 2600/6416, lr 0.000398, loss 3.750132
+INFO 2021-11-06 07:26:01 train.py: 88] Epoch 5, iter 2800/6416, lr 0.000397, loss 3.745643
+INFO 2021-11-06 07:32:56 train.py: 101] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-11-06 07:32:57 train.py: 88] Epoch 5, iter 3000/6416, lr 0.000396, loss 3.713246
+INFO 2021-11-06 07:39:46 train.py: 88] Epoch 5, iter 3200/6416, lr 0.000395, loss 3.665213
+INFO 2021-11-06 07:46:34 train.py: 88] Epoch 5, iter 3400/6416, lr 0.000393, loss 3.661409
+INFO 2021-11-06 07:53:22 train.py: 88] Epoch 5, iter 3600/6416, lr 0.000392, loss 3.668914
+INFO 2021-11-06 08:00:11 train.py: 88] Epoch 5, iter 3800/6416, lr 0.000391, loss 3.600594
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-11-06 08:06:59 train.py: 88] Epoch 5, iter 4000/6416, lr 0.000390, loss 3.632864
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 08:13:49 train.py: 88] Epoch 5, iter 4200/6416, lr 0.000389, loss 3.538527
+INFO 2021-11-06 08:20:42 train.py: 88] Epoch 5, iter 4400/6416, lr 0.000388, loss 3.595721
+INFO 2021-11-06 08:27:37 train.py: 88] Epoch 5, iter 4600/6416, lr 0.000387, loss 3.579950
+INFO 2021-11-06 08:34:22 train.py: 88] Epoch 5, iter 4800/6416, lr 0.000386, loss 3.528899
+INFO 2021-11-06 08:41:16 train.py: 88] Epoch 5, iter 5000/6416, lr 0.000384, loss 3.554575
+INFO 2021-11-06 08:48:08 train.py: 88] Epoch 5, iter 5200/6416, lr 0.000383, loss 3.550355
+INFO 2021-11-06 08:54:51 train.py: 88] Epoch 5, iter 5400/6416, lr 0.000382, loss 3.480871
+INFO 2021-11-06 09:01:45 train.py: 88] Epoch 5, iter 5600/6416, lr 0.000381, loss 3.502277
+INFO 2021-11-06 09:08:37 train.py: 88] Epoch 5, iter 5800/6416, lr 0.000380, loss 3.456603
+INFO 2021-11-06 09:15:34 train.py: 101] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-11-06 09:15:36 train.py: 88] Epoch 5, iter 6000/6416, lr 0.000379, loss 3.454415
+INFO 2021-11-06 09:22:26 train.py: 88] Epoch 5, iter 6200/6416, lr 0.000378, loss 3.462584
+INFO 2021-11-06 09:29:13 train.py: 88] Epoch 5, iter 6400/6416, lr 0.000376, loss 3.434576
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 09:29:46 train.py: 108] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-11-06 09:29:48 train.py: 88] Epoch 6, iter 0/6416, lr 0.000376, loss 3.415257
+INFO 2021-11-06 09:36:40 train.py: 88] Epoch 6, iter 200/6416, lr 0.000375, loss 3.398793
+INFO 2021-11-06 09:43:33 train.py: 88] Epoch 6, iter 400/6416, lr 0.000374, loss 3.426743
+INFO 2021-11-06 09:50:27 train.py: 88] Epoch 6, iter 600/6416, lr 0.000373, loss 3.421948
+INFO 2021-11-06 09:57:20 train.py: 88] Epoch 6, iter 800/6416, lr 0.000372, loss 3.386297
+INFO 2021-11-06 10:04:17 train.py: 88] Epoch 6, iter 1000/6416, lr 0.000370, loss 3.435088
+INFO 2021-11-06 10:11:10 train.py: 88] Epoch 6, iter 1200/6416, lr 0.000369, loss 3.350302
+INFO 2021-11-06 10:18:04 train.py: 88] Epoch 6, iter 1400/6416, lr 0.000368, loss 3.306239
+INFO 2021-11-06 10:24:51 train.py: 88] Epoch 6, iter 1600/6416, lr 0.000367, loss 3.372015
+INFO 2021-11-06 10:31:36 train.py: 88] Epoch 6, iter 1800/6416, lr 0.000366, loss 3.270989
+INFO 2021-11-06 10:38:31 train.py: 88] Epoch 6, iter 2000/6416, lr 0.000364, loss 3.266378
+INFO 2021-11-06 10:45:24 train.py: 88] Epoch 6, iter 2200/6416, lr 0.000363, loss 3.264796
+INFO 2021-11-06 10:52:19 train.py: 88] Epoch 6, iter 2400/6416, lr 0.000362, loss 3.223176
+INFO 2021-11-06 10:59:10 train.py: 88] Epoch 6, iter 2600/6416, lr 0.000361, loss 3.282731
+INFO 2021-11-06 11:05:57 train.py: 88] Epoch 6, iter 2800/6416, lr 0.000360, loss 3.263794
+INFO 2021-11-06 11:12:47 train.py: 101] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-11-06 11:12:49 train.py: 88] Epoch 6, iter 3000/6416, lr 0.000358, loss 3.240334
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 11:19:37 train.py: 88] Epoch 6, iter 3200/6416, lr 0.000357, loss 3.195197
+INFO 2021-11-06 11:26:26 train.py: 88] Epoch 6, iter 3400/6416, lr 0.000356, loss 3.179522
+INFO 2021-11-06 11:33:15 train.py: 88] Epoch 6, iter 3600/6416, lr 0.000355, loss 3.195714
+INFO 2021-11-06 11:40:04 train.py: 88] Epoch 6, iter 3800/6416, lr 0.000353, loss 3.132245
+INFO 2021-11-06 11:46:48 train.py: 88] Epoch 6, iter 4000/6416, lr 0.000352, loss 3.182484
+INFO 2021-11-06 11:53:46 train.py: 88] Epoch 6, iter 4200/6416, lr 0.000351, loss 3.099346
+INFO 2021-11-06 12:00:38 train.py: 88] Epoch 6, iter 4400/6416, lr 0.000350, loss 3.141692
+INFO 2021-11-06 12:07:30 train.py: 88] Epoch 6, iter 4600/6416, lr 0.000349, loss 3.126413
+INFO 2021-11-06 12:14:20 train.py: 88] Epoch 6, iter 4800/6416, lr 0.000347, loss 3.080857
+INFO 2021-11-06 12:21:16 train.py: 88] Epoch 6, iter 5000/6416, lr 0.000346, loss 3.132587
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 12:28:08 train.py: 88] Epoch 6, iter 5200/6416, lr 0.000345, loss 3.106201
+INFO 2021-11-06 12:35:02 train.py: 88] Epoch 6, iter 5400/6416, lr 0.000344, loss 3.064733
+INFO 2021-11-06 12:41:52 train.py: 88] Epoch 6, iter 5600/6416, lr 0.000342, loss 3.069506
+INFO 2021-11-06 12:48:43 train.py: 88] Epoch 6, iter 5800/6416, lr 0.000341, loss 3.029465
+INFO 2021-11-06 12:55:36 train.py: 101] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-11-06 12:55:38 train.py: 88] Epoch 6, iter 6000/6416, lr 0.000340, loss 3.035294
+INFO 2021-11-06 13:02:29 train.py: 88] Epoch 6, iter 6200/6416, lr 0.000339, loss 3.042550
+INFO 2021-11-06 13:09:25 train.py: 88] Epoch 6, iter 6400/6416, lr 0.000337, loss 3.020754
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-06 13:09:58 train.py: 108] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-11-06 13:10:00 train.py: 88] Epoch 7, iter 0/6416, lr 0.000337, loss 2.999156
+INFO 2021-11-06 13:16:55 train.py: 88] Epoch 7, iter 200/6416, lr 0.000336, loss 2.985685
+INFO 2021-11-06 13:23:47 train.py: 88] Epoch 7, iter 400/6416, lr 0.000335, loss 2.992421
+INFO 2021-11-06 13:30:36 train.py: 88] Epoch 7, iter 600/6416, lr 0.000333, loss 3.019848
+INFO 2021-11-06 13:37:29 train.py: 88] Epoch 7, iter 800/6416, lr 0.000332, loss 2.986077
+INFO 2021-11-06 13:44:21 train.py: 88] Epoch 7, iter 1000/6416, lr 0.000331, loss 3.009750
+INFO 2021-11-06 13:51:09 train.py: 88] Epoch 7, iter 1200/6416, lr 0.000330, loss 2.954818
+INFO 2021-11-06 13:57:56 train.py: 88] Epoch 7, iter 1400/6416, lr 0.000328, loss 2.903799
+INFO 2021-11-06 14:04:43 train.py: 88] Epoch 7, iter 1600/6416, lr 0.000327, loss 2.949513
+INFO 2021-11-06 14:11:32 train.py: 88] Epoch 7, iter 1800/6416, lr 0.000326, loss 2.864682
+INFO 2021-11-06 14:18:21 train.py: 88] Epoch 7, iter 2000/6416, lr 0.000324, loss 2.884796
+INFO 2021-11-06 14:25:19 train.py: 88] Epoch 7, iter 2200/6416, lr 0.000323, loss 2.894090
+INFO 2021-11-06 14:32:00 train.py: 88] Epoch 7, iter 2400/6416, lr 0.000322, loss 2.848664
+INFO 2021-11-06 14:38:50 train.py: 88] Epoch 7, iter 2600/6416, lr 0.000321, loss 2.875024
+INFO 2021-11-06 14:45:43 train.py: 88] Epoch 7, iter 2800/6416, lr 0.000319, loss 2.882685
+INFO 2021-11-06 14:52:34 train.py: 101] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-11-06 14:52:36 train.py: 88] Epoch 7, iter 3000/6416, lr 0.000318, loss 2.871736
+INFO 2021-11-06 14:59:23 train.py: 88] Epoch 7, iter 3200/6416, lr 0.000317, loss 2.824761
+INFO 2021-11-06 15:06:14 train.py: 88] Epoch 7, iter 3400/6416, lr 0.000315, loss 2.818186
+INFO 2021-11-06 15:13:08 train.py: 88] Epoch 7, iter 3600/6416, lr 0.000314, loss 2.813882
+INFO 2021-11-06 15:20:05 train.py: 88] Epoch 7, iter 3800/6416, lr 0.000313, loss 2.756519
+INFO 2021-11-06 15:26:55 train.py: 88] Epoch 7, iter 4000/6416, lr 0.000311, loss 2.799995
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 15:33:46 train.py: 88] Epoch 7, iter 4200/6416, lr 0.000310, loss 2.729271
+INFO 2021-11-06 15:40:37 train.py: 88] Epoch 7, iter 4400/6416, lr 0.000309, loss 2.779209
+INFO 2021-11-06 15:47:29 train.py: 88] Epoch 7, iter 4600/6416, lr 0.000307, loss 2.770328
+INFO 2021-11-06 15:54:26 train.py: 88] Epoch 7, iter 4800/6416, lr 0.000306, loss 2.729695
+INFO 2021-11-06 16:01:19 train.py: 88] Epoch 7, iter 5000/6416, lr 0.000305, loss 2.780403
+INFO 2021-11-06 16:08:12 train.py: 88] Epoch 7, iter 5200/6416, lr 0.000304, loss 2.746901
+INFO 2021-11-06 16:15:02 train.py: 88] Epoch 7, iter 5400/6416, lr 0.000302, loss 2.696840
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-06 16:21:58 train.py: 88] Epoch 7, iter 5600/6416, lr 0.000301, loss 2.718681
+INFO 2021-11-06 16:28:54 train.py: 88] Epoch 7, iter 5800/6416, lr 0.000300, loss 2.677439
+INFO 2021-11-06 16:35:52 train.py: 101] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-11-06 16:35:54 train.py: 88] Epoch 7, iter 6000/6416, lr 0.000298, loss 2.675162
+INFO 2021-11-06 16:42:50 train.py: 88] Epoch 7, iter 6200/6416, lr 0.000297, loss 2.690852
+INFO 2021-11-06 16:49:43 train.py: 88] Epoch 7, iter 6400/6416, lr 0.000296, loss 2.684395
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+INFO 2021-11-06 16:50:16 train.py: 108] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-11-06 16:50:18 train.py: 88] Epoch 8, iter 0/6416, lr 0.000295, loss 2.669463
+INFO 2021-11-06 16:57:14 train.py: 88] Epoch 8, iter 200/6416, lr 0.000294, loss 2.644270
+INFO 2021-11-06 17:04:07 train.py: 88] Epoch 8, iter 400/6416, lr 0.000293, loss 2.665808
+INFO 2021-11-06 17:11:02 train.py: 88] Epoch 8, iter 600/6416, lr 0.000291, loss 2.675715
+INFO 2021-11-06 17:18:00 train.py: 88] Epoch 8, iter 800/6416, lr 0.000290, loss 2.638412
+INFO 2021-11-06 17:24:58 train.py: 88] Epoch 8, iter 1000/6416, lr 0.000289, loss 2.676718
+INFO 2021-11-06 17:31:52 train.py: 88] Epoch 8, iter 1200/6416, lr 0.000287, loss 2.616763
+INFO 2021-11-06 17:38:44 train.py: 88] Epoch 8, iter 1400/6416, lr 0.000286, loss 2.564106
+INFO 2021-11-06 17:45:39 train.py: 88] Epoch 8, iter 1600/6416, lr 0.000285, loss 2.622060
+INFO 2021-11-06 17:52:42 train.py: 88] Epoch 8, iter 1800/6416, lr 0.000283, loss 2.540723
+INFO 2021-11-06 17:59:39 train.py: 88] Epoch 8, iter 2000/6416, lr 0.000282, loss 2.553938
+INFO 2021-11-06 18:06:40 train.py: 88] Epoch 8, iter 2200/6416, lr 0.000281, loss 2.549358
+INFO 2021-11-06 18:13:36 train.py: 88] Epoch 8, iter 2400/6416, lr 0.000279, loss 2.516490
+INFO 2021-11-06 18:20:40 train.py: 88] Epoch 8, iter 2600/6416, lr 0.000278, loss 2.546188
+INFO 2021-11-06 18:27:38 train.py: 88] Epoch 8, iter 2800/6416, lr 0.000277, loss 2.571529
+INFO 2021-11-06 18:34:45 train.py: 101] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-11-06 18:34:47 train.py: 88] Epoch 8, iter 3000/6416, lr 0.000275, loss 2.548459
+INFO 2021-11-06 18:41:42 train.py: 88] Epoch 8, iter 3200/6416, lr 0.000274, loss 2.497961
+INFO 2021-11-06 18:48:39 train.py: 88] Epoch 8, iter 3400/6416, lr 0.000273, loss 2.494363
+INFO 2021-11-06 18:55:39 train.py: 88] Epoch 8, iter 3600/6416, lr 0.000271, loss 2.513804
+INFO 2021-11-06 19:02:41 train.py: 88] Epoch 8, iter 3800/6416, lr 0.000270, loss 2.446475
+INFO 2021-11-06 19:09:46 train.py: 88] Epoch 8, iter 4000/6416, lr 0.000269, loss 2.500524
+INFO 2021-11-06 19:16:49 train.py: 88] Epoch 8, iter 4200/6416, lr 0.000267, loss 2.407533
+INFO 2021-11-06 19:23:50 train.py: 88] Epoch 8, iter 4400/6416, lr 0.000266, loss 2.460430
+INFO 2021-11-06 19:30:47 train.py: 88] Epoch 8, iter 4600/6416, lr 0.000265, loss 2.446950
+INFO 2021-11-06 19:37:47 train.py: 88] Epoch 8, iter 4800/6416, lr 0.000263, loss 2.423938
+INFO 2021-11-06 19:44:51 train.py: 88] Epoch 8, iter 5000/6416, lr 0.000262, loss 2.455807
+INFO 2021-11-06 19:51:48 train.py: 88] Epoch 8, iter 5200/6416, lr 0.000261, loss 2.424536
+INFO 2021-11-06 19:58:56 train.py: 88] Epoch 8, iter 5400/6416, lr 0.000259, loss 2.408734
+INFO 2021-11-06 20:05:54 train.py: 88] Epoch 8, iter 5600/6416, lr 0.000258, loss 2.401375
+INFO 2021-11-06 20:12:54 train.py: 88] Epoch 8, iter 5800/6416, lr 0.000257, loss 2.383591
+INFO 2021-11-06 20:19:52 train.py: 101] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-11-06 20:19:54 train.py: 88] Epoch 8, iter 6000/6416, lr 0.000255, loss 2.397859
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 20:26:54 train.py: 88] Epoch 8, iter 6200/6416, lr 0.000254, loss 2.396250
+INFO 2021-11-06 20:33:54 train.py: 88] Epoch 8, iter 6400/6416, lr 0.000253, loss 2.365418
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-06 20:34:29 train.py: 108] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-11-06 20:34:31 train.py: 88] Epoch 9, iter 0/6416, lr 0.000253, loss 2.369200
+INFO 2021-11-06 20:42:20 train.py: 88] Epoch 9, iter 200/6416, lr 0.000251, loss 2.352967
+INFO 2021-11-06 20:50:12 train.py: 88] Epoch 9, iter 400/6416, lr 0.000250, loss 2.359061
+INFO 2021-11-06 20:58:09 train.py: 88] Epoch 9, iter 600/6416, lr 0.000248, loss 2.375999
+INFO 2021-11-06 21:05:42 train.py: 88] Epoch 9, iter 800/6416, lr 0.000247, loss 2.344395
+INFO 2021-11-06 21:13:36 train.py: 88] Epoch 9, iter 1000/6416, lr 0.000246, loss 2.400711
+INFO 2021-11-06 21:21:18 train.py: 88] Epoch 9, iter 1200/6416, lr 0.000244, loss 2.325335
+INFO 2021-11-06 21:29:02 train.py: 88] Epoch 9, iter 1400/6416, lr 0.000243, loss 2.266354
+INFO 2021-11-06 21:36:36 train.py: 88] Epoch 9, iter 1600/6416, lr 0.000242, loss 2.338721
+INFO 2021-11-06 21:44:13 train.py: 88] Epoch 9, iter 1800/6416, lr 0.000240, loss 2.260129
+INFO 2021-11-06 21:51:45 train.py: 88] Epoch 9, iter 2000/6416, lr 0.000239, loss 2.262438
+INFO 2021-11-06 21:59:34 train.py: 88] Epoch 9, iter 2200/6416, lr 0.000238, loss 2.262319
+INFO 2021-11-06 22:07:28 train.py: 88] Epoch 9, iter 2400/6416, lr 0.000236, loss 2.239283
+INFO 2021-11-06 22:15:19 train.py: 88] Epoch 9, iter 2600/6416, lr 0.000235, loss 2.267560
+INFO 2021-11-06 22:22:51 train.py: 88] Epoch 9, iter 2800/6416, lr 0.000234, loss 2.262765
+INFO 2021-11-06 22:30:01 train.py: 101] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-11-06 22:30:03 train.py: 88] Epoch 9, iter 3000/6416, lr 0.000232, loss 2.263784
+INFO 2021-11-06 22:37:23 train.py: 88] Epoch 9, iter 3200/6416, lr 0.000231, loss 2.213395
+INFO 2021-11-06 22:45:09 train.py: 88] Epoch 9, iter 3400/6416, lr 0.000230, loss 2.218716
+INFO 2021-11-06 22:52:39 train.py: 88] Epoch 9, iter 3600/6416, lr 0.000228, loss 2.221739
+INFO 2021-11-06 22:59:52 train.py: 88] Epoch 9, iter 3800/6416, lr 0.000227, loss 2.157623
+INFO 2021-11-06 23:07:04 train.py: 88] Epoch 9, iter 4000/6416, lr 0.000226, loss 2.221776
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-06 23:14:32 train.py: 88] Epoch 9, iter 4200/6416, lr 0.000224, loss 2.145635
+INFO 2021-11-06 23:22:11 train.py: 88] Epoch 9, iter 4400/6416, lr 0.000223, loss 2.170148
+INFO 2021-11-06 23:29:35 train.py: 88] Epoch 9, iter 4600/6416, lr 0.000222, loss 2.175371
+INFO 2021-11-06 23:37:02 train.py: 88] Epoch 9, iter 4800/6416, lr 0.000220, loss 2.161735
+INFO 2021-11-06 23:44:16 train.py: 88] Epoch 9, iter 5000/6416, lr 0.000219, loss 2.186107
+INFO 2021-11-06 23:51:47 train.py: 88] Epoch 9, iter 5200/6416, lr 0.000218, loss 2.147041
+INFO 2021-11-06 23:59:15 train.py: 88] Epoch 9, iter 5400/6416, lr 0.000216, loss 2.134002
+INFO 2021-11-07 00:06:33 train.py: 88] Epoch 9, iter 5600/6416, lr 0.000215, loss 2.133852
+INFO 2021-11-07 00:13:53 train.py: 88] Epoch 9, iter 5800/6416, lr 0.000214, loss 2.095229
+INFO 2021-11-07 00:21:10 train.py: 101] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-11-07 00:21:12 train.py: 88] Epoch 9, iter 6000/6416, lr 0.000212, loss 2.109279
+INFO 2021-11-07 00:28:32 train.py: 88] Epoch 9, iter 6200/6416, lr 0.000211, loss 2.134581
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 00:36:09 train.py: 88] Epoch 9, iter 6400/6416, lr 0.000210, loss 2.109606
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-07 00:36:42 train.py: 108] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-11-07 00:36:44 train.py: 88] Epoch 10, iter 0/6416, lr 0.000210, loss 2.086751
+INFO 2021-11-07 00:43:33 train.py: 88] Epoch 10, iter 200/6416, lr 0.000208, loss 2.099767
+INFO 2021-11-07 00:50:25 train.py: 88] Epoch 10, iter 400/6416, lr 0.000207, loss 2.098388
+INFO 2021-11-07 00:57:11 train.py: 88] Epoch 10, iter 600/6416, lr 0.000206, loss 2.101931
+INFO 2021-11-07 01:04:02 train.py: 88] Epoch 10, iter 800/6416, lr 0.000204, loss 2.082929
+INFO 2021-11-07 01:10:54 train.py: 88] Epoch 10, iter 1000/6416, lr 0.000203, loss 2.109219
+INFO 2021-11-07 01:17:54 train.py: 88] Epoch 10, iter 1200/6416, lr 0.000202, loss 2.060676
+INFO 2021-11-07 01:24:43 train.py: 88] Epoch 10, iter 1400/6416, lr 0.000200, loss 2.013363
+INFO 2021-11-07 01:31:35 train.py: 88] Epoch 10, iter 1600/6416, lr 0.000199, loss 2.060205
+INFO 2021-11-07 01:38:27 train.py: 88] Epoch 10, iter 1800/6416, lr 0.000198, loss 2.001704
+INFO 2021-11-07 01:45:22 train.py: 88] Epoch 10, iter 2000/6416, lr 0.000196, loss 2.006452
+INFO 2021-11-07 01:52:15 train.py: 88] Epoch 10, iter 2200/6416, lr 0.000195, loss 1.997370
+INFO 2021-11-07 01:59:12 train.py: 88] Epoch 10, iter 2400/6416, lr 0.000194, loss 1.979525
+INFO 2021-11-07 02:06:07 train.py: 88] Epoch 10, iter 2600/6416, lr 0.000192, loss 1.987469
+INFO 2021-11-07 02:12:54 train.py: 88] Epoch 10, iter 2800/6416, lr 0.000191, loss 2.022716
+INFO 2021-11-07 02:19:43 train.py: 101] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-11-07 02:19:45 train.py: 88] Epoch 10, iter 3000/6416, lr 0.000190, loss 2.000076
+INFO 2021-11-07 02:26:35 train.py: 88] Epoch 10, iter 3200/6416, lr 0.000188, loss 1.966920
+INFO 2021-11-07 02:33:20 train.py: 88] Epoch 10, iter 3400/6416, lr 0.000187, loss 1.956708
+INFO 2021-11-07 02:40:12 train.py: 88] Epoch 10, iter 3600/6416, lr 0.000186, loss 1.966688
+INFO 2021-11-07 02:47:00 train.py: 88] Epoch 10, iter 3800/6416, lr 0.000185, loss 1.898777
+INFO 2021-11-07 02:53:54 train.py: 88] Epoch 10, iter 4000/6416, lr 0.000183, loss 1.966120
+INFO 2021-11-07 03:00:41 train.py: 88] Epoch 10, iter 4200/6416, lr 0.000182, loss 1.894511
+INFO 2021-11-07 03:07:30 train.py: 88] Epoch 10, iter 4400/6416, lr 0.000181, loss 1.924982
+INFO 2021-11-07 03:14:23 train.py: 88] Epoch 10, iter 4600/6416, lr 0.000179, loss 1.931672
+INFO 2021-11-07 03:21:15 train.py: 88] Epoch 10, iter 4800/6416, lr 0.000178, loss 1.905511
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 03:28:06 train.py: 88] Epoch 10, iter 5000/6416, lr 0.000177, loss 1.935831
+INFO 2021-11-07 03:34:56 train.py: 88] Epoch 10, iter 5200/6416, lr 0.000176, loss 1.904486
+INFO 2021-11-07 03:41:49 train.py: 88] Epoch 10, iter 5400/6416, lr 0.000174, loss 1.877272
+INFO 2021-11-07 03:48:40 train.py: 88] Epoch 10, iter 5600/6416, lr 0.000173, loss 1.888271
+INFO 2021-11-07 03:55:39 train.py: 88] Epoch 10, iter 5800/6416, lr 0.000172, loss 1.863815
+INFO 2021-11-07 04:02:33 train.py: 101] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-11-07 04:02:35 train.py: 88] Epoch 10, iter 6000/6416, lr 0.000170, loss 1.881692
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-07 04:09:28 train.py: 88] Epoch 10, iter 6200/6416, lr 0.000169, loss 1.885417
+INFO 2021-11-07 04:16:20 train.py: 88] Epoch 10, iter 6400/6416, lr 0.000168, loss 1.872222
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+INFO 2021-11-07 04:16:52 train.py: 108] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-11-07 04:16:54 train.py: 88] Epoch 11, iter 0/6416, lr 0.000168, loss 1.871913
+INFO 2021-11-07 04:23:49 train.py: 88] Epoch 11, iter 200/6416, lr 0.000167, loss 1.848261
+INFO 2021-11-07 04:30:41 train.py: 88] Epoch 11, iter 400/6416, lr 0.000165, loss 1.833693
+INFO 2021-11-07 04:37:31 train.py: 88] Epoch 11, iter 600/6416, lr 0.000164, loss 1.859055
+INFO 2021-11-07 04:44:25 train.py: 88] Epoch 11, iter 800/6416, lr 0.000163, loss 1.829220
+INFO 2021-11-07 04:51:23 train.py: 88] Epoch 11, iter 1000/6416, lr 0.000162, loss 1.854977
+INFO 2021-11-07 04:58:20 train.py: 88] Epoch 11, iter 1200/6416, lr 0.000160, loss 1.823928
+INFO 2021-11-07 05:05:16 train.py: 88] Epoch 11, iter 1400/6416, lr 0.000159, loss 1.766915
+INFO 2021-11-07 05:12:06 train.py: 88] Epoch 11, iter 1600/6416, lr 0.000158, loss 1.828711
+INFO 2021-11-07 05:18:58 train.py: 88] Epoch 11, iter 1800/6416, lr 0.000157, loss 1.766249
+INFO 2021-11-07 05:25:52 train.py: 88] Epoch 11, iter 2000/6416, lr 0.000155, loss 1.760022
+INFO 2021-11-07 05:32:49 train.py: 88] Epoch 11, iter 2200/6416, lr 0.000154, loss 1.759000
+INFO 2021-11-07 05:39:38 train.py: 88] Epoch 11, iter 2400/6416, lr 0.000153, loss 1.750986
+INFO 2021-11-07 05:46:30 train.py: 88] Epoch 11, iter 2600/6416, lr 0.000152, loss 1.769970
+INFO 2021-11-07 05:53:18 train.py: 88] Epoch 11, iter 2800/6416, lr 0.000150, loss 1.772384
+INFO 2021-11-07 06:00:09 train.py: 101] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-11-07 06:00:11 train.py: 88] Epoch 11, iter 3000/6416, lr 0.000149, loss 1.770960
+INFO 2021-11-07 06:06:58 train.py: 88] Epoch 11, iter 3200/6416, lr 0.000148, loss 1.729162
+INFO 2021-11-07 06:13:49 train.py: 88] Epoch 11, iter 3400/6416, lr 0.000147, loss 1.742616
+INFO 2021-11-07 06:20:39 train.py: 88] Epoch 11, iter 3600/6416, lr 0.000146, loss 1.732729
+INFO 2021-11-07 06:27:27 train.py: 88] Epoch 11, iter 3800/6416, lr 0.000144, loss 1.679888
+INFO 2021-11-07 06:34:18 train.py: 88] Epoch 11, iter 4000/6416, lr 0.000143, loss 1.744265
+INFO 2021-11-07 06:41:05 train.py: 88] Epoch 11, iter 4200/6416, lr 0.000142, loss 1.667845
+INFO 2021-11-07 06:47:53 train.py: 88] Epoch 11, iter 4400/6416, lr 0.000141, loss 1.692872
+INFO 2021-11-07 06:54:41 train.py: 88] Epoch 11, iter 4600/6416, lr 0.000139, loss 1.700547
+INFO 2021-11-07 07:01:30 train.py: 88] Epoch 11, iter 4800/6416, lr 0.000138, loss 1.686599
+INFO 2021-11-07 07:08:24 train.py: 88] Epoch 11, iter 5000/6416, lr 0.000137, loss 1.699951
+INFO 2021-11-07 07:15:17 train.py: 88] Epoch 11, iter 5200/6416, lr 0.000136, loss 1.681832
+INFO 2021-11-07 07:22:06 train.py: 88] Epoch 11, iter 5400/6416, lr 0.000135, loss 1.657159
+INFO 2021-11-07 07:28:52 train.py: 88] Epoch 11, iter 5600/6416, lr 0.000134, loss 1.661121
+INFO 2021-11-07 07:35:44 train.py: 88] Epoch 11, iter 5800/6416, lr 0.000132, loss 1.640726
+INFO 2021-11-07 07:42:34 train.py: 101] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-11-07 07:42:37 train.py: 88] Epoch 11, iter 6000/6416, lr 0.000131, loss 1.661956
+INFO 2021-11-07 07:49:26 train.py: 88] Epoch 11, iter 6200/6416, lr 0.000130, loss 1.664726
+INFO 2021-11-07 07:56:13 train.py: 88] Epoch 11, iter 6400/6416, lr 0.000129, loss 1.641801
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 07:56:45 train.py: 108] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-11-07 07:56:47 train.py: 88] Epoch 12, iter 0/6416, lr 0.000129, loss 1.635231
+INFO 2021-11-07 08:03:30 train.py: 88] Epoch 12, iter 200/6416, lr 0.000128, loss 1.641065
+INFO 2021-11-07 08:10:17 train.py: 88] Epoch 12, iter 400/6416, lr 0.000126, loss 1.622131
+INFO 2021-11-07 08:17:05 train.py: 88] Epoch 12, iter 600/6416, lr 0.000125, loss 1.645248
+INFO 2021-11-07 08:23:55 train.py: 88] Epoch 12, iter 800/6416, lr 0.000124, loss 1.608267
+INFO 2021-11-07 08:30:47 train.py: 88] Epoch 12, iter 1000/6416, lr 0.000123, loss 1.629783
+INFO 2021-11-07 08:37:39 train.py: 88] Epoch 12, iter 1200/6416, lr 0.000122, loss 1.608606
+INFO 2021-11-07 08:44:22 train.py: 88] Epoch 12, iter 1400/6416, lr 0.000121, loss 1.558950
+INFO 2021-11-07 08:51:15 train.py: 88] Epoch 12, iter 1600/6416, lr 0.000120, loss 1.615986
+INFO 2021-11-07 08:58:03 train.py: 88] Epoch 12, iter 1800/6416, lr 0.000118, loss 1.550936
+INFO 2021-11-07 09:04:49 train.py: 88] Epoch 12, iter 2000/6416, lr 0.000117, loss 1.556552
+INFO 2021-11-07 09:11:38 train.py: 88] Epoch 12, iter 2200/6416, lr 0.000116, loss 1.546049
+INFO 2021-11-07 09:18:27 train.py: 88] Epoch 12, iter 2400/6416, lr 0.000115, loss 1.540118
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 09:25:23 train.py: 88] Epoch 12, iter 2600/6416, lr 0.000114, loss 1.554523
+INFO 2021-11-07 09:32:09 train.py: 88] Epoch 12, iter 2800/6416, lr 0.000113, loss 1.582650
+INFO 2021-11-07 09:39:03 train.py: 101] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-11-07 09:39:05 train.py: 88] Epoch 12, iter 3000/6416, lr 0.000112, loss 1.538779
+INFO 2021-11-07 09:45:51 train.py: 88] Epoch 12, iter 3200/6416, lr 0.000111, loss 1.521763
+INFO 2021-11-07 09:52:42 train.py: 88] Epoch 12, iter 3400/6416, lr 0.000109, loss 1.541065
+INFO 2021-11-07 09:59:33 train.py: 88] Epoch 12, iter 3600/6416, lr 0.000108, loss 1.531970
+INFO 2021-11-07 10:06:15 train.py: 88] Epoch 12, iter 3800/6416, lr 0.000107, loss 1.485012
+INFO 2021-11-07 10:13:03 train.py: 88] Epoch 12, iter 4000/6416, lr 0.000106, loss 1.548506
+INFO 2021-11-07 10:19:52 train.py: 88] Epoch 12, iter 4200/6416, lr 0.000105, loss 1.476993
+INFO 2021-11-07 10:26:40 train.py: 88] Epoch 12, iter 4400/6416, lr 0.000104, loss 1.486938
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 10:33:26 train.py: 88] Epoch 12, iter 4600/6416, lr 0.000103, loss 1.489733
+INFO 2021-11-07 10:40:16 train.py: 88] Epoch 12, iter 4800/6416, lr 0.000102, loss 1.477284
+INFO 2021-11-07 10:46:59 train.py: 88] Epoch 12, iter 5000/6416, lr 0.000101, loss 1.500890
+INFO 2021-11-07 10:53:50 train.py: 88] Epoch 12, iter 5200/6416, lr 0.000100, loss 1.485094
+INFO 2021-11-07 11:00:33 train.py: 88] Epoch 12, iter 5400/6416, lr 0.000099, loss 1.449405
+INFO 2021-11-07 11:07:27 train.py: 88] Epoch 12, iter 5600/6416, lr 0.000098, loss 1.457030
+INFO 2021-11-07 11:14:18 train.py: 88] Epoch 12, iter 5800/6416, lr 0.000097, loss 1.442909
+INFO 2021-11-07 11:21:04 train.py: 101] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-11-07 11:21:07 train.py: 88] Epoch 12, iter 6000/6416, lr 0.000096, loss 1.456157
+INFO 2021-11-07 11:27:55 train.py: 88] Epoch 12, iter 6200/6416, lr 0.000095, loss 1.453426
+INFO 2021-11-07 11:34:40 train.py: 88] Epoch 12, iter 6400/6416, lr 0.000093, loss 1.476685
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-07 11:35:12 train.py: 108] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-11-07 11:35:14 train.py: 88] Epoch 13, iter 0/6416, lr 0.000093, loss 1.430794
+INFO 2021-11-07 11:42:13 train.py: 88] Epoch 13, iter 200/6416, lr 0.000092, loss 1.446974
+INFO 2021-11-07 11:49:08 train.py: 88] Epoch 13, iter 400/6416, lr 0.000091, loss 1.429229
+INFO 2021-11-07 11:55:58 train.py: 88] Epoch 13, iter 600/6416, lr 0.000090, loss 1.456690
+INFO 2021-11-07 12:02:51 train.py: 88] Epoch 13, iter 800/6416, lr 0.000089, loss 1.425573
+INFO 2021-11-07 12:09:36 train.py: 88] Epoch 13, iter 1000/6416, lr 0.000088, loss 1.450044
+INFO 2021-11-07 12:16:22 train.py: 88] Epoch 13, iter 1200/6416, lr 0.000087, loss 1.426263
+INFO 2021-11-07 12:23:07 train.py: 88] Epoch 13, iter 1400/6416, lr 0.000086, loss 1.373677
+INFO 2021-11-07 12:29:48 train.py: 88] Epoch 13, iter 1600/6416, lr 0.000085, loss 1.428967
+INFO 2021-11-07 12:36:36 train.py: 88] Epoch 13, iter 1800/6416, lr 0.000084, loss 1.377697
+INFO 2021-11-07 12:43:17 train.py: 88] Epoch 13, iter 2000/6416, lr 0.000083, loss 1.379910
+INFO 2021-11-07 12:49:58 train.py: 88] Epoch 13, iter 2200/6416, lr 0.000082, loss 1.373246
+INFO 2021-11-07 12:56:43 train.py: 88] Epoch 13, iter 2400/6416, lr 0.000081, loss 1.366846
+INFO 2021-11-07 13:03:33 train.py: 88] Epoch 13, iter 2600/6416, lr 0.000080, loss 1.371527
+INFO 2021-11-07 13:10:14 train.py: 88] Epoch 13, iter 2800/6416, lr 0.000079, loss 1.396763
+INFO 2021-11-07 13:17:01 train.py: 101] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-11-07 13:17:03 train.py: 88] Epoch 13, iter 3000/6416, lr 0.000078, loss 1.356363
+INFO 2021-11-07 13:23:50 train.py: 88] Epoch 13, iter 3200/6416, lr 0.000078, loss 1.349428
+INFO 2021-11-07 13:30:42 train.py: 88] Epoch 13, iter 3400/6416, lr 0.000077, loss 1.342755
+INFO 2021-11-07 13:37:31 train.py: 88] Epoch 13, iter 3600/6416, lr 0.000076, loss 1.339078
+INFO 2021-11-07 13:44:22 train.py: 88] Epoch 13, iter 3800/6416, lr 0.000075, loss 1.307975
+INFO 2021-11-07 13:51:15 train.py: 88] Epoch 13, iter 4000/6416, lr 0.000074, loss 1.361762
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 13:58:04 train.py: 88] Epoch 13, iter 4200/6416, lr 0.000073, loss 1.308019
+INFO 2021-11-07 14:04:47 train.py: 88] Epoch 13, iter 4400/6416, lr 0.000072, loss 1.315427
+INFO 2021-11-07 14:11:32 train.py: 88] Epoch 13, iter 4600/6416, lr 0.000071, loss 1.328882
+INFO 2021-11-07 14:18:25 train.py: 88] Epoch 13, iter 4800/6416, lr 0.000070, loss 1.294228
+INFO 2021-11-07 14:25:13 train.py: 88] Epoch 13, iter 5000/6416, lr 0.000069, loss 1.329374
+INFO 2021-11-07 14:31:58 train.py: 88] Epoch 13, iter 5200/6416, lr 0.000068, loss 1.317093
+INFO 2021-11-07 14:38:43 train.py: 88] Epoch 13, iter 5400/6416, lr 0.000067, loss 1.291690
+INFO 2021-11-07 14:45:26 train.py: 88] Epoch 13, iter 5600/6416, lr 0.000066, loss 1.305910
+INFO 2021-11-07 14:52:19 train.py: 88] Epoch 13, iter 5800/6416, lr 0.000066, loss 1.283950
+INFO 2021-11-07 14:59:03 train.py: 101] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-11-07 14:59:05 train.py: 88] Epoch 13, iter 6000/6416, lr 0.000065, loss 1.298038
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 15:05:59 train.py: 88] Epoch 13, iter 6200/6416, lr 0.000064, loss 1.293516
+INFO 2021-11-07 15:12:44 train.py: 88] Epoch 13, iter 6400/6416, lr 0.000063, loss 1.296222
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-07 15:13:16 train.py: 108] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-11-07 15:13:18 train.py: 88] Epoch 14, iter 0/6416, lr 0.000063, loss 1.259523
+INFO 2021-11-07 15:20:06 train.py: 88] Epoch 14, iter 200/6416, lr 0.000062, loss 1.282866
+INFO 2021-11-07 15:26:51 train.py: 88] Epoch 14, iter 400/6416, lr 0.000061, loss 1.278429
+INFO 2021-11-07 15:33:42 train.py: 88] Epoch 14, iter 600/6416, lr 0.000060, loss 1.304096
+INFO 2021-11-07 15:40:27 train.py: 88] Epoch 14, iter 800/6416, lr 0.000059, loss 1.263985
+INFO 2021-11-07 15:47:20 train.py: 88] Epoch 14, iter 1000/6416, lr 0.000059, loss 1.286399
+INFO 2021-11-07 15:54:11 train.py: 88] Epoch 14, iter 1200/6416, lr 0.000058, loss 1.272722
+INFO 2021-11-07 16:01:06 train.py: 88] Epoch 14, iter 1400/6416, lr 0.000057, loss 1.219060
+INFO 2021-11-07 16:07:57 train.py: 88] Epoch 14, iter 1600/6416, lr 0.000056, loss 1.276336
+INFO 2021-11-07 16:14:45 train.py: 88] Epoch 14, iter 1800/6416, lr 0.000055, loss 1.226711
+INFO 2021-11-07 16:21:32 train.py: 88] Epoch 14, iter 2000/6416, lr 0.000055, loss 1.234672
+INFO 2021-11-07 16:28:21 train.py: 88] Epoch 14, iter 2200/6416, lr 0.000054, loss 1.226858
+INFO 2021-11-07 16:35:14 train.py: 88] Epoch 14, iter 2400/6416, lr 0.000053, loss 1.217566
+INFO 2021-11-07 16:42:04 train.py: 88] Epoch 14, iter 2600/6416, lr 0.000052, loss 1.210259
+INFO 2021-11-07 16:49:00 train.py: 88] Epoch 14, iter 2800/6416, lr 0.000051, loss 1.233618
+INFO 2021-11-07 16:55:52 train.py: 101] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-11-07 16:55:53 train.py: 88] Epoch 14, iter 3000/6416, lr 0.000051, loss 1.224489
+INFO 2021-11-07 17:02:46 train.py: 88] Epoch 14, iter 3200/6416, lr 0.000050, loss 1.206357
+INFO 2021-11-07 17:09:39 train.py: 88] Epoch 14, iter 3400/6416, lr 0.000049, loss 1.206064
+INFO 2021-11-07 17:16:30 train.py: 88] Epoch 14, iter 3600/6416, lr 0.000048, loss 1.199699
+INFO 2021-11-07 17:23:20 train.py: 88] Epoch 14, iter 3800/6416, lr 0.000048, loss 1.161058
+INFO 2021-11-07 17:30:14 train.py: 88] Epoch 14, iter 4000/6416, lr 0.000047, loss 1.217640
+INFO 2021-11-07 17:37:04 train.py: 88] Epoch 14, iter 4200/6416, lr 0.000046, loss 1.174307
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 17:43:55 train.py: 88] Epoch 14, iter 4400/6416, lr 0.000045, loss 1.188214
+INFO 2021-11-07 17:50:45 train.py: 88] Epoch 14, iter 4600/6416, lr 0.000045, loss 1.196620
+INFO 2021-11-07 17:57:41 train.py: 88] Epoch 14, iter 4800/6416, lr 0.000044, loss 1.170810
+INFO 2021-11-07 18:04:36 train.py: 88] Epoch 14, iter 5000/6416, lr 0.000043, loss 1.196106
+INFO 2021-11-07 18:11:30 train.py: 88] Epoch 14, iter 5200/6416, lr 0.000042, loss 1.179906
+INFO 2021-11-07 18:18:27 train.py: 88] Epoch 14, iter 5400/6416, lr 0.000042, loss 1.148387
+INFO 2021-11-07 18:25:23 train.py: 88] Epoch 14, iter 5600/6416, lr 0.000041, loss 1.161857
+INFO 2021-11-07 18:32:16 train.py: 88] Epoch 14, iter 5800/6416, lr 0.000040, loss 1.148571
+INFO 2021-11-07 18:39:08 train.py: 101] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-11-07 18:39:10 train.py: 88] Epoch 14, iter 6000/6416, lr 0.000040, loss 1.171037
+INFO 2021-11-07 18:46:05 train.py: 88] Epoch 14, iter 6200/6416, lr 0.000039, loss 1.177743
+INFO 2021-11-07 18:53:00 train.py: 88] Epoch 14, iter 6400/6416, lr 0.000038, loss 1.167124
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 18:53:32 train.py: 108] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-11-07 18:53:35 train.py: 88] Epoch 15, iter 0/6416, lr 0.000038, loss 1.140286
+INFO 2021-11-07 19:00:29 train.py: 88] Epoch 15, iter 200/6416, lr 0.000037, loss 1.159364
+INFO 2021-11-07 19:07:25 train.py: 88] Epoch 15, iter 400/6416, lr 0.000037, loss 1.141635
+INFO 2021-11-07 19:14:20 train.py: 88] Epoch 15, iter 600/6416, lr 0.000036, loss 1.168675
+INFO 2021-11-07 19:21:15 train.py: 88] Epoch 15, iter 800/6416, lr 0.000036, loss 1.148305
+INFO 2021-11-07 19:28:09 train.py: 88] Epoch 15, iter 1000/6416, lr 0.000035, loss 1.169405
+INFO 2021-11-07 19:35:11 train.py: 88] Epoch 15, iter 1200/6416, lr 0.000034, loss 1.149939
+INFO 2021-11-07 19:42:08 train.py: 88] Epoch 15, iter 1400/6416, lr 0.000034, loss 1.109310
+INFO 2021-11-07 19:49:06 train.py: 88] Epoch 15, iter 1600/6416, lr 0.000033, loss 1.149165
+INFO 2021-11-07 19:56:02 train.py: 88] Epoch 15, iter 1800/6416, lr 0.000032, loss 1.103462
+INFO 2021-11-07 20:03:09 train.py: 88] Epoch 15, iter 2000/6416, lr 0.000032, loss 1.108025
+INFO 2021-11-07 20:10:03 train.py: 88] Epoch 15, iter 2200/6416, lr 0.000031, loss 1.112653
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 20:17:03 train.py: 88] Epoch 15, iter 2400/6416, lr 0.000031, loss 1.108122
+INFO 2021-11-07 20:24:01 train.py: 88] Epoch 15, iter 2600/6416, lr 0.000030, loss 1.101269
+INFO 2021-11-07 20:31:02 train.py: 88] Epoch 15, iter 2800/6416, lr 0.000029, loss 1.133066
+INFO 2021-11-07 20:37:57 train.py: 101] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-11-07 20:37:59 train.py: 88] Epoch 15, iter 3000/6416, lr 0.000029, loss 1.103648
+INFO 2021-11-07 20:44:50 train.py: 88] Epoch 15, iter 3200/6416, lr 0.000028, loss 1.097744
+INFO 2021-11-07 20:51:53 train.py: 88] Epoch 15, iter 3400/6416, lr 0.000028, loss 1.100097
+INFO 2021-11-07 20:58:47 train.py: 88] Epoch 15, iter 3600/6416, lr 0.000027, loss 1.093784
+INFO 2021-11-07 21:05:43 train.py: 88] Epoch 15, iter 3800/6416, lr 0.000027, loss 1.056752
+INFO 2021-11-07 21:12:42 train.py: 88] Epoch 15, iter 4000/6416, lr 0.000026, loss 1.106672
+INFO 2021-11-07 21:19:40 train.py: 88] Epoch 15, iter 4200/6416, lr 0.000025, loss 1.071153
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 21:26:41 train.py: 88] Epoch 15, iter 4400/6416, lr 0.000025, loss 1.081791
+INFO 2021-11-07 21:33:43 train.py: 88] Epoch 15, iter 4600/6416, lr 0.000024, loss 1.094822
+INFO 2021-11-07 21:40:42 train.py: 88] Epoch 15, iter 4800/6416, lr 0.000024, loss 1.068287
+INFO 2021-11-07 21:47:36 train.py: 88] Epoch 15, iter 5000/6416, lr 0.000023, loss 1.094929
+INFO 2021-11-07 21:54:25 train.py: 88] Epoch 15, iter 5200/6416, lr 0.000023, loss 1.083551
+INFO 2021-11-07 22:01:18 train.py: 88] Epoch 15, iter 5400/6416, lr 0.000022, loss 1.053417
+INFO 2021-11-07 22:08:12 train.py: 88] Epoch 15, iter 5600/6416, lr 0.000022, loss 1.066749
+INFO 2021-11-07 22:15:06 train.py: 88] Epoch 15, iter 5800/6416, lr 0.000021, loss 1.060041
+INFO 2021-11-07 22:21:57 train.py: 101] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-11-07 22:21:59 train.py: 88] Epoch 15, iter 6000/6416, lr 0.000021, loss 1.078213
+INFO 2021-11-07 22:28:55 train.py: 88] Epoch 15, iter 6200/6416, lr 0.000020, loss 1.080530
+INFO 2021-11-07 22:35:46 train.py: 88] Epoch 15, iter 6400/6416, lr 0.000020, loss 1.077898
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-07 22:36:18 train.py: 108] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-11-07 22:36:20 train.py: 88] Epoch 16, iter 0/6416, lr 0.000020, loss 1.017158
+INFO 2021-11-07 22:44:38 train.py: 88] Epoch 16, iter 200/6416, lr 0.000019, loss 1.068190
+INFO 2021-11-07 22:52:30 train.py: 88] Epoch 16, iter 400/6416, lr 0.000019, loss 1.050526
+INFO 2021-11-07 22:59:56 train.py: 88] Epoch 16, iter 600/6416, lr 0.000019, loss 1.089178
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-07 23:08:12 train.py: 88] Epoch 16, iter 800/6416, lr 0.000018, loss 1.044952
+INFO 2021-11-07 23:15:59 train.py: 88] Epoch 16, iter 1000/6416, lr 0.000018, loss 1.074805
+INFO 2021-11-07 23:23:25 train.py: 88] Epoch 16, iter 1200/6416, lr 0.000017, loss 1.056154
+INFO 2021-11-07 23:30:46 train.py: 88] Epoch 16, iter 1400/6416, lr 0.000017, loss 1.021646
+INFO 2021-11-07 23:37:54 train.py: 88] Epoch 16, iter 1600/6416, lr 0.000016, loss 1.062962
+INFO 2021-11-07 23:45:05 train.py: 88] Epoch 16, iter 1800/6416, lr 0.000016, loss 1.031611
+INFO 2021-11-07 23:52:15 train.py: 88] Epoch 16, iter 2000/6416, lr 0.000016, loss 1.042913
+INFO 2021-11-07 23:59:46 train.py: 88] Epoch 16, iter 2200/6416, lr 0.000015, loss 1.020936
+INFO 2021-11-08 00:07:12 train.py: 88] Epoch 16, iter 2400/6416, lr 0.000015, loss 1.034180
+INFO 2021-11-08 00:14:38 train.py: 88] Epoch 16, iter 2600/6416, lr 0.000015, loss 1.031583
+INFO 2021-11-08 00:21:34 train.py: 88] Epoch 16, iter 2800/6416, lr 0.000014, loss 1.057401
+INFO 2021-11-08 00:28:45 train.py: 101] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-11-08 00:28:47 train.py: 88] Epoch 16, iter 3000/6416, lr 0.000014, loss 1.036593
+INFO 2021-11-08 00:36:19 train.py: 88] Epoch 16, iter 3200/6416, lr 0.000013, loss 1.019005
+INFO 2021-11-08 00:43:42 train.py: 88] Epoch 16, iter 3400/6416, lr 0.000013, loss 1.033647
+INFO 2021-11-08 00:51:07 train.py: 88] Epoch 16, iter 3600/6416, lr 0.000013, loss 1.026655
+INFO 2021-11-08 00:58:36 train.py: 88] Epoch 16, iter 3800/6416, lr 0.000012, loss 0.996125
+INFO 2021-11-08 01:05:55 train.py: 88] Epoch 16, iter 4000/6416, lr 0.000012, loss 1.045330
+INFO 2021-11-08 01:13:14 train.py: 88] Epoch 16, iter 4200/6416, lr 0.000012, loss 1.010357
+INFO 2021-11-08 01:20:54 train.py: 88] Epoch 16, iter 4400/6416, lr 0.000011, loss 1.025883
+INFO 2021-11-08 01:28:20 train.py: 88] Epoch 16, iter 4600/6416, lr 0.000011, loss 1.022122
+INFO 2021-11-08 01:35:30 train.py: 88] Epoch 16, iter 4800/6416, lr 0.000011, loss 1.006844
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-08 01:43:00 train.py: 88] Epoch 16, iter 5000/6416, lr 0.000011, loss 1.026209
+INFO 2021-11-08 01:50:51 train.py: 88] Epoch 16, iter 5200/6416, lr 0.000010, loss 1.023270
+INFO 2021-11-08 01:58:05 train.py: 88] Epoch 16, iter 5400/6416, lr 0.000010, loss 0.994332
+INFO 2021-11-08 02:05:13 train.py: 88] Epoch 16, iter 5600/6416, lr 0.000010, loss 1.003676
+INFO 2021-11-08 02:12:42 train.py: 88] Epoch 16, iter 5800/6416, lr 0.000010, loss 1.003138
+INFO 2021-11-08 02:20:40 train.py: 101] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-11-08 02:20:42 train.py: 88] Epoch 16, iter 6000/6416, lr 0.000009, loss 1.017384
+INFO 2021-11-08 02:28:14 train.py: 88] Epoch 16, iter 6200/6416, lr 0.000009, loss 1.011799
+INFO 2021-11-08 02:35:28 train.py: 88] Epoch 16, iter 6400/6416, lr 0.000009, loss 1.028454
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-08 02:36:03 train.py: 108] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-11-08 02:36:05 train.py: 88] Epoch 17, iter 0/6416, lr 0.000009, loss 1.017879
+INFO 2021-11-08 02:42:54 train.py: 88] Epoch 17, iter 200/6416, lr 0.000009, loss 1.022501
+INFO 2021-11-08 02:49:43 train.py: 88] Epoch 17, iter 400/6416, lr 0.000008, loss 0.992820
+INFO 2021-11-08 02:56:36 train.py: 88] Epoch 17, iter 600/6416, lr 0.000008, loss 1.030340
+INFO 2021-11-08 03:03:38 train.py: 88] Epoch 17, iter 800/6416, lr 0.000008, loss 0.992674
+INFO 2021-11-08 03:10:45 train.py: 88] Epoch 17, iter 1000/6416, lr 0.000008, loss 1.009708
+INFO 2021-11-08 03:17:48 train.py: 88] Epoch 17, iter 1200/6416, lr 0.000007, loss 1.012903
+INFO 2021-11-08 03:24:49 train.py: 88] Epoch 17, iter 1400/6416, lr 0.000007, loss 0.966573
+INFO 2021-11-08 03:31:53 train.py: 88] Epoch 17, iter 1600/6416, lr 0.000007, loss 1.007244
+INFO 2021-11-08 03:38:58 train.py: 88] Epoch 17, iter 1800/6416, lr 0.000007, loss 0.977010
+INFO 2021-11-08 03:46:04 train.py: 88] Epoch 17, iter 2000/6416, lr 0.000007, loss 0.982413
+INFO 2021-11-08 03:53:07 train.py: 88] Epoch 17, iter 2200/6416, lr 0.000007, loss 0.980783
+INFO 2021-11-08 04:00:11 train.py: 88] Epoch 17, iter 2400/6416, lr 0.000006, loss 0.986870
+INFO 2021-11-08 04:07:14 train.py: 88] Epoch 17, iter 2600/6416, lr 0.000006, loss 0.982425
+INFO 2021-11-08 04:14:13 train.py: 88] Epoch 17, iter 2800/6416, lr 0.000006, loss 1.019400
+INFO 2021-11-08 04:21:14 train.py: 101] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-11-08 04:21:15 train.py: 88] Epoch 17, iter 3000/6416, lr 0.000006, loss 0.992717
+INFO 2021-11-08 04:28:16 train.py: 88] Epoch 17, iter 3200/6416, lr 0.000006, loss 0.988175
+INFO 2021-11-08 04:35:23 train.py: 88] Epoch 17, iter 3400/6416, lr 0.000006, loss 0.990665
+INFO 2021-11-08 04:42:27 train.py: 88] Epoch 17, iter 3600/6416, lr 0.000006, loss 0.991908
+INFO 2021-11-08 04:49:32 train.py: 88] Epoch 17, iter 3800/6416, lr 0.000006, loss 0.952285
+INFO 2021-11-08 04:56:37 train.py: 88] Epoch 17, iter 4000/6416, lr 0.000006, loss 1.004719
+INFO 2021-11-08 05:03:47 train.py: 88] Epoch 17, iter 4200/6416, lr 0.000005, loss 0.964050
+INFO 2021-11-08 05:10:44 train.py: 88] Epoch 17, iter 4400/6416, lr 0.000005, loss 0.990251
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-11-08 05:17:49 train.py: 88] Epoch 17, iter 4600/6416, lr 0.000005, loss 0.992029
+INFO 2021-11-08 05:24:55 train.py: 88] Epoch 17, iter 4800/6416, lr 0.000005, loss 0.984346
+INFO 2021-11-08 05:31:58 train.py: 88] Epoch 17, iter 5000/6416, lr 0.000005, loss 0.991912
+INFO 2021-11-08 05:39:02 train.py: 88] Epoch 17, iter 5200/6416, lr 0.000005, loss 0.987374
+INFO 2021-11-08 05:46:05 train.py: 88] Epoch 17, iter 5400/6416, lr 0.000005, loss 0.970941
+INFO 2021-11-08 05:53:13 train.py: 88] Epoch 17, iter 5600/6416, lr 0.000005, loss 0.988832
+INFO 2021-11-08 06:00:22 train.py: 88] Epoch 17, iter 5800/6416, lr 0.000005, loss 0.964585
+INFO 2021-11-08 06:07:28 train.py: 101] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-11-08 06:07:30 train.py: 88] Epoch 17, iter 6000/6416, lr 0.000005, loss 0.984010
+INFO 2021-11-08 06:14:30 train.py: 88] Epoch 17, iter 6200/6416, lr 0.000005, loss 0.994161
+INFO 2021-11-08 06:21:30 train.py: 88] Epoch 17, iter 6400/6416, lr 0.000005, loss 0.993996
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-11-08 06:22:04 train.py: 108] Save checkpoint Epoch_17.pt to disk...
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/.gitkeep b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/Swin_Transformer.py b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/Swin_Transformer.py
new file mode 100644
index 0000000000000000000000000000000000000000..f42c1a96ca445fa8e236c953bb63458fee2690ee
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/Swin_Transformer.py
@@ -0,0 +1,593 @@
+# --------------------------------------------------------
+# Swin Transformer
+# Copyright (c) 2021 Microsoft
+# Licensed under The MIT License [see LICENSE for details]
+# Written by Ze Liu
+# --------------------------------------------------------
+
+import torch
+import torch.nn as nn
+import torch.utils.checkpoint as checkpoint
+from timm.models.layers import DropPath, to_2tuple, trunc_normal_
+
+
+class Flatten(nn.Module):
+ def forward(self, input):
+ return input.view(input.size(0), -1)
+
+
+class Mlp(nn.Module):
+ def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
+ super().__init__()
+ out_features = out_features or in_features
+ hidden_features = hidden_features or in_features
+ self.fc1 = nn.Linear(in_features, hidden_features)
+ self.act = act_layer()
+ self.fc2 = nn.Linear(hidden_features, out_features)
+ self.drop = nn.Dropout(drop)
+
+ def forward(self, x):
+ x = self.fc1(x)
+ x = self.act(x)
+ x = self.drop(x)
+ x = self.fc2(x)
+ x = self.drop(x)
+ return x
+
+
+def window_partition(x, window_size):
+ """
+ Args:
+ x: (B, H, W, C)
+ window_size (int): window size
+
+ Returns:
+ windows: (num_windows*B, window_size, window_size, C)
+ """
+ B, H, W, C = x.shape
+ x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
+ windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
+ return windows
+
+
+def window_reverse(windows, window_size, H, W):
+ """
+ Args:
+ windows: (num_windows*B, window_size, window_size, C)
+ window_size (int): Window size
+ H (int): Height of image
+ W (int): Width of image
+
+ Returns:
+ x: (B, H, W, C)
+ """
+ B = int(windows.shape[0] / (H * W / window_size / window_size))
+ x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
+ x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
+ return x
+
+
+class WindowAttention(nn.Module):
+ r""" Window based multi-head self attention (W-MSA) module with relative position bias.
+ It supports both of shifted and non-shifted window.
+
+ Args:
+ dim (int): Number of input channels.
+ window_size (tuple[int]): The height and width of the window.
+ num_heads (int): Number of attention heads.
+ qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set
+ attn_drop (float, optional): Dropout ratio of attention weight. Default: 0.0
+ proj_drop (float, optional): Dropout ratio of output. Default: 0.0
+ """
+
+ def __init__(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0.):
+
+ super().__init__()
+ self.dim = dim
+ self.window_size = window_size # Wh, Ww
+ self.num_heads = num_heads
+ head_dim = dim // num_heads
+ self.scale = qk_scale or head_dim ** -0.5
+
+ # define a parameter table of relative position bias
+ self.relative_position_bias_table = nn.Parameter(
+ torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads)) # 2*Wh-1 * 2*Ww-1, nH
+
+ # get pair-wise relative position index for each token inside the window
+ coords_h = torch.arange(self.window_size[0])
+ coords_w = torch.arange(self.window_size[1])
+ coords = torch.stack(torch.meshgrid([coords_h, coords_w])) # 2, Wh, Ww
+ coords_flatten = torch.flatten(coords, 1) # 2, Wh*Ww
+ relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :] # 2, Wh*Ww, Wh*Ww
+ relative_coords = relative_coords.permute(1, 2, 0).contiguous() # Wh*Ww, Wh*Ww, 2
+ relative_coords[:, :, 0] += self.window_size[0] - 1 # shift to start from 0
+ relative_coords[:, :, 1] += self.window_size[1] - 1
+ relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1
+ relative_position_index = relative_coords.sum(-1) # Wh*Ww, Wh*Ww
+ self.register_buffer("relative_position_index", relative_position_index)
+
+ self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
+ self.attn_drop = nn.Dropout(attn_drop)
+ self.proj = nn.Linear(dim, dim)
+ self.proj_drop = nn.Dropout(proj_drop)
+
+ trunc_normal_(self.relative_position_bias_table, std=.02)
+ self.softmax = nn.Softmax(dim=-1)
+
+ def forward(self, x, mask=None):
+ """
+ Args:
+ x: input features with shape of (num_windows*B, N, C)
+ mask: (0/-inf) mask with shape of (num_windows, Wh*Ww, Wh*Ww) or None
+ """
+ B_, N, C = x.shape
+ qkv = self.qkv(x).reshape(B_, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
+ q, k, v = qkv[0], qkv[1], qkv[2] # make torchscript happy (cannot use tensor as tuple)
+
+ q = q * self.scale
+ attn = (q @ k.transpose(-2, -1))
+
+ relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(
+ self.window_size[0] * self.window_size[1], self.window_size[0] * self.window_size[1], -1) # Wh*Ww,Wh*Ww,nH
+ relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous() # nH, Wh*Ww, Wh*Ww
+ attn = attn + relative_position_bias.unsqueeze(0)
+
+ if mask is not None:
+ nW = mask.shape[0]
+ attn = attn.view(B_ // nW, nW, self.num_heads, N, N) + mask.unsqueeze(1).unsqueeze(0)
+ attn = attn.view(-1, self.num_heads, N, N)
+ attn = self.softmax(attn)
+ else:
+ attn = self.softmax(attn)
+
+ attn = self.attn_drop(attn)
+
+ x = (attn @ v).transpose(1, 2).reshape(B_, N, C)
+ x = self.proj(x)
+ x = self.proj_drop(x)
+ return x
+
+ def extra_repr(self) -> str:
+ return f'dim={self.dim}, window_size={self.window_size}, num_heads={self.num_heads}'
+
+ def flops(self, N):
+ # calculate flops for 1 window with token length of N
+ flops = 0
+ # qkv = self.qkv(x)
+ flops += N * self.dim * 3 * self.dim
+ # attn = (q @ k.transpose(-2, -1))
+ flops += self.num_heads * N * (self.dim // self.num_heads) * N
+ # x = (attn @ v)
+ flops += self.num_heads * N * N * (self.dim // self.num_heads)
+ # x = self.proj(x)
+ flops += N * self.dim * self.dim
+ return flops
+
+
+class SwinTransformerBlock(nn.Module):
+ r""" Swin Transformer Block.
+
+ Args:
+ dim (int): Number of input channels.
+ input_resolution (tuple[int]): Input resulotion.
+ num_heads (int): Number of attention heads.
+ window_size (int): Window size.
+ shift_size (int): Shift size for SW-MSA.
+ mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+ qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
+ drop (float, optional): Dropout rate. Default: 0.0
+ attn_drop (float, optional): Attention dropout rate. Default: 0.0
+ drop_path (float, optional): Stochastic depth rate. Default: 0.0
+ act_layer (nn.Module, optional): Activation layer. Default: nn.GELU
+ norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+ """
+
+ def __init__(self, dim, input_resolution, num_heads, window_size=7, shift_size=0,
+ mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0., drop_path=0.,
+ act_layer=nn.GELU, norm_layer=nn.LayerNorm):
+ super().__init__()
+ self.dim = dim
+ self.input_resolution = input_resolution
+ self.num_heads = num_heads
+ self.window_size = window_size
+ self.shift_size = shift_size
+ self.mlp_ratio = mlp_ratio
+ if min(self.input_resolution) <= self.window_size:
+ # if window size is larger than input resolution, we don't partition windows
+ self.shift_size = 0
+ self.window_size = min(self.input_resolution)
+ assert 0 <= self.shift_size < self.window_size, "shift_size must in 0-window_size"
+
+ self.norm1 = norm_layer(dim)
+ self.attn = WindowAttention(
+ dim, window_size=to_2tuple(self.window_size), num_heads=num_heads,
+ qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop)
+
+ self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
+ self.norm2 = norm_layer(dim)
+ mlp_hidden_dim = int(dim * mlp_ratio)
+ self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)
+
+ if self.shift_size > 0:
+ # calculate attention mask for SW-MSA
+ H, W = self.input_resolution
+ img_mask = torch.zeros((1, H, W, 1)) # 1 H W 1
+ h_slices = (slice(0, -self.window_size),
+ slice(-self.window_size, -self.shift_size),
+ slice(-self.shift_size, None))
+ w_slices = (slice(0, -self.window_size),
+ slice(-self.window_size, -self.shift_size),
+ slice(-self.shift_size, None))
+ cnt = 0
+ for h in h_slices:
+ for w in w_slices:
+ img_mask[:, h, w, :] = cnt
+ cnt += 1
+
+ mask_windows = window_partition(img_mask, self.window_size) # nW, window_size, window_size, 1
+ mask_windows = mask_windows.view(-1, self.window_size * self.window_size)
+ attn_mask = mask_windows.unsqueeze(1) - mask_windows.unsqueeze(2)
+ attn_mask = attn_mask.masked_fill(attn_mask != 0, float(-100.0)).masked_fill(attn_mask == 0, float(0.0))
+ else:
+ attn_mask = None
+
+ self.register_buffer("attn_mask", attn_mask)
+
+ def forward(self, x):
+ H, W = self.input_resolution
+ B, L, C = x.shape
+ assert L == H * W, "input feature has wrong size"
+
+ shortcut = x
+ x = self.norm1(x)
+ x = x.view(B, H, W, C)
+
+ # cyclic shift
+ if self.shift_size > 0:
+ shifted_x = torch.roll(x, shifts=(-self.shift_size, -self.shift_size), dims=(1, 2))
+ else:
+ shifted_x = x
+
+ # partition windows
+ x_windows = window_partition(shifted_x, self.window_size) # nW*B, window_size, window_size, C
+ x_windows = x_windows.view(-1, self.window_size * self.window_size, C) # nW*B, window_size*window_size, C
+
+ # W-MSA/SW-MSA
+ attn_windows = self.attn(x_windows, mask=self.attn_mask) # nW*B, window_size*window_size, C
+
+ # merge windows
+ attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
+ shifted_x = window_reverse(attn_windows, self.window_size, H, W) # B H' W' C
+
+ # reverse cyclic shift
+ if self.shift_size > 0:
+ x = torch.roll(shifted_x, shifts=(self.shift_size, self.shift_size), dims=(1, 2))
+ else:
+ x = shifted_x
+ x = x.view(B, H * W, C)
+
+ # FFN
+ x = shortcut + self.drop_path(x)
+ x = x + self.drop_path(self.mlp(self.norm2(x)))
+
+ return x
+
+ def extra_repr(self) -> str:
+ return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
+ f"window_size={self.window_size}, shift_size={self.shift_size}, mlp_ratio={self.mlp_ratio}"
+
+ def flops(self):
+ flops = 0
+ H, W = self.input_resolution
+ # norm1
+ flops += self.dim * H * W
+ # W-MSA/SW-MSA
+ nW = H * W / self.window_size / self.window_size
+ flops += nW * self.attn.flops(self.window_size * self.window_size)
+ # mlp
+ flops += 2 * H * W * self.dim * self.dim * self.mlp_ratio
+ # norm2
+ flops += self.dim * H * W
+ return flops
+
+
+class PatchMerging(nn.Module):
+ r""" Patch Merging Layer.
+
+ Args:
+ input_resolution (tuple[int]): Resolution of input feature.
+ dim (int): Number of input channels.
+ norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+ """
+
+ def __init__(self, input_resolution, dim, norm_layer=nn.LayerNorm):
+ super().__init__()
+ self.input_resolution = input_resolution
+ self.dim = dim
+ self.reduction = nn.Linear(4 * dim, 2 * dim, bias=False)
+ self.norm = norm_layer(4 * dim)
+
+ def forward(self, x):
+ """
+ x: B, H*W, C
+ """
+ H, W = self.input_resolution
+ B, L, C = x.shape
+ assert L == H * W, "input feature has wrong size"
+ assert H % 2 == 0 and W % 2 == 0, f"x size ({H}*{W}) are not even."
+
+ x = x.view(B, H, W, C)
+
+ x0 = x[:, 0::2, 0::2, :] # B H/2 W/2 C
+ x1 = x[:, 1::2, 0::2, :] # B H/2 W/2 C
+ x2 = x[:, 0::2, 1::2, :] # B H/2 W/2 C
+ x3 = x[:, 1::2, 1::2, :] # B H/2 W/2 C
+ x = torch.cat([x0, x1, x2, x3], -1) # B H/2 W/2 4*C
+ x = x.view(B, -1, 4 * C) # B H/2*W/2 4*C
+
+ x = self.norm(x)
+ x = self.reduction(x)
+
+ return x
+
+ def extra_repr(self) -> str:
+ return f"input_resolution={self.input_resolution}, dim={self.dim}"
+
+ def flops(self):
+ H, W = self.input_resolution
+ flops = H * W * self.dim
+ flops += (H // 2) * (W // 2) * 4 * self.dim * 2 * self.dim
+ return flops
+
+
+class BasicLayer(nn.Module):
+ """ A basic Swin Transformer layer for one stage.
+
+ Args:
+ dim (int): Number of input channels.
+ input_resolution (tuple[int]): Input resolution.
+ depth (int): Number of blocks.
+ num_heads (int): Number of attention heads.
+ window_size (int): Local window size.
+ mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+ qkv_bias (bool, optional): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float | None, optional): Override default qk scale of head_dim ** -0.5 if set.
+ drop (float, optional): Dropout rate. Default: 0.0
+ attn_drop (float, optional): Attention dropout rate. Default: 0.0
+ drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
+ norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
+ downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
+ use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
+ """
+
+ def __init__(self, dim, input_resolution, depth, num_heads, window_size,
+ mlp_ratio=4., qkv_bias=True, qk_scale=None, drop=0., attn_drop=0.,
+ drop_path=0., norm_layer=nn.LayerNorm, downsample=None, use_checkpoint=False):
+
+ super().__init__()
+ self.dim = dim
+ self.input_resolution = input_resolution
+ self.depth = depth
+ self.use_checkpoint = use_checkpoint
+
+ # build blocks
+ self.blocks = nn.ModuleList([
+ SwinTransformerBlock(dim=dim, input_resolution=input_resolution,
+ num_heads=num_heads, window_size=window_size,
+ shift_size=0 if (i % 2 == 0) else window_size // 2,
+ mlp_ratio=mlp_ratio,
+ qkv_bias=qkv_bias, qk_scale=qk_scale,
+ drop=drop, attn_drop=attn_drop,
+ drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path,
+ norm_layer=norm_layer)
+ for i in range(depth)])
+
+ # patch merging layer
+ if downsample is not None:
+ self.downsample = downsample(input_resolution, dim=dim, norm_layer=norm_layer)
+ else:
+ self.downsample = None
+
+ def forward(self, x):
+ for blk in self.blocks:
+ if self.use_checkpoint:
+ x = checkpoint.checkpoint(blk, x)
+ else:
+ x = blk(x)
+ if self.downsample is not None:
+ x = self.downsample(x)
+ return x
+
+ def extra_repr(self) -> str:
+ return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
+
+ def flops(self):
+ flops = 0
+ for blk in self.blocks:
+ flops += blk.flops()
+ if self.downsample is not None:
+ flops += self.downsample.flops()
+ return flops
+
+
+class PatchEmbed(nn.Module):
+ r""" Image to Patch Embedding
+
+ Args:
+ img_size (int): Image size. Default: 224.
+ patch_size (int): Patch token size. Default: 4.
+ in_chans (int): Number of input image channels. Default: 3.
+ embed_dim (int): Number of linear projection output channels. Default: 96.
+ norm_layer (nn.Module, optional): Normalization layer. Default: None
+ """
+
+ def __init__(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None):
+ super().__init__()
+ img_size = to_2tuple(img_size)
+ patch_size = to_2tuple(patch_size)
+ patches_resolution = [img_size[0] // patch_size[0], img_size[1] // patch_size[1]]
+ self.img_size = img_size
+ self.patch_size = patch_size
+ self.patches_resolution = patches_resolution
+ self.num_patches = patches_resolution[0] * patches_resolution[1]
+
+ self.in_chans = in_chans
+ self.embed_dim = embed_dim
+
+ self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)
+ if norm_layer is not None:
+ self.norm = norm_layer(embed_dim)
+ else:
+ self.norm = None
+
+ def forward(self, x):
+ B, C, H, W = x.shape
+ # FIXME look at relaxing size constraints
+ assert H == self.img_size[0] and W == self.img_size[1], \
+ f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})."
+ x = self.proj(x).flatten(2).transpose(1, 2) # B Ph*Pw C
+ if self.norm is not None:
+ x = self.norm(x)
+ return x
+
+ def flops(self):
+ Ho, Wo = self.patches_resolution
+ flops = Ho * Wo * self.embed_dim * self.in_chans * (self.patch_size[0] * self.patch_size[1])
+ if self.norm is not None:
+ flops += Ho * Wo * self.embed_dim
+ return flops
+
+
+class SwinTransformer(nn.Module):
+ r""" Swin Transformer
+ A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` -
+ https://arxiv.org/pdf/2103.14030
+
+ Args:
+ img_size (int | tuple(int)): Input image size. Default 224
+ patch_size (int | tuple(int)): Patch size. Default: 4
+ in_chans (int): Number of input image channels. Default: 3
+ num_classes (int): Number of classes for classification head. Default: 1000
+ embed_dim (int): Patch embedding dimension. Default: 96
+ depths (tuple(int)): Depth of each Swin Transformer layer.
+ num_heads (tuple(int)): Number of attention heads in different layers.
+ window_size (int): Window size. Default: 7
+ mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4
+ qkv_bias (bool): If True, add a learnable bias to query, key, value. Default: True
+ qk_scale (float): Override default qk scale of head_dim ** -0.5 if set. Default: None
+ drop_rate (float): Dropout rate. Default: 0
+ attn_drop_rate (float): Attention dropout rate. Default: 0
+ drop_path_rate (float): Stochastic depth rate. Default: 0.1
+ norm_layer (nn.Module): Normalization layer. Default: nn.LayerNorm.
+ ape (bool): If True, add absolute position embedding to the patch embedding. Default: False
+ patch_norm (bool): If True, add normalization after patch embedding. Default: True
+ use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False
+ """
+
+ def __init__(self, img_size=224, patch_size=4, in_chans=3,
+ embed_dim=96, depths=[2, 2, 6, 2], num_heads=[3, 6, 12, 24],
+ window_size=7, mlp_ratio=4., qkv_bias=True, qk_scale=None,
+ drop_rate=0., attn_drop_rate=0., drop_path_rate=0.1,
+ norm_layer=nn.LayerNorm, ape=False, patch_norm=True,
+ use_checkpoint=False, **kwargs):
+ super().__init__()
+
+ self.num_layers = len(depths)
+ self.embed_dim = embed_dim
+ self.ape = ape
+ self.patch_norm = patch_norm
+ self.num_features = int(embed_dim * 2 ** (self.num_layers - 1))
+ self.mlp_ratio = mlp_ratio
+
+ # split image into non-overlapping patches
+ self.patch_embed = PatchEmbed(
+ img_size=img_size, patch_size=patch_size, in_chans=in_chans, embed_dim=embed_dim,
+ norm_layer=norm_layer if self.patch_norm else None)
+ num_patches = self.patch_embed.num_patches
+ patches_resolution = self.patch_embed.patches_resolution
+ self.patches_resolution = patches_resolution
+
+ # absolute position embedding
+ if self.ape:
+ self.absolute_pos_embed = nn.Parameter(torch.zeros(1, num_patches, embed_dim))
+ trunc_normal_(self.absolute_pos_embed, std=.02)
+
+ self.pos_drop = nn.Dropout(p=drop_rate)
+
+ # stochastic depth
+ dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depths))] # stochastic depth decay rule
+
+ # build layers
+ self.layers = nn.ModuleList()
+ for i_layer in range(self.num_layers):
+ layer = BasicLayer(dim=int(embed_dim * 2 ** i_layer),
+ input_resolution=(patches_resolution[0] // (2 ** i_layer),
+ patches_resolution[1] // (2 ** i_layer)),
+ depth=depths[i_layer],
+ num_heads=num_heads[i_layer],
+ window_size=window_size,
+ mlp_ratio=self.mlp_ratio,
+ qkv_bias=qkv_bias, qk_scale=qk_scale,
+ drop=drop_rate, attn_drop=attn_drop_rate,
+ drop_path=dpr[sum(depths[:i_layer]):sum(depths[:i_layer + 1])],
+ norm_layer=norm_layer,
+ downsample=PatchMerging if (i_layer < self.num_layers - 1) else None,
+ use_checkpoint=use_checkpoint)
+ self.layers.append(layer)
+
+ #self.norm = norm_layer(self.num_features)
+ #self.avgpool = nn.AdaptiveAvgPool1d(1)
+ self.output_layer = nn.Sequential(norm_layer(self.num_features),
+ Flatten(),
+ nn.Linear(49*768, 512),
+ nn.BatchNorm1d(512))
+
+ self.apply(self._init_weights)
+
+ def _init_weights(self, m):
+ if isinstance(m, nn.Linear):
+ trunc_normal_(m.weight, std=.02)
+ if isinstance(m, nn.Linear) and m.bias is not None:
+ nn.init.constant_(m.bias, 0)
+ elif isinstance(m, nn.LayerNorm):
+ nn.init.constant_(m.bias, 0)
+ nn.init.constant_(m.weight, 1.0)
+
+ @torch.jit.ignore
+ def no_weight_decay(self):
+ return {'absolute_pos_embed'}
+
+ @torch.jit.ignore
+ def no_weight_decay_keywords(self):
+ return {'relative_position_bias_table'}
+
+ def forward_features(self, x):
+ x = self.patch_embed(x)
+ if self.ape:
+ x = x + self.absolute_pos_embed
+ x = self.pos_drop(x)
+
+ for layer in self.layers:
+ x = layer(x)
+
+ #x = self.norm(x) # B L C --> [128, 49, 768]
+ #x = self.avgpool(x.transpose(1, 2)) # B C 1 --> [128, 768, 1]
+ #x = torch.flatten(x, 1)
+ x = self.output_layer(x)
+ return x
+
+ def forward(self, x):
+ x = self.forward_features(x) #[128,768]
+ return x
+ '''
+ def flops(self):
+ flops = 0
+ flops += self.patch_embed.flops()
+ for i, layer in enumerate(self.layers):
+ flops += layer.flops()
+ flops += self.num_features * self.patches_resolution[0] * self.patches_resolution[1] // (2 ** self.num_layers)
+ flops += self.num_features * self.num_classes
+ return flops
+ '''
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..13438aa8be7f3b742edcf76ae1c6eb6da8c14ab7
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13.pt | 0.9789999999999999 | 0.0020964402515681337 |
+| Epoch_15_batch_2999.pt | 0.9789999999999999 | 0.002051798368068822 |
+| Epoch_13_batch_5999.pt | 0.9788333333333332 | 0.002166666666666662 |
+| Epoch_15_batch_5999.pt | 0.9785 | 0.0023233915184807195 |
+| Epoch_17_batch_5999.pt | 0.9784999999999998 | 0.0023366378716458977 |
+| Epoch_14_batch_5999.pt | 0.9783333333333333 | 0.002151657414559674 |
+| Epoch_12_batch_2999.pt | 0.9776666666666667 | 0.002320068113091231 |
+| Epoch_17.pt | 0.9775 | 0.002386303510546058 |
+| Epoch_16_batch_5999.pt | 0.9774999999999998 | 0.0023207331749117116 |
+| Epoch_14_batch_2999.pt | 0.9773333333333334 | 0.002346523564660313 |
+| Epoch_15.pt | 0.9773333333333334 | 0.002536158269002962 |
+| Epoch_16_batch_2999.pt | 0.9771666666666668 | 0.0022641870969238665 |
+| Epoch_16.pt | 0.9771666666666666 | 0.0023313483620396964 |
+| Epoch_14.pt | 0.977 | 0.002380476142847612 |
+| Epoch_12.pt | 0.9766666666666666 | 0.001956312984628775 |
+| Epoch_12_batch_5999.pt | 0.9765 | 0.002452637783849342 |
+| Epoch_17_batch_2999.pt | 0.9765 | 0.0021293075440818616 |
+| Epoch_13_batch_2999.pt | 0.9763333333333334 | 0.002233305693582418 |
+| Epoch_11_batch_5999.pt | 0.9763333333333334 | 0.002662033011269096 |
+| Epoch_10_batch_5999.pt | 0.9763333333333332 | 0.00227438772116207 |
+| Epoch_11_batch_2999.pt | 0.9761666666666666 | 0.0021379868227659796 |
+| Epoch_10_batch_2999.pt | 0.9758333333333333 | 0.0021118419787498 |
+| Epoch_9_batch_2999.pt | 0.9756666666666666 | 0.0020816659994661296 |
+| Epoch_9_batch_5999.pt | 0.9756666666666666 | 0.002735229694464701 |
+| Epoch_8_batch_5999.pt | 0.9754999999999999 | 0.0021808651584878298 |
+| Epoch_11.pt | 0.9754999999999999 | 0.002422247706287954 |
+| Epoch_7_batch_2999.pt | 0.975 | 0.0018425693279752213 |
+| Epoch_7_batch_5999.pt | 0.9746666666666666 | 0.0025190631219454704 |
+| Epoch_7.pt | 0.9744999999999999 | 0.002472690342696427 |
+| Epoch_10.pt | 0.9744999999999999 | 0.002357677241546987 |
+| Epoch_8.pt | 0.9741666666666667 | 0.0025849755831404055 |
+| Epoch_9.pt | 0.9736666666666667 | 0.0024695678634325405 |
+| Epoch_8_batch_2999.pt | 0.9733333333333334 | 0.0017033010796395469 |
+| Epoch_6_batch_2999.pt | 0.9726666666666665 | 0.0018790593916986357 |
+| Epoch_5_batch_5999.pt | 0.9716666666666667 | 0.002434322477800738 |
+| Epoch_6_batch_5999.pt | 0.9715 | 0.0022284634577923994 |
+| Epoch_5_batch_2999.pt | 0.9713333333333333 | 0.0023147407395555193 |
+| Epoch_6.pt | 0.9709999999999999 | 0.0028674417556808773 |
+| Epoch_5.pt | 0.9704999999999998 | 0.0032207851201412783 |
+| Epoch_4.pt | 0.9693333333333334 | 0.002619961360567027 |
+| Epoch_4_batch_5999.pt | 0.9691666666666668 | 0.0025123153454655596 |
+| Epoch_4_batch_2999.pt | 0.9676666666666666 | 0.002372684056006962 |
+| Epoch_3_batch_5999.pt | 0.9656666666666667 | 0.002320068113091238 |
+| Epoch_3.pt | 0.9645000000000001 | 0.002534332161768239 |
+| Epoch_3_batch_2999.pt | 0.9621666666666666 | 0.002019197982820704 |
+| Epoch_2_batch_5999.pt | 0.9603333333333332 | 0.0022743877211620816 |
+| Epoch_2.pt | 0.9586666666666666 | 0.0026851213274654622 |
+| Epoch_2_batch_2999.pt | 0.9526666666666666 | 0.0034084137000395453 |
+| Epoch_1_batch_5999.pt | 0.942 | 0.005245809615161278 |
+| Epoch_1.pt | 0.9388333333333334 | 0.004694690986294 |
+| Epoch_1_batch_2999.pt | 0.9116666666666665 | 0.005962847939999437 |
+| Epoch_0.pt | 0.7963333333333333 | 0.005504207145111483 |
+| Epoch_0_batch_5999.pt | 0.7736666666666666 | 0.006911254017815107 |
+| Epoch_0_batch_2999.pt | 0.537 | 0.008351092188494912 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..8723cf0334fff7f7e44a1e815d727cb054853235
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9555000000000001 | 0.0038733817807896347 |
+| Epoch_14_batch_2999.pt | 0.9551666666666667 | 0.003595693583396535 |
+| Epoch_12.pt | 0.9550000000000001 | 0.0035918288611654714 |
+| Epoch_11.pt | 0.9548333333333334 | 0.003978724282176413 |
+| Epoch_14.pt | 0.9546666666666667 | 0.003632840605393736 |
+| Epoch_13_batch_2999.pt | 0.9546666666666667 | 0.003546864377669419 |
+| Epoch_13.pt | 0.9546666666666667 | 0.003815174380753196 |
+| Epoch_14_batch_5999.pt | 0.9546666666666667 | 0.0038232556742411714 |
+| Epoch_12_batch_5999.pt | 0.9543333333333333 | 0.003882534491034705 |
+| Epoch_10_batch_2999.pt | 0.9543333333333333 | 0.0038425814368423573 |
+| Epoch_16_batch_2999.pt | 0.9541666666666668 | 0.0037205634778340506 |
+| Epoch_17_batch_5999.pt | 0.9541666666666666 | 0.003593976442141304 |
+| Epoch_16.pt | 0.9541666666666666 | 0.003745367509040709 |
+| Epoch_15_batch_2999.pt | 0.9540000000000001 | 0.0036447154370792666 |
+| Epoch_11_batch_5999.pt | 0.9540000000000001 | 0.0038184089505421286 |
+| Epoch_11_batch_2999.pt | 0.9538333333333334 | 0.0039678491857042435 |
+| Epoch_10.pt | 0.9538333333333334 | 0.004105476619298288 |
+| Epoch_17_batch_2999.pt | 0.9533333333333334 | 0.003896817314833374 |
+| Epoch_12_batch_2999.pt | 0.9533333333333334 | 0.003975231959999623 |
+| Epoch_17.pt | 0.9533333333333331 | 0.003936219907537408 |
+| Epoch_13_batch_5999.pt | 0.9531666666666666 | 0.003939746811442535 |
+| Epoch_15_batch_5999.pt | 0.9531666666666666 | 0.003939746811442537 |
+| Epoch_10_batch_5999.pt | 0.9531666666666666 | 0.004017323597731322 |
+| Epoch_16_batch_5999.pt | 0.9528333333333332 | 0.0038813418832685733 |
+| Epoch_9.pt | 0.9525 | 0.004091922184071176 |
+| Epoch_8_batch_5999.pt | 0.9523333333333334 | 0.003559026084010433 |
+| Epoch_9_batch_5999.pt | 0.9523333333333334 | 0.0038586123009300773 |
+| Epoch_6.pt | 0.9521666666666666 | 0.004388888888888891 |
+| Epoch_7.pt | 0.952 | 0.004163331998932267 |
+| Epoch_7_batch_5999.pt | 0.952 | 0.003926799343793843 |
+| Epoch_8.pt | 0.9516666666666665 | 0.0044025806124797645 |
+| Epoch_8_batch_2999.pt | 0.9506666666666668 | 0.00447351603093276 |
+| Epoch_9_batch_2999.pt | 0.9506666666666665 | 0.004015402444353923 |
+| Epoch_7_batch_2999.pt | 0.9504999999999999 | 0.0038733817807896356 |
+| Epoch_6_batch_2999.pt | 0.9496666666666667 | 0.004309048762147846 |
+| Epoch_5.pt | 0.9494999999999999 | 0.004075294430020738 |
+| Epoch_5_batch_5999.pt | 0.9494999999999999 | 0.003651906317676626 |
+| Epoch_6_batch_5999.pt | 0.9488333333333333 | 0.0033797691498344703 |
+| Epoch_4.pt | 0.9476666666666669 | 0.004076430295076477 |
+| Epoch_4_batch_5999.pt | 0.9471666666666666 | 0.0042167362020324286 |
+| Epoch_5_batch_2999.pt | 0.9470000000000001 | 0.004222222222222223 |
+| Epoch_4_batch_2999.pt | 0.9465 | 0.00439170092754926 |
+| Epoch_3_batch_5999.pt | 0.9450000000000001 | 0.004082482904638631 |
+| Epoch_3.pt | 0.9431666666666667 | 0.003955383891406127 |
+| Epoch_3_batch_2999.pt | 0.9423333333333334 | 0.004487293445846371 |
+| Epoch_2_batch_5999.pt | 0.9391666666666667 | 0.004084372505713955 |
+| Epoch_2.pt | 0.9388333333333334 | 0.003514081022415213 |
+| Epoch_2_batch_2999.pt | 0.9355 | 0.004157768276015032 |
+| Epoch_1_batch_5999.pt | 0.9248333333333333 | 0.004577521567803228 |
+| Epoch_1.pt | 0.9245000000000001 | 0.004097951912424451 |
+| Epoch_1_batch_2999.pt | 0.9048333333333334 | 0.0036213529972738347 |
+| Epoch_0.pt | 0.8025 | 0.00537053001039866 |
+| Epoch_0_batch_5999.pt | 0.7623333333333333 | 0.00561193613296639 |
+| Epoch_0_batch_2999.pt | 0.5131666666666668 | 0.004833333333333333 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..716edbf8efdc7c731f990da3742facdc2b7879d6
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8856666666666666 | 0.005551109331909688 |
+| Epoch_14_batch_2999.pt | 0.8853333333333333 | 0.005734883511361748 |
+| Epoch_17_batch_5999.pt | 0.8845000000000001 | 0.005352108157389486 |
+| Epoch_16.pt | 0.8845000000000001 | 0.005544711639075007 |
+| Epoch_15.pt | 0.8845000000000001 | 0.005714665992637489 |
+| Epoch_16_batch_5999.pt | 0.8843333333333334 | 0.0054557705320848505 |
+| Epoch_14_batch_5999.pt | 0.8836666666666668 | 0.005408041566061312 |
+| Epoch_15_batch_5999.pt | 0.8835 | 0.005462837412242163 |
+| Epoch_15_batch_2999.pt | 0.8835 | 0.00569518965313581 |
+| Epoch_16_batch_2999.pt | 0.8833333333333334 | 0.005555555555555556 |
+| Epoch_14.pt | 0.8831666666666667 | 0.005290627632024463 |
+| Epoch_13_batch_5999.pt | 0.8823333333333332 | 0.0052422782720379785 |
+| Epoch_17_batch_2999.pt | 0.8823333333333332 | 0.00537024265493092 |
+| Epoch_12.pt | 0.8818333333333334 | 0.005684340629753082 |
+| Epoch_13.pt | 0.8816666666666666 | 0.005465943944999482 |
+| Epoch_13_batch_2999.pt | 0.8803333333333334 | 0.0052044615238357175 |
+| Epoch_12_batch_2999.pt | 0.8796666666666667 | 0.0057133156143033 |
+| Epoch_11_batch_2999.pt | 0.8795 | 0.005200011870831654 |
+| Epoch_11.pt | 0.8793333333333333 | 0.00571547606649408 |
+| Epoch_9.pt | 0.8788333333333334 | 0.005826980667867682 |
+| Epoch_10.pt | 0.8788333333333334 | 0.005499999999999995 |
+| Epoch_11_batch_5999.pt | 0.8781666666666667 | 0.0056734708604996185 |
+| Epoch_12_batch_5999.pt | 0.8778333333333335 | 0.005352108157389483 |
+| Epoch_9_batch_2999.pt | 0.8776666666666667 | 0.004741464065189298 |
+| Epoch_10_batch_2999.pt | 0.8773333333333333 | 0.005715476066494082 |
+| Epoch_9_batch_5999.pt | 0.8770000000000001 | 0.005593205754956988 |
+| Epoch_7.pt | 0.8766666666666666 | 0.004950370979987884 |
+| Epoch_10_batch_5999.pt | 0.8758333333333335 | 0.00524845656324755 |
+| Epoch_8_batch_5999.pt | 0.8755000000000001 | 0.005369380496256552 |
+| Epoch_8.pt | 0.875 | 0.005978356023250932 |
+| Epoch_7_batch_5999.pt | 0.8741666666666668 | 0.004636476418601404 |
+| Epoch_8_batch_2999.pt | 0.8734999999999999 | 0.004846087896565083 |
+| Epoch_7_batch_2999.pt | 0.8728333333333333 | 0.005471869700997213 |
+| Epoch_5.pt | 0.8693333333333333 | 0.00567754946268498 |
+| Epoch_6_batch_5999.pt | 0.8673333333333334 | 0.004951617767812703 |
+| Epoch_6.pt | 0.8673333333333334 | 0.005347204184673765 |
+| Epoch_6_batch_2999.pt | 0.8658333333333333 | 0.005721143371869073 |
+| Epoch_5_batch_2999.pt | 0.8638333333333333 | 0.006863078890108422 |
+| Epoch_5_batch_5999.pt | 0.8626666666666667 | 0.0049203532945521124 |
+| Epoch_4.pt | 0.8620000000000001 | 0.006544246365612095 |
+| Epoch_4_batch_5999.pt | 0.8603333333333334 | 0.005920251913225428 |
+| Epoch_3_batch_5999.pt | 0.8581666666666665 | 0.006153639894454119 |
+| Epoch_3.pt | 0.8566666666666667 | 0.006974379922877868 |
+| Epoch_4_batch_2999.pt | 0.8559999999999999 | 0.006315765107165472 |
+| Epoch_2_batch_5999.pt | 0.85 | 0.006740333728045818 |
+| Epoch_3_batch_2999.pt | 0.8460000000000001 | 0.00634501846011966 |
+| Epoch_2.pt | 0.8403333333333333 | 0.007753135566408666 |
+| Epoch_2_batch_2999.pt | 0.834 | 0.005461424767753113 |
+| Epoch_1.pt | 0.8218333333333334 | 0.0067194668365802065 |
+| Epoch_1_batch_5999.pt | 0.8153333333333332 | 0.00611515018036312 |
+| Epoch_1_batch_2999.pt | 0.7763333333333333 | 0.007808668813695153 |
+| Epoch_0.pt | 0.6738333333333333 | 0.009529194905948533 |
+| Epoch_0_batch_5999.pt | 0.6609999999999999 | 0.009353153558104427 |
+| Epoch_0_batch_2999.pt | 0.5210000000000001 | 0.004812535072823931 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3ebdf5fd3102aae6c01335a2acc61888c1338fac
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9986666666666666 | 0.0005443310539518171 |
+| Epoch_13_batch_5999.pt | 0.9984999999999999 | 0.000524110062892033 |
+| Epoch_15.pt | 0.9984999999999999 | 0.0005800170282728065 |
+| Epoch_16_batch_5999.pt | 0.9984999999999999 | 0.000524110062892033 |
+| Epoch_13_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_11_batch_2999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_12_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_12.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_14_batch_5999.pt | 0.9983333333333334 | 0.0005555555555555536 |
+| Epoch_14.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_15_batch_5999.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_14_batch_2999.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_13.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_17.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_17_batch_5999.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_16_batch_2999.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_16.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_17_batch_2999.pt | 0.9983333333333333 | 0.0005555555555555536 |
+| Epoch_11.pt | 0.9983333333333333 | 0.0004969039949999524 |
+| Epoch_8_batch_5999.pt | 0.9981666666666668 | 0.0005241100628920312 |
+| Epoch_9.pt | 0.9981666666666668 | 0.0005241100628920312 |
+| Epoch_10_batch_2999.pt | 0.9981666666666665 | 0.0005800170282728054 |
+| Epoch_6_batch_5999.pt | 0.998 | 0.0005983516452371637 |
+| Epoch_7_batch_5999.pt | 0.998 | 0.000544331053951814 |
+| Epoch_10.pt | 0.998 | 0.000544331053951814 |
+| Epoch_8.pt | 0.9978333333333333 | 0.0006596856715021059 |
+| Epoch_10_batch_5999.pt | 0.9978333333333333 | 0.0005583264233956013 |
+| Epoch_5.pt | 0.9978333333333333 | 0.0004999999999999959 |
+| Epoch_6_batch_2999.pt | 0.9978333333333333 | 0.0006596856715021046 |
+| Epoch_4.pt | 0.9978333333333333 | 0.0005583264233956013 |
+| Epoch_12_batch_2999.pt | 0.9978333333333333 | 0.000611111111111109 |
+| Epoch_4_batch_2999.pt | 0.9978333333333333 | 0.0006596856715021046 |
+| Epoch_5_batch_5999.pt | 0.9978333333333333 | 0.0006596856715021046 |
+| Epoch_3.pt | 0.9976666666666667 | 0.0007114582486036489 |
+| Epoch_8_batch_2999.pt | 0.9976666666666667 | 0.0006186404847588889 |
+| Epoch_9_batch_2999.pt | 0.9976666666666667 | 0.0006186404847588889 |
+| Epoch_9_batch_5999.pt | 0.9976666666666667 | 0.0006186404847588889 |
+| Epoch_6.pt | 0.9976666666666667 | 0.0006186404847588889 |
+| Epoch_11_batch_5999.pt | 0.9976666666666667 | 0.0006186404847588889 |
+| Epoch_7.pt | 0.9974999999999999 | 0.0007136240321480618 |
+| Epoch_7_batch_2999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_3_batch_5999.pt | 0.9974999999999999 | 0.0007136240321480632 |
+| Epoch_4_batch_5999.pt | 0.9973333333333333 | 0.0006186404847588889 |
+| Epoch_5_batch_2999.pt | 0.9971666666666665 | 0.0005583264233956028 |
+| Epoch_2_batch_5999.pt | 0.9968333333333333 | 0.0008031573497111664 |
+| Epoch_2.pt | 0.9963333333333333 | 0.0009558139185602936 |
+| Epoch_3_batch_2999.pt | 0.9963333333333333 | 0.0009558139185602873 |
+| Epoch_2_batch_2999.pt | 0.9955 | 0.0008975274678557484 |
+| Epoch_1_batch_5999.pt | 0.994 | 0.0010000000000000028 |
+| Epoch_1.pt | 0.9938333333333335 | 0.0014497764834110998 |
+| Epoch_1_batch_2999.pt | 0.9894999999999999 | 0.001572330188676104 |
+| Epoch_0.pt | 0.9651666666666667 | 0.0027492984514797 |
+| Epoch_0_batch_5999.pt | 0.9511666666666667 | 0.0038091021592048745 |
+| Epoch_0_batch_2999.pt | 0.7126666666666666 | 0.006546132587402417 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..2805140e4f6af6d9ad203dededcf52b447fe1ea4
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.9782990744236302 |
+| 2 | 0.9840790450030592 |
+| 3 | 0.9860642826457685 |
+| 4 | 0.9872554252313941 |
+| 5 | 0.9880039574573337 |
+| 6 | 0.9886418370933517 |
+| 7 | 0.9891885910670815 |
+| 8 | 0.989592147571501 |
+| 9 | 0.9899110873895102 |
+| 10 | 0.9902039913040083 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/backbone_conf.yaml b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/backbone_conf.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..6af5e97b33060573c8adb6149d42e1c8c5ef513a
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/accu_files/backbone_conf.yaml
@@ -0,0 +1,158 @@
+MobileFaceNet:
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+ResNet:
+ depth: 152
+ drop_ratio: 0.4
+ net_mode: ir_se
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+EfficientNet:
+ width: 1.0
+ depth: 1.0
+ image_size: 112
+ drop_ratio: 0.2
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+
+HRNet:
+ NAME: cls_hrnet
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+ IMAGE_SIZE:
+ - 112
+ - 112
+ EXTRA:
+ STAGE1:
+ NUM_MODULES: 1
+ NUM_RANCHES: 1
+ BLOCK: BOTTLENECK
+ NUM_BLOCKS:
+ - 4
+ NUM_CHANNELS:
+ - 64
+ FUSE_METHOD: SUM
+ STAGE2:
+ NUM_MODULES: 1
+ NUM_BRANCHES: 2
+ BLOCK: BASIC
+ NUM_BLOCKS:
+ - 4
+ - 4
+ NUM_CHANNELS:
+ - 18
+ - 36
+ FUSE_METHOD: SUM
+ STAGE3:
+ NUM_MODULES: 4
+ NUM_BRANCHES: 3
+ BLOCK: BASIC
+ NUM_BLOCKS:
+ - 4
+ - 4
+ - 4
+ NUM_CHANNELS:
+ - 18
+ - 36
+ - 72
+ FUSE_METHOD: SUM
+ STAGE4:
+ NUM_MODULES: 3
+ NUM_BRANCHES: 4
+ BLOCK: BASIC
+ NUM_BLOCKS:
+ - 4
+ - 4
+ - 4
+ - 4
+ NUM_CHANNELS:
+ - 18
+ - 36
+ - 72
+ - 144
+ FUSE_METHOD: SUM
+
+GhostNet:
+ width: 1.0
+ drop_ratio: 0.2
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+
+AttentionNet:
+ stage1_modules: 1
+ stage2_modules: 1
+ stage3_modules: 1
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+TF-NAS:
+ feat_dim: 512
+ drop_ratio: 0.2
+ out_h: 7
+ out_w: 7
+
+ResNeSt:
+ depth: 50
+ drop_ratio: 0.4
+ feat_dim: 512
+ out_h: 7
+ out_w: 7
+
+ReXNet:
+ input_ch: 16
+ final_ch: 180
+ width_mult: 1.0
+ depth_mult: 1.0
+ use_se: 0
+ se_ratio: 12
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+ dropout_ratio: 0.2
+
+LightCNN:
+ depth: 29
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+ dropout_ratio: 0.2
+
+RepVGG:
+ blocks1: 4
+ blocks2: 6
+ blocks3: 16
+ blocks4: 1
+ width1: 2
+ width2: 2
+ width3: 2
+ width4: 4
+ out_h: 7
+ out_w: 7
+ feat_dim: 512
+SwinTransformer:
+ img_size: 224
+ patch_size: 4
+ in_chans: 3
+ embed_dim: 96
+ depths:
+ - 2
+ - 2
+ - 6
+ - 2
+ num_heads:
+ - 3
+ - 6
+ - 12
+ - 24
+ window_size: 7
+ mlp_ratio: 4.0
+ drop_rate: 0.0
+ drop_path_rate: 0.2
\ No newline at end of file
diff --git a/bob/bio/facexzoo/models/backbones/SwinTransformer_T/log.log_ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/log.log_
new file mode 100644
index 0000000000000000000000000000000000000000..9e2452b06e3deacfa16cb0c82c1fc312940092f1
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/SwinTransformer_T/log.log_
@@ -0,0 +1,871 @@
+Use GPU: 2 for training
+Use GPU: 3 for training
+Use GPU: 0 for training
+Use GPU: 1 for training
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 6, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.2}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 6, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.2}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 6, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.2}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+backbone param:
+{'img_size': 224, 'patch_size': 4, 'in_chans': 3, 'embed_dim': 96, 'depths': [2, 2, 6, 2], 'num_heads': [3, 6, 12, 24], 'window_size': 7, 'mlp_ratio': 4.0, 'drop_rate': 0.0, 'drop_path_rate': 0.2}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+Selected optimization level O1: Insert automatic casts around Pytorch functions and Tensor methods.
+
+Defaults for this optimization level are:
+enabled : True
+opt_level : O1
+cast_model_type : None
+patch_torch_functions : True
+keep_batchnorm_fp32 : None
+master_weights : None
+loss_scale : dynamic
+Processing user overrides (additional kwargs that are not None)...
+After processing overrides, optimization options are:
+enabled : True
+opt_level : O1
+cast_model_type : None
+patch_torch_functions : True
+keep_batchnorm_fp32 : None
+master_weights : None
+loss_scale : dynamic
+INFO 2021-10-27 19:45:30 train.py: 89] Epoch 0, iter 0/6416, lr 0.000000, loss 16.308924
+INFO 2021-10-27 19:45:30 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-10-27 19:45:30 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-10-27 19:45:30 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-10-27 19:45:30 distributed.py: 607] Reducer buckets have been rebuilt in this iteration.
+INFO 2021-10-27 19:49:00 train.py: 89] Epoch 0, iter 200/6416, lr 0.000016, loss 16.205033
+INFO 2021-10-27 19:52:43 train.py: 89] Epoch 0, iter 400/6416, lr 0.000032, loss 15.805516
+INFO 2021-10-27 20:02:18 train.py: 89] Epoch 0, iter 600/6416, lr 0.000047, loss 15.338665
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-10-27 20:10:37 train.py: 89] Epoch 0, iter 800/6416, lr 0.000063, loss 15.030042
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+INFO 2021-10-27 20:19:06 train.py: 89] Epoch 0, iter 1000/6416, lr 0.000078, loss 14.953949
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 4096.0
+INFO 2021-10-27 20:28:00 train.py: 89] Epoch 0, iter 1200/6416, lr 0.000094, loss 14.956697
+INFO 2021-10-27 20:36:02 train.py: 89] Epoch 0, iter 1400/6416, lr 0.000109, loss 14.965466
+INFO 2021-10-27 20:44:48 train.py: 89] Epoch 0, iter 1600/6416, lr 0.000125, loss 14.979692
+INFO 2021-10-27 20:52:39 train.py: 89] Epoch 0, iter 1800/6416, lr 0.000141, loss 14.992545
+INFO 2021-10-27 21:01:05 train.py: 89] Epoch 0, iter 2000/6416, lr 0.000156, loss 14.996153
+INFO 2021-10-27 21:09:37 train.py: 89] Epoch 0, iter 2200/6416, lr 0.000172, loss 15.012541
+INFO 2021-10-27 21:17:55 train.py: 89] Epoch 0, iter 2400/6416, lr 0.000187, loss 14.999665
+INFO 2021-10-27 21:26:33 train.py: 89] Epoch 0, iter 2600/6416, lr 0.000203, loss 14.988519
+INFO 2021-10-27 21:35:05 train.py: 89] Epoch 0, iter 2800/6416, lr 0.000218, loss 14.942005
+INFO 2021-10-27 21:43:07 train.py: 101] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-10-27 21:43:08 train.py: 89] Epoch 0, iter 3000/6416, lr 0.000234, loss 14.865995
+INFO 2021-10-27 21:51:16 train.py: 89] Epoch 0, iter 3200/6416, lr 0.000250, loss 14.720523
+INFO 2021-10-27 21:59:47 train.py: 89] Epoch 0, iter 3400/6416, lr 0.000265, loss 14.571263
+INFO 2021-10-27 22:07:55 train.py: 89] Epoch 0, iter 3600/6416, lr 0.000281, loss 14.371213
+INFO 2021-10-27 22:16:10 train.py: 89] Epoch 0, iter 3800/6416, lr 0.000296, loss 14.128887
+INFO 2021-10-27 22:24:19 train.py: 89] Epoch 0, iter 4000/6416, lr 0.000312, loss 13.889157
+INFO 2021-10-27 22:32:54 train.py: 89] Epoch 0, iter 4200/6416, lr 0.000327, loss 13.596615
+INFO 2021-10-27 22:41:09 train.py: 89] Epoch 0, iter 4400/6416, lr 0.000343, loss 13.282345
+INFO 2021-10-27 22:49:16 train.py: 89] Epoch 0, iter 4600/6416, lr 0.000359, loss 12.950714
+INFO 2021-10-27 22:57:55 train.py: 89] Epoch 0, iter 4800/6416, lr 0.000374, loss 12.586951
+INFO 2021-10-27 23:06:24 train.py: 89] Epoch 0, iter 5000/6416, lr 0.000390, loss 12.181788
+INFO 2021-10-27 23:14:45 train.py: 89] Epoch 0, iter 5200/6416, lr 0.000405, loss 11.810964
+INFO 2021-10-27 23:23:25 train.py: 89] Epoch 0, iter 5400/6416, lr 0.000421, loss 11.537682
+INFO 2021-10-27 23:31:45 train.py: 89] Epoch 0, iter 5600/6416, lr 0.000436, loss 11.368209
+INFO 2021-10-27 23:40:32 train.py: 89] Epoch 0, iter 5800/6416, lr 0.000452, loss 11.323373
+INFO 2021-10-27 23:48:40 train.py: 101] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-10-27 23:48:41 train.py: 89] Epoch 0, iter 6000/6416, lr 0.000468, loss 11.393962
+INFO 2021-10-27 23:57:36 train.py: 89] Epoch 0, iter 6200/6416, lr 0.000483, loss 11.585024
+INFO 2021-10-28 00:05:37 train.py: 89] Epoch 0, iter 6400/6416, lr 0.000499, loss 11.859979
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 8192.0
+INFO 2021-10-28 00:06:15 train.py: 108] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-10-28 00:06:16 train.py: 89] Epoch 1, iter 0/6416, lr 0.000496, loss 11.960120
+INFO 2021-10-28 00:09:44 train.py: 89] Epoch 1, iter 200/6416, lr 0.000496, loss 12.161499
+INFO 2021-10-28 00:13:11 train.py: 89] Epoch 1, iter 400/6416, lr 0.000496, loss 12.452237
+INFO 2021-10-28 00:16:38 train.py: 89] Epoch 1, iter 600/6416, lr 0.000496, loss 12.631589
+INFO 2021-10-28 00:20:04 train.py: 89] Epoch 1, iter 800/6416, lr 0.000495, loss 12.769428
+INFO 2021-10-28 00:23:30 train.py: 89] Epoch 1, iter 1000/6416, lr 0.000495, loss 12.910998
+INFO 2021-10-28 00:26:54 train.py: 89] Epoch 1, iter 1200/6416, lr 0.000495, loss 12.902616
+INFO 2021-10-28 00:30:17 train.py: 89] Epoch 1, iter 1400/6416, lr 0.000494, loss 12.860602
+INFO 2021-10-28 00:33:40 train.py: 89] Epoch 1, iter 1600/6416, lr 0.000494, loss 12.762770
+INFO 2021-10-28 00:37:02 train.py: 89] Epoch 1, iter 1800/6416, lr 0.000494, loss 12.690929
+INFO 2021-10-28 00:40:24 train.py: 89] Epoch 1, iter 2000/6416, lr 0.000494, loss 12.452563
+INFO 2021-10-28 00:43:44 train.py: 89] Epoch 1, iter 2200/6416, lr 0.000493, loss 12.270658
+INFO 2021-10-28 00:47:06 train.py: 89] Epoch 1, iter 2400/6416, lr 0.000493, loss 12.026586
+INFO 2021-10-28 00:50:28 train.py: 89] Epoch 1, iter 2600/6416, lr 0.000493, loss 11.823563
+INFO 2021-10-28 00:53:49 train.py: 89] Epoch 1, iter 2800/6416, lr 0.000492, loss 11.601190
+INFO 2021-10-28 00:57:10 train.py: 101] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-10-28 00:57:11 train.py: 89] Epoch 1, iter 3000/6416, lr 0.000492, loss 11.326194
+INFO 2021-10-28 01:00:31 train.py: 89] Epoch 1, iter 3200/6416, lr 0.000492, loss 11.030585
+INFO 2021-10-28 01:03:50 train.py: 89] Epoch 1, iter 3400/6416, lr 0.000491, loss 10.827321
+INFO 2021-10-28 01:07:10 train.py: 89] Epoch 1, iter 3600/6416, lr 0.000491, loss 10.609490
+INFO 2021-10-28 01:10:29 train.py: 89] Epoch 1, iter 3800/6416, lr 0.000491, loss 10.353394
+INFO 2021-10-28 01:13:49 train.py: 89] Epoch 1, iter 4000/6416, lr 0.000490, loss 10.128440
+INFO 2021-10-28 01:17:09 train.py: 89] Epoch 1, iter 4200/6416, lr 0.000490, loss 9.852172
+INFO 2021-10-28 01:20:31 train.py: 89] Epoch 1, iter 4400/6416, lr 0.000489, loss 9.687043
+INFO 2021-10-28 01:23:49 train.py: 89] Epoch 1, iter 4600/6416, lr 0.000489, loss 9.468055
+INFO 2021-10-28 01:27:11 train.py: 89] Epoch 1, iter 4800/6416, lr 0.000489, loss 9.270857
+INFO 2021-10-28 01:30:33 train.py: 89] Epoch 1, iter 5000/6416, lr 0.000488, loss 9.079371
+INFO 2021-10-28 01:33:54 train.py: 89] Epoch 1, iter 5200/6416, lr 0.000488, loss 8.894046
+INFO 2021-10-28 01:37:14 train.py: 89] Epoch 1, iter 5400/6416, lr 0.000487, loss 8.731286
+INFO 2021-10-28 01:40:35 train.py: 89] Epoch 1, iter 5600/6416, lr 0.000487, loss 8.599285
+INFO 2021-10-28 01:43:57 train.py: 89] Epoch 1, iter 5800/6416, lr 0.000486, loss 8.439380
+INFO 2021-10-28 01:47:18 train.py: 101] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-10-28 01:47:19 train.py: 89] Epoch 1, iter 6000/6416, lr 0.000486, loss 8.233267
+INFO 2021-10-28 01:50:40 train.py: 89] Epoch 1, iter 6200/6416, lr 0.000486, loss 8.157759
+INFO 2021-10-28 01:54:01 train.py: 89] Epoch 1, iter 6400/6416, lr 0.000485, loss 7.971182
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 01:54:17 train.py: 108] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-10-28 01:54:18 train.py: 89] Epoch 2, iter 0/6416, lr 0.000485, loss 7.945005
+INFO 2021-10-28 01:57:38 train.py: 89] Epoch 2, iter 200/6416, lr 0.000485, loss 7.874197
+INFO 2021-10-28 02:00:59 train.py: 89] Epoch 2, iter 400/6416, lr 0.000484, loss 7.727359
+INFO 2021-10-28 02:04:18 train.py: 89] Epoch 2, iter 600/6416, lr 0.000484, loss 7.657552
+INFO 2021-10-28 02:07:38 train.py: 89] Epoch 2, iter 800/6416, lr 0.000483, loss 7.560459
+INFO 2021-10-28 02:10:58 train.py: 89] Epoch 2, iter 1000/6416, lr 0.000483, loss 7.466606
+INFO 2021-10-28 02:14:19 train.py: 89] Epoch 2, iter 1200/6416, lr 0.000482, loss 7.344927
+INFO 2021-10-28 02:17:41 train.py: 89] Epoch 2, iter 1400/6416, lr 0.000482, loss 7.279073
+INFO 2021-10-28 02:21:02 train.py: 89] Epoch 2, iter 1600/6416, lr 0.000481, loss 7.179081
+INFO 2021-10-28 02:24:22 train.py: 89] Epoch 2, iter 1800/6416, lr 0.000481, loss 7.030001
+INFO 2021-10-28 02:27:43 train.py: 89] Epoch 2, iter 2000/6416, lr 0.000480, loss 6.925757
+INFO 2021-10-28 02:31:02 train.py: 89] Epoch 2, iter 2200/6416, lr 0.000480, loss 6.895507
+INFO 2021-10-28 02:34:20 train.py: 89] Epoch 2, iter 2400/6416, lr 0.000479, loss 6.804717
+INFO 2021-10-28 02:37:39 train.py: 89] Epoch 2, iter 2600/6416, lr 0.000479, loss 6.774853
+INFO 2021-10-28 02:40:59 train.py: 89] Epoch 2, iter 2800/6416, lr 0.000478, loss 6.731346
+INFO 2021-10-28 02:44:19 train.py: 101] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-10-28 02:44:20 train.py: 89] Epoch 2, iter 3000/6416, lr 0.000477, loss 6.601882
+INFO 2021-10-28 02:47:39 train.py: 89] Epoch 2, iter 3200/6416, lr 0.000477, loss 6.513037
+INFO 2021-10-28 02:50:58 train.py: 89] Epoch 2, iter 3400/6416, lr 0.000476, loss 6.489415
+INFO 2021-10-28 02:54:18 train.py: 89] Epoch 2, iter 3600/6416, lr 0.000476, loss 6.449724
+INFO 2021-10-28 02:57:37 train.py: 89] Epoch 2, iter 3800/6416, lr 0.000475, loss 6.345946
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 03:00:57 train.py: 89] Epoch 2, iter 4000/6416, lr 0.000475, loss 6.313829
+INFO 2021-10-28 03:04:16 train.py: 89] Epoch 2, iter 4200/6416, lr 0.000474, loss 6.215627
+INFO 2021-10-28 03:07:34 train.py: 89] Epoch 2, iter 4400/6416, lr 0.000473, loss 6.190373
+INFO 2021-10-28 03:10:54 train.py: 89] Epoch 2, iter 4600/6416, lr 0.000473, loss 6.158196
+INFO 2021-10-28 03:14:13 train.py: 89] Epoch 2, iter 4800/6416, lr 0.000472, loss 6.060778
+INFO 2021-10-28 03:17:32 train.py: 89] Epoch 2, iter 5000/6416, lr 0.000471, loss 6.031413
+INFO 2021-10-28 03:20:52 train.py: 89] Epoch 2, iter 5200/6416, lr 0.000471, loss 5.983041
+INFO 2021-10-28 03:24:12 train.py: 89] Epoch 2, iter 5400/6416, lr 0.000470, loss 5.902842
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 03:27:31 train.py: 89] Epoch 2, iter 5600/6416, lr 0.000470, loss 5.901454
+INFO 2021-10-28 03:30:51 train.py: 89] Epoch 2, iter 5800/6416, lr 0.000469, loss 5.850066
+INFO 2021-10-28 03:34:11 train.py: 101] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-10-28 03:34:12 train.py: 89] Epoch 2, iter 6000/6416, lr 0.000468, loss 5.772196
+INFO 2021-10-28 03:37:31 train.py: 89] Epoch 2, iter 6200/6416, lr 0.000468, loss 5.763667
+INFO 2021-10-28 03:40:50 train.py: 89] Epoch 2, iter 6400/6416, lr 0.000467, loss 5.689551
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-10-28 03:41:06 train.py: 108] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-10-28 03:41:07 train.py: 89] Epoch 3, iter 0/6416, lr 0.000467, loss 5.721676
+INFO 2021-10-28 03:44:26 train.py: 89] Epoch 3, iter 200/6416, lr 0.000466, loss 5.679465
+INFO 2021-10-28 03:47:45 train.py: 89] Epoch 3, iter 400/6416, lr 0.000465, loss 5.610471
+INFO 2021-10-28 03:51:05 train.py: 89] Epoch 3, iter 600/6416, lr 0.000465, loss 5.618477
+INFO 2021-10-28 03:54:25 train.py: 89] Epoch 3, iter 800/6416, lr 0.000464, loss 5.579800
+INFO 2021-10-28 03:57:45 train.py: 89] Epoch 3, iter 1000/6416, lr 0.000463, loss 5.564950
+INFO 2021-10-28 04:01:07 train.py: 89] Epoch 3, iter 1200/6416, lr 0.000463, loss 5.478641
+INFO 2021-10-28 04:04:28 train.py: 89] Epoch 3, iter 1400/6416, lr 0.000462, loss 5.480149
+INFO 2021-10-28 04:07:48 train.py: 89] Epoch 3, iter 1600/6416, lr 0.000461, loss 5.442156
+INFO 2021-10-28 04:11:08 train.py: 89] Epoch 3, iter 1800/6416, lr 0.000461, loss 5.344044
+INFO 2021-10-28 04:14:30 train.py: 89] Epoch 3, iter 2000/6416, lr 0.000460, loss 5.308675
+INFO 2021-10-28 04:17:49 train.py: 89] Epoch 3, iter 2200/6416, lr 0.000459, loss 5.287351
+INFO 2021-10-28 04:21:08 train.py: 89] Epoch 3, iter 2400/6416, lr 0.000458, loss 5.278140
+INFO 2021-10-28 04:24:28 train.py: 89] Epoch 3, iter 2600/6416, lr 0.000458, loss 5.264993
+INFO 2021-10-28 04:27:47 train.py: 89] Epoch 3, iter 2800/6416, lr 0.000457, loss 5.252692
+INFO 2021-10-28 04:31:06 train.py: 101] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-10-28 04:31:07 train.py: 89] Epoch 3, iter 3000/6416, lr 0.000456, loss 5.195883
+INFO 2021-10-28 04:34:26 train.py: 89] Epoch 3, iter 3200/6416, lr 0.000455, loss 5.144471
+INFO 2021-10-28 04:37:45 train.py: 89] Epoch 3, iter 3400/6416, lr 0.000454, loss 5.122191
+INFO 2021-10-28 04:41:05 train.py: 89] Epoch 3, iter 3600/6416, lr 0.000454, loss 5.127404
+INFO 2021-10-28 04:44:25 train.py: 89] Epoch 3, iter 3800/6416, lr 0.000453, loss 5.040150
+INFO 2021-10-28 04:47:45 train.py: 89] Epoch 3, iter 4000/6416, lr 0.000452, loss 5.053704
+INFO 2021-10-28 04:51:06 train.py: 89] Epoch 3, iter 4200/6416, lr 0.000451, loss 5.003899
+INFO 2021-10-28 04:54:26 train.py: 89] Epoch 3, iter 4400/6416, lr 0.000451, loss 4.972808
+INFO 2021-10-28 04:57:46 train.py: 89] Epoch 3, iter 4600/6416, lr 0.000450, loss 4.966010
+INFO 2021-10-28 05:01:06 train.py: 89] Epoch 3, iter 4800/6416, lr 0.000449, loss 4.902362
+INFO 2021-10-28 05:04:26 train.py: 89] Epoch 3, iter 5000/6416, lr 0.000448, loss 4.902580
+INFO 2021-10-28 05:07:47 train.py: 89] Epoch 3, iter 5200/6416, lr 0.000447, loss 4.894835
+INFO 2021-10-28 05:11:06 train.py: 89] Epoch 3, iter 5400/6416, lr 0.000446, loss 4.813634
+INFO 2021-10-28 05:14:26 train.py: 89] Epoch 3, iter 5600/6416, lr 0.000446, loss 4.815355
+INFO 2021-10-28 05:17:47 train.py: 89] Epoch 3, iter 5800/6416, lr 0.000445, loss 4.802002
+INFO 2021-10-28 05:21:08 train.py: 101] Save checkpoint Epoch_3_batch_5999.pt to disk.
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 05:21:09 train.py: 89] Epoch 3, iter 6000/6416, lr 0.000444, loss 4.739605
+INFO 2021-10-28 05:24:28 train.py: 89] Epoch 3, iter 6200/6416, lr 0.000443, loss 4.760175
+INFO 2021-10-28 05:27:48 train.py: 89] Epoch 3, iter 6400/6416, lr 0.000442, loss 4.738929
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 05:28:04 train.py: 108] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-10-28 05:28:06 train.py: 89] Epoch 4, iter 0/6416, lr 0.000442, loss 4.712062
+INFO 2021-10-28 05:31:26 train.py: 89] Epoch 4, iter 200/6416, lr 0.000441, loss 4.701935
+INFO 2021-10-28 05:34:47 train.py: 89] Epoch 4, iter 400/6416, lr 0.000440, loss 4.674354
+INFO 2021-10-28 05:38:05 train.py: 89] Epoch 4, iter 600/6416, lr 0.000439, loss 4.697434
+INFO 2021-10-28 05:41:26 train.py: 89] Epoch 4, iter 800/6416, lr 0.000439, loss 4.677425
+INFO 2021-10-28 05:44:46 train.py: 89] Epoch 4, iter 1000/6416, lr 0.000438, loss 4.665173
+INFO 2021-10-28 05:48:05 train.py: 89] Epoch 4, iter 1200/6416, lr 0.000437, loss 4.594268
+INFO 2021-10-28 05:51:28 train.py: 89] Epoch 4, iter 1400/6416, lr 0.000436, loss 4.593345
+INFO 2021-10-28 05:54:48 train.py: 89] Epoch 4, iter 1600/6416, lr 0.000435, loss 4.590036
+INFO 2021-10-28 05:58:08 train.py: 89] Epoch 4, iter 1800/6416, lr 0.000434, loss 4.499436
+INFO 2021-10-28 06:01:28 train.py: 89] Epoch 4, iter 2000/6416, lr 0.000433, loss 4.463434
+INFO 2021-10-28 06:04:47 train.py: 89] Epoch 4, iter 2200/6416, lr 0.000432, loss 4.499488
+INFO 2021-10-28 06:08:08 train.py: 89] Epoch 4, iter 2400/6416, lr 0.000431, loss 4.476484
+INFO 2021-10-28 06:11:27 train.py: 89] Epoch 4, iter 2600/6416, lr 0.000430, loss 4.474912
+INFO 2021-10-28 06:14:47 train.py: 89] Epoch 4, iter 2800/6416, lr 0.000429, loss 4.470611
+INFO 2021-10-28 06:18:08 train.py: 101] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-10-28 06:18:09 train.py: 89] Epoch 4, iter 3000/6416, lr 0.000428, loss 4.437052
+INFO 2021-10-28 06:21:27 train.py: 89] Epoch 4, iter 3200/6416, lr 0.000428, loss 4.383440
+INFO 2021-10-28 06:24:47 train.py: 89] Epoch 4, iter 3400/6416, lr 0.000427, loss 4.366967
+INFO 2021-10-28 06:28:08 train.py: 89] Epoch 4, iter 3600/6416, lr 0.000426, loss 4.397078
+INFO 2021-10-28 06:31:28 train.py: 89] Epoch 4, iter 3800/6416, lr 0.000425, loss 4.307374
+INFO 2021-10-28 06:34:46 train.py: 89] Epoch 4, iter 4000/6416, lr 0.000424, loss 4.333356
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 06:38:05 train.py: 89] Epoch 4, iter 4200/6416, lr 0.000423, loss 4.299256
+INFO 2021-10-28 06:41:24 train.py: 89] Epoch 4, iter 4400/6416, lr 0.000422, loss 4.279603
+INFO 2021-10-28 06:44:44 train.py: 89] Epoch 4, iter 4600/6416, lr 0.000421, loss 4.297715
+INFO 2021-10-28 06:48:03 train.py: 89] Epoch 4, iter 4800/6416, lr 0.000420, loss 4.235661
+INFO 2021-10-28 06:51:26 train.py: 89] Epoch 4, iter 5000/6416, lr 0.000419, loss 4.245424
+INFO 2021-10-28 06:54:46 train.py: 89] Epoch 4, iter 5200/6416, lr 0.000418, loss 4.227683
+INFO 2021-10-28 06:58:06 train.py: 89] Epoch 4, iter 5400/6416, lr 0.000417, loss 4.168996
+INFO 2021-10-28 07:01:26 train.py: 89] Epoch 4, iter 5600/6416, lr 0.000416, loss 4.189912
+INFO 2021-10-28 07:04:45 train.py: 89] Epoch 4, iter 5800/6416, lr 0.000415, loss 4.156153
+INFO 2021-10-28 07:08:05 train.py: 101] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-10-28 07:08:06 train.py: 89] Epoch 4, iter 6000/6416, lr 0.000414, loss 4.144393
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 07:11:26 train.py: 89] Epoch 4, iter 6200/6416, lr 0.000413, loss 4.149570
+INFO 2021-10-28 07:14:46 train.py: 89] Epoch 4, iter 6400/6416, lr 0.000412, loss 4.118135
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 07:15:02 train.py: 108] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-10-28 07:15:03 train.py: 89] Epoch 5, iter 0/6416, lr 0.000412, loss 4.106445
+INFO 2021-10-28 07:18:23 train.py: 89] Epoch 5, iter 200/6416, lr 0.000411, loss 4.103658
+INFO 2021-10-28 07:21:43 train.py: 89] Epoch 5, iter 400/6416, lr 0.000410, loss 4.069567
+INFO 2021-10-28 07:25:02 train.py: 89] Epoch 5, iter 600/6416, lr 0.000408, loss 4.116186
+INFO 2021-10-28 07:28:22 train.py: 89] Epoch 5, iter 800/6416, lr 0.000407, loss 4.094658
+INFO 2021-10-28 07:31:41 train.py: 89] Epoch 5, iter 1000/6416, lr 0.000406, loss 4.083572
+INFO 2021-10-28 07:35:02 train.py: 89] Epoch 5, iter 1200/6416, lr 0.000405, loss 4.029658
+INFO 2021-10-28 07:38:21 train.py: 89] Epoch 5, iter 1400/6416, lr 0.000404, loss 4.022752
+INFO 2021-10-28 07:41:40 train.py: 89] Epoch 5, iter 1600/6416, lr 0.000403, loss 4.024136
+INFO 2021-10-28 07:45:00 train.py: 89] Epoch 5, iter 1800/6416, lr 0.000402, loss 3.950005
+INFO 2021-10-28 07:48:19 train.py: 89] Epoch 5, iter 2000/6416, lr 0.000401, loss 3.927258
+INFO 2021-10-28 07:51:38 train.py: 89] Epoch 5, iter 2200/6416, lr 0.000400, loss 3.961413
+INFO 2021-10-28 07:54:57 train.py: 89] Epoch 5, iter 2400/6416, lr 0.000399, loss 3.932422
+INFO 2021-10-28 07:58:18 train.py: 89] Epoch 5, iter 2600/6416, lr 0.000398, loss 3.949419
+INFO 2021-10-28 08:01:39 train.py: 89] Epoch 5, iter 2800/6416, lr 0.000397, loss 3.934371
+INFO 2021-10-28 08:04:58 train.py: 101] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-10-28 08:04:59 train.py: 89] Epoch 5, iter 3000/6416, lr 0.000396, loss 3.913483
+INFO 2021-10-28 08:08:20 train.py: 89] Epoch 5, iter 3200/6416, lr 0.000395, loss 3.871430
+INFO 2021-10-28 08:11:38 train.py: 89] Epoch 5, iter 3400/6416, lr 0.000393, loss 3.881252
+INFO 2021-10-28 08:14:59 train.py: 89] Epoch 5, iter 3600/6416, lr 0.000392, loss 3.894050
+INFO 2021-10-28 08:18:18 train.py: 89] Epoch 5, iter 3800/6416, lr 0.000391, loss 3.827464
+INFO 2021-10-28 08:21:37 train.py: 89] Epoch 5, iter 4000/6416, lr 0.000390, loss 3.841538
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 08:24:56 train.py: 89] Epoch 5, iter 4200/6416, lr 0.000389, loss 3.802447
+INFO 2021-10-28 08:28:15 train.py: 89] Epoch 5, iter 4400/6416, lr 0.000388, loss 3.790499
+INFO 2021-10-28 08:31:34 train.py: 89] Epoch 5, iter 4600/6416, lr 0.000387, loss 3.819048
+INFO 2021-10-28 08:34:53 train.py: 89] Epoch 5, iter 4800/6416, lr 0.000386, loss 3.738784
+INFO 2021-10-28 08:38:13 train.py: 89] Epoch 5, iter 5000/6416, lr 0.000384, loss 3.775999
+INFO 2021-10-28 08:41:33 train.py: 89] Epoch 5, iter 5200/6416, lr 0.000383, loss 3.751282
+INFO 2021-10-28 08:44:52 train.py: 89] Epoch 5, iter 5400/6416, lr 0.000382, loss 3.698726
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 08:48:12 train.py: 89] Epoch 5, iter 5600/6416, lr 0.000381, loss 3.717989
+INFO 2021-10-28 08:51:33 train.py: 89] Epoch 5, iter 5800/6416, lr 0.000380, loss 3.713844
+INFO 2021-10-28 08:54:54 train.py: 101] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-10-28 08:54:55 train.py: 89] Epoch 5, iter 6000/6416, lr 0.000379, loss 3.690298
+INFO 2021-10-28 08:58:15 train.py: 89] Epoch 5, iter 6200/6416, lr 0.000378, loss 3.697322
+INFO 2021-10-28 09:01:34 train.py: 89] Epoch 5, iter 6400/6416, lr 0.000376, loss 3.670910
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-10-28 09:01:50 train.py: 108] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-10-28 09:01:51 train.py: 89] Epoch 6, iter 0/6416, lr 0.000376, loss 3.712390
+INFO 2021-10-28 09:05:11 train.py: 89] Epoch 6, iter 200/6416, lr 0.000375, loss 3.670777
+INFO 2021-10-28 09:08:31 train.py: 89] Epoch 6, iter 400/6416, lr 0.000374, loss 3.631060
+INFO 2021-10-28 09:11:50 train.py: 89] Epoch 6, iter 600/6416, lr 0.000373, loss 3.680791
+INFO 2021-10-28 09:15:10 train.py: 89] Epoch 6, iter 800/6416, lr 0.000372, loss 3.658321
+INFO 2021-10-28 09:18:29 train.py: 89] Epoch 6, iter 1000/6416, lr 0.000370, loss 3.655869
+INFO 2021-10-28 09:21:49 train.py: 89] Epoch 6, iter 1200/6416, lr 0.000369, loss 3.607072
+INFO 2021-10-28 09:25:08 train.py: 89] Epoch 6, iter 1400/6416, lr 0.000368, loss 3.594422
+INFO 2021-10-28 09:28:29 train.py: 89] Epoch 6, iter 1600/6416, lr 0.000367, loss 3.610397
+INFO 2021-10-28 09:31:48 train.py: 89] Epoch 6, iter 1800/6416, lr 0.000366, loss 3.553388
+INFO 2021-10-28 09:35:06 train.py: 89] Epoch 6, iter 2000/6416, lr 0.000364, loss 3.517313
+INFO 2021-10-28 09:38:25 train.py: 89] Epoch 6, iter 2200/6416, lr 0.000363, loss 3.547160
+INFO 2021-10-28 09:41:45 train.py: 89] Epoch 6, iter 2400/6416, lr 0.000362, loss 3.544881
+INFO 2021-10-28 09:45:04 train.py: 89] Epoch 6, iter 2600/6416, lr 0.000361, loss 3.546773
+INFO 2021-10-28 09:48:24 train.py: 89] Epoch 6, iter 2800/6416, lr 0.000360, loss 3.531729
+INFO 2021-10-28 09:51:44 train.py: 101] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-10-28 09:51:45 train.py: 89] Epoch 6, iter 3000/6416, lr 0.000358, loss 3.524637
+INFO 2021-10-28 09:55:04 train.py: 89] Epoch 6, iter 3200/6416, lr 0.000357, loss 3.478864
+INFO 2021-10-28 09:58:23 train.py: 89] Epoch 6, iter 3400/6416, lr 0.000356, loss 3.470553
+INFO 2021-10-28 10:01:42 train.py: 89] Epoch 6, iter 3600/6416, lr 0.000355, loss 3.496517
+INFO 2021-10-28 10:05:01 train.py: 89] Epoch 6, iter 3800/6416, lr 0.000353, loss 3.432307
+INFO 2021-10-28 10:08:21 train.py: 89] Epoch 6, iter 4000/6416, lr 0.000352, loss 3.446772
+INFO 2021-10-28 10:11:40 train.py: 89] Epoch 6, iter 4200/6416, lr 0.000351, loss 3.428196
+INFO 2021-10-28 10:14:58 train.py: 89] Epoch 6, iter 4400/6416, lr 0.000350, loss 3.412509
+INFO 2021-10-28 10:18:18 train.py: 89] Epoch 6, iter 4600/6416, lr 0.000349, loss 3.428365
+INFO 2021-10-28 10:21:38 train.py: 89] Epoch 6, iter 4800/6416, lr 0.000347, loss 3.372309
+INFO 2021-10-28 10:24:56 train.py: 89] Epoch 6, iter 5000/6416, lr 0.000346, loss 3.405291
+INFO 2021-10-28 10:28:15 train.py: 89] Epoch 6, iter 5200/6416, lr 0.000345, loss 3.384060
+INFO 2021-10-28 10:31:35 train.py: 89] Epoch 6, iter 5400/6416, lr 0.000344, loss 3.331648
+INFO 2021-10-28 10:34:55 train.py: 89] Epoch 6, iter 5600/6416, lr 0.000342, loss 3.351478
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 10:38:15 train.py: 89] Epoch 6, iter 5800/6416, lr 0.000341, loss 3.343926
+INFO 2021-10-28 10:41:34 train.py: 101] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-10-28 10:41:35 train.py: 89] Epoch 6, iter 6000/6416, lr 0.000340, loss 3.342539
+INFO 2021-10-28 10:44:54 train.py: 89] Epoch 6, iter 6200/6416, lr 0.000339, loss 3.323385
+INFO 2021-10-28 10:48:14 train.py: 89] Epoch 6, iter 6400/6416, lr 0.000337, loss 3.329841
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-10-28 10:48:30 train.py: 108] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-10-28 10:48:31 train.py: 89] Epoch 7, iter 0/6416, lr 0.000337, loss 3.344971
+INFO 2021-10-28 10:51:50 train.py: 89] Epoch 7, iter 200/6416, lr 0.000336, loss 3.302571
+INFO 2021-10-28 10:55:11 train.py: 89] Epoch 7, iter 400/6416, lr 0.000335, loss 3.287852
+INFO 2021-10-28 10:58:30 train.py: 89] Epoch 7, iter 600/6416, lr 0.000333, loss 3.334028
+INFO 2021-10-28 11:01:48 train.py: 89] Epoch 7, iter 800/6416, lr 0.000332, loss 3.294619
+INFO 2021-10-28 11:05:06 train.py: 89] Epoch 7, iter 1000/6416, lr 0.000331, loss 3.302449
+INFO 2021-10-28 11:08:27 train.py: 89] Epoch 7, iter 1200/6416, lr 0.000330, loss 3.249785
+INFO 2021-10-28 11:11:46 train.py: 89] Epoch 7, iter 1400/6416, lr 0.000328, loss 3.255981
+INFO 2021-10-28 11:15:06 train.py: 89] Epoch 7, iter 1600/6416, lr 0.000327, loss 3.274777
+INFO 2021-10-28 11:18:26 train.py: 89] Epoch 7, iter 1800/6416, lr 0.000326, loss 3.203267
+INFO 2021-10-28 11:21:45 train.py: 89] Epoch 7, iter 2000/6416, lr 0.000324, loss 3.186509
+INFO 2021-10-28 11:25:06 train.py: 89] Epoch 7, iter 2200/6416, lr 0.000323, loss 3.229299
+INFO 2021-10-28 11:28:26 train.py: 89] Epoch 7, iter 2400/6416, lr 0.000322, loss 3.212539
+INFO 2021-10-28 11:31:46 train.py: 89] Epoch 7, iter 2600/6416, lr 0.000321, loss 3.215443
+INFO 2021-10-28 11:35:05 train.py: 89] Epoch 7, iter 2800/6416, lr 0.000319, loss 3.197864
+INFO 2021-10-28 11:38:25 train.py: 101] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-10-28 11:38:26 train.py: 89] Epoch 7, iter 3000/6416, lr 0.000318, loss 3.179902
+INFO 2021-10-28 11:41:44 train.py: 89] Epoch 7, iter 3200/6416, lr 0.000317, loss 3.153305
+INFO 2021-10-28 11:45:05 train.py: 89] Epoch 7, iter 3400/6416, lr 0.000315, loss 3.143844
+INFO 2021-10-28 11:48:25 train.py: 89] Epoch 7, iter 3600/6416, lr 0.000314, loss 3.175027
+INFO 2021-10-28 11:51:44 train.py: 89] Epoch 7, iter 3800/6416, lr 0.000313, loss 3.111098
+INFO 2021-10-28 11:55:04 train.py: 89] Epoch 7, iter 4000/6416, lr 0.000311, loss 3.121993
+INFO 2021-10-28 11:58:24 train.py: 89] Epoch 7, iter 4200/6416, lr 0.000310, loss 3.106656
+INFO 2021-10-28 12:01:44 train.py: 89] Epoch 7, iter 4400/6416, lr 0.000309, loss 3.091902
+INFO 2021-10-28 12:05:04 train.py: 89] Epoch 7, iter 4600/6416, lr 0.000307, loss 3.122038
+INFO 2021-10-28 12:08:23 train.py: 89] Epoch 7, iter 4800/6416, lr 0.000306, loss 3.055590
+INFO 2021-10-28 12:11:43 train.py: 89] Epoch 7, iter 5000/6416, lr 0.000305, loss 3.088864
+INFO 2021-10-28 12:15:03 train.py: 89] Epoch 7, iter 5200/6416, lr 0.000304, loss 3.074753
+INFO 2021-10-28 12:18:23 train.py: 89] Epoch 7, iter 5400/6416, lr 0.000302, loss 3.023677
+INFO 2021-10-28 12:21:42 train.py: 89] Epoch 7, iter 5600/6416, lr 0.000301, loss 3.042343
+INFO 2021-10-28 12:25:01 train.py: 89] Epoch 7, iter 5800/6416, lr 0.000300, loss 3.039879
+INFO 2021-10-28 12:28:20 train.py: 101] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-10-28 12:28:21 train.py: 89] Epoch 7, iter 6000/6416, lr 0.000298, loss 3.016025
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 12:31:41 train.py: 89] Epoch 7, iter 6200/6416, lr 0.000297, loss 3.023507
+INFO 2021-10-28 12:34:59 train.py: 89] Epoch 7, iter 6400/6416, lr 0.000296, loss 3.021032
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 12:35:15 train.py: 108] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-10-28 12:35:16 train.py: 89] Epoch 8, iter 0/6416, lr 0.000295, loss 3.034137
+INFO 2021-10-28 12:38:36 train.py: 89] Epoch 8, iter 200/6416, lr 0.000294, loss 3.011912
+INFO 2021-10-28 12:41:55 train.py: 89] Epoch 8, iter 400/6416, lr 0.000293, loss 2.986379
+INFO 2021-10-28 12:45:16 train.py: 89] Epoch 8, iter 600/6416, lr 0.000291, loss 3.010782
+INFO 2021-10-28 12:48:37 train.py: 89] Epoch 8, iter 800/6416, lr 0.000290, loss 2.988311
+INFO 2021-10-28 12:51:57 train.py: 89] Epoch 8, iter 1000/6416, lr 0.000289, loss 2.996471
+INFO 2021-10-28 12:55:16 train.py: 89] Epoch 8, iter 1200/6416, lr 0.000287, loss 2.945526
+INFO 2021-10-28 12:58:36 train.py: 89] Epoch 8, iter 1400/6416, lr 0.000286, loss 2.957708
+INFO 2021-10-28 13:01:56 train.py: 89] Epoch 8, iter 1600/6416, lr 0.000285, loss 2.966754
+INFO 2021-10-28 13:05:14 train.py: 89] Epoch 8, iter 1800/6416, lr 0.000283, loss 2.917854
+INFO 2021-10-28 13:08:33 train.py: 89] Epoch 8, iter 2000/6416, lr 0.000282, loss 2.886562
+INFO 2021-10-28 13:11:52 train.py: 89] Epoch 8, iter 2200/6416, lr 0.000281, loss 2.930095
+INFO 2021-10-28 13:15:12 train.py: 89] Epoch 8, iter 2400/6416, lr 0.000279, loss 2.904232
+INFO 2021-10-28 13:18:31 train.py: 89] Epoch 8, iter 2600/6416, lr 0.000278, loss 2.918377
+INFO 2021-10-28 13:21:53 train.py: 89] Epoch 8, iter 2800/6416, lr 0.000277, loss 2.899135
+INFO 2021-10-28 13:25:14 train.py: 101] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-10-28 13:25:14 train.py: 89] Epoch 8, iter 3000/6416, lr 0.000275, loss 2.898554
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 13:28:35 train.py: 89] Epoch 8, iter 3200/6416, lr 0.000274, loss 2.866500
+INFO 2021-10-28 13:31:55 train.py: 89] Epoch 8, iter 3400/6416, lr 0.000273, loss 2.862508
+INFO 2021-10-28 13:35:14 train.py: 89] Epoch 8, iter 3600/6416, lr 0.000271, loss 2.878355
+INFO 2021-10-28 13:38:33 train.py: 89] Epoch 8, iter 3800/6416, lr 0.000270, loss 2.822844
+INFO 2021-10-28 13:41:52 train.py: 89] Epoch 8, iter 4000/6416, lr 0.000269, loss 2.858157
+INFO 2021-10-28 13:45:11 train.py: 89] Epoch 8, iter 4200/6416, lr 0.000267, loss 2.833464
+INFO 2021-10-28 13:48:31 train.py: 89] Epoch 8, iter 4400/6416, lr 0.000266, loss 2.814546
+INFO 2021-10-28 13:51:50 train.py: 89] Epoch 8, iter 4600/6416, lr 0.000265, loss 2.829329
+INFO 2021-10-28 13:55:09 train.py: 89] Epoch 8, iter 4800/6416, lr 0.000263, loss 2.787915
+INFO 2021-10-28 13:58:30 train.py: 89] Epoch 8, iter 5000/6416, lr 0.000262, loss 2.815309
+INFO 2021-10-28 14:01:48 train.py: 89] Epoch 8, iter 5200/6416, lr 0.000261, loss 2.797039
+INFO 2021-10-28 14:05:08 train.py: 89] Epoch 8, iter 5400/6416, lr 0.000259, loss 2.748207
+INFO 2021-10-28 14:08:27 train.py: 89] Epoch 8, iter 5600/6416, lr 0.000258, loss 2.782927
+INFO 2021-10-28 14:11:47 train.py: 89] Epoch 8, iter 5800/6416, lr 0.000257, loss 2.787326
+INFO 2021-10-28 14:15:08 train.py: 101] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-10-28 14:15:09 train.py: 89] Epoch 8, iter 6000/6416, lr 0.000255, loss 2.749455
+INFO 2021-10-28 14:18:29 train.py: 89] Epoch 8, iter 6200/6416, lr 0.000254, loss 2.739642
+INFO 2021-10-28 14:21:48 train.py: 89] Epoch 8, iter 6400/6416, lr 0.000253, loss 2.740048
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 14:22:04 train.py: 108] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-10-28 14:22:05 train.py: 89] Epoch 9, iter 0/6416, lr 0.000253, loss 2.784332
+INFO 2021-10-28 14:25:23 train.py: 89] Epoch 9, iter 200/6416, lr 0.000251, loss 2.744047
+INFO 2021-10-28 14:28:41 train.py: 89] Epoch 9, iter 400/6416, lr 0.000250, loss 2.712745
+INFO 2021-10-28 14:32:00 train.py: 89] Epoch 9, iter 600/6416, lr 0.000248, loss 2.746047
+INFO 2021-10-28 14:35:19 train.py: 89] Epoch 9, iter 800/6416, lr 0.000247, loss 2.714920
+INFO 2021-10-28 14:38:38 train.py: 89] Epoch 9, iter 1000/6416, lr 0.000246, loss 2.720971
+INFO 2021-10-28 14:41:58 train.py: 89] Epoch 9, iter 1200/6416, lr 0.000244, loss 2.686197
+INFO 2021-10-28 14:45:16 train.py: 89] Epoch 9, iter 1400/6416, lr 0.000243, loss 2.683106
+INFO 2021-10-28 14:48:37 train.py: 89] Epoch 9, iter 1600/6416, lr 0.000242, loss 2.693393
+INFO 2021-10-28 14:51:57 train.py: 89] Epoch 9, iter 1800/6416, lr 0.000240, loss 2.633042
+INFO 2021-10-28 14:55:16 train.py: 89] Epoch 9, iter 2000/6416, lr 0.000239, loss 2.633354
+INFO 2021-10-28 14:58:35 train.py: 89] Epoch 9, iter 2200/6416, lr 0.000238, loss 2.666727
+INFO 2021-10-28 15:01:54 train.py: 89] Epoch 9, iter 2400/6416, lr 0.000236, loss 2.643372
+INFO 2021-10-28 15:05:14 train.py: 89] Epoch 9, iter 2600/6416, lr 0.000235, loss 2.659169
+INFO 2021-10-28 15:08:34 train.py: 89] Epoch 9, iter 2800/6416, lr 0.000234, loss 2.636055
+INFO 2021-10-28 15:11:55 train.py: 101] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-10-28 15:11:56 train.py: 89] Epoch 9, iter 3000/6416, lr 0.000232, loss 2.627908
+INFO 2021-10-28 15:15:13 train.py: 89] Epoch 9, iter 3200/6416, lr 0.000231, loss 2.600309
+INFO 2021-10-28 15:18:33 train.py: 89] Epoch 9, iter 3400/6416, lr 0.000230, loss 2.605260
+INFO 2021-10-28 15:21:53 train.py: 89] Epoch 9, iter 3600/6416, lr 0.000228, loss 2.630783
+INFO 2021-10-28 15:25:13 train.py: 89] Epoch 9, iter 3800/6416, lr 0.000227, loss 2.553965
+INFO 2021-10-28 15:28:33 train.py: 89] Epoch 9, iter 4000/6416, lr 0.000226, loss 2.591634
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 15:31:52 train.py: 89] Epoch 9, iter 4200/6416, lr 0.000224, loss 2.570424
+INFO 2021-10-28 15:35:11 train.py: 89] Epoch 9, iter 4400/6416, lr 0.000223, loss 2.543214
+INFO 2021-10-28 15:38:31 train.py: 89] Epoch 9, iter 4600/6416, lr 0.000222, loss 2.574642
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 15:41:50 train.py: 89] Epoch 9, iter 4800/6416, lr 0.000220, loss 2.524163
+INFO 2021-10-28 15:45:08 train.py: 89] Epoch 9, iter 5000/6416, lr 0.000219, loss 2.556397
+INFO 2021-10-28 15:48:28 train.py: 89] Epoch 9, iter 5200/6416, lr 0.000218, loss 2.530255
+INFO 2021-10-28 15:51:47 train.py: 89] Epoch 9, iter 5400/6416, lr 0.000216, loss 2.501259
+INFO 2021-10-28 15:55:08 train.py: 89] Epoch 9, iter 5600/6416, lr 0.000215, loss 2.529827
+INFO 2021-10-28 15:58:28 train.py: 89] Epoch 9, iter 5800/6416, lr 0.000214, loss 2.532809
+INFO 2021-10-28 16:01:48 train.py: 101] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-10-28 16:01:48 train.py: 89] Epoch 9, iter 6000/6416, lr 0.000212, loss 2.504846
+INFO 2021-10-28 16:05:06 train.py: 89] Epoch 9, iter 6200/6416, lr 0.000211, loss 2.493524
+INFO 2021-10-28 16:08:26 train.py: 89] Epoch 9, iter 6400/6416, lr 0.000210, loss 2.490000
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-10-28 16:08:42 train.py: 108] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-10-28 16:08:43 train.py: 89] Epoch 10, iter 0/6416, lr 0.000210, loss 2.525472
+INFO 2021-10-28 16:12:03 train.py: 89] Epoch 10, iter 200/6416, lr 0.000208, loss 2.486344
+INFO 2021-10-28 16:15:23 train.py: 89] Epoch 10, iter 400/6416, lr 0.000207, loss 2.469940
+INFO 2021-10-28 16:18:43 train.py: 89] Epoch 10, iter 600/6416, lr 0.000206, loss 2.489966
+INFO 2021-10-28 16:22:02 train.py: 89] Epoch 10, iter 800/6416, lr 0.000204, loss 2.483110
+INFO 2021-10-28 16:25:22 train.py: 89] Epoch 10, iter 1000/6416, lr 0.000203, loss 2.486994
+INFO 2021-10-28 16:28:43 train.py: 89] Epoch 10, iter 1200/6416, lr 0.000202, loss 2.435380
+INFO 2021-10-28 16:32:02 train.py: 89] Epoch 10, iter 1400/6416, lr 0.000200, loss 2.439600
+INFO 2021-10-28 16:35:20 train.py: 89] Epoch 10, iter 1600/6416, lr 0.000199, loss 2.455162
+INFO 2021-10-28 16:38:40 train.py: 89] Epoch 10, iter 1800/6416, lr 0.000198, loss 2.405398
+INFO 2021-10-28 16:41:59 train.py: 89] Epoch 10, iter 2000/6416, lr 0.000196, loss 2.378120
+INFO 2021-10-28 16:45:18 train.py: 89] Epoch 10, iter 2200/6416, lr 0.000195, loss 2.420234
+INFO 2021-10-28 16:48:38 train.py: 89] Epoch 10, iter 2400/6416, lr 0.000194, loss 2.418369
+INFO 2021-10-28 16:51:58 train.py: 89] Epoch 10, iter 2600/6416, lr 0.000192, loss 2.414467
+INFO 2021-10-28 16:55:16 train.py: 89] Epoch 10, iter 2800/6416, lr 0.000191, loss 2.412433
+INFO 2021-10-28 16:58:36 train.py: 101] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-10-28 16:58:37 train.py: 89] Epoch 10, iter 3000/6416, lr 0.000190, loss 2.391040
+INFO 2021-10-28 17:01:55 train.py: 89] Epoch 10, iter 3200/6416, lr 0.000188, loss 2.360694
+INFO 2021-10-28 17:05:16 train.py: 89] Epoch 10, iter 3400/6416, lr 0.000187, loss 2.370609
+INFO 2021-10-28 17:08:35 train.py: 89] Epoch 10, iter 3600/6416, lr 0.000186, loss 2.376378
+INFO 2021-10-28 17:11:53 train.py: 89] Epoch 10, iter 3800/6416, lr 0.000185, loss 2.322251
+INFO 2021-10-28 17:15:13 train.py: 89] Epoch 10, iter 4000/6416, lr 0.000183, loss 2.361513
+INFO 2021-10-28 17:18:33 train.py: 89] Epoch 10, iter 4200/6416, lr 0.000182, loss 2.327074
+INFO 2021-10-28 17:21:52 train.py: 89] Epoch 10, iter 4400/6416, lr 0.000181, loss 2.309213
+INFO 2021-10-28 17:25:12 train.py: 89] Epoch 10, iter 4600/6416, lr 0.000179, loss 2.355231
+INFO 2021-10-28 17:28:32 train.py: 89] Epoch 10, iter 4800/6416, lr 0.000178, loss 2.307535
+INFO 2021-10-28 17:31:51 train.py: 89] Epoch 10, iter 5000/6416, lr 0.000177, loss 2.328193
+INFO 2021-10-28 17:35:11 train.py: 89] Epoch 10, iter 5200/6416, lr 0.000176, loss 2.297734
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 17:38:31 train.py: 89] Epoch 10, iter 5400/6416, lr 0.000174, loss 2.263825
+INFO 2021-10-28 17:41:50 train.py: 89] Epoch 10, iter 5600/6416, lr 0.000173, loss 2.282416
+INFO 2021-10-28 17:45:10 train.py: 89] Epoch 10, iter 5800/6416, lr 0.000172, loss 2.277420
+INFO 2021-10-28 17:48:29 train.py: 101] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-10-28 17:48:30 train.py: 89] Epoch 10, iter 6000/6416, lr 0.000170, loss 2.269950
+INFO 2021-10-28 17:51:51 train.py: 89] Epoch 10, iter 6200/6416, lr 0.000169, loss 2.250397
+INFO 2021-10-28 17:55:10 train.py: 89] Epoch 10, iter 6400/6416, lr 0.000168, loss 2.266215
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 16384.0
+INFO 2021-10-28 17:55:26 train.py: 108] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-10-28 17:55:27 train.py: 89] Epoch 11, iter 0/6416, lr 0.000168, loss 2.287239
+INFO 2021-10-28 17:58:47 train.py: 89] Epoch 11, iter 200/6416, lr 0.000167, loss 2.263867
+INFO 2021-10-28 18:02:06 train.py: 89] Epoch 11, iter 400/6416, lr 0.000165, loss 2.240852
+INFO 2021-10-28 18:05:27 train.py: 89] Epoch 11, iter 600/6416, lr 0.000164, loss 2.262677
+INFO 2021-10-28 18:08:47 train.py: 89] Epoch 11, iter 800/6416, lr 0.000163, loss 2.248080
+INFO 2021-10-28 18:12:06 train.py: 89] Epoch 11, iter 1000/6416, lr 0.000162, loss 2.258660
+INFO 2021-10-28 18:15:25 train.py: 89] Epoch 11, iter 1200/6416, lr 0.000160, loss 2.210815
+INFO 2021-10-28 18:18:44 train.py: 89] Epoch 11, iter 1400/6416, lr 0.000159, loss 2.204155
+INFO 2021-10-28 18:22:03 train.py: 89] Epoch 11, iter 1600/6416, lr 0.000158, loss 2.227258
+INFO 2021-10-28 18:25:22 train.py: 89] Epoch 11, iter 1800/6416, lr 0.000157, loss 2.182560
+INFO 2021-10-28 18:28:41 train.py: 89] Epoch 11, iter 2000/6416, lr 0.000155, loss 2.170454
+INFO 2021-10-28 18:32:01 train.py: 89] Epoch 11, iter 2200/6416, lr 0.000154, loss 2.204617
+INFO 2021-10-28 18:35:20 train.py: 89] Epoch 11, iter 2400/6416, lr 0.000153, loss 2.170376
+INFO 2021-10-28 18:38:40 train.py: 89] Epoch 11, iter 2600/6416, lr 0.000152, loss 2.185701
+INFO 2021-10-28 18:41:59 train.py: 89] Epoch 11, iter 2800/6416, lr 0.000150, loss 2.184631
+INFO 2021-10-28 18:45:19 train.py: 101] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-10-28 18:45:20 train.py: 89] Epoch 11, iter 3000/6416, lr 0.000149, loss 2.170771
+INFO 2021-10-28 18:48:40 train.py: 89] Epoch 11, iter 3200/6416, lr 0.000148, loss 2.151940
+INFO 2021-10-28 18:52:00 train.py: 89] Epoch 11, iter 3400/6416, lr 0.000147, loss 2.139898
+INFO 2021-10-28 18:55:19 train.py: 89] Epoch 11, iter 3600/6416, lr 0.000146, loss 2.159178
+INFO 2021-10-28 18:58:39 train.py: 89] Epoch 11, iter 3800/6416, lr 0.000144, loss 2.116504
+INFO 2021-10-28 19:01:57 train.py: 89] Epoch 11, iter 4000/6416, lr 0.000143, loss 2.155296
+INFO 2021-10-28 19:05:17 train.py: 89] Epoch 11, iter 4200/6416, lr 0.000142, loss 2.114795
+INFO 2021-10-28 19:08:36 train.py: 89] Epoch 11, iter 4400/6416, lr 0.000141, loss 2.105746
+INFO 2021-10-28 19:11:56 train.py: 89] Epoch 11, iter 4600/6416, lr 0.000139, loss 2.129044
+INFO 2021-10-28 19:15:16 train.py: 89] Epoch 11, iter 4800/6416, lr 0.000138, loss 2.089457
+INFO 2021-10-28 19:18:36 train.py: 89] Epoch 11, iter 5000/6416, lr 0.000137, loss 2.097527
+INFO 2021-10-28 19:21:55 train.py: 89] Epoch 11, iter 5200/6416, lr 0.000136, loss 2.081058
+INFO 2021-10-28 19:25:15 train.py: 89] Epoch 11, iter 5400/6416, lr 0.000135, loss 2.052979
+INFO 2021-10-28 19:28:35 train.py: 89] Epoch 11, iter 5600/6416, lr 0.000134, loss 2.062901
+INFO 2021-10-28 19:31:55 train.py: 89] Epoch 11, iter 5800/6416, lr 0.000132, loss 2.059637
+INFO 2021-10-28 19:35:15 train.py: 101] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-10-28 19:35:16 train.py: 89] Epoch 11, iter 6000/6416, lr 0.000131, loss 2.049424
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-28 19:38:36 train.py: 89] Epoch 11, iter 6200/6416, lr 0.000130, loss 2.046439
+INFO 2021-10-28 19:41:54 train.py: 89] Epoch 11, iter 6400/6416, lr 0.000129, loss 2.057654
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 19:42:10 train.py: 108] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-10-28 19:42:11 train.py: 89] Epoch 12, iter 0/6416, lr 0.000129, loss 2.092508
+INFO 2021-10-28 19:49:59 train.py: 89] Epoch 12, iter 200/6416, lr 0.000128, loss 2.060046
+INFO 2021-10-28 19:59:25 train.py: 89] Epoch 12, iter 400/6416, lr 0.000126, loss 2.035100
+INFO 2021-10-28 20:07:36 train.py: 89] Epoch 12, iter 600/6416, lr 0.000125, loss 2.070074
+INFO 2021-10-28 20:16:01 train.py: 89] Epoch 12, iter 800/6416, lr 0.000124, loss 2.038943
+INFO 2021-10-28 20:24:04 train.py: 89] Epoch 12, iter 1000/6416, lr 0.000123, loss 2.039655
+INFO 2021-10-28 20:32:21 train.py: 89] Epoch 12, iter 1200/6416, lr 0.000122, loss 2.007822
+INFO 2021-10-28 20:40:36 train.py: 89] Epoch 12, iter 1400/6416, lr 0.000121, loss 2.000669
+INFO 2021-10-28 20:48:41 train.py: 89] Epoch 12, iter 1600/6416, lr 0.000120, loss 2.024840
+INFO 2021-10-28 20:56:57 train.py: 89] Epoch 12, iter 1800/6416, lr 0.000118, loss 1.973704
+INFO 2021-10-28 21:04:52 train.py: 89] Epoch 12, iter 2000/6416, lr 0.000117, loss 1.958400
+INFO 2021-10-28 21:13:31 train.py: 89] Epoch 12, iter 2200/6416, lr 0.000116, loss 1.986368
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-28 21:21:42 train.py: 89] Epoch 12, iter 2400/6416, lr 0.000115, loss 1.980485
+INFO 2021-10-28 21:29:45 train.py: 89] Epoch 12, iter 2600/6416, lr 0.000114, loss 1.981403
+INFO 2021-10-28 21:37:52 train.py: 89] Epoch 12, iter 2800/6416, lr 0.000113, loss 1.976436
+INFO 2021-10-28 21:46:08 train.py: 101] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-10-28 21:46:22 train.py: 89] Epoch 12, iter 3000/6416, lr 0.000112, loss 1.963808
+INFO 2021-10-28 21:54:34 train.py: 89] Epoch 12, iter 3200/6416, lr 0.000111, loss 1.947435
+INFO 2021-10-28 22:02:37 train.py: 89] Epoch 12, iter 3400/6416, lr 0.000109, loss 1.942628
+INFO 2021-10-28 22:11:04 train.py: 89] Epoch 12, iter 3600/6416, lr 0.000108, loss 1.947086
+INFO 2021-10-28 22:19:15 train.py: 89] Epoch 12, iter 3800/6416, lr 0.000107, loss 1.910440
+INFO 2021-10-28 22:27:20 train.py: 89] Epoch 12, iter 4000/6416, lr 0.000106, loss 1.941989
+INFO 2021-10-28 22:35:23 train.py: 89] Epoch 12, iter 4200/6416, lr 0.000105, loss 1.913958
+INFO 2021-10-28 22:44:07 train.py: 89] Epoch 12, iter 4400/6416, lr 0.000104, loss 1.895913
+INFO 2021-10-28 22:52:13 train.py: 89] Epoch 12, iter 4600/6416, lr 0.000103, loss 1.920565
+INFO 2021-10-28 23:01:00 train.py: 89] Epoch 12, iter 4800/6416, lr 0.000102, loss 1.893225
+INFO 2021-10-28 23:09:21 train.py: 89] Epoch 12, iter 5000/6416, lr 0.000101, loss 1.909615
+INFO 2021-10-28 23:17:48 train.py: 89] Epoch 12, iter 5200/6416, lr 0.000100, loss 1.892094
+INFO 2021-10-28 23:25:48 train.py: 89] Epoch 12, iter 5400/6416, lr 0.000099, loss 1.861759
+INFO 2021-10-28 23:33:47 train.py: 89] Epoch 12, iter 5600/6416, lr 0.000098, loss 1.873800
+INFO 2021-10-28 23:42:25 train.py: 89] Epoch 12, iter 5800/6416, lr 0.000097, loss 1.875867
+INFO 2021-10-28 23:50:21 train.py: 101] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-10-28 23:50:27 train.py: 89] Epoch 12, iter 6000/6416, lr 0.000096, loss 1.862276
+INFO 2021-10-28 23:58:39 train.py: 89] Epoch 12, iter 6200/6416, lr 0.000095, loss 1.860097
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-29 00:06:45 train.py: 89] Epoch 12, iter 6400/6416, lr 0.000093, loss 1.856626
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 00:07:35 train.py: 108] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-10-29 00:07:37 train.py: 89] Epoch 13, iter 0/6416, lr 0.000093, loss 1.884216
+INFO 2021-10-29 00:10:56 train.py: 89] Epoch 13, iter 200/6416, lr 0.000092, loss 1.860971
+INFO 2021-10-29 00:14:15 train.py: 89] Epoch 13, iter 400/6416, lr 0.000091, loss 1.839763
+INFO 2021-10-29 00:17:35 train.py: 89] Epoch 13, iter 600/6416, lr 0.000090, loss 1.864118
+INFO 2021-10-29 00:20:54 train.py: 89] Epoch 13, iter 800/6416, lr 0.000089, loss 1.843837
+INFO 2021-10-29 00:24:14 train.py: 89] Epoch 13, iter 1000/6416, lr 0.000088, loss 1.848595
+INFO 2021-10-29 00:27:34 train.py: 89] Epoch 13, iter 1200/6416, lr 0.000087, loss 1.837939
+INFO 2021-10-29 00:30:54 train.py: 89] Epoch 13, iter 1400/6416, lr 0.000086, loss 1.811434
+INFO 2021-10-29 00:34:14 train.py: 89] Epoch 13, iter 1600/6416, lr 0.000085, loss 1.830203
+INFO 2021-10-29 00:37:33 train.py: 89] Epoch 13, iter 1800/6416, lr 0.000084, loss 1.783875
+INFO 2021-10-29 00:40:53 train.py: 89] Epoch 13, iter 2000/6416, lr 0.000083, loss 1.780625
+INFO 2021-10-29 00:44:12 train.py: 89] Epoch 13, iter 2200/6416, lr 0.000082, loss 1.812669
+INFO 2021-10-29 00:47:31 train.py: 89] Epoch 13, iter 2400/6416, lr 0.000081, loss 1.800422
+INFO 2021-10-29 00:50:50 train.py: 89] Epoch 13, iter 2600/6416, lr 0.000080, loss 1.795023
+INFO 2021-10-29 00:54:10 train.py: 89] Epoch 13, iter 2800/6416, lr 0.000079, loss 1.795762
+INFO 2021-10-29 00:57:31 train.py: 101] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-10-29 00:57:31 train.py: 89] Epoch 13, iter 3000/6416, lr 0.000078, loss 1.779563
+INFO 2021-10-29 01:00:51 train.py: 89] Epoch 13, iter 3200/6416, lr 0.000078, loss 1.766444
+INFO 2021-10-29 01:04:11 train.py: 89] Epoch 13, iter 3400/6416, lr 0.000077, loss 1.776283
+INFO 2021-10-29 01:07:29 train.py: 89] Epoch 13, iter 3600/6416, lr 0.000076, loss 1.780677
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 01:10:49 train.py: 89] Epoch 13, iter 3800/6416, lr 0.000075, loss 1.731199
+INFO 2021-10-29 01:14:09 train.py: 89] Epoch 13, iter 4000/6416, lr 0.000074, loss 1.775170
+INFO 2021-10-29 01:17:29 train.py: 89] Epoch 13, iter 4200/6416, lr 0.000073, loss 1.738853
+INFO 2021-10-29 01:20:48 train.py: 89] Epoch 13, iter 4400/6416, lr 0.000072, loss 1.730463
+INFO 2021-10-29 01:24:08 train.py: 89] Epoch 13, iter 4600/6416, lr 0.000071, loss 1.757649
+INFO 2021-10-29 01:27:26 train.py: 89] Epoch 13, iter 4800/6416, lr 0.000070, loss 1.721404
+INFO 2021-10-29 01:30:45 train.py: 89] Epoch 13, iter 5000/6416, lr 0.000069, loss 1.737237
+INFO 2021-10-29 01:34:04 train.py: 89] Epoch 13, iter 5200/6416, lr 0.000068, loss 1.723513
+INFO 2021-10-29 01:37:23 train.py: 89] Epoch 13, iter 5400/6416, lr 0.000067, loss 1.682793
+INFO 2021-10-29 01:40:45 train.py: 89] Epoch 13, iter 5600/6416, lr 0.000066, loss 1.715812
+INFO 2021-10-29 01:44:04 train.py: 89] Epoch 13, iter 5800/6416, lr 0.000066, loss 1.710610
+INFO 2021-10-29 01:47:24 train.py: 101] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-10-29 01:47:25 train.py: 89] Epoch 13, iter 6000/6416, lr 0.000065, loss 1.698509
+INFO 2021-10-29 01:50:44 train.py: 89] Epoch 13, iter 6200/6416, lr 0.000064, loss 1.694478
+INFO 2021-10-29 01:54:04 train.py: 89] Epoch 13, iter 6400/6416, lr 0.000063, loss 1.701797
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 01:54:21 train.py: 108] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-10-29 01:54:22 train.py: 89] Epoch 14, iter 0/6416, lr 0.000063, loss 1.719011
+INFO 2021-10-29 01:57:41 train.py: 89] Epoch 14, iter 200/6416, lr 0.000062, loss 1.711288
+INFO 2021-10-29 02:01:01 train.py: 89] Epoch 14, iter 400/6416, lr 0.000061, loss 1.684729
+INFO 2021-10-29 02:04:20 train.py: 89] Epoch 14, iter 600/6416, lr 0.000060, loss 1.707041
+INFO 2021-10-29 02:07:40 train.py: 89] Epoch 14, iter 800/6416, lr 0.000059, loss 1.676569
+INFO 2021-10-29 02:11:02 train.py: 89] Epoch 14, iter 1000/6416, lr 0.000059, loss 1.684303
+INFO 2021-10-29 02:14:21 train.py: 89] Epoch 14, iter 1200/6416, lr 0.000058, loss 1.678350
+INFO 2021-10-29 02:17:41 train.py: 89] Epoch 14, iter 1400/6416, lr 0.000057, loss 1.649548
+INFO 2021-10-29 02:21:01 train.py: 89] Epoch 14, iter 1600/6416, lr 0.000056, loss 1.672499
+INFO 2021-10-29 02:24:21 train.py: 89] Epoch 14, iter 1800/6416, lr 0.000055, loss 1.636551
+INFO 2021-10-29 02:27:39 train.py: 89] Epoch 14, iter 2000/6416, lr 0.000055, loss 1.626832
+INFO 2021-10-29 02:30:58 train.py: 89] Epoch 14, iter 2200/6416, lr 0.000054, loss 1.652834
+INFO 2021-10-29 02:34:16 train.py: 89] Epoch 14, iter 2400/6416, lr 0.000053, loss 1.652167
+INFO 2021-10-29 02:37:36 train.py: 89] Epoch 14, iter 2600/6416, lr 0.000052, loss 1.641388
+INFO 2021-10-29 02:40:54 train.py: 89] Epoch 14, iter 2800/6416, lr 0.000051, loss 1.646133
+INFO 2021-10-29 02:44:13 train.py: 101] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-10-29 02:44:14 train.py: 89] Epoch 14, iter 3000/6416, lr 0.000051, loss 1.627837
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 02:47:35 train.py: 89] Epoch 14, iter 3200/6416, lr 0.000050, loss 1.615743
+INFO 2021-10-29 02:50:55 train.py: 89] Epoch 14, iter 3400/6416, lr 0.000049, loss 1.621257
+INFO 2021-10-29 02:54:13 train.py: 89] Epoch 14, iter 3600/6416, lr 0.000048, loss 1.625576
+INFO 2021-10-29 02:57:33 train.py: 89] Epoch 14, iter 3800/6416, lr 0.000048, loss 1.589151
+INFO 2021-10-29 03:00:53 train.py: 89] Epoch 14, iter 4000/6416, lr 0.000047, loss 1.636877
+INFO 2021-10-29 03:04:13 train.py: 89] Epoch 14, iter 4200/6416, lr 0.000046, loss 1.589705
+INFO 2021-10-29 03:07:32 train.py: 89] Epoch 14, iter 4400/6416, lr 0.000045, loss 1.580660
+INFO 2021-10-29 03:10:52 train.py: 89] Epoch 14, iter 4600/6416, lr 0.000045, loss 1.604650
+INFO 2021-10-29 03:14:11 train.py: 89] Epoch 14, iter 4800/6416, lr 0.000044, loss 1.585943
+INFO 2021-10-29 03:17:31 train.py: 89] Epoch 14, iter 5000/6416, lr 0.000043, loss 1.602329
+INFO 2021-10-29 03:20:51 train.py: 89] Epoch 14, iter 5200/6416, lr 0.000042, loss 1.578424
+INFO 2021-10-29 03:24:11 train.py: 89] Epoch 14, iter 5400/6416, lr 0.000042, loss 1.552562
+INFO 2021-10-29 03:27:30 train.py: 89] Epoch 14, iter 5600/6416, lr 0.000041, loss 1.567654
+INFO 2021-10-29 03:30:49 train.py: 89] Epoch 14, iter 5800/6416, lr 0.000040, loss 1.573535
+INFO 2021-10-29 03:34:11 train.py: 101] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-10-29 03:34:12 train.py: 89] Epoch 14, iter 6000/6416, lr 0.000040, loss 1.568791
+INFO 2021-10-29 03:37:31 train.py: 89] Epoch 14, iter 6200/6416, lr 0.000039, loss 1.557410
+INFO 2021-10-29 03:40:50 train.py: 89] Epoch 14, iter 6400/6416, lr 0.000038, loss 1.566284
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 03:41:05 train.py: 108] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-10-29 03:41:07 train.py: 89] Epoch 15, iter 0/6416, lr 0.000038, loss 1.591660
+INFO 2021-10-29 03:44:27 train.py: 89] Epoch 15, iter 200/6416, lr 0.000037, loss 1.578093
+INFO 2021-10-29 03:47:47 train.py: 89] Epoch 15, iter 400/6416, lr 0.000037, loss 1.550428
+INFO 2021-10-29 03:51:07 train.py: 89] Epoch 15, iter 600/6416, lr 0.000036, loss 1.582690
+INFO 2021-10-29 03:54:28 train.py: 89] Epoch 15, iter 800/6416, lr 0.000036, loss 1.546944
+INFO 2021-10-29 03:57:46 train.py: 89] Epoch 15, iter 1000/6416, lr 0.000035, loss 1.563195
+INFO 2021-10-29 04:01:07 train.py: 89] Epoch 15, iter 1200/6416, lr 0.000034, loss 1.548260
+INFO 2021-10-29 04:04:26 train.py: 89] Epoch 15, iter 1400/6416, lr 0.000034, loss 1.523338
+INFO 2021-10-29 04:07:47 train.py: 89] Epoch 15, iter 1600/6416, lr 0.000033, loss 1.551978
+INFO 2021-10-29 04:11:07 train.py: 89] Epoch 15, iter 1800/6416, lr 0.000032, loss 1.524203
+INFO 2021-10-29 04:14:28 train.py: 89] Epoch 15, iter 2000/6416, lr 0.000032, loss 1.509108
+INFO 2021-10-29 04:17:48 train.py: 89] Epoch 15, iter 2200/6416, lr 0.000031, loss 1.535697
+INFO 2021-10-29 04:21:07 train.py: 89] Epoch 15, iter 2400/6416, lr 0.000031, loss 1.531258
+INFO 2021-10-29 04:24:24 train.py: 89] Epoch 15, iter 2600/6416, lr 0.000030, loss 1.528870
+INFO 2021-10-29 04:27:44 train.py: 89] Epoch 15, iter 2800/6416, lr 0.000029, loss 1.535915
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 04:31:04 train.py: 101] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-10-29 04:31:05 train.py: 89] Epoch 15, iter 3000/6416, lr 0.000029, loss 1.516979
+INFO 2021-10-29 04:34:26 train.py: 89] Epoch 15, iter 3200/6416, lr 0.000028, loss 1.496399
+INFO 2021-10-29 04:37:46 train.py: 89] Epoch 15, iter 3400/6416, lr 0.000028, loss 1.506587
+INFO 2021-10-29 04:41:07 train.py: 89] Epoch 15, iter 3600/6416, lr 0.000027, loss 1.505540
+INFO 2021-10-29 04:44:26 train.py: 89] Epoch 15, iter 3800/6416, lr 0.000027, loss 1.484397
+INFO 2021-10-29 04:47:46 train.py: 89] Epoch 15, iter 4000/6416, lr 0.000026, loss 1.521520
+INFO 2021-10-29 04:51:06 train.py: 89] Epoch 15, iter 4200/6416, lr 0.000025, loss 1.489612
+INFO 2021-10-29 04:54:26 train.py: 89] Epoch 15, iter 4400/6416, lr 0.000025, loss 1.481175
+INFO 2021-10-29 04:57:45 train.py: 89] Epoch 15, iter 4600/6416, lr 0.000024, loss 1.513087
+INFO 2021-10-29 05:01:05 train.py: 89] Epoch 15, iter 4800/6416, lr 0.000024, loss 1.473940
+INFO 2021-10-29 05:04:25 train.py: 89] Epoch 15, iter 5000/6416, lr 0.000023, loss 1.491677
+INFO 2021-10-29 05:07:45 train.py: 89] Epoch 15, iter 5200/6416, lr 0.000023, loss 1.480122
+INFO 2021-10-29 05:11:06 train.py: 89] Epoch 15, iter 5400/6416, lr 0.000022, loss 1.459458
+INFO 2021-10-29 05:14:25 train.py: 89] Epoch 15, iter 5600/6416, lr 0.000022, loss 1.476707
+INFO 2021-10-29 05:17:46 train.py: 89] Epoch 15, iter 5800/6416, lr 0.000021, loss 1.475208
+INFO 2021-10-29 05:21:06 train.py: 101] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-10-29 05:21:07 train.py: 89] Epoch 15, iter 6000/6416, lr 0.000021, loss 1.477782
+INFO 2021-10-29 05:24:27 train.py: 89] Epoch 15, iter 6200/6416, lr 0.000020, loss 1.459215
+INFO 2021-10-29 05:27:48 train.py: 89] Epoch 15, iter 6400/6416, lr 0.000020, loss 1.477508
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 05:28:04 train.py: 108] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-10-29 05:28:05 train.py: 89] Epoch 16, iter 0/6416, lr 0.000020, loss 1.467504
+INFO 2021-10-29 05:31:26 train.py: 89] Epoch 16, iter 200/6416, lr 0.000019, loss 1.483658
+INFO 2021-10-29 05:34:48 train.py: 89] Epoch 16, iter 400/6416, lr 0.000019, loss 1.460342
+INFO 2021-10-29 05:38:10 train.py: 89] Epoch 16, iter 600/6416, lr 0.000019, loss 1.491294
+INFO 2021-10-29 05:41:30 train.py: 89] Epoch 16, iter 800/6416, lr 0.000018, loss 1.448856
+INFO 2021-10-29 05:44:50 train.py: 89] Epoch 16, iter 1000/6416, lr 0.000018, loss 1.456664
+INFO 2021-10-29 05:48:09 train.py: 89] Epoch 16, iter 1200/6416, lr 0.000017, loss 1.452737
+INFO 2021-10-29 05:51:28 train.py: 89] Epoch 16, iter 1400/6416, lr 0.000017, loss 1.442130
+INFO 2021-10-29 05:54:46 train.py: 89] Epoch 16, iter 1600/6416, lr 0.000016, loss 1.467303
+INFO 2021-10-29 05:58:06 train.py: 89] Epoch 16, iter 1800/6416, lr 0.000016, loss 1.435211
+INFO 2021-10-29 06:01:25 train.py: 89] Epoch 16, iter 2000/6416, lr 0.000016, loss 1.426680
+INFO 2021-10-29 06:04:43 train.py: 89] Epoch 16, iter 2200/6416, lr 0.000015, loss 1.445586
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 06:08:02 train.py: 89] Epoch 16, iter 2400/6416, lr 0.000015, loss 1.449211
+INFO 2021-10-29 06:11:22 train.py: 89] Epoch 16, iter 2600/6416, lr 0.000015, loss 1.445514
+INFO 2021-10-29 06:14:42 train.py: 89] Epoch 16, iter 2800/6416, lr 0.000014, loss 1.445417
+INFO 2021-10-29 06:18:02 train.py: 101] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-10-29 06:18:03 train.py: 89] Epoch 16, iter 3000/6416, lr 0.000014, loss 1.441951
+INFO 2021-10-29 06:21:24 train.py: 89] Epoch 16, iter 3200/6416, lr 0.000013, loss 1.435696
+INFO 2021-10-29 06:24:42 train.py: 89] Epoch 16, iter 3400/6416, lr 0.000013, loss 1.424956
+INFO 2021-10-29 06:28:02 train.py: 89] Epoch 16, iter 3600/6416, lr 0.000013, loss 1.444388
+INFO 2021-10-29 06:31:21 train.py: 89] Epoch 16, iter 3800/6416, lr 0.000012, loss 1.407472
+INFO 2021-10-29 06:34:41 train.py: 89] Epoch 16, iter 4000/6416, lr 0.000012, loss 1.460436
+INFO 2021-10-29 06:38:01 train.py: 89] Epoch 16, iter 4200/6416, lr 0.000012, loss 1.411899
+INFO 2021-10-29 06:41:22 train.py: 89] Epoch 16, iter 4400/6416, lr 0.000011, loss 1.408492
+INFO 2021-10-29 06:44:42 train.py: 89] Epoch 16, iter 4600/6416, lr 0.000011, loss 1.432907
+INFO 2021-10-29 06:48:01 train.py: 89] Epoch 16, iter 4800/6416, lr 0.000011, loss 1.410086
+INFO 2021-10-29 06:51:20 train.py: 89] Epoch 16, iter 5000/6416, lr 0.000011, loss 1.428360
+INFO 2021-10-29 06:54:40 train.py: 89] Epoch 16, iter 5200/6416, lr 0.000010, loss 1.409050
+INFO 2021-10-29 06:58:01 train.py: 89] Epoch 16, iter 5400/6416, lr 0.000010, loss 1.385086
+INFO 2021-10-29 07:01:21 train.py: 89] Epoch 16, iter 5600/6416, lr 0.000010, loss 1.404112
+INFO 2021-10-29 07:04:39 train.py: 89] Epoch 16, iter 5800/6416, lr 0.000010, loss 1.418528
+INFO 2021-10-29 07:08:00 train.py: 101] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-10-29 07:08:01 train.py: 89] Epoch 16, iter 6000/6416, lr 0.000009, loss 1.409249
+INFO 2021-10-29 07:11:21 train.py: 89] Epoch 16, iter 6200/6416, lr 0.000009, loss 1.412503
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-29 07:14:41 train.py: 89] Epoch 16, iter 6400/6416, lr 0.000009, loss 1.410205
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 07:14:57 train.py: 108] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-10-29 07:14:58 train.py: 89] Epoch 17, iter 0/6416, lr 0.000009, loss 1.458197
+INFO 2021-10-29 07:18:18 train.py: 89] Epoch 17, iter 200/6416, lr 0.000009, loss 1.430702
+INFO 2021-10-29 07:21:38 train.py: 89] Epoch 17, iter 400/6416, lr 0.000008, loss 1.407919
+INFO 2021-10-29 07:24:57 train.py: 89] Epoch 17, iter 600/6416, lr 0.000008, loss 1.429480
+INFO 2021-10-29 07:28:16 train.py: 89] Epoch 17, iter 800/6416, lr 0.000008, loss 1.404181
+INFO 2021-10-29 07:31:36 train.py: 89] Epoch 17, iter 1000/6416, lr 0.000008, loss 1.416104
+INFO 2021-10-29 07:34:58 train.py: 89] Epoch 17, iter 1200/6416, lr 0.000007, loss 1.396715
+INFO 2021-10-29 07:38:17 train.py: 89] Epoch 17, iter 1400/6416, lr 0.000007, loss 1.386170
+INFO 2021-10-29 07:41:36 train.py: 89] Epoch 17, iter 1600/6416, lr 0.000007, loss 1.409991
+INFO 2021-10-29 07:44:54 train.py: 89] Epoch 17, iter 1800/6416, lr 0.000007, loss 1.378998
+INFO 2021-10-29 07:48:15 train.py: 89] Epoch 17, iter 2000/6416, lr 0.000007, loss 1.384717
+INFO 2021-10-29 07:51:34 train.py: 89] Epoch 17, iter 2200/6416, lr 0.000007, loss 1.398329
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 07:54:51 train.py: 89] Epoch 17, iter 2400/6416, lr 0.000006, loss 1.396066
+INFO 2021-10-29 07:58:11 train.py: 89] Epoch 17, iter 2600/6416, lr 0.000006, loss 1.404444
+INFO 2021-10-29 08:01:29 train.py: 89] Epoch 17, iter 2800/6416, lr 0.000006, loss 1.409084
+INFO 2021-10-29 08:04:50 train.py: 101] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-10-29 08:04:51 train.py: 89] Epoch 17, iter 3000/6416, lr 0.000006, loss 1.398937
+INFO 2021-10-29 08:08:11 train.py: 89] Epoch 17, iter 3200/6416, lr 0.000006, loss 1.387884
+INFO 2021-10-29 08:11:30 train.py: 89] Epoch 17, iter 3400/6416, lr 0.000006, loss 1.386974
+INFO 2021-10-29 08:14:50 train.py: 89] Epoch 17, iter 3600/6416, lr 0.000006, loss 1.391950
+INFO 2021-10-29 08:18:10 train.py: 89] Epoch 17, iter 3800/6416, lr 0.000006, loss 1.366216
+INFO 2021-10-29 08:21:29 train.py: 89] Epoch 17, iter 4000/6416, lr 0.000006, loss 1.401237
+INFO 2021-10-29 08:24:49 train.py: 89] Epoch 17, iter 4200/6416, lr 0.000005, loss 1.384945
+INFO 2021-10-29 08:28:08 train.py: 89] Epoch 17, iter 4400/6416, lr 0.000005, loss 1.372791
+INFO 2021-10-29 08:31:28 train.py: 89] Epoch 17, iter 4600/6416, lr 0.000005, loss 1.392185
+INFO 2021-10-29 08:34:47 train.py: 89] Epoch 17, iter 4800/6416, lr 0.000005, loss 1.375438
+INFO 2021-10-29 08:38:06 train.py: 89] Epoch 17, iter 5000/6416, lr 0.000005, loss 1.393828
+INFO 2021-10-29 08:41:25 train.py: 89] Epoch 17, iter 5200/6416, lr 0.000005, loss 1.375953
+INFO 2021-10-29 08:44:43 train.py: 89] Epoch 17, iter 5400/6416, lr 0.000005, loss 1.355519
+INFO 2021-10-29 08:48:02 train.py: 89] Epoch 17, iter 5600/6416, lr 0.000005, loss 1.369040
+INFO 2021-10-29 08:51:23 train.py: 89] Epoch 17, iter 5800/6416, lr 0.000005, loss 1.393012
+INFO 2021-10-29 08:54:43 train.py: 101] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-10-29 08:54:44 train.py: 89] Epoch 17, iter 6000/6416, lr 0.000005, loss 1.379958
+INFO 2021-10-29 08:58:03 train.py: 89] Epoch 17, iter 6200/6416, lr 0.000005, loss 1.375570
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 65536.0
+INFO 2021-10-29 09:01:23 train.py: 89] Epoch 17, iter 6400/6416, lr 0.000005, loss 1.388825
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+Gradient overflow. Skipping step, loss scaler 0 reducing loss scale to 32768.0
+INFO 2021-10-29 09:01:39 train.py: 108] Save checkpoint Epoch_17.pt to disk...
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/.gitkeep b/bob/bio/facexzoo/models/backbones/TF_NAS_A/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..64afb59008657584114c9e974256c2978de3bf6c
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_agedb.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9723333333333333 | 0.0024745619390355578 |
+| Epoch_13.pt | 0.9713333333333335 | 0.0029999999999999988 |
+| Epoch_16_batch_2999.pt | 0.9713333333333335 | 0.0031011746082117465 |
+| Epoch_11_batch_5999.pt | 0.9711666666666667 | 0.0020344259359556154 |
+| Epoch_16.pt | 0.9711666666666666 | 0.0027448042948968097 |
+| Epoch_16_batch_5999.pt | 0.9710000000000001 | 0.0028021156028707733 |
+| Epoch_17_batch_5999.pt | 0.9706666666666667 | 0.0024241582476968258 |
+| Epoch_15_batch_5999.pt | 0.9704999999999998 | 0.002710860644167974 |
+| Epoch_14_batch_2999.pt | 0.9703333333333333 | 0.002708012801545322 |
+| Epoch_13_batch_5999.pt | 0.9701666666666668 | 0.002870131411797662 |
+| Epoch_14_batch_5999.pt | 0.9701666666666668 | 0.002891558595482913 |
+| Epoch_15_batch_2999.pt | 0.9701666666666668 | 0.0025873624493766715 |
+| Epoch_14.pt | 0.9700000000000001 | 0.0031525024353580267 |
+| Epoch_15.pt | 0.9696666666666667 | 0.002958561545709856 |
+| Epoch_13_batch_2999.pt | 0.9695 | 0.00298194153340437 |
+| Epoch_17_batch_2999.pt | 0.9693333333333334 | 0.0030651364942519423 |
+| Epoch_12_batch_5999.pt | 0.9693333333333332 | 0.0026199613605670247 |
+| Epoch_12_batch_2999.pt | 0.9691666666666666 | 0.002644000896509064 |
+| Epoch_11_batch_2999.pt | 0.9691666666666666 | 0.002573008039313202 |
+| Epoch_10_batch_5999.pt | 0.9685000000000002 | 0.0025754059969998327 |
+| Epoch_11.pt | 0.9681666666666668 | 0.0023100692095879807 |
+| Epoch_12.pt | 0.9678333333333333 | 0.0025825865109265506 |
+| Epoch_10_batch_2999.pt | 0.9673333333333336 | 0.0030550504633038923 |
+| Epoch_10.pt | 0.9663333333333334 | 0.00289529204906564 |
+| Epoch_9_batch_5999.pt | 0.959 | 0.003362832433427014 |
+| Epoch_9_batch_2999.pt | 0.958 | 0.0026965913554470207 |
+| Epoch_7_batch_5999.pt | 0.9550000000000001 | 0.002290614236454262 |
+| Epoch_8_batch_2999.pt | 0.9535 | 0.0038365525934710115 |
+| Epoch_7_batch_2999.pt | 0.9516666666666668 | 0.004230985058813283 |
+| Epoch_9.pt | 0.9510000000000002 | 0.00344444444444444 |
+| Epoch_7.pt | 0.951 | 0.003559026084010438 |
+| Epoch_8.pt | 0.9506666666666665 | 0.0034533933928711214 |
+| Epoch_8_batch_5999.pt | 0.9486666666666667 | 0.002219442706159799 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0ca66c3efa924882f3eea8d440e8f2c6a826239f
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_calfw.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9486666666666667 | 0.003476552854924892 |
+| Epoch_12.pt | 0.9486666666666667 | 0.0033957126199858066 |
+| Epoch_15_batch_2999.pt | 0.9486666666666667 | 0.003275422882885261 |
+| Epoch_17_batch_5999.pt | 0.9483333333333333 | 0.003191423692521125 |
+| Epoch_15_batch_5999.pt | 0.9476666666666667 | 0.0029731307022799183 |
+| Epoch_10.pt | 0.9476666666666667 | 0.0033259176771323912 |
+| Epoch_11_batch_2999.pt | 0.9475 | 0.003307772365912045 |
+| Epoch_14_batch_5999.pt | 0.9471666666666667 | 0.0034251250315107456 |
+| Epoch_16_batch_5999.pt | 0.9471666666666666 | 0.0032871804872193354 |
+| Epoch_14_batch_2999.pt | 0.9471666666666666 | 0.003452052529534663 |
+| Epoch_13_batch_2999.pt | 0.9470000000000001 | 0.0034676636742949378 |
+| Epoch_10_batch_5999.pt | 0.9466666666666667 | 0.0033609963249453738 |
+| Epoch_13.pt | 0.9466666666666667 | 0.003513641844631532 |
+| Epoch_13_batch_5999.pt | 0.9465 | 0.003672133971035721 |
+| Epoch_12_batch_2999.pt | 0.9465 | 0.0030877096070200546 |
+| Epoch_16.pt | 0.9461666666666668 | 0.003549039167485431 |
+| Epoch_11.pt | 0.9461666666666668 | 0.0033979841518606718 |
+| Epoch_11_batch_5999.pt | 0.9456666666666667 | 0.003408413700039547 |
+| Epoch_15.pt | 0.9455000000000002 | 0.0033245254000838173 |
+| Epoch_12_batch_5999.pt | 0.9455 | 0.0036855573979159965 |
+| Epoch_10_batch_2999.pt | 0.9455 | 0.003857412297075545 |
+| Epoch_17_batch_2999.pt | 0.9455 | 0.003478771600989613 |
+| Epoch_14.pt | 0.9445 | 0.0036349639562123512 |
+| Epoch_16_batch_2999.pt | 0.9445 | 0.003514081022415214 |
+| Epoch_7_batch_5999.pt | 0.9404999999999999 | 0.003702268240343582 |
+| Epoch_8_batch_5999.pt | 0.9395000000000001 | 0.0039287638240526985 |
+| Epoch_8_batch_2999.pt | 0.9385 | 0.0043422273399264695 |
+| Epoch_9_batch_5999.pt | 0.9381666666666668 | 0.0039942087706708205 |
+| Epoch_9_batch_2999.pt | 0.9375 | 0.004196191689062601 |
+| Epoch_7_batch_2999.pt | 0.9353333333333333 | 0.003973678831610597 |
+| Epoch_7.pt | 0.9343333333333333 | 0.0047088044667593495 |
+| Epoch_8.pt | 0.9341666666666667 | 0.0036955929710139074 |
+| Epoch_9.pt | 0.9280000000000002 | 0.004265856404544987 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0ff4db3eb549c5090b4f2b4c6745de550826a07d
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_cplfw.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.859 | 0.00723161782012132 |
+| Epoch_14_batch_5999.pt | 0.8584999999999999 | 0.007215595250207341 |
+| Epoch_12_batch_2999.pt | 0.8583333333333332 | 0.006898738478344618 |
+| Epoch_15_batch_2999.pt | 0.858 | 0.007469278643670357 |
+| Epoch_15_batch_5999.pt | 0.8576666666666666 | 0.007282653159188418 |
+| Epoch_13_batch_5999.pt | 0.8573333333333334 | 0.007450246495217864 |
+| Epoch_17_batch_5999.pt | 0.8573333333333334 | 0.007528546496215526 |
+| Epoch_16_batch_5999.pt | 0.857 | 0.007242705994547676 |
+| Epoch_13_batch_2999.pt | 0.8568333333333333 | 0.007500823000112205 |
+| Epoch_14_batch_2999.pt | 0.8563333333333334 | 0.0073190095801664206 |
+| Epoch_11_batch_5999.pt | 0.8563333333333333 | 0.00699558943942097 |
+| Epoch_17_batch_2999.pt | 0.8561666666666667 | 0.007260794433288209 |
+| Epoch_15.pt | 0.8556666666666667 | 0.0073961952180633915 |
+| Epoch_13.pt | 0.8553333333333333 | 0.007144021069190452 |
+| Epoch_16_batch_2999.pt | 0.8548333333333333 | 0.007397238390268971 |
+| Epoch_17.pt | 0.8546666666666667 | 0.0072085340370260384 |
+| Epoch_12.pt | 0.8546666666666667 | 0.006274581290442431 |
+| Epoch_14.pt | 0.8539999999999999 | 0.007516237566794556 |
+| Epoch_12_batch_5999.pt | 0.8535 | 0.006380187311000883 |
+| Epoch_11.pt | 0.8531666666666666 | 0.006001285870441881 |
+| Epoch_11_batch_2999.pt | 0.8528333333333332 | 0.006601767440112783 |
+| Epoch_10_batch_5999.pt | 0.8521666666666666 | 0.00639757863145729 |
+| Epoch_10.pt | 0.8503333333333334 | 0.005804425727381765 |
+| Epoch_10_batch_2999.pt | 0.8491666666666667 | 0.006935549859373218 |
+| Epoch_9_batch_2999.pt | 0.825 | 0.007969850595746357 |
+| Epoch_8_batch_5999.pt | 0.8241666666666667 | 0.006291528696058958 |
+| Epoch_7_batch_2999.pt | 0.8233333333333335 | 0.007175059684872417 |
+| Epoch_9_batch_5999.pt | 0.8196666666666668 | 0.00763277533137582 |
+| Epoch_7_batch_5999.pt | 0.8183333333333334 | 0.006378010061584961 |
+| Epoch_8_batch_2999.pt | 0.8178333333333333 | 0.006657632768039557 |
+| Epoch_9.pt | 0.8139999999999998 | 0.00755718936583642 |
+| Epoch_7.pt | 0.8126666666666666 | 0.005779914134951089 |
+| Epoch_8.pt | 0.8126666666666666 | 0.005958705634574188 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..77ffe3e5d01805483ca0fa61027939e78ec1560f
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_lfw.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11.pt | 0.9974999999999999 | 0.000668977476599572 |
+| Epoch_17.pt | 0.9973333333333333 | 0.0007934920476158734 |
+| Epoch_16.pt | 0.9971666666666665 | 0.0007474235581707618 |
+| Epoch_12_batch_5999.pt | 0.9971666666666665 | 0.0006111111111111101 |
+| Epoch_16_batch_5999.pt | 0.9971666666666665 | 0.0006111111111111077 |
+| Epoch_15.pt | 0.9971666666666665 | 0.0007474235581707618 |
+| Epoch_13.pt | 0.9971666666666665 | 0.0006596856715021055 |
+| Epoch_14_batch_2999.pt | 0.9971666666666665 | 0.0005583264233956013 |
+| Epoch_17_batch_5999.pt | 0.9969999999999999 | 0.0006938886664887096 |
+| Epoch_15_batch_5999.pt | 0.9969999999999999 | 0.0006938886664887096 |
+| Epoch_12.pt | 0.9968333333333333 | 0.0005800170282728086 |
+| Epoch_13_batch_2999.pt | 0.9968333333333333 | 0.00063098981620003 |
+| Epoch_16_batch_2999.pt | 0.9968333333333332 | 0.0007637626158259738 |
+| Epoch_10_batch_5999.pt | 0.9966666666666667 | 0.0007453559924999305 |
+| Epoch_15_batch_2999.pt | 0.9966666666666667 | 0.0007453559924999305 |
+| Epoch_11_batch_5999.pt | 0.9966666666666667 | 0.0005555555555555536 |
+| Epoch_14.pt | 0.9964999999999999 | 0.0007222222222222236 |
+| Epoch_14_batch_5999.pt | 0.9964999999999999 | 0.0007222222222222234 |
+| Epoch_12_batch_2999.pt | 0.9964999999999999 | 0.0008407081083567494 |
+| Epoch_9_batch_2999.pt | 0.9963333333333335 | 0.0011600340565456157 |
+| Epoch_10.pt | 0.9963333333333333 | 0.0008534606386520708 |
+| Epoch_13_batch_5999.pt | 0.9961666666666666 | 0.000747423558170758 |
+| Epoch_17_batch_2999.pt | 0.9961666666666666 | 0.0008624541497922222 |
+| Epoch_10_batch_2999.pt | 0.9959999999999999 | 0.0008314794192830944 |
+| Epoch_11_batch_2999.pt | 0.9959999999999999 | 0.0007535922203472518 |
+| Epoch_8_batch_5999.pt | 0.9953333333333335 | 0.0010482201257840705 |
+| Epoch_8.pt | 0.9953333333333333 | 0.0010183501544346308 |
+| Epoch_8_batch_2999.pt | 0.9953333333333332 | 0.0008534606386520666 |
+| Epoch_7.pt | 0.9951666666666666 | 0.0009444444444444424 |
+| Epoch_7_batch_5999.pt | 0.9946666666666666 | 0.0011331154474650675 |
+| Epoch_9_batch_5999.pt | 0.9944999999999998 | 0.0011399046960379607 |
+| Epoch_9.pt | 0.9943333333333333 | 0.0010599324460188284 |
+| Epoch_7_batch_2999.pt | 0.9935 | 0.001203133768205986 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d946630259aa9fff6a9597cc738e604373d1ff20
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_African.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.9196666666666665 | 0.0048291808471334925 |
+| Epoch_16_batch_5999.pt | 0.9191666666666667 | 0.00503230305864761 |
+| Epoch_16_batch_2999.pt | 0.9189999999999999 | 0.004514722145313344 |
+| Epoch_14_batch_2999.pt | 0.9173333333333333 | 0.005242278272037973 |
+| Epoch_16.pt | 0.9173333333333332 | 0.004247003300803643 |
+| Epoch_13.pt | 0.9171666666666667 | 0.00447248101390639 |
+| Epoch_15.pt | 0.9165000000000001 | 0.005255508573903228 |
+| Epoch_12.pt | 0.9164999999999999 | 0.004509591971906817 |
+| Epoch_15_batch_5999.pt | 0.9163333333333334 | 0.005066228051190224 |
+| Epoch_17.pt | 0.9163333333333332 | 0.005078397725345078 |
+| Epoch_11_batch_5999.pt | 0.9161666666666667 | 0.004720914851703747 |
+| Epoch_13_batch_2999.pt | 0.9161666666666666 | 0.004437842318564781 |
+| Epoch_14_batch_5999.pt | 0.9158333333333335 | 0.004629814811111267 |
+| Epoch_15_batch_2999.pt | 0.9158333333333333 | 0.00447386098339739 |
+| Epoch_11_batch_2999.pt | 0.9158333333333333 | 0.003703935177951845 |
+| Epoch_17_batch_2999.pt | 0.9153333333333334 | 0.004712735568988558 |
+| Epoch_13_batch_5999.pt | 0.9146666666666668 | 0.005150811993999666 |
+| Epoch_14.pt | 0.9146666666666666 | 0.004491418429201723 |
+| Epoch_12_batch_2999.pt | 0.9145 | 0.004868961914356778 |
+| Epoch_10_batch_5999.pt | 0.9139999999999999 | 0.004368800722519449 |
+| Epoch_10_batch_2999.pt | 0.9120000000000001 | 0.004065816547451557 |
+| Epoch_10.pt | 0.9119999999999999 | 0.003208784239598587 |
+| Epoch_11.pt | 0.9113333333333333 | 0.00449141842920172 |
+| Epoch_12_batch_5999.pt | 0.9093333333333333 | 0.005277485372016858 |
+| Epoch_9_batch_5999.pt | 0.8931666666666667 | 0.004342227339926476 |
+| Epoch_7_batch_5999.pt | 0.8845000000000001 | 0.0044239109003596545 |
+| Epoch_8_batch_2999.pt | 0.883 | 0.0050356751971665165 |
+| Epoch_7_batch_2999.pt | 0.8781666666666667 | 0.005208314814781896 |
+| Epoch_9_batch_2999.pt | 0.8781666666666667 | 0.005190506528720616 |
+| Epoch_8_batch_5999.pt | 0.8774999999999998 | 0.006644639536340649 |
+| Epoch_9.pt | 0.8681666666666666 | 0.003215030288051179 |
+| Epoch_7.pt | 0.8665 | 0.004858808978685827 |
+| Epoch_8.pt | 0.8641666666666665 | 0.005512331854046502 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..85b2d9f860af6c4c9ce43ec28ed2b8a825da7f7e
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9161666666666667 | 0.003991116679035958 |
+| Epoch_14.pt | 0.915 | 0.003912625969257557 |
+| Epoch_16.pt | 0.9146666666666666 | 0.003911047979288456 |
+| Epoch_17_batch_5999.pt | 0.9141666666666668 | 0.003712258638286253 |
+| Epoch_15_batch_5999.pt | 0.9141666666666668 | 0.004076808846434962 |
+| Epoch_16_batch_2999.pt | 0.9138333333333334 | 0.003452052529534663 |
+| Epoch_13_batch_2999.pt | 0.9136666666666666 | 0.00368346321205573 |
+| Epoch_12.pt | 0.913 | 0.004822785425380163 |
+| Epoch_14_batch_2999.pt | 0.9128333333333334 | 0.0033979841518606713 |
+| Epoch_17_batch_2999.pt | 0.9128333333333334 | 0.0035490391674854334 |
+| Epoch_17.pt | 0.9126666666666667 | 0.003802208584948592 |
+| Epoch_13_batch_5999.pt | 0.9125 | 0.0035246048723470945 |
+| Epoch_13.pt | 0.9123333333333333 | 0.003842581436842349 |
+| Epoch_14_batch_5999.pt | 0.9119999999999999 | 0.003581502546952488 |
+| Epoch_15_batch_2999.pt | 0.9113333333333333 | 0.0036582394740342955 |
+| Epoch_15.pt | 0.9091666666666665 | 0.004362084109434135 |
+| Epoch_12_batch_5999.pt | 0.909 | 0.004368800722519447 |
+| Epoch_10_batch_2999.pt | 0.9088333333333333 | 0.0039051248379533246 |
+| Epoch_11_batch_5999.pt | 0.9066666666666666 | 0.0034336749200805485 |
+| Epoch_12_batch_2999.pt | 0.9044999999999999 | 0.003928763824052703 |
+| Epoch_10_batch_5999.pt | 0.9043333333333333 | 0.003744955454745048 |
+| Epoch_11.pt | 0.9038333333333333 | 0.0035140810224152147 |
+| Epoch_11_batch_2999.pt | 0.9031666666666667 | 0.003939746811442536 |
+| Epoch_10.pt | 0.9031666666666667 | 0.004509591971906822 |
+| Epoch_8_batch_5999.pt | 0.8818333333333334 | 0.004814138129086656 |
+| Epoch_9_batch_5999.pt | 0.8811666666666668 | 0.004707821178306751 |
+| Epoch_8_batch_2999.pt | 0.8775000000000001 | 0.004787135538781684 |
+| Epoch_9_batch_2999.pt | 0.8753333333333334 | 0.004841946348777985 |
+| Epoch_7_batch_2999.pt | 0.8734999999999999 | 0.004670963630191093 |
+| Epoch_9.pt | 0.873 | 0.00428750697368327 |
+| Epoch_7_batch_5999.pt | 0.8700000000000001 | 0.004274529791482523 |
+| Epoch_7.pt | 0.8634999999999999 | 0.003939746811442535 |
+| Epoch_8.pt | 0.8588333333333333 | 0.004849907724717204 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6a2e7da44e599822b650321a633915be2cb58907
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.9743333333333334 | 0.001845916413981794 |
+| Epoch_13_batch_5999.pt | 0.974 | 0.0018954135676924459 |
+| Epoch_14.pt | 0.974 | 0.0018121673811444593 |
+| Epoch_15_batch_5999.pt | 0.9734999999999999 | 0.0015605079894653485 |
+| Epoch_17_batch_2999.pt | 0.9734999999999999 | 0.0016933056282364591 |
+| Epoch_16_batch_2999.pt | 0.9733333333333333 | 0.001490711984999854 |
+| Epoch_17.pt | 0.9730000000000001 | 0.0017881641043812288 |
+| Epoch_14_batch_5999.pt | 0.9730000000000001 | 0.001905158688831364 |
+| Epoch_13.pt | 0.9728333333333335 | 0.001809610830544709 |
+| Epoch_15.pt | 0.9726666666666667 | 0.0017427096823731166 |
+| Epoch_13_batch_2999.pt | 0.9724999999999999 | 0.0019444444444444422 |
+| Epoch_12_batch_5999.pt | 0.9721666666666666 | 0.0016489802310728674 |
+| Epoch_16_batch_5999.pt | 0.9721666666666666 | 0.0015918387535438184 |
+| Epoch_15_batch_2999.pt | 0.9718333333333333 | 0.001540602735984666 |
+| Epoch_16.pt | 0.9716666666666667 | 0.001337954953199144 |
+| Epoch_17_batch_5999.pt | 0.9708333333333335 | 0.0016339379077614214 |
+| Epoch_12.pt | 0.97 | 0.0014487116456005894 |
+| Epoch_10_batch_2999.pt | 0.97 | 0.0016850834320114576 |
+| Epoch_11_batch_5999.pt | 0.9698333333333332 | 0.001963399671896923 |
+| Epoch_12_batch_2999.pt | 0.9696666666666667 | 0.0016442942874387541 |
+| Epoch_10_batch_5999.pt | 0.9690000000000001 | 0.00227980939207591 |
+| Epoch_11_batch_2999.pt | 0.9690000000000001 | 0.0016329931618554569 |
+| Epoch_10.pt | 0.9685 | 0.0018500917561496378 |
+| Epoch_11.pt | 0.9665000000000001 | 0.0015605079894653493 |
+| Epoch_9_batch_2999.pt | 0.9516666666666665 | 0.0028544961285922508 |
+| Epoch_9_batch_5999.pt | 0.9501666666666665 | 0.0031471131196152335 |
+| Epoch_8_batch_5999.pt | 0.9486666666666667 | 0.0028738927014172405 |
+| Epoch_8_batch_2999.pt | 0.9483333333333335 | 0.0025215123817578303 |
+| Epoch_7_batch_2999.pt | 0.9461666666666668 | 0.0032683480177961607 |
+| Epoch_9.pt | 0.9418333333333333 | 0.0016933056282364598 |
+| Epoch_8.pt | 0.9404999999999999 | 0.0029507270499754407 |
+| Epoch_7_batch_5999.pt | 0.9399999999999998 | 0.00272165526975909 |
+| Epoch_7.pt | 0.9375 | 0.003480545579483793 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ee07ebaa66d459e7738ac9196e714d2b65c4f7ea
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9333333333333333 | 0.0029917582261858294 |
+| Epoch_13.pt | 0.9323333333333332 | 0.0027239223715847236 |
+| Epoch_13_batch_2999.pt | 0.9318333333333333 | 0.0025754059969998358 |
+| Epoch_17_batch_2999.pt | 0.9315 | 0.002323391518480717 |
+| Epoch_17.pt | 0.9313333333333332 | 0.002506780927261885 |
+| Epoch_12_batch_5999.pt | 0.9311666666666667 | 0.0024851410273716693 |
+| Epoch_16_batch_5999.pt | 0.9311666666666667 | 0.0024975296436654405 |
+| Epoch_14_batch_2999.pt | 0.931 | 0.001943650631615096 |
+| Epoch_17_batch_5999.pt | 0.9308333333333334 | 0.0028246052930858117 |
+| Epoch_14_batch_5999.pt | 0.9305 | 0.0030025709148603684 |
+| Epoch_16_batch_2999.pt | 0.9303333333333335 | 0.0025067809272618802 |
+| Epoch_12.pt | 0.9301666666666668 | 0.003290934048804412 |
+| Epoch_15_batch_5999.pt | 0.93 | 0.002675909906398288 |
+| Epoch_16.pt | 0.9291666666666668 | 0.00209717623201965 |
+| Epoch_13_batch_5999.pt | 0.9291666666666668 | 0.0027357938338322487 |
+| Epoch_11_batch_2999.pt | 0.9291666666666666 | 0.003213109720204717 |
+| Epoch_15.pt | 0.9289999999999999 | 0.0022249982660556412 |
+| Epoch_14.pt | 0.9288333333333334 | 0.0026879934229784327 |
+| Epoch_10_batch_2999.pt | 0.9278333333333334 | 0.0031234872881934677 |
+| Epoch_11_batch_5999.pt | 0.9271666666666667 | 0.003626463093415017 |
+| Epoch_10_batch_5999.pt | 0.9266666666666667 | 0.0024969116726938022 |
+| Epoch_10.pt | 0.9263333333333333 | 0.003061106069767977 |
+| Epoch_12_batch_2999.pt | 0.9259999999999998 | 0.0029313124351717573 |
+| Epoch_11.pt | 0.9256666666666667 | 0.0032697642154582594 |
+| Epoch_9_batch_2999.pt | 0.9123333333333333 | 0.0040307460327149 |
+| Epoch_9_batch_5999.pt | 0.9118333333333334 | 0.004085883557377088 |
+| Epoch_8_batch_2999.pt | 0.9078333333333333 | 0.002485141027371676 |
+| Epoch_8_batch_5999.pt | 0.9055 | 0.002733536577809458 |
+| Epoch_7_batch_2999.pt | 0.9040000000000001 | 0.003390254733864971 |
+| Epoch_7_batch_5999.pt | 0.8996666666666666 | 0.0038232556742411597 |
+| Epoch_7.pt | 0.8969999999999999 | 0.003855411460643881 |
+| Epoch_9.pt | 0.8939999999999999 | 0.004361730316975672 |
+| Epoch_8.pt | 0.8924999999999998 | 0.003014880789158913 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/backbones/TF_NAS_A/log.log b/bob/bio/facexzoo/models/backbones/TF_NAS_A/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..ff57aa19ba751263adada903119a7d002f168182
--- /dev/null
+++ b/bob/bio/facexzoo/models/backbones/TF_NAS_A/log.log
@@ -0,0 +1,657 @@
+INFO 2020-12-16 10:57:33 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2020-12-16 10:57:33 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2020-12-16 10:57:33 train.py: 177] Start optimization.
+INFO 2020-12-16 10:57:33 train.py: 178] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='TF-NAS-A', batch_size=512, data_root='/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MV-Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10,13,16', tensorboardx_logdir='mv-tfnas', train_file='/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f3484206b00>)
+backbone param:
+{'feat_dim': 512, 'drop_ratio': 0.2, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-12-16 10:57:58 train.py: 79] Epoch 0, iter 0/6416, lr 0.100000, loss 16.294960
+INFO 2020-12-16 11:01:08 train.py: 79] Epoch 0, iter 200/6416, lr 0.100000, loss 15.628962
+INFO 2020-12-16 11:04:18 train.py: 79] Epoch 0, iter 400/6416, lr 0.100000, loss 15.347841
+INFO 2020-12-16 11:07:28 train.py: 79] Epoch 0, iter 600/6416, lr 0.100000, loss 15.309812
+INFO 2020-12-16 11:10:38 train.py: 79] Epoch 0, iter 800/6416, lr 0.100000, loss 15.309494
+INFO 2020-12-16 11:13:48 train.py: 79] Epoch 0, iter 1000/6416, lr 0.100000, loss 15.286847
+INFO 2020-12-16 11:16:58 train.py: 79] Epoch 0, iter 1200/6416, lr 0.100000, loss 15.272687
+INFO 2020-12-16 11:20:09 train.py: 79] Epoch 0, iter 1400/6416, lr 0.100000, loss 15.239681
+INFO 2020-12-16 11:23:19 train.py: 79] Epoch 0, iter 1600/6416, lr 0.100000, loss 15.144505
+INFO 2020-12-16 11:26:29 train.py: 79] Epoch 0, iter 1800/6416, lr 0.100000, loss 14.907998
+INFO 2020-12-16 11:29:39 train.py: 79] Epoch 0, iter 2000/6416, lr 0.100000, loss 14.635004
+INFO 2020-12-16 11:32:49 train.py: 79] Epoch 0, iter 2200/6416, lr 0.100000, loss 14.340544
+INFO 2020-12-16 11:35:59 train.py: 79] Epoch 0, iter 2400/6416, lr 0.100000, loss 14.012194
+INFO 2020-12-16 11:39:09 train.py: 79] Epoch 0, iter 2600/6416, lr 0.100000, loss 13.669523
+INFO 2020-12-16 11:42:19 train.py: 79] Epoch 0, iter 2800/6416, lr 0.100000, loss 13.328470
+INFO 2020-12-16 11:45:28 train.py: 92] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-12-16 11:45:29 train.py: 79] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.987305
+INFO 2020-12-16 11:48:38 train.py: 79] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.652747
+INFO 2020-12-16 11:51:48 train.py: 79] Epoch 0, iter 3400/6416, lr 0.100000, loss 12.352968
+INFO 2020-12-16 11:54:57 train.py: 79] Epoch 0, iter 3600/6416, lr 0.100000, loss 12.072103
+INFO 2020-12-16 11:58:06 train.py: 79] Epoch 0, iter 3800/6416, lr 0.100000, loss 11.911163
+INFO 2020-12-16 12:01:15 train.py: 79] Epoch 0, iter 4000/6416, lr 0.100000, loss 11.857078
+INFO 2020-12-16 12:04:23 train.py: 79] Epoch 0, iter 4200/6416, lr 0.100000, loss 11.922731
+INFO 2020-12-16 12:07:31 train.py: 79] Epoch 0, iter 4400/6416, lr 0.100000, loss 12.120648
+INFO 2020-12-16 12:10:39 train.py: 79] Epoch 0, iter 4600/6416, lr 0.100000, loss 12.409902
+INFO 2020-12-16 12:13:46 train.py: 79] Epoch 0, iter 4800/6416, lr 0.100000, loss 12.760713
+INFO 2020-12-16 12:16:53 train.py: 79] Epoch 0, iter 5000/6416, lr 0.100000, loss 13.104537
+INFO 2020-12-16 12:19:59 train.py: 79] Epoch 0, iter 5200/6416, lr 0.100000, loss 13.451702
+INFO 2020-12-16 12:23:05 train.py: 79] Epoch 0, iter 5400/6416, lr 0.100000, loss 13.739899
+INFO 2020-12-16 12:26:11 train.py: 79] Epoch 0, iter 5600/6416, lr 0.100000, loss 14.008614
+INFO 2020-12-16 12:29:16 train.py: 79] Epoch 0, iter 5800/6416, lr 0.100000, loss 14.199121
+INFO 2020-12-16 12:32:21 train.py: 92] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-12-16 12:32:22 train.py: 79] Epoch 0, iter 6000/6416, lr 0.100000, loss 14.361762
+INFO 2020-12-16 12:35:27 train.py: 79] Epoch 0, iter 6200/6416, lr 0.100000, loss 14.418223
+INFO 2020-12-16 12:38:31 train.py: 79] Epoch 0, iter 6400/6416, lr 0.100000, loss 14.439503
+INFO 2020-12-16 12:38:45 train.py: 97] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-12-16 12:38:47 train.py: 79] Epoch 1, iter 0/6416, lr 0.100000, loss 14.445541
+INFO 2020-12-16 12:41:52 train.py: 79] Epoch 1, iter 200/6416, lr 0.100000, loss 14.368868
+INFO 2020-12-16 12:44:57 train.py: 79] Epoch 1, iter 400/6416, lr 0.100000, loss 14.274350
+INFO 2020-12-16 12:48:02 train.py: 79] Epoch 1, iter 600/6416, lr 0.100000, loss 14.169469
+INFO 2020-12-16 12:51:06 train.py: 79] Epoch 1, iter 800/6416, lr 0.100000, loss 13.976274
+INFO 2020-12-16 12:54:10 train.py: 79] Epoch 1, iter 1000/6416, lr 0.100000, loss 13.831864
+INFO 2020-12-16 12:57:14 train.py: 79] Epoch 1, iter 1200/6416, lr 0.100000, loss 13.624695
+INFO 2020-12-16 13:00:18 train.py: 79] Epoch 1, iter 1400/6416, lr 0.100000, loss 13.438048
+INFO 2020-12-16 13:03:22 train.py: 79] Epoch 1, iter 1600/6416, lr 0.100000, loss 13.208371
+INFO 2020-12-16 13:06:26 train.py: 79] Epoch 1, iter 1800/6416, lr 0.100000, loss 13.018888
+INFO 2020-12-16 13:09:30 train.py: 79] Epoch 1, iter 2000/6416, lr 0.100000, loss 12.782446
+INFO 2020-12-16 13:12:34 train.py: 79] Epoch 1, iter 2200/6416, lr 0.100000, loss 12.595565
+INFO 2020-12-16 13:15:38 train.py: 79] Epoch 1, iter 2400/6416, lr 0.100000, loss 12.374886
+INFO 2020-12-16 13:18:43 train.py: 79] Epoch 1, iter 2600/6416, lr 0.100000, loss 12.191778
+INFO 2020-12-16 13:21:47 train.py: 79] Epoch 1, iter 2800/6416, lr 0.100000, loss 12.001840
+INFO 2020-12-16 13:24:50 train.py: 92] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-12-16 13:24:51 train.py: 79] Epoch 1, iter 3000/6416, lr 0.100000, loss 11.820816
+INFO 2020-12-16 13:27:55 train.py: 79] Epoch 1, iter 3200/6416, lr 0.100000, loss 11.634691
+INFO 2020-12-16 13:30:59 train.py: 79] Epoch 1, iter 3400/6416, lr 0.100000, loss 11.460390
+INFO 2020-12-16 13:34:03 train.py: 79] Epoch 1, iter 3600/6416, lr 0.100000, loss 11.296575
+INFO 2020-12-16 13:37:07 train.py: 79] Epoch 1, iter 3800/6416, lr 0.100000, loss 11.133909
+INFO 2020-12-16 13:40:11 train.py: 79] Epoch 1, iter 4000/6416, lr 0.100000, loss 10.998391
+INFO 2020-12-16 13:43:14 train.py: 79] Epoch 1, iter 4200/6416, lr 0.100000, loss 10.859886
+INFO 2020-12-16 13:46:18 train.py: 79] Epoch 1, iter 4400/6416, lr 0.100000, loss 10.732823
+INFO 2020-12-16 13:49:22 train.py: 79] Epoch 1, iter 4600/6416, lr 0.100000, loss 10.622386
+INFO 2020-12-16 13:52:25 train.py: 79] Epoch 1, iter 4800/6416, lr 0.100000, loss 10.526137
+INFO 2020-12-16 13:55:29 train.py: 79] Epoch 1, iter 5000/6416, lr 0.100000, loss 10.388890
+INFO 2020-12-16 13:58:33 train.py: 79] Epoch 1, iter 5200/6416, lr 0.100000, loss 10.319600
+INFO 2020-12-16 14:01:37 train.py: 79] Epoch 1, iter 5400/6416, lr 0.100000, loss 10.169680
+INFO 2020-12-16 14:04:40 train.py: 79] Epoch 1, iter 5600/6416, lr 0.100000, loss 10.059016
+INFO 2020-12-16 14:07:44 train.py: 79] Epoch 1, iter 5800/6416, lr 0.100000, loss 9.972363
+INFO 2020-12-16 14:10:47 train.py: 92] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-12-16 14:10:48 train.py: 79] Epoch 1, iter 6000/6416, lr 0.100000, loss 9.893893
+INFO 2020-12-16 14:13:52 train.py: 79] Epoch 1, iter 6200/6416, lr 0.100000, loss 9.828813
+INFO 2020-12-16 14:16:55 train.py: 79] Epoch 1, iter 6400/6416, lr 0.100000, loss 9.750626
+INFO 2020-12-16 14:17:09 train.py: 97] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-12-16 14:17:11 train.py: 79] Epoch 2, iter 0/6416, lr 0.100000, loss 9.628840
+INFO 2020-12-16 14:20:15 train.py: 79] Epoch 2, iter 200/6416, lr 0.100000, loss 9.140671
+INFO 2020-12-16 14:23:19 train.py: 79] Epoch 2, iter 400/6416, lr 0.100000, loss 9.117125
+INFO 2020-12-16 14:26:23 train.py: 79] Epoch 2, iter 600/6416, lr 0.100000, loss 9.107641
+INFO 2020-12-16 14:29:27 train.py: 79] Epoch 2, iter 800/6416, lr 0.100000, loss 9.125537
+INFO 2020-12-16 14:32:30 train.py: 79] Epoch 2, iter 1000/6416, lr 0.100000, loss 9.127542
+INFO 2020-12-16 14:35:34 train.py: 79] Epoch 2, iter 1200/6416, lr 0.100000, loss 9.128590
+INFO 2020-12-16 14:38:38 train.py: 79] Epoch 2, iter 1400/6416, lr 0.100000, loss 9.100560
+INFO 2020-12-16 14:41:41 train.py: 79] Epoch 2, iter 1600/6416, lr 0.100000, loss 9.047722
+INFO 2020-12-16 14:44:45 train.py: 79] Epoch 2, iter 1800/6416, lr 0.100000, loss 9.017468
+INFO 2020-12-16 14:47:48 train.py: 79] Epoch 2, iter 2000/6416, lr 0.100000, loss 8.979833
+INFO 2020-12-16 14:50:52 train.py: 79] Epoch 2, iter 2200/6416, lr 0.100000, loss 8.965562
+INFO 2020-12-16 14:53:56 train.py: 79] Epoch 2, iter 2400/6416, lr 0.100000, loss 8.879372
+INFO 2020-12-16 14:57:00 train.py: 79] Epoch 2, iter 2600/6416, lr 0.100000, loss 8.840598
+INFO 2020-12-16 15:00:03 train.py: 79] Epoch 2, iter 2800/6416, lr 0.100000, loss 8.810855
+INFO 2020-12-16 15:03:07 train.py: 92] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-12-16 15:03:08 train.py: 79] Epoch 2, iter 3000/6416, lr 0.100000, loss 8.737037
+INFO 2020-12-16 15:06:11 train.py: 79] Epoch 2, iter 3200/6416, lr 0.100000, loss 8.690072
+INFO 2020-12-16 15:09:15 train.py: 79] Epoch 2, iter 3400/6416, lr 0.100000, loss 8.631141
+INFO 2020-12-16 15:12:18 train.py: 79] Epoch 2, iter 3600/6416, lr 0.100000, loss 8.581275
+INFO 2020-12-16 15:15:22 train.py: 79] Epoch 2, iter 3800/6416, lr 0.100000, loss 8.567976
+INFO 2020-12-16 15:18:26 train.py: 79] Epoch 2, iter 4000/6416, lr 0.100000, loss 8.555046
+INFO 2020-12-16 15:21:29 train.py: 79] Epoch 2, iter 4200/6416, lr 0.100000, loss 8.484894
+INFO 2020-12-16 15:24:33 train.py: 79] Epoch 2, iter 4400/6416, lr 0.100000, loss 8.455214
+INFO 2020-12-16 15:27:37 train.py: 79] Epoch 2, iter 4600/6416, lr 0.100000, loss 8.422606
+INFO 2020-12-16 15:30:40 train.py: 79] Epoch 2, iter 4800/6416, lr 0.100000, loss 8.361818
+INFO 2020-12-16 15:33:44 train.py: 79] Epoch 2, iter 5000/6416, lr 0.100000, loss 8.353954
+INFO 2020-12-16 15:36:48 train.py: 79] Epoch 2, iter 5200/6416, lr 0.100000, loss 8.275327
+INFO 2020-12-16 15:39:51 train.py: 79] Epoch 2, iter 5400/6416, lr 0.100000, loss 8.234590
+INFO 2020-12-16 15:42:55 train.py: 79] Epoch 2, iter 5600/6416, lr 0.100000, loss 8.235253
+INFO 2020-12-16 15:45:58 train.py: 79] Epoch 2, iter 5800/6416, lr 0.100000, loss 8.185953
+INFO 2020-12-16 15:49:02 train.py: 92] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-12-16 15:49:02 train.py: 79] Epoch 2, iter 6000/6416, lr 0.100000, loss 8.157096
+INFO 2020-12-16 15:52:06 train.py: 79] Epoch 2, iter 6200/6416, lr 0.100000, loss 8.127464
+INFO 2020-12-16 15:55:10 train.py: 79] Epoch 2, iter 6400/6416, lr 0.100000, loss 8.068287
+INFO 2020-12-16 15:55:24 train.py: 97] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-12-16 15:55:26 train.py: 79] Epoch 3, iter 0/6416, lr 0.100000, loss 8.097464
+INFO 2020-12-16 15:58:30 train.py: 79] Epoch 3, iter 200/6416, lr 0.100000, loss 7.533775
+INFO 2020-12-16 16:01:33 train.py: 79] Epoch 3, iter 400/6416, lr 0.100000, loss 7.509258
+INFO 2020-12-16 16:04:37 train.py: 79] Epoch 3, iter 600/6416, lr 0.100000, loss 7.622495
+INFO 2020-12-16 16:07:41 train.py: 79] Epoch 3, iter 800/6416, lr 0.100000, loss 7.665167
+INFO 2020-12-16 16:10:44 train.py: 79] Epoch 3, iter 1000/6416, lr 0.100000, loss 7.698643
+INFO 2020-12-16 16:13:48 train.py: 79] Epoch 3, iter 1200/6416, lr 0.100000, loss 7.713717
+INFO 2020-12-16 16:16:51 train.py: 79] Epoch 3, iter 1400/6416, lr 0.100000, loss 7.722271
+INFO 2020-12-16 16:19:55 train.py: 79] Epoch 3, iter 1600/6416, lr 0.100000, loss 7.747544
+INFO 2020-12-16 16:22:58 train.py: 79] Epoch 3, iter 1800/6416, lr 0.100000, loss 7.709758
+INFO 2020-12-16 16:26:02 train.py: 79] Epoch 3, iter 2000/6416, lr 0.100000, loss 7.683169
+INFO 2020-12-16 16:29:05 train.py: 79] Epoch 3, iter 2200/6416, lr 0.100000, loss 7.743547
+INFO 2020-12-16 16:32:09 train.py: 79] Epoch 3, iter 2400/6416, lr 0.100000, loss 7.704084
+INFO 2020-12-16 16:35:12 train.py: 79] Epoch 3, iter 2600/6416, lr 0.100000, loss 7.653470
+INFO 2020-12-16 16:38:16 train.py: 79] Epoch 3, iter 2800/6416, lr 0.100000, loss 7.676525
+INFO 2020-12-16 16:41:20 train.py: 92] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-12-16 16:41:20 train.py: 79] Epoch 3, iter 3000/6416, lr 0.100000, loss 7.665897
+INFO 2020-12-16 16:44:24 train.py: 79] Epoch 3, iter 3200/6416, lr 0.100000, loss 7.595931
+INFO 2020-12-16 16:47:28 train.py: 79] Epoch 3, iter 3400/6416, lr 0.100000, loss 7.579141
+INFO 2020-12-16 16:50:31 train.py: 79] Epoch 3, iter 3600/6416, lr 0.100000, loss 7.594003
+INFO 2020-12-16 16:53:35 train.py: 79] Epoch 3, iter 3800/6416, lr 0.100000, loss 7.554896
+INFO 2020-12-16 16:56:38 train.py: 79] Epoch 3, iter 4000/6416, lr 0.100000, loss 7.532942
+INFO 2020-12-16 16:59:42 train.py: 79] Epoch 3, iter 4200/6416, lr 0.100000, loss 7.528585
+INFO 2020-12-16 17:02:45 train.py: 79] Epoch 3, iter 4400/6416, lr 0.100000, loss 7.520380
+INFO 2020-12-16 17:05:49 train.py: 79] Epoch 3, iter 4600/6416, lr 0.100000, loss 7.500985
+INFO 2020-12-16 17:08:53 train.py: 79] Epoch 3, iter 4800/6416, lr 0.100000, loss 7.484528
+INFO 2020-12-16 17:11:56 train.py: 79] Epoch 3, iter 5000/6416, lr 0.100000, loss 7.464162
+INFO 2020-12-16 17:15:00 train.py: 79] Epoch 3, iter 5200/6416, lr 0.100000, loss 7.440261
+INFO 2020-12-16 17:18:04 train.py: 79] Epoch 3, iter 5400/6416, lr 0.100000, loss 7.417208
+INFO 2020-12-16 17:21:07 train.py: 79] Epoch 3, iter 5600/6416, lr 0.100000, loss 7.410865
+INFO 2020-12-16 17:24:11 train.py: 79] Epoch 3, iter 5800/6416, lr 0.100000, loss 7.364992
+INFO 2020-12-16 17:27:14 train.py: 92] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-12-16 17:27:15 train.py: 79] Epoch 3, iter 6000/6416, lr 0.100000, loss 7.353658
+INFO 2020-12-16 17:30:19 train.py: 79] Epoch 3, iter 6200/6416, lr 0.100000, loss 7.362351
+INFO 2020-12-16 17:33:22 train.py: 79] Epoch 3, iter 6400/6416, lr 0.100000, loss 7.285520
+INFO 2020-12-16 17:33:36 train.py: 97] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-12-16 17:33:38 train.py: 79] Epoch 4, iter 0/6416, lr 0.100000, loss 7.220638
+INFO 2020-12-16 17:36:42 train.py: 79] Epoch 4, iter 200/6416, lr 0.100000, loss 6.813221
+INFO 2020-12-16 17:39:46 train.py: 79] Epoch 4, iter 400/6416, lr 0.100000, loss 6.796029
+INFO 2020-12-16 17:42:50 train.py: 79] Epoch 4, iter 600/6416, lr 0.100000, loss 6.908998
+INFO 2020-12-16 17:45:54 train.py: 79] Epoch 4, iter 800/6416, lr 0.100000, loss 6.986903
+INFO 2020-12-16 17:48:57 train.py: 79] Epoch 4, iter 1000/6416, lr 0.100000, loss 6.994841
+INFO 2020-12-16 17:52:01 train.py: 79] Epoch 4, iter 1200/6416, lr 0.100000, loss 7.036995
+INFO 2020-12-16 17:55:04 train.py: 79] Epoch 4, iter 1400/6416, lr 0.100000, loss 7.033270
+INFO 2020-12-16 17:58:08 train.py: 79] Epoch 4, iter 1600/6416, lr 0.100000, loss 7.053071
+INFO 2020-12-16 18:01:11 train.py: 79] Epoch 4, iter 1800/6416, lr 0.100000, loss 7.053654
+INFO 2020-12-16 18:04:15 train.py: 79] Epoch 4, iter 2000/6416, lr 0.100000, loss 7.041008
+INFO 2020-12-16 18:07:18 train.py: 79] Epoch 4, iter 2200/6416, lr 0.100000, loss 7.034623
+INFO 2020-12-16 18:10:22 train.py: 79] Epoch 4, iter 2400/6416, lr 0.100000, loss 7.047747
+INFO 2020-12-16 18:13:25 train.py: 79] Epoch 4, iter 2600/6416, lr 0.100000, loss 7.061938
+INFO 2020-12-16 18:16:29 train.py: 79] Epoch 4, iter 2800/6416, lr 0.100000, loss 7.041687
+INFO 2020-12-16 18:19:32 train.py: 92] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-12-16 18:19:33 train.py: 79] Epoch 4, iter 3000/6416, lr 0.100000, loss 7.010068
+INFO 2020-12-16 18:22:37 train.py: 79] Epoch 4, iter 3200/6416, lr 0.100000, loss 7.032898
+INFO 2020-12-16 18:25:41 train.py: 79] Epoch 4, iter 3400/6416, lr 0.100000, loss 6.985824
+INFO 2020-12-16 18:28:44 train.py: 79] Epoch 4, iter 3600/6416, lr 0.100000, loss 7.010521
+INFO 2020-12-16 18:31:48 train.py: 79] Epoch 4, iter 3800/6416, lr 0.100000, loss 6.984254
+INFO 2020-12-16 18:34:52 train.py: 79] Epoch 4, iter 4000/6416, lr 0.100000, loss 6.971621
+INFO 2020-12-16 18:37:55 train.py: 79] Epoch 4, iter 4200/6416, lr 0.100000, loss 6.998433
+INFO 2020-12-16 18:40:59 train.py: 79] Epoch 4, iter 4400/6416, lr 0.100000, loss 6.958774
+INFO 2020-12-16 18:44:03 train.py: 79] Epoch 4, iter 4600/6416, lr 0.100000, loss 6.944697
+INFO 2020-12-16 18:47:06 train.py: 79] Epoch 4, iter 4800/6416, lr 0.100000, loss 6.948095
+INFO 2020-12-16 18:50:10 train.py: 79] Epoch 4, iter 5000/6416, lr 0.100000, loss 6.935915
+INFO 2020-12-16 18:53:14 train.py: 79] Epoch 4, iter 5200/6416, lr 0.100000, loss 6.932230
+INFO 2020-12-16 18:56:17 train.py: 79] Epoch 4, iter 5400/6416, lr 0.100000, loss 6.942065
+INFO 2020-12-16 18:59:21 train.py: 79] Epoch 4, iter 5600/6416, lr 0.100000, loss 6.880897
+INFO 2020-12-16 19:02:25 train.py: 79] Epoch 4, iter 5800/6416, lr 0.100000, loss 6.858976
+INFO 2020-12-16 19:05:28 train.py: 92] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-12-16 19:05:29 train.py: 79] Epoch 4, iter 6000/6416, lr 0.100000, loss 6.882915
+INFO 2020-12-16 19:08:33 train.py: 79] Epoch 4, iter 6200/6416, lr 0.100000, loss 6.854021
+INFO 2020-12-16 19:11:36 train.py: 79] Epoch 4, iter 6400/6416, lr 0.100000, loss 6.868132
+INFO 2020-12-16 19:11:50 train.py: 97] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-12-16 19:11:53 train.py: 79] Epoch 5, iter 0/6416, lr 0.100000, loss 6.687169
+INFO 2020-12-16 19:14:56 train.py: 79] Epoch 5, iter 200/6416, lr 0.100000, loss 6.330479
+INFO 2020-12-16 19:18:00 train.py: 79] Epoch 5, iter 400/6416, lr 0.100000, loss 6.332625
+INFO 2020-12-16 19:21:04 train.py: 79] Epoch 5, iter 600/6416, lr 0.100000, loss 6.413189
+INFO 2020-12-16 19:24:07 train.py: 79] Epoch 5, iter 800/6416, lr 0.100000, loss 6.457377
+INFO 2020-12-16 19:27:11 train.py: 79] Epoch 5, iter 1000/6416, lr 0.100000, loss 6.523624
+INFO 2020-12-16 19:30:14 train.py: 79] Epoch 5, iter 1200/6416, lr 0.100000, loss 6.593338
+INFO 2020-12-16 19:33:18 train.py: 79] Epoch 5, iter 1400/6416, lr 0.100000, loss 6.596045
+INFO 2020-12-16 19:36:21 train.py: 79] Epoch 5, iter 1600/6416, lr 0.100000, loss 6.599684
+INFO 2020-12-16 19:39:25 train.py: 79] Epoch 5, iter 1800/6416, lr 0.100000, loss 6.650840
+INFO 2020-12-16 19:42:28 train.py: 79] Epoch 5, iter 2000/6416, lr 0.100000, loss 6.614480
+INFO 2020-12-16 19:45:32 train.py: 79] Epoch 5, iter 2200/6416, lr 0.100000, loss 6.666180
+INFO 2020-12-16 19:48:36 train.py: 79] Epoch 5, iter 2400/6416, lr 0.100000, loss 6.652113
+INFO 2020-12-16 19:51:39 train.py: 79] Epoch 5, iter 2600/6416, lr 0.100000, loss 6.670007
+INFO 2020-12-16 19:54:43 train.py: 79] Epoch 5, iter 2800/6416, lr 0.100000, loss 6.655893
+INFO 2020-12-16 19:57:46 train.py: 92] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-12-16 19:57:47 train.py: 79] Epoch 5, iter 3000/6416, lr 0.100000, loss 6.664842
+INFO 2020-12-16 20:00:51 train.py: 79] Epoch 5, iter 3200/6416, lr 0.100000, loss 6.657407
+INFO 2020-12-16 20:03:55 train.py: 79] Epoch 5, iter 3400/6416, lr 0.100000, loss 6.619209
+INFO 2020-12-16 20:06:59 train.py: 79] Epoch 5, iter 3600/6416, lr 0.100000, loss 6.687481
+INFO 2020-12-16 20:10:02 train.py: 79] Epoch 5, iter 3800/6416, lr 0.100000, loss 6.582684
+INFO 2020-12-16 20:13:06 train.py: 79] Epoch 5, iter 4000/6416, lr 0.100000, loss 6.588120
+INFO 2020-12-16 20:16:10 train.py: 79] Epoch 5, iter 4200/6416, lr 0.100000, loss 6.629578
+INFO 2020-12-16 20:19:14 train.py: 79] Epoch 5, iter 4400/6416, lr 0.100000, loss 6.611664
+INFO 2020-12-16 20:22:18 train.py: 79] Epoch 5, iter 4600/6416, lr 0.100000, loss 6.604345
+INFO 2020-12-16 20:25:21 train.py: 79] Epoch 5, iter 4800/6416, lr 0.100000, loss 6.581388
+INFO 2020-12-16 20:28:25 train.py: 79] Epoch 5, iter 5000/6416, lr 0.100000, loss 6.587704
+INFO 2020-12-16 20:31:29 train.py: 79] Epoch 5, iter 5200/6416, lr 0.100000, loss 6.566637
+INFO 2020-12-16 20:34:33 train.py: 79] Epoch 5, iter 5400/6416, lr 0.100000, loss 6.560541
+INFO 2020-12-16 20:37:37 train.py: 79] Epoch 5, iter 5600/6416, lr 0.100000, loss 6.570186
+INFO 2020-12-16 20:40:41 train.py: 79] Epoch 5, iter 5800/6416, lr 0.100000, loss 6.543714
+INFO 2020-12-16 20:43:44 train.py: 92] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-12-16 20:43:45 train.py: 79] Epoch 5, iter 6000/6416, lr 0.100000, loss 6.559604
+INFO 2020-12-16 20:46:49 train.py: 79] Epoch 5, iter 6200/6416, lr 0.100000, loss 6.551549
+INFO 2020-12-16 20:49:53 train.py: 79] Epoch 5, iter 6400/6416, lr 0.100000, loss 6.527905
+INFO 2020-12-16 20:50:06 train.py: 97] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-12-16 20:50:09 train.py: 79] Epoch 6, iter 0/6416, lr 0.100000, loss 6.550037
+INFO 2020-12-16 20:53:12 train.py: 79] Epoch 6, iter 200/6416, lr 0.100000, loss 5.998583
+INFO 2020-12-16 20:56:17 train.py: 79] Epoch 6, iter 400/6416, lr 0.100000, loss 6.006238
+INFO 2020-12-16 20:59:20 train.py: 79] Epoch 6, iter 600/6416, lr 0.100000, loss 6.111885
+INFO 2020-12-16 21:02:24 train.py: 79] Epoch 6, iter 800/6416, lr 0.100000, loss 6.202286
+INFO 2020-12-16 21:05:27 train.py: 79] Epoch 6, iter 1000/6416, lr 0.100000, loss 6.234919
+INFO 2020-12-16 21:08:31 train.py: 79] Epoch 6, iter 1200/6416, lr 0.100000, loss 6.274875
+INFO 2020-12-16 21:11:35 train.py: 79] Epoch 6, iter 1400/6416, lr 0.100000, loss 6.264605
+INFO 2020-12-16 21:14:38 train.py: 79] Epoch 6, iter 1600/6416, lr 0.100000, loss 6.301119
+INFO 2020-12-16 21:17:42 train.py: 79] Epoch 6, iter 1800/6416, lr 0.100000, loss 6.318995
+INFO 2020-12-16 21:20:46 train.py: 79] Epoch 6, iter 2000/6416, lr 0.100000, loss 6.324353
+INFO 2020-12-16 21:23:49 train.py: 79] Epoch 6, iter 2200/6416, lr 0.100000, loss 6.354859
+INFO 2020-12-16 21:26:53 train.py: 79] Epoch 6, iter 2400/6416, lr 0.100000, loss 6.373282
+INFO 2020-12-16 21:29:57 train.py: 79] Epoch 6, iter 2600/6416, lr 0.100000, loss 6.391004
+INFO 2020-12-16 21:33:00 train.py: 79] Epoch 6, iter 2800/6416, lr 0.100000, loss 6.355699
+INFO 2020-12-16 21:36:04 train.py: 92] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-12-16 21:36:05 train.py: 79] Epoch 6, iter 3000/6416, lr 0.100000, loss 6.374998
+INFO 2020-12-16 21:39:09 train.py: 79] Epoch 6, iter 3200/6416, lr 0.100000, loss 6.332537
+INFO 2020-12-16 21:42:13 train.py: 79] Epoch 6, iter 3400/6416, lr 0.100000, loss 6.375002
+INFO 2020-12-16 21:45:16 train.py: 79] Epoch 6, iter 3600/6416, lr 0.100000, loss 6.355496
+INFO 2020-12-16 21:48:20 train.py: 79] Epoch 6, iter 3800/6416, lr 0.100000, loss 6.362530
+INFO 2020-12-16 21:51:24 train.py: 79] Epoch 6, iter 4000/6416, lr 0.100000, loss 6.365120
+INFO 2020-12-16 21:54:28 train.py: 79] Epoch 6, iter 4200/6416, lr 0.100000, loss 6.342688
+INFO 2020-12-16 21:57:32 train.py: 79] Epoch 6, iter 4400/6416, lr 0.100000, loss 6.328660
+INFO 2020-12-16 22:00:36 train.py: 79] Epoch 6, iter 4600/6416, lr 0.100000, loss 6.347540
+INFO 2020-12-16 22:03:39 train.py: 79] Epoch 6, iter 4800/6416, lr 0.100000, loss 6.342310
+INFO 2020-12-16 22:06:43 train.py: 79] Epoch 6, iter 5000/6416, lr 0.100000, loss 6.342663
+INFO 2020-12-16 22:09:47 train.py: 79] Epoch 6, iter 5200/6416, lr 0.100000, loss 6.331906
+INFO 2020-12-16 22:12:51 train.py: 79] Epoch 6, iter 5400/6416, lr 0.100000, loss 6.304072
+INFO 2020-12-16 22:15:55 train.py: 79] Epoch 6, iter 5600/6416, lr 0.100000, loss 6.285095
+INFO 2020-12-16 22:18:59 train.py: 79] Epoch 6, iter 5800/6416, lr 0.100000, loss 6.282518
+INFO 2020-12-16 22:22:02 train.py: 92] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-12-16 22:22:03 train.py: 79] Epoch 6, iter 6000/6416, lr 0.100000, loss 6.259123
+INFO 2020-12-16 22:25:08 train.py: 79] Epoch 6, iter 6200/6416, lr 0.100000, loss 6.293265
+INFO 2020-12-16 22:28:11 train.py: 79] Epoch 6, iter 6400/6416, lr 0.100000, loss 6.283082
+INFO 2020-12-16 22:28:25 train.py: 97] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-12-16 22:28:27 train.py: 79] Epoch 7, iter 0/6416, lr 0.100000, loss 6.194263
+INFO 2020-12-16 22:31:31 train.py: 79] Epoch 7, iter 200/6416, lr 0.100000, loss 5.780380
+INFO 2020-12-16 22:34:35 train.py: 79] Epoch 7, iter 400/6416, lr 0.100000, loss 5.745367
+INFO 2020-12-16 22:37:39 train.py: 79] Epoch 7, iter 600/6416, lr 0.100000, loss 5.883271
+INFO 2020-12-16 22:40:43 train.py: 79] Epoch 7, iter 800/6416, lr 0.100000, loss 5.921427
+INFO 2020-12-16 22:43:47 train.py: 79] Epoch 7, iter 1000/6416, lr 0.100000, loss 5.963688
+INFO 2020-12-16 22:46:50 train.py: 79] Epoch 7, iter 1200/6416, lr 0.100000, loss 6.054396
+INFO 2020-12-16 22:49:54 train.py: 79] Epoch 7, iter 1400/6416, lr 0.100000, loss 6.047568
+INFO 2020-12-16 22:52:57 train.py: 79] Epoch 7, iter 1600/6416, lr 0.100000, loss 6.087585
+INFO 2020-12-16 22:56:01 train.py: 79] Epoch 7, iter 1800/6416, lr 0.100000, loss 6.113140
+INFO 2020-12-16 22:59:05 train.py: 79] Epoch 7, iter 2000/6416, lr 0.100000, loss 6.111894
+INFO 2020-12-16 23:02:09 train.py: 79] Epoch 7, iter 2200/6416, lr 0.100000, loss 6.125291
+INFO 2020-12-16 23:05:13 train.py: 79] Epoch 7, iter 2400/6416, lr 0.100000, loss 6.127479
+INFO 2020-12-16 23:08:17 train.py: 79] Epoch 7, iter 2600/6416, lr 0.100000, loss 6.141241
+INFO 2020-12-16 23:11:20 train.py: 79] Epoch 7, iter 2800/6416, lr 0.100000, loss 6.150876
+INFO 2020-12-16 23:14:24 train.py: 92] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-12-16 23:14:25 train.py: 79] Epoch 7, iter 3000/6416, lr 0.100000, loss 6.135567
+INFO 2020-12-16 23:17:29 train.py: 79] Epoch 7, iter 3200/6416, lr 0.100000, loss 6.122425
+INFO 2020-12-16 23:20:33 train.py: 79] Epoch 7, iter 3400/6416, lr 0.100000, loss 6.179848
+INFO 2020-12-16 23:23:37 train.py: 79] Epoch 7, iter 3600/6416, lr 0.100000, loss 6.112599
+INFO 2020-12-16 23:26:41 train.py: 79] Epoch 7, iter 3800/6416, lr 0.100000, loss 6.131287
+INFO 2020-12-16 23:29:45 train.py: 79] Epoch 7, iter 4000/6416, lr 0.100000, loss 6.169256
+INFO 2020-12-16 23:32:49 train.py: 79] Epoch 7, iter 4200/6416, lr 0.100000, loss 6.100719
+INFO 2020-12-16 23:35:53 train.py: 79] Epoch 7, iter 4400/6416, lr 0.100000, loss 6.111884
+INFO 2020-12-16 23:38:57 train.py: 79] Epoch 7, iter 4600/6416, lr 0.100000, loss 6.138353
+INFO 2020-12-16 23:42:01 train.py: 79] Epoch 7, iter 4800/6416, lr 0.100000, loss 6.122168
+INFO 2020-12-16 23:45:05 train.py: 79] Epoch 7, iter 5000/6416, lr 0.100000, loss 6.119304
+INFO 2020-12-16 23:48:09 train.py: 79] Epoch 7, iter 5200/6416, lr 0.100000, loss 6.130274
+INFO 2020-12-16 23:51:13 train.py: 79] Epoch 7, iter 5400/6416, lr 0.100000, loss 6.115806
+INFO 2020-12-16 23:54:17 train.py: 79] Epoch 7, iter 5600/6416, lr 0.100000, loss 6.110932
+INFO 2020-12-16 23:57:20 train.py: 79] Epoch 7, iter 5800/6416, lr 0.100000, loss 6.104609
+INFO 2020-12-17 00:00:24 train.py: 92] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-12-17 00:00:25 train.py: 79] Epoch 7, iter 6000/6416, lr 0.100000, loss 6.125439
+INFO 2020-12-17 00:03:29 train.py: 79] Epoch 7, iter 6200/6416, lr 0.100000, loss 6.079707
+INFO 2020-12-17 00:06:33 train.py: 79] Epoch 7, iter 6400/6416, lr 0.100000, loss 6.081316
+INFO 2020-12-17 00:06:47 train.py: 97] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-12-17 00:06:49 train.py: 79] Epoch 8, iter 0/6416, lr 0.100000, loss 6.092733
+INFO 2020-12-17 00:09:53 train.py: 79] Epoch 8, iter 200/6416, lr 0.100000, loss 5.553806
+INFO 2020-12-17 00:12:57 train.py: 79] Epoch 8, iter 400/6416, lr 0.100000, loss 5.557566
+INFO 2020-12-17 00:16:00 train.py: 79] Epoch 8, iter 600/6416, lr 0.100000, loss 5.667213
+INFO 2020-12-17 00:19:04 train.py: 79] Epoch 8, iter 800/6416, lr 0.100000, loss 5.740584
+INFO 2020-12-17 00:22:08 train.py: 79] Epoch 8, iter 1000/6416, lr 0.100000, loss 5.824890
+INFO 2020-12-17 00:25:12 train.py: 79] Epoch 8, iter 1200/6416, lr 0.100000, loss 5.858538
+INFO 2020-12-17 00:28:15 train.py: 79] Epoch 8, iter 1400/6416, lr 0.100000, loss 5.878859
+INFO 2020-12-17 00:31:19 train.py: 79] Epoch 8, iter 1600/6416, lr 0.100000, loss 5.925390
+INFO 2020-12-17 00:34:23 train.py: 79] Epoch 8, iter 1800/6416, lr 0.100000, loss 5.907194
+INFO 2020-12-17 00:37:27 train.py: 79] Epoch 8, iter 2000/6416, lr 0.100000, loss 5.959864
+INFO 2020-12-17 00:40:31 train.py: 79] Epoch 8, iter 2200/6416, lr 0.100000, loss 5.964619
+INFO 2020-12-17 00:43:34 train.py: 79] Epoch 8, iter 2400/6416, lr 0.100000, loss 5.939840
+INFO 2020-12-17 00:46:38 train.py: 79] Epoch 8, iter 2600/6416, lr 0.100000, loss 5.942517
+INFO 2020-12-17 00:49:42 train.py: 79] Epoch 8, iter 2800/6416, lr 0.100000, loss 6.001136
+INFO 2020-12-17 00:52:46 train.py: 92] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-12-17 00:52:47 train.py: 79] Epoch 8, iter 3000/6416, lr 0.100000, loss 5.995145
+INFO 2020-12-17 00:55:51 train.py: 79] Epoch 8, iter 3200/6416, lr 0.100000, loss 5.972148
+INFO 2020-12-17 00:58:55 train.py: 79] Epoch 8, iter 3400/6416, lr 0.100000, loss 5.972411
+INFO 2020-12-17 01:01:59 train.py: 79] Epoch 8, iter 3600/6416, lr 0.100000, loss 5.974967
+INFO 2020-12-17 01:05:03 train.py: 79] Epoch 8, iter 3800/6416, lr 0.100000, loss 5.949742
+INFO 2020-12-17 01:08:07 train.py: 79] Epoch 8, iter 4000/6416, lr 0.100000, loss 5.986706
+INFO 2020-12-17 01:11:11 train.py: 79] Epoch 8, iter 4200/6416, lr 0.100000, loss 5.954821
+INFO 2020-12-17 01:14:15 train.py: 79] Epoch 8, iter 4400/6416, lr 0.100000, loss 5.985813
+INFO 2020-12-17 01:17:19 train.py: 79] Epoch 8, iter 4600/6416, lr 0.100000, loss 5.952517
+INFO 2020-12-17 01:20:23 train.py: 79] Epoch 8, iter 4800/6416, lr 0.100000, loss 5.955298
+INFO 2020-12-17 01:23:27 train.py: 79] Epoch 8, iter 5000/6416, lr 0.100000, loss 5.966541
+INFO 2020-12-17 01:26:31 train.py: 79] Epoch 8, iter 5200/6416, lr 0.100000, loss 5.967509
+INFO 2020-12-17 01:29:35 train.py: 79] Epoch 8, iter 5400/6416, lr 0.100000, loss 5.932452
+INFO 2020-12-17 01:32:39 train.py: 79] Epoch 8, iter 5600/6416, lr 0.100000, loss 5.928681
+INFO 2020-12-17 01:35:43 train.py: 79] Epoch 8, iter 5800/6416, lr 0.100000, loss 5.972631
+INFO 2020-12-17 01:38:47 train.py: 92] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-12-17 01:38:48 train.py: 79] Epoch 8, iter 6000/6416, lr 0.100000, loss 5.937896
+INFO 2020-12-17 01:41:51 train.py: 79] Epoch 8, iter 6200/6416, lr 0.100000, loss 5.930704
+INFO 2020-12-17 01:44:55 train.py: 79] Epoch 8, iter 6400/6416, lr 0.100000, loss 5.927029
+INFO 2020-12-17 01:45:09 train.py: 97] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-12-17 01:45:11 train.py: 79] Epoch 9, iter 0/6416, lr 0.100000, loss 5.869226
+INFO 2020-12-17 01:48:15 train.py: 79] Epoch 9, iter 200/6416, lr 0.100000, loss 5.419071
+INFO 2020-12-17 01:51:19 train.py: 79] Epoch 9, iter 400/6416, lr 0.100000, loss 5.440582
+INFO 2020-12-17 01:54:23 train.py: 79] Epoch 9, iter 600/6416, lr 0.100000, loss 5.507678
+INFO 2020-12-17 01:57:27 train.py: 79] Epoch 9, iter 800/6416, lr 0.100000, loss 5.640536
+INFO 2020-12-17 02:00:30 train.py: 79] Epoch 9, iter 1000/6416, lr 0.100000, loss 5.660732
+INFO 2020-12-17 02:03:34 train.py: 79] Epoch 9, iter 1200/6416, lr 0.100000, loss 5.705915
+INFO 2020-12-17 02:06:38 train.py: 79] Epoch 9, iter 1400/6416, lr 0.100000, loss 5.732998
+INFO 2020-12-17 02:09:41 train.py: 79] Epoch 9, iter 1600/6416, lr 0.100000, loss 5.780678
+INFO 2020-12-17 02:12:45 train.py: 79] Epoch 9, iter 1800/6416, lr 0.100000, loss 5.761457
+INFO 2020-12-17 02:15:49 train.py: 79] Epoch 9, iter 2000/6416, lr 0.100000, loss 5.801741
+INFO 2020-12-17 02:18:53 train.py: 79] Epoch 9, iter 2200/6416, lr 0.100000, loss 5.787668
+INFO 2020-12-17 02:21:57 train.py: 79] Epoch 9, iter 2400/6416, lr 0.100000, loss 5.785251
+INFO 2020-12-17 02:25:01 train.py: 79] Epoch 9, iter 2600/6416, lr 0.100000, loss 5.852773
+INFO 2020-12-17 02:28:05 train.py: 79] Epoch 9, iter 2800/6416, lr 0.100000, loss 5.823755
+INFO 2020-12-17 02:31:08 train.py: 92] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-12-17 02:31:09 train.py: 79] Epoch 9, iter 3000/6416, lr 0.100000, loss 5.808953
+INFO 2020-12-17 02:34:13 train.py: 79] Epoch 9, iter 3200/6416, lr 0.100000, loss 5.833811
+INFO 2020-12-17 02:37:17 train.py: 79] Epoch 9, iter 3400/6416, lr 0.100000, loss 5.830485
+INFO 2020-12-17 02:40:21 train.py: 79] Epoch 9, iter 3600/6416, lr 0.100000, loss 5.844789
+INFO 2020-12-17 02:43:25 train.py: 79] Epoch 9, iter 3800/6416, lr 0.100000, loss 5.834277
+INFO 2020-12-17 02:46:29 train.py: 79] Epoch 9, iter 4000/6416, lr 0.100000, loss 5.850949
+INFO 2020-12-17 02:49:33 train.py: 79] Epoch 9, iter 4200/6416, lr 0.100000, loss 5.809742
+INFO 2020-12-17 02:52:37 train.py: 79] Epoch 9, iter 4400/6416, lr 0.100000, loss 5.834941
+INFO 2020-12-17 02:55:41 train.py: 79] Epoch 9, iter 4600/6416, lr 0.100000, loss 5.822046
+INFO 2020-12-17 02:58:45 train.py: 79] Epoch 9, iter 4800/6416, lr 0.100000, loss 5.821672
+INFO 2020-12-17 03:01:49 train.py: 79] Epoch 9, iter 5000/6416, lr 0.100000, loss 5.831677
+INFO 2020-12-17 03:04:53 train.py: 79] Epoch 9, iter 5200/6416, lr 0.100000, loss 5.825064
+INFO 2020-12-17 03:07:57 train.py: 79] Epoch 9, iter 5400/6416, lr 0.100000, loss 5.805704
+INFO 2020-12-17 03:11:01 train.py: 79] Epoch 9, iter 5600/6416, lr 0.100000, loss 5.826465
+INFO 2020-12-17 03:14:05 train.py: 79] Epoch 9, iter 5800/6416, lr 0.100000, loss 5.820823
+INFO 2020-12-17 03:17:09 train.py: 92] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-12-17 03:17:10 train.py: 79] Epoch 9, iter 6000/6416, lr 0.100000, loss 5.841555
+INFO 2020-12-17 03:20:14 train.py: 79] Epoch 9, iter 6200/6416, lr 0.100000, loss 5.823700
+INFO 2020-12-17 03:23:18 train.py: 79] Epoch 9, iter 6400/6416, lr 0.100000, loss 5.830068
+INFO 2020-12-17 03:23:32 train.py: 97] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-12-17 03:23:34 train.py: 79] Epoch 10, iter 0/6416, lr 0.010000, loss 5.826842
+INFO 2020-12-17 03:26:38 train.py: 79] Epoch 10, iter 200/6416, lr 0.010000, loss 4.548205
+INFO 2020-12-17 03:29:42 train.py: 79] Epoch 10, iter 400/6416, lr 0.010000, loss 4.211603
+INFO 2020-12-17 03:32:45 train.py: 79] Epoch 10, iter 600/6416, lr 0.010000, loss 4.126667
+INFO 2020-12-17 03:35:49 train.py: 79] Epoch 10, iter 800/6416, lr 0.010000, loss 4.055592
+INFO 2020-12-17 03:38:53 train.py: 79] Epoch 10, iter 1000/6416, lr 0.010000, loss 4.005355
+INFO 2020-12-17 03:41:57 train.py: 79] Epoch 10, iter 1200/6416, lr 0.010000, loss 3.973728
+INFO 2020-12-17 03:45:00 train.py: 79] Epoch 10, iter 1400/6416, lr 0.010000, loss 3.918247
+INFO 2020-12-17 03:48:04 train.py: 79] Epoch 10, iter 1600/6416, lr 0.010000, loss 3.893333
+INFO 2020-12-17 03:51:08 train.py: 79] Epoch 10, iter 1800/6416, lr 0.010000, loss 3.884688
+INFO 2020-12-17 03:54:12 train.py: 79] Epoch 10, iter 2000/6416, lr 0.010000, loss 3.809224
+INFO 2020-12-17 03:57:16 train.py: 79] Epoch 10, iter 2200/6416, lr 0.010000, loss 3.818752
+INFO 2020-12-17 04:00:20 train.py: 79] Epoch 10, iter 2400/6416, lr 0.010000, loss 3.792871
+INFO 2020-12-17 04:03:24 train.py: 79] Epoch 10, iter 2600/6416, lr 0.010000, loss 3.750689
+INFO 2020-12-17 04:06:28 train.py: 79] Epoch 10, iter 2800/6416, lr 0.010000, loss 3.748104
+INFO 2020-12-17 04:09:31 train.py: 92] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-12-17 04:09:32 train.py: 79] Epoch 10, iter 3000/6416, lr 0.010000, loss 3.714453
+INFO 2020-12-17 04:12:36 train.py: 79] Epoch 10, iter 3200/6416, lr 0.010000, loss 3.671531
+INFO 2020-12-17 04:15:40 train.py: 79] Epoch 10, iter 3400/6416, lr 0.010000, loss 3.664450
+INFO 2020-12-17 04:18:44 train.py: 79] Epoch 10, iter 3600/6416, lr 0.010000, loss 3.644666
+INFO 2020-12-17 04:21:47 train.py: 79] Epoch 10, iter 3800/6416, lr 0.010000, loss 3.634291
+INFO 2020-12-17 04:24:52 train.py: 79] Epoch 10, iter 4000/6416, lr 0.010000, loss 3.615768
+INFO 2020-12-17 04:27:55 train.py: 79] Epoch 10, iter 4200/6416, lr 0.010000, loss 3.607924
+INFO 2020-12-17 04:30:59 train.py: 79] Epoch 10, iter 4400/6416, lr 0.010000, loss 3.626286
+INFO 2020-12-17 04:34:03 train.py: 79] Epoch 10, iter 4600/6416, lr 0.010000, loss 3.587321
+INFO 2020-12-17 04:37:07 train.py: 79] Epoch 10, iter 4800/6416, lr 0.010000, loss 3.568344
+INFO 2020-12-17 04:40:11 train.py: 79] Epoch 10, iter 5000/6416, lr 0.010000, loss 3.561552
+INFO 2020-12-17 04:43:15 train.py: 79] Epoch 10, iter 5200/6416, lr 0.010000, loss 3.562436
+INFO 2020-12-17 04:46:19 train.py: 79] Epoch 10, iter 5400/6416, lr 0.010000, loss 3.546892
+INFO 2020-12-17 04:49:23 train.py: 79] Epoch 10, iter 5600/6416, lr 0.010000, loss 3.522081
+INFO 2020-12-17 04:52:27 train.py: 79] Epoch 10, iter 5800/6416, lr 0.010000, loss 3.495743
+INFO 2020-12-17 04:55:30 train.py: 92] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-12-17 04:55:31 train.py: 79] Epoch 10, iter 6000/6416, lr 0.010000, loss 3.492026
+INFO 2020-12-17 04:58:35 train.py: 79] Epoch 10, iter 6200/6416, lr 0.010000, loss 3.484539
+INFO 2020-12-17 05:01:39 train.py: 79] Epoch 10, iter 6400/6416, lr 0.010000, loss 3.469855
+INFO 2020-12-17 05:01:53 train.py: 97] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-12-17 05:01:55 train.py: 79] Epoch 11, iter 0/6416, lr 0.010000, loss 3.514514
+INFO 2020-12-17 05:04:59 train.py: 79] Epoch 11, iter 200/6416, lr 0.010000, loss 3.160030
+INFO 2020-12-17 05:08:03 train.py: 79] Epoch 11, iter 400/6416, lr 0.010000, loss 3.146543
+INFO 2020-12-17 05:11:07 train.py: 79] Epoch 11, iter 600/6416, lr 0.010000, loss 3.166033
+INFO 2020-12-17 05:14:10 train.py: 79] Epoch 11, iter 800/6416, lr 0.010000, loss 3.168596
+INFO 2020-12-17 05:17:14 train.py: 79] Epoch 11, iter 1000/6416, lr 0.010000, loss 3.177543
+INFO 2020-12-17 05:20:17 train.py: 79] Epoch 11, iter 1200/6416, lr 0.010000, loss 3.128748
+INFO 2020-12-17 05:23:21 train.py: 79] Epoch 11, iter 1400/6416, lr 0.010000, loss 3.164774
+INFO 2020-12-17 05:26:25 train.py: 79] Epoch 11, iter 1600/6416, lr 0.010000, loss 3.163165
+INFO 2020-12-17 05:29:29 train.py: 79] Epoch 11, iter 1800/6416, lr 0.010000, loss 3.175456
+INFO 2020-12-17 05:32:32 train.py: 79] Epoch 11, iter 2000/6416, lr 0.010000, loss 3.198288
+INFO 2020-12-17 05:35:36 train.py: 79] Epoch 11, iter 2200/6416, lr 0.010000, loss 3.167570
+INFO 2020-12-17 05:38:40 train.py: 79] Epoch 11, iter 2400/6416, lr 0.010000, loss 3.155572
+INFO 2020-12-17 05:41:43 train.py: 79] Epoch 11, iter 2600/6416, lr 0.010000, loss 3.179223
+INFO 2020-12-17 05:44:47 train.py: 79] Epoch 11, iter 2800/6416, lr 0.010000, loss 3.169990
+INFO 2020-12-17 05:47:51 train.py: 92] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-12-17 05:47:52 train.py: 79] Epoch 11, iter 3000/6416, lr 0.010000, loss 3.185505
+INFO 2020-12-17 05:50:55 train.py: 79] Epoch 11, iter 3200/6416, lr 0.010000, loss 3.146870
+INFO 2020-12-17 05:53:59 train.py: 79] Epoch 11, iter 3400/6416, lr 0.010000, loss 3.173579
+INFO 2020-12-17 05:57:03 train.py: 79] Epoch 11, iter 3600/6416, lr 0.010000, loss 3.168758
+INFO 2020-12-17 06:00:07 train.py: 79] Epoch 11, iter 3800/6416, lr 0.010000, loss 3.160528
+INFO 2020-12-17 06:03:10 train.py: 79] Epoch 11, iter 4000/6416, lr 0.010000, loss 3.180473
+INFO 2020-12-17 06:06:14 train.py: 79] Epoch 11, iter 4200/6416, lr 0.010000, loss 3.171116
+INFO 2020-12-17 06:09:18 train.py: 79] Epoch 11, iter 4400/6416, lr 0.010000, loss 3.154475
+INFO 2020-12-17 06:12:22 train.py: 79] Epoch 11, iter 4600/6416, lr 0.010000, loss 3.176243
+INFO 2020-12-17 06:15:26 train.py: 79] Epoch 11, iter 4800/6416, lr 0.010000, loss 3.171883
+INFO 2020-12-17 06:18:30 train.py: 79] Epoch 11, iter 5000/6416, lr 0.010000, loss 3.195679
+INFO 2020-12-17 06:21:34 train.py: 79] Epoch 11, iter 5200/6416, lr 0.010000, loss 3.175597
+INFO 2020-12-17 06:24:38 train.py: 79] Epoch 11, iter 5400/6416, lr 0.010000, loss 3.163461
+INFO 2020-12-17 06:27:42 train.py: 79] Epoch 11, iter 5600/6416, lr 0.010000, loss 3.163557
+INFO 2020-12-17 06:30:46 train.py: 79] Epoch 11, iter 5800/6416, lr 0.010000, loss 3.162324
+INFO 2020-12-17 06:33:50 train.py: 92] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-12-17 06:33:50 train.py: 79] Epoch 11, iter 6000/6416, lr 0.010000, loss 3.158371
+INFO 2020-12-17 06:36:54 train.py: 79] Epoch 11, iter 6200/6416, lr 0.010000, loss 3.154266
+INFO 2020-12-17 06:39:58 train.py: 79] Epoch 11, iter 6400/6416, lr 0.010000, loss 3.171731
+INFO 2020-12-17 06:40:12 train.py: 97] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-12-17 06:40:14 train.py: 79] Epoch 12, iter 0/6416, lr 0.010000, loss 3.129189
+INFO 2020-12-17 06:43:18 train.py: 79] Epoch 12, iter 200/6416, lr 0.010000, loss 2.860576
+INFO 2020-12-17 06:46:22 train.py: 79] Epoch 12, iter 400/6416, lr 0.010000, loss 2.850822
+INFO 2020-12-17 06:49:26 train.py: 79] Epoch 12, iter 600/6416, lr 0.010000, loss 2.881905
+INFO 2020-12-17 06:52:29 train.py: 79] Epoch 12, iter 800/6416, lr 0.010000, loss 2.882144
+INFO 2020-12-17 06:55:33 train.py: 79] Epoch 12, iter 1000/6416, lr 0.010000, loss 2.895373
+INFO 2020-12-17 06:58:36 train.py: 79] Epoch 12, iter 1200/6416, lr 0.010000, loss 2.908096
+INFO 2020-12-17 07:01:40 train.py: 79] Epoch 12, iter 1400/6416, lr 0.010000, loss 2.925051
+INFO 2020-12-17 07:04:43 train.py: 79] Epoch 12, iter 1600/6416, lr 0.010000, loss 2.935689
+INFO 2020-12-17 07:07:47 train.py: 79] Epoch 12, iter 1800/6416, lr 0.010000, loss 2.928721
+INFO 2020-12-17 07:10:51 train.py: 79] Epoch 12, iter 2000/6416, lr 0.010000, loss 2.924306
+INFO 2020-12-17 07:13:54 train.py: 79] Epoch 12, iter 2200/6416, lr 0.010000, loss 2.971523
+INFO 2020-12-17 07:16:58 train.py: 79] Epoch 12, iter 2400/6416, lr 0.010000, loss 2.974434
+INFO 2020-12-17 07:20:02 train.py: 79] Epoch 12, iter 2600/6416, lr 0.010000, loss 2.981440
+INFO 2020-12-17 07:23:06 train.py: 79] Epoch 12, iter 2800/6416, lr 0.010000, loss 2.982280
+INFO 2020-12-17 07:26:09 train.py: 92] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-12-17 07:26:10 train.py: 79] Epoch 12, iter 3000/6416, lr 0.010000, loss 2.991981
+INFO 2020-12-17 07:29:14 train.py: 79] Epoch 12, iter 3200/6416, lr 0.010000, loss 2.962047
+INFO 2020-12-17 07:32:18 train.py: 79] Epoch 12, iter 3400/6416, lr 0.010000, loss 2.991310
+INFO 2020-12-17 07:35:22 train.py: 79] Epoch 12, iter 3600/6416, lr 0.010000, loss 2.976563
+INFO 2020-12-17 07:38:25 train.py: 79] Epoch 12, iter 3800/6416, lr 0.010000, loss 3.016750
+INFO 2020-12-17 07:41:29 train.py: 79] Epoch 12, iter 4000/6416, lr 0.010000, loss 3.025505
+INFO 2020-12-17 07:44:33 train.py: 79] Epoch 12, iter 4200/6416, lr 0.010000, loss 3.000748
+INFO 2020-12-17 07:47:37 train.py: 79] Epoch 12, iter 4400/6416, lr 0.010000, loss 3.024861
+INFO 2020-12-17 07:50:41 train.py: 79] Epoch 12, iter 4600/6416, lr 0.010000, loss 3.037012
+INFO 2020-12-17 07:53:45 train.py: 79] Epoch 12, iter 4800/6416, lr 0.010000, loss 3.035897
+INFO 2020-12-17 07:56:49 train.py: 79] Epoch 12, iter 5000/6416, lr 0.010000, loss 3.052873
+INFO 2020-12-17 07:59:52 train.py: 79] Epoch 12, iter 5200/6416, lr 0.010000, loss 3.025014
+INFO 2020-12-17 08:02:56 train.py: 79] Epoch 12, iter 5400/6416, lr 0.010000, loss 3.047568
+INFO 2020-12-17 08:06:00 train.py: 79] Epoch 12, iter 5600/6416, lr 0.010000, loss 3.047303
+INFO 2020-12-17 08:09:04 train.py: 79] Epoch 12, iter 5800/6416, lr 0.010000, loss 3.069041
+INFO 2020-12-17 08:12:08 train.py: 92] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-12-17 08:12:09 train.py: 79] Epoch 12, iter 6000/6416, lr 0.010000, loss 3.057256
+INFO 2020-12-17 08:15:13 train.py: 79] Epoch 12, iter 6200/6416, lr 0.010000, loss 3.045071
+INFO 2020-12-17 08:18:16 train.py: 79] Epoch 12, iter 6400/6416, lr 0.010000, loss 3.071860
+INFO 2020-12-17 08:18:30 train.py: 97] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-12-17 08:18:33 train.py: 79] Epoch 13, iter 0/6416, lr 0.001000, loss 3.004036
+INFO 2020-12-17 08:21:36 train.py: 79] Epoch 13, iter 200/6416, lr 0.001000, loss 2.667893
+INFO 2020-12-17 08:24:40 train.py: 79] Epoch 13, iter 400/6416, lr 0.001000, loss 2.608168
+INFO 2020-12-17 08:27:44 train.py: 79] Epoch 13, iter 600/6416, lr 0.001000, loss 2.594166
+INFO 2020-12-17 08:30:48 train.py: 79] Epoch 13, iter 800/6416, lr 0.001000, loss 2.590228
+INFO 2020-12-17 08:33:51 train.py: 79] Epoch 13, iter 1000/6416, lr 0.001000, loss 2.588210
+INFO 2020-12-17 08:36:55 train.py: 79] Epoch 13, iter 1200/6416, lr 0.001000, loss 2.595686
+INFO 2020-12-17 08:39:58 train.py: 79] Epoch 13, iter 1400/6416, lr 0.001000, loss 2.577873
+INFO 2020-12-17 08:43:02 train.py: 79] Epoch 13, iter 1600/6416, lr 0.001000, loss 2.592464
+INFO 2020-12-17 08:46:06 train.py: 79] Epoch 13, iter 1800/6416, lr 0.001000, loss 2.583104
+INFO 2020-12-17 08:49:09 train.py: 79] Epoch 13, iter 2000/6416, lr 0.001000, loss 2.597395
+INFO 2020-12-17 08:52:13 train.py: 79] Epoch 13, iter 2200/6416, lr 0.001000, loss 2.590685
+INFO 2020-12-17 08:55:16 train.py: 79] Epoch 13, iter 2400/6416, lr 0.001000, loss 2.586631
+INFO 2020-12-17 08:58:20 train.py: 79] Epoch 13, iter 2600/6416, lr 0.001000, loss 2.571441
+INFO 2020-12-17 09:01:24 train.py: 79] Epoch 13, iter 2800/6416, lr 0.001000, loss 2.577548
+INFO 2020-12-17 09:04:27 train.py: 92] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-12-17 09:04:28 train.py: 79] Epoch 13, iter 3000/6416, lr 0.001000, loss 2.585470
+INFO 2020-12-17 09:07:32 train.py: 79] Epoch 13, iter 3200/6416, lr 0.001000, loss 2.580309
+INFO 2020-12-17 09:10:36 train.py: 79] Epoch 13, iter 3400/6416, lr 0.001000, loss 2.586393
+INFO 2020-12-17 09:13:40 train.py: 79] Epoch 13, iter 3600/6416, lr 0.001000, loss 2.589329
+INFO 2020-12-17 09:16:43 train.py: 79] Epoch 13, iter 3800/6416, lr 0.001000, loss 2.582097
+INFO 2020-12-17 09:19:47 train.py: 79] Epoch 13, iter 4000/6416, lr 0.001000, loss 2.582904
+INFO 2020-12-17 09:22:51 train.py: 79] Epoch 13, iter 4200/6416, lr 0.001000, loss 2.578104
+INFO 2020-12-17 09:25:55 train.py: 79] Epoch 13, iter 4400/6416, lr 0.001000, loss 2.596582
+INFO 2020-12-17 09:28:58 train.py: 79] Epoch 13, iter 4600/6416, lr 0.001000, loss 2.568332
+INFO 2020-12-17 09:32:02 train.py: 79] Epoch 13, iter 4800/6416, lr 0.001000, loss 2.572608
+INFO 2020-12-17 09:35:06 train.py: 79] Epoch 13, iter 5000/6416, lr 0.001000, loss 2.590468
+INFO 2020-12-17 09:38:10 train.py: 79] Epoch 13, iter 5200/6416, lr 0.001000, loss 2.592332
+INFO 2020-12-17 09:41:14 train.py: 79] Epoch 13, iter 5400/6416, lr 0.001000, loss 2.587129
+INFO 2020-12-17 09:44:18 train.py: 79] Epoch 13, iter 5600/6416, lr 0.001000, loss 2.601162
+INFO 2020-12-17 09:47:22 train.py: 79] Epoch 13, iter 5800/6416, lr 0.001000, loss 2.573307
+INFO 2020-12-17 09:50:26 train.py: 92] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-12-17 09:50:26 train.py: 79] Epoch 13, iter 6000/6416, lr 0.001000, loss 2.592083
+INFO 2020-12-17 09:53:30 train.py: 79] Epoch 13, iter 6200/6416, lr 0.001000, loss 2.587068
+INFO 2020-12-17 09:56:34 train.py: 79] Epoch 13, iter 6400/6416, lr 0.001000, loss 2.572470
+INFO 2020-12-17 09:56:48 train.py: 97] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-12-17 09:56:50 train.py: 79] Epoch 14, iter 0/6416, lr 0.001000, loss 2.583252
+INFO 2020-12-17 09:59:54 train.py: 79] Epoch 14, iter 200/6416, lr 0.001000, loss 2.510700
+INFO 2020-12-17 10:02:58 train.py: 79] Epoch 14, iter 400/6416, lr 0.001000, loss 2.536795
+INFO 2020-12-17 10:06:02 train.py: 79] Epoch 14, iter 600/6416, lr 0.001000, loss 2.539800
+INFO 2020-12-17 10:09:06 train.py: 79] Epoch 14, iter 800/6416, lr 0.001000, loss 2.521127
+INFO 2020-12-17 10:12:09 train.py: 79] Epoch 14, iter 1000/6416, lr 0.001000, loss 2.519733
+INFO 2020-12-17 10:15:13 train.py: 79] Epoch 14, iter 1200/6416, lr 0.001000, loss 2.540676
+INFO 2020-12-17 10:18:16 train.py: 79] Epoch 14, iter 1400/6416, lr 0.001000, loss 2.520579
+INFO 2020-12-17 10:21:20 train.py: 79] Epoch 14, iter 1600/6416, lr 0.001000, loss 2.537068
+INFO 2020-12-17 10:24:24 train.py: 79] Epoch 14, iter 1800/6416, lr 0.001000, loss 2.546813
+INFO 2020-12-17 10:27:27 train.py: 79] Epoch 14, iter 2000/6416, lr 0.001000, loss 2.556447
+INFO 2020-12-17 10:30:31 train.py: 79] Epoch 14, iter 2200/6416, lr 0.001000, loss 2.553732
+INFO 2020-12-17 10:33:35 train.py: 79] Epoch 14, iter 2400/6416, lr 0.001000, loss 2.555915
+INFO 2020-12-17 10:36:39 train.py: 79] Epoch 14, iter 2600/6416, lr 0.001000, loss 2.548276
+INFO 2020-12-17 10:39:43 train.py: 79] Epoch 14, iter 2800/6416, lr 0.001000, loss 2.545036
+INFO 2020-12-17 10:42:46 train.py: 92] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-12-17 10:42:47 train.py: 79] Epoch 14, iter 3000/6416, lr 0.001000, loss 2.560245
+INFO 2020-12-17 10:45:51 train.py: 79] Epoch 14, iter 3200/6416, lr 0.001000, loss 2.551805
+INFO 2020-12-17 10:48:55 train.py: 79] Epoch 14, iter 3400/6416, lr 0.001000, loss 2.553714
+INFO 2020-12-17 10:51:58 train.py: 79] Epoch 14, iter 3600/6416, lr 0.001000, loss 2.535960
+INFO 2020-12-17 10:55:02 train.py: 79] Epoch 14, iter 3800/6416, lr 0.001000, loss 2.540353
+INFO 2020-12-17 10:58:06 train.py: 79] Epoch 14, iter 4000/6416, lr 0.001000, loss 2.554041
+INFO 2020-12-17 11:01:10 train.py: 79] Epoch 14, iter 4200/6416, lr 0.001000, loss 2.564377
+INFO 2020-12-17 11:04:14 train.py: 79] Epoch 14, iter 4400/6416, lr 0.001000, loss 2.545202
+INFO 2020-12-17 11:07:18 train.py: 79] Epoch 14, iter 4600/6416, lr 0.001000, loss 2.564473
+INFO 2020-12-17 11:10:21 train.py: 79] Epoch 14, iter 4800/6416, lr 0.001000, loss 2.559647
+INFO 2020-12-17 11:13:25 train.py: 79] Epoch 14, iter 5000/6416, lr 0.001000, loss 2.546554
+INFO 2020-12-17 11:16:29 train.py: 79] Epoch 14, iter 5200/6416, lr 0.001000, loss 2.550266
+INFO 2020-12-17 11:19:33 train.py: 79] Epoch 14, iter 5400/6416, lr 0.001000, loss 2.553600
+INFO 2020-12-17 11:22:37 train.py: 79] Epoch 14, iter 5600/6416, lr 0.001000, loss 2.565816
+INFO 2020-12-17 11:25:41 train.py: 79] Epoch 14, iter 5800/6416, lr 0.001000, loss 2.548201
+INFO 2020-12-17 11:28:44 train.py: 92] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-12-17 11:28:45 train.py: 79] Epoch 14, iter 6000/6416, lr 0.001000, loss 2.562819
+INFO 2020-12-17 11:31:49 train.py: 79] Epoch 14, iter 6200/6416, lr 0.001000, loss 2.565471
+INFO 2020-12-17 11:34:53 train.py: 79] Epoch 14, iter 6400/6416, lr 0.001000, loss 2.579433
+INFO 2020-12-17 11:35:07 train.py: 97] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-12-17 11:35:09 train.py: 79] Epoch 15, iter 0/6416, lr 0.001000, loss 2.552524
+INFO 2020-12-17 11:38:13 train.py: 79] Epoch 15, iter 200/6416, lr 0.001000, loss 2.501346
+INFO 2020-12-17 11:41:17 train.py: 79] Epoch 15, iter 400/6416, lr 0.001000, loss 2.505867
+INFO 2020-12-17 11:44:20 train.py: 79] Epoch 15, iter 600/6416, lr 0.001000, loss 2.507808
+INFO 2020-12-17 11:47:24 train.py: 79] Epoch 15, iter 800/6416, lr 0.001000, loss 2.506381
+INFO 2020-12-17 11:50:28 train.py: 79] Epoch 15, iter 1000/6416, lr 0.001000, loss 2.520504
+INFO 2020-12-17 11:53:31 train.py: 79] Epoch 15, iter 1200/6416, lr 0.001000, loss 2.532223
+INFO 2020-12-17 11:56:35 train.py: 79] Epoch 15, iter 1400/6416, lr 0.001000, loss 2.500966
+INFO 2020-12-17 11:59:39 train.py: 79] Epoch 15, iter 1600/6416, lr 0.001000, loss 2.513043
+INFO 2020-12-17 12:02:42 train.py: 79] Epoch 15, iter 1800/6416, lr 0.001000, loss 2.507298
+INFO 2020-12-17 12:05:46 train.py: 79] Epoch 15, iter 2000/6416, lr 0.001000, loss 2.517797
+INFO 2020-12-17 12:08:49 train.py: 79] Epoch 15, iter 2200/6416, lr 0.001000, loss 2.539955
+INFO 2020-12-17 12:11:53 train.py: 79] Epoch 15, iter 2400/6416, lr 0.001000, loss 2.514410
+INFO 2020-12-17 12:14:57 train.py: 79] Epoch 15, iter 2600/6416, lr 0.001000, loss 2.520550
+INFO 2020-12-17 12:18:01 train.py: 79] Epoch 15, iter 2800/6416, lr 0.001000, loss 2.533944
+INFO 2020-12-17 12:21:04 train.py: 92] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-12-17 12:21:05 train.py: 79] Epoch 15, iter 3000/6416, lr 0.001000, loss 2.502836
+INFO 2020-12-17 12:24:09 train.py: 79] Epoch 15, iter 3200/6416, lr 0.001000, loss 2.525814
+INFO 2020-12-17 12:27:13 train.py: 79] Epoch 15, iter 3400/6416, lr 0.001000, loss 2.538511
+INFO 2020-12-17 12:30:16 train.py: 79] Epoch 15, iter 3600/6416, lr 0.001000, loss 2.548587
+INFO 2020-12-17 12:33:20 train.py: 79] Epoch 15, iter 3800/6416, lr 0.001000, loss 2.522662
+INFO 2020-12-17 12:36:24 train.py: 79] Epoch 15, iter 4000/6416, lr 0.001000, loss 2.538134
+INFO 2020-12-17 12:39:28 train.py: 79] Epoch 15, iter 4200/6416, lr 0.001000, loss 2.531150
+INFO 2020-12-17 12:42:32 train.py: 79] Epoch 15, iter 4400/6416, lr 0.001000, loss 2.524360
+INFO 2020-12-17 12:45:36 train.py: 79] Epoch 15, iter 4600/6416, lr 0.001000, loss 2.549406
+INFO 2020-12-17 12:48:40 train.py: 79] Epoch 15, iter 4800/6416, lr 0.001000, loss 2.528144
+INFO 2020-12-17 12:51:44 train.py: 79] Epoch 15, iter 5000/6416, lr 0.001000, loss 2.529887
+INFO 2020-12-17 12:54:48 train.py: 79] Epoch 15, iter 5200/6416, lr 0.001000, loss 2.521092
+INFO 2020-12-17 12:57:51 train.py: 79] Epoch 15, iter 5400/6416, lr 0.001000, loss 2.524156
+INFO 2020-12-17 13:00:55 train.py: 79] Epoch 15, iter 5600/6416, lr 0.001000, loss 2.532197
+INFO 2020-12-17 13:03:59 train.py: 79] Epoch 15, iter 5800/6416, lr 0.001000, loss 2.527621
+INFO 2020-12-17 13:07:03 train.py: 92] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-12-17 13:07:04 train.py: 79] Epoch 15, iter 6000/6416, lr 0.001000, loss 2.554056
+INFO 2020-12-17 13:10:08 train.py: 79] Epoch 15, iter 6200/6416, lr 0.001000, loss 2.539484
+INFO 2020-12-17 13:13:11 train.py: 79] Epoch 15, iter 6400/6416, lr 0.001000, loss 2.531550
+INFO 2020-12-17 13:13:25 train.py: 97] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-12-17 13:13:27 train.py: 79] Epoch 16, iter 0/6416, lr 0.000100, loss 2.556221
+INFO 2020-12-17 13:16:31 train.py: 79] Epoch 16, iter 200/6416, lr 0.000100, loss 2.476342
+INFO 2020-12-17 13:19:35 train.py: 79] Epoch 16, iter 400/6416, lr 0.000100, loss 2.477670
+INFO 2020-12-17 13:22:39 train.py: 79] Epoch 16, iter 600/6416, lr 0.000100, loss 2.482082
+INFO 2020-12-17 13:25:42 train.py: 79] Epoch 16, iter 800/6416, lr 0.000100, loss 2.466461
+INFO 2020-12-17 13:28:46 train.py: 79] Epoch 16, iter 1000/6416, lr 0.000100, loss 2.457624
+INFO 2020-12-17 13:31:49 train.py: 79] Epoch 16, iter 1200/6416, lr 0.000100, loss 2.492245
+INFO 2020-12-17 13:34:53 train.py: 79] Epoch 16, iter 1400/6416, lr 0.000100, loss 2.485995
+INFO 2020-12-17 13:37:57 train.py: 79] Epoch 16, iter 1600/6416, lr 0.000100, loss 2.482639
+INFO 2020-12-17 13:41:01 train.py: 79] Epoch 16, iter 1800/6416, lr 0.000100, loss 2.478015
+INFO 2020-12-17 13:44:04 train.py: 79] Epoch 16, iter 2000/6416, lr 0.000100, loss 2.480291
+INFO 2020-12-17 13:47:08 train.py: 79] Epoch 16, iter 2200/6416, lr 0.000100, loss 2.486277
+INFO 2020-12-17 13:50:12 train.py: 79] Epoch 16, iter 2400/6416, lr 0.000100, loss 2.472441
+INFO 2020-12-17 13:53:15 train.py: 79] Epoch 16, iter 2600/6416, lr 0.000100, loss 2.476300
+INFO 2020-12-17 13:56:19 train.py: 79] Epoch 16, iter 2800/6416, lr 0.000100, loss 2.463268
+INFO 2020-12-17 13:59:22 train.py: 92] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-12-17 13:59:23 train.py: 79] Epoch 16, iter 3000/6416, lr 0.000100, loss 2.488549
+INFO 2020-12-17 14:02:27 train.py: 79] Epoch 16, iter 3200/6416, lr 0.000100, loss 2.453056
+INFO 2020-12-17 14:05:31 train.py: 79] Epoch 16, iter 3400/6416, lr 0.000100, loss 2.476695
+INFO 2020-12-17 14:08:35 train.py: 79] Epoch 16, iter 3600/6416, lr 0.000100, loss 2.465886
+INFO 2020-12-17 14:11:39 train.py: 79] Epoch 16, iter 3800/6416, lr 0.000100, loss 2.471524
+INFO 2020-12-17 14:14:43 train.py: 79] Epoch 16, iter 4000/6416, lr 0.000100, loss 2.471558
+INFO 2020-12-17 14:17:46 train.py: 79] Epoch 16, iter 4200/6416, lr 0.000100, loss 2.480813
+INFO 2020-12-17 14:20:50 train.py: 79] Epoch 16, iter 4400/6416, lr 0.000100, loss 2.481037
+INFO 2020-12-17 14:23:54 train.py: 79] Epoch 16, iter 4600/6416, lr 0.000100, loss 2.467424
+INFO 2020-12-17 14:26:58 train.py: 79] Epoch 16, iter 4800/6416, lr 0.000100, loss 2.475681
+INFO 2020-12-17 14:30:02 train.py: 79] Epoch 16, iter 5000/6416, lr 0.000100, loss 2.470315
+INFO 2020-12-17 14:33:06 train.py: 79] Epoch 16, iter 5200/6416, lr 0.000100, loss 2.481494
+INFO 2020-12-17 14:36:09 train.py: 79] Epoch 16, iter 5400/6416, lr 0.000100, loss 2.473234
+INFO 2020-12-17 14:39:13 train.py: 79] Epoch 16, iter 5600/6416, lr 0.000100, loss 2.492355
+INFO 2020-12-17 14:42:17 train.py: 79] Epoch 16, iter 5800/6416, lr 0.000100, loss 2.479858
+INFO 2020-12-17 14:45:21 train.py: 92] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-12-17 14:45:22 train.py: 79] Epoch 16, iter 6000/6416, lr 0.000100, loss 2.476148
+INFO 2020-12-17 14:48:25 train.py: 79] Epoch 16, iter 6200/6416, lr 0.000100, loss 2.478741
+INFO 2020-12-17 14:51:29 train.py: 79] Epoch 16, iter 6400/6416, lr 0.000100, loss 2.479260
+INFO 2020-12-17 14:51:43 train.py: 97] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-12-17 14:51:45 train.py: 79] Epoch 17, iter 0/6416, lr 0.000100, loss 2.472111
+INFO 2020-12-17 14:54:49 train.py: 79] Epoch 17, iter 200/6416, lr 0.000100, loss 2.469867
+INFO 2020-12-17 14:57:53 train.py: 79] Epoch 17, iter 400/6416, lr 0.000100, loss 2.455912
+INFO 2020-12-17 15:00:57 train.py: 79] Epoch 17, iter 600/6416, lr 0.000100, loss 2.465573
+INFO 2020-12-17 15:04:00 train.py: 79] Epoch 17, iter 800/6416, lr 0.000100, loss 2.467211
+INFO 2020-12-17 15:07:04 train.py: 79] Epoch 17, iter 1000/6416, lr 0.000100, loss 2.473948
+INFO 2020-12-17 15:10:08 train.py: 79] Epoch 17, iter 1200/6416, lr 0.000100, loss 2.475249
+INFO 2020-12-17 15:13:11 train.py: 79] Epoch 17, iter 1400/6416, lr 0.000100, loss 2.472363
+INFO 2020-12-17 15:16:15 train.py: 79] Epoch 17, iter 1600/6416, lr 0.000100, loss 2.462654
+INFO 2020-12-17 15:19:19 train.py: 79] Epoch 17, iter 1800/6416, lr 0.000100, loss 2.477459
+INFO 2020-12-17 15:22:22 train.py: 79] Epoch 17, iter 2000/6416, lr 0.000100, loss 2.470119
+INFO 2020-12-17 15:25:26 train.py: 79] Epoch 17, iter 2200/6416, lr 0.000100, loss 2.500324
+INFO 2020-12-17 15:28:30 train.py: 79] Epoch 17, iter 2400/6416, lr 0.000100, loss 2.482650
+INFO 2020-12-17 15:31:34 train.py: 79] Epoch 17, iter 2600/6416, lr 0.000100, loss 2.462124
+INFO 2020-12-17 15:34:37 train.py: 79] Epoch 17, iter 2800/6416, lr 0.000100, loss 2.453094
+INFO 2020-12-17 15:37:41 train.py: 92] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-12-17 15:37:42 train.py: 79] Epoch 17, iter 3000/6416, lr 0.000100, loss 2.467537
+INFO 2020-12-17 15:40:46 train.py: 79] Epoch 17, iter 3200/6416, lr 0.000100, loss 2.460593
+INFO 2020-12-17 15:43:49 train.py: 79] Epoch 17, iter 3400/6416, lr 0.000100, loss 2.492480
+INFO 2020-12-17 15:46:53 train.py: 79] Epoch 17, iter 3600/6416, lr 0.000100, loss 2.493263
+INFO 2020-12-17 15:49:57 train.py: 79] Epoch 17, iter 3800/6416, lr 0.000100, loss 2.477044
+INFO 2020-12-17 15:53:01 train.py: 79] Epoch 17, iter 4000/6416, lr 0.000100, loss 2.453467
+INFO 2020-12-17 15:56:05 train.py: 79] Epoch 17, iter 4200/6416, lr 0.000100, loss 2.478751
+INFO 2020-12-17 15:59:09 train.py: 79] Epoch 17, iter 4400/6416, lr 0.000100, loss 2.472310
+INFO 2020-12-17 16:02:12 train.py: 79] Epoch 17, iter 4600/6416, lr 0.000100, loss 2.478181
+INFO 2020-12-17 16:05:16 train.py: 79] Epoch 17, iter 4800/6416, lr 0.000100, loss 2.477833
+INFO 2020-12-17 16:08:20 train.py: 79] Epoch 17, iter 5000/6416, lr 0.000100, loss 2.467330
+INFO 2020-12-17 16:11:24 train.py: 79] Epoch 17, iter 5200/6416, lr 0.000100, loss 2.464586
+INFO 2020-12-17 16:14:28 train.py: 79] Epoch 17, iter 5400/6416, lr 0.000100, loss 2.457134
+INFO 2020-12-17 16:17:32 train.py: 79] Epoch 17, iter 5600/6416, lr 0.000100, loss 2.487210
+INFO 2020-12-17 16:20:35 train.py: 79] Epoch 17, iter 5800/6416, lr 0.000100, loss 2.486697
+INFO 2020-12-17 16:23:39 train.py: 92] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-12-17 16:23:40 train.py: 79] Epoch 17, iter 6000/6416, lr 0.000100, loss 2.474871
+INFO 2020-12-17 16:26:44 train.py: 79] Epoch 17, iter 6200/6416, lr 0.000100, loss 2.468082
+INFO 2020-12-17 16:29:48 train.py: 79] Epoch 17, iter 6400/6416, lr 0.000100, loss 2.470948
+INFO 2020-12-17 16:30:01 train.py: 97] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-12-17 16:30:02 train.py: 180] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/.gitkeep b/bob/bio/facexzoo/models/heads/AM-Softmax/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..21ed5a5c79f2593b9d13259d8c4ef10a18b0e06d
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_agedb.txt
@@ -0,0 +1,38 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_2999.pt | 0.9584999999999999 | 0.002465189748037095 |
+| Epoch_13.pt | 0.9583333333333334 | 0.0026988795114424712 |
+| Epoch_16_batch_2999.pt | 0.9581666666666667 | 0.00211476292340825 |
+| Epoch_16_batch_5999.pt | 0.9580000000000002 | 0.002431785403137706 |
+| Epoch_15.pt | 0.9571666666666667 | 0.002472690342696438 |
+| Epoch_17.pt | 0.9571666666666667 | 0.0024094720491334904 |
+| Epoch_10_batch_5999.pt | 0.9570000000000001 | 0.002406267536411978 |
+| Epoch_16.pt | 0.9570000000000001 | 0.0024695678634325466 |
+| Epoch_15_batch_5999.pt | 0.9566666666666667 | 0.0026988795114424678 |
+| Epoch_11.pt | 0.9564999999999999 | 0.0023366378716459025 |
+| Epoch_13_batch_5999.pt | 0.9564999999999999 | 0.00238888888888889 |
+| Epoch_10.pt | 0.9561666666666667 | 0.002733536577809457 |
+| Epoch_11_batch_2999.pt | 0.9561666666666667 | 0.0024222477062879593 |
+| Epoch_12_batch_5999.pt | 0.9561666666666667 | 0.0017042068500197742 |
+| Epoch_17_batch_5999.pt | 0.9559999999999998 | 0.0027239223715847267 |
+| Epoch_12_batch_2999.pt | 0.9558333333333333 | 0.003115572198402213 |
+| Epoch_14.pt | 0.9558333333333332 | 0.003065639924738091 |
+| Epoch_11_batch_5999.pt | 0.9556666666666667 | 0.0028995529668222023 |
+| Epoch_14_batch_5999.pt | 0.9555 | 0.002811462428640004 |
+| Epoch_17_batch_2999.pt | 0.9555 | 0.0025706078447242844 |
+| Epoch_12.pt | 0.9550000000000001 | 0.003531166351571275 |
+| Epoch_15_batch_2999.pt | 0.9549999999999998 | 0.0026527414191807454 |
+| Epoch_10_batch_2999.pt | 0.9546666666666667 | 0.0031991511219751083 |
+| Epoch_13_batch_2999.pt | 0.9546666666666667 | 0.0025795970591653158 |
+| Epoch_9_batch_5999.pt | 0.9458333333333334 | 0.0031647167485835785 |
+| Epoch_9.pt | 0.9446666666666668 | 0.003855411460643884 |
+| Epoch_8_batch_5999.pt | 0.9406666666666667 | 0.00412908981690268 |
+| Epoch_6_batch_2999.pt | 0.9404999999999999 | 0.004875296762097512 |
+| Epoch_9_batch_2999.pt | 0.9401666666666666 | 0.0028158501994387073 |
+| Epoch_7_batch_5999.pt | 0.9400000000000001 | 0.004310481053615731 |
+| Epoch_8_batch_2999.pt | 0.9393333333333332 | 0.003506607519568777 |
+| Epoch_5_batch_2999.pt | 0.9388333333333334 | 0.004135438531988093 |
+| Epoch_6_batch_5999.pt | 0.9375 | 0.004643128468533759 |
+| Epoch_7_batch_2999.pt | 0.9368333333333334 | 0.003908284963769236 |
+| Epoch_7.pt | 0.9343333333333333 | 0.004247003300803639 |
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..76dca7cb9efd3f50bb18e7ecbd75ddcd0b6ee265
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_calfw.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9393333333333332 | 0.0037118429085533458 |
+| Epoch_15_batch_5999.pt | 0.9393333333333332 | 0.0036106837353937593 |
+| Epoch_14_batch_5999.pt | 0.9386666666666666 | 0.0038554114606438807 |
+| Epoch_15.pt | 0.9386666666666666 | 0.0033591592128513212 |
+| Epoch_17_batch_2999.pt | 0.9383333333333332 | 0.00405973909032114 |
+| Epoch_13.pt | 0.9378333333333334 | 0.00396784918570424 |
+| Epoch_16.pt | 0.9378333333333332 | 0.003668770440244238 |
+| Epoch_15_batch_2999.pt | 0.9376666666666666 | 0.003914203323068559 |
+| Epoch_16_batch_2999.pt | 0.9373333333333334 | 0.0040000000000000036 |
+| Epoch_11_batch_2999.pt | 0.9371666666666666 | 0.004339383241150777 |
+| Epoch_14_batch_2999.pt | 0.9371666666666666 | 0.003685557397915994 |
+| Epoch_17_batch_5999.pt | 0.9371666666666666 | 0.0036349639562123543 |
+| Epoch_14.pt | 0.9369999999999999 | 0.0037745083892140084 |
+| Epoch_16_batch_5999.pt | 0.9366666666666665 | 0.0036260375271290504 |
+| Epoch_17.pt | 0.9366666666666665 | 0.0039047296424012416 |
+| Epoch_12.pt | 0.9365 | 0.004123479892003003 |
+| Epoch_12_batch_2999.pt | 0.9363333333333334 | 0.004148478822798771 |
+| Epoch_10_batch_5999.pt | 0.9353333333333333 | 0.0038473977095957253 |
+| Epoch_11.pt | 0.9353333333333333 | 0.004586614986788948 |
+| Epoch_13_batch_5999.pt | 0.9353333333333331 | 0.0037498971179302674 |
+| Epoch_11_batch_5999.pt | 0.9351666666666667 | 0.004078322700496211 |
+| Epoch_10.pt | 0.9348333333333334 | 0.004270556675772805 |
+| Epoch_10_batch_2999.pt | 0.9340000000000002 | 0.004217833976911618 |
+| Epoch_12_batch_5999.pt | 0.933 | 0.0040960685758148286 |
+| Epoch_9_batch_2999.pt | 0.9256666666666666 | 0.005129195061148349 |
+| Epoch_9_batch_5999.pt | 0.9243333333333332 | 0.003670032125536389 |
+| Epoch_8_batch_5999.pt | 0.9238333333333332 | 0.0029085866917129824 |
+| Epoch_6_batch_5999.pt | 0.9235000000000001 | 0.00414587404595681 |
+| Epoch_7_batch_5999.pt | 0.9233333333333332 | 0.0031426968052735457 |
+| Epoch_5_batch_2999.pt | 0.9225 | 0.00415925266316551 |
+| Epoch_9.pt | 0.9213333333333334 | 0.0038713891978187473 |
+| Epoch_8.pt | 0.9213333333333333 | 0.004784233364802444 |
+| Epoch_7_batch_2999.pt | 0.9211666666666668 | 0.0047078211783067574 |
+| Epoch_5_batch_5999.pt | 0.921 | 0.00362773949273656 |
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d9a882beb99e0ff02ada6c31ff75c1c08f86a255
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_cplfw.txt
@@ -0,0 +1,37 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.8363333333333334 | 0.006028737762778021 |
+| Epoch_14_batch_5999.pt | 0.8361666666666666 | 0.0054831391387429535 |
+| Epoch_17_batch_5999.pt | 0.8361666666666666 | 0.006024896905014171 |
+| Epoch_16_batch_5999.pt | 0.8358333333333332 | 0.005382011648325939 |
+| Epoch_16.pt | 0.8358333333333332 | 0.005512331854046501 |
+| Epoch_17_batch_2999.pt | 0.8356666666666668 | 0.0056338921484832055 |
+| Epoch_13_batch_5999.pt | 0.8356666666666666 | 0.0061121211286555895 |
+| Epoch_15.pt | 0.8351666666666666 | 0.006612978261077343 |
+| Epoch_16_batch_2999.pt | 0.8341666666666667 | 0.005927806414592915 |
+| Epoch_17.pt | 0.8341666666666667 | 0.005694105680980364 |
+| Epoch_15_batch_5999.pt | 0.8338333333333334 | 0.005816377473443434 |
+| Epoch_11_batch_5999.pt | 0.8338333333333333 | 0.006662267066782587 |
+| Epoch_12_batch_5999.pt | 0.8335000000000001 | 0.006233372945011793 |
+| Epoch_11.pt | 0.833 | 0.005659036256998601 |
+| Epoch_13.pt | 0.833 | 0.005680810236072985 |
+| Epoch_13_batch_2999.pt | 0.8324999999999999 | 0.006152636692638772 |
+| Epoch_14_batch_2999.pt | 0.8314999999999999 | 0.005897530347317501 |
+| Epoch_10.pt | 0.8313333333333333 | 0.005543319859229525 |
+| Epoch_10_batch_2999.pt | 0.8306666666666667 | 0.005937950762616611 |
+| Epoch_12_batch_2999.pt | 0.8305 | 0.005789784532974571 |
+| Epoch_11_batch_2999.pt | 0.8301666666666666 | 0.005325515102890156 |
+| Epoch_14.pt | 0.829 | 0.005859465277082311 |
+| Epoch_12.pt | 0.8276666666666668 | 0.005539978161030178 |
+| Epoch_10_batch_5999.pt | 0.8248333333333333 | 0.004419722911821956 |
+| Epoch_9_batch_2999.pt | 0.8066666666666669 | 0.006324555320336763 |
+| Epoch_8_batch_5999.pt | 0.8059999999999998 | 0.007892800282301347 |
+| Epoch_7_batch_5999.pt | 0.8013333333333333 | 0.007832348243812055 |
+| Epoch_5_batch_2999.pt | 0.8011666666666667 | 0.007457906567554363 |
+| Epoch_9.pt | 0.7998333333333334 | 0.008334999833366661 |
+| Epoch_9_batch_5999.pt | 0.7998333333333332 | 0.007667673041838907 |
+| Epoch_7.pt | 0.7995 | 0.007718224384812934 |
+| Epoch_8_batch_2999.pt | 0.7995 | 0.006275810901320704 |
+| Epoch_6_batch_5999.pt | 0.7993333333333335 | 0.007102425250887509 |
+| Epoch_7_batch_2999.pt | 0.7989999999999999 | 0.006992941767804253 |
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e55447cb0049f3f48c92e2133025491a34d3cf33
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13.pt | 0.9958333333333332 | 0.0011453071182271274 |
+| Epoch_16.pt | 0.9958333333333332 | 0.0011180339887498902 |
+| Epoch_17.pt | 0.9958333333333332 | 0.0011979921473804322 |
+| Epoch_12_batch_5999.pt | 0.9956666666666667 | 0.0011706281947614144 |
+| Epoch_14.pt | 0.9956666666666665 | 0.0011967032904743316 |
+| Epoch_15_batch_5999.pt | 0.9956666666666665 | 0.0011967032904743316 |
+| Epoch_16_batch_2999.pt | 0.9956666666666665 | 0.0011967032904743316 |
+| Epoch_10_batch_5999.pt | 0.9955 | 0.0011928283640879936 |
+| Epoch_13_batch_2999.pt | 0.9955 | 0.0012184284555256265 |
+| Epoch_15_batch_2999.pt | 0.9955 | 0.0012184284555256265 |
+| Epoch_15.pt | 0.9955 | 0.0012921892610681118 |
+| Epoch_11_batch_5999.pt | 0.9953333333333335 | 0.001160034056545619 |
+| Epoch_11.pt | 0.9953333333333335 | 0.001160034056545619 |
+| Epoch_12.pt | 0.9953333333333335 | 0.0010772621905369623 |
+| Epoch_10.pt | 0.9953333333333333 | 0.0011863420280034786 |
+| Epoch_11_batch_2999.pt | 0.9951666666666666 | 0.0013017082793177807 |
+| Epoch_10_batch_2999.pt | 0.9950000000000001 | 0.0011653431646335038 |
+| Epoch_12_batch_2999.pt | 0.9950000000000001 | 0.001337954953199147 |
+| Epoch_14_batch_2999.pt | 0.9950000000000001 | 0.0013833221775543059 |
+| Epoch_16_batch_5999.pt | 0.9950000000000001 | 0.0013833221775543059 |
+| Epoch_17_batch_5999.pt | 0.9950000000000001 | 0.001337954953199147 |
+| Epoch_14_batch_5999.pt | 0.9949999999999999 | 0.0013146843962443596 |
+| Epoch_13_batch_5999.pt | 0.9948333333333335 | 0.0012777777777777798 |
+| Epoch_17_batch_2999.pt | 0.9948333333333335 | 0.0013482956777235145 |
+| Epoch_9.pt | 0.9945 | 0.0010555555555555537 |
+| Epoch_6_batch_5999.pt | 0.9943333333333333 | 0.001143958904554113 |
+| Epoch_8.pt | 0.994 | 0.0015355861067872475 |
+| Epoch_8_batch_2999.pt | 0.9938333333333335 | 0.0013391078659104375 |
+| Epoch_4_batch_2999.pt | 0.9933333333333334 | 0.0012668615834434832 |
+| Epoch_5_batch_2999.pt | 0.9933333333333334 | 0.0012422599874998784 |
+| Epoch_8_batch_5999.pt | 0.9933333333333334 | 0.0012909944487358084 |
+| Epoch_9_batch_5999.pt | 0.9933333333333334 | 0.0014698618394803228 |
+| Epoch_6_batch_2999.pt | 0.9931666666666666 | 0.0016377114414426284 |
+| Epoch_7_batch_5999.pt | 0.9930000000000001 | 0.001378852627332321 |
+| Epoch_9_batch_2999.pt | 0.993 | 0.0016254154264808637 |
+| Epoch_4_batch_5999.pt | 0.9926666666666666 | 0.0010886621079036368 |
+| Epoch_7_batch_2999.pt | 0.9926666666666666 | 0.0018121673811444558 |
+| Epoch_5.pt | 0.9925 | 0.001223484196974741 |
+| Epoch_7.pt | 0.9923333333333334 | 0.0019594657876164877 |
+| Epoch_3.pt | 0.9921666666666666 | 0.001192828364087998 |
+| Epoch_5_batch_5999.pt | 0.992 | 0.0015475986974649056 |
+| Epoch_2_batch_5999.pt | 0.9911666666666668 | 0.001166666666666673 |
+| Epoch_6.pt | 0.9911666666666665 | 0.0017924739783224046 |
+| Epoch_3_batch_5999.pt | 0.991 | 0.0017950549357115021 |
+| Epoch_3_batch_2999.pt | 0.9904999999999999 | 0.0016301556390134716 |
+| Epoch_4.pt | 0.9904999999999999 | 0.0015525765124980132 |
+| Epoch_2.pt | 0.9888333333333333 | 0.001722222222222219 |
+| Epoch_2_batch_2999.pt | 0.9871666666666667 | 0.0013391078659104408 |
+| Epoch_1_batch_5999.pt | 0.9863333333333333 | 0.0016442942874387437 |
+| Epoch_1.pt | 0.9844999999999999 | 0.0020038543107634894 |
+| Epoch_1_batch_2999.pt | 0.9835 | 0.0023366378716459003 |
+| Epoch_0_batch_5999.pt | 0.9726666666666667 | 0.0027910792637956612 |
+| Epoch_0.pt | 0.9700000000000001 | 0.0028760398012321717 |
+| Epoch_0_batch_2999.pt | 0.9458333333333334 | 0.0028787214240070195 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..fd622b0c68a76c44a893385b86fd9a2a1ca6d660
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_African.txt
@@ -0,0 +1,40 @@
++------------------------+--------------------+-----------------------+ [19/294]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.8838333333333332 | 0.004661703710177375 |
+| Epoch_15_batch_2999.pt | 0.8833333333333334 | 0.005127991448539701 |
+| Epoch_17.pt | 0.8833333333333332 | 0.005258737584977436 |
+| Epoch_17_batch_5999.pt | 0.882 | 0.004552844721899531 |
+| Epoch_14_batch_2999.pt | 0.8818333333333334 | 0.004826943413387952 |
+| Epoch_17_batch_2999.pt | 0.8818333333333334 | 0.004270556675772805 |
+| Epoch_13_batch_5999.pt | 0.8816666666666666 | 0.004260064336151301 |
+| Epoch_16_batch_5999.pt | 0.8813333333333334 | 0.004207576939724776 |
+| Epoch_14_batch_5999.pt | 0.8813333333333333 | 0.004764840364975905 |
+| Epoch_13_batch_2999.pt | 0.8811666666666668 | 0.004568072235704827 |
+| Epoch_15.pt | 0.8811666666666668 | 0.004388888888888883 |
+| Epoch_16_batch_2999.pt | 0.8800000000000001 | 0.004267303193260341 |
+| Epoch_14.pt | 0.8788333333333334 | 0.0049131343243278905 |
+| Epoch_16.pt | 0.8783333333333333 | 0.003824869884013002 |
+| Epoch_13.pt | 0.8771666666666667 | 0.004540965797093208 |
+| Epoch_12_batch_5999.pt | 0.8768333333333335 | 0.00523196490719533 |
+| Epoch_11_batch_2999.pt | 0.8766666666666667 | 0.005778846055087042 |
+| Epoch_11_batch_5999.pt | 0.8763333333333334 | 0.004640136226182472 |
+| Epoch_11.pt | 0.8763333333333334 | 0.005333333333333333 |
+| Epoch_10_batch_5999.pt | 0.8758333333333332 | 0.004857538369621558 |
+| Epoch_12.pt | 0.8741666666666668 | 0.004939448162310054 |
+| Epoch_10_batch_2999.pt | 0.874 | 0.005289168993516396 |
+| Epoch_10.pt | 0.8738333333333334 | 0.00466831981301016 |
+| Epoch_12_batch_2999.pt | 0.8716666666666667 | 0.00571979452277055 |
+| Epoch_9_batch_2999.pt | 0.8418333333333333 | 0.007355396104347599 |
+| Epoch_6_batch_5999.pt | 0.8408333333333333 | 0.0047025735008465516 |
+| Epoch_7_batch_2999.pt | 0.8406666666666667 | 0.004708804466759353 |
+| Epoch_7_batch_5999.pt | 0.8388333333333333 | 0.0053057737011581235 |
+| Epoch_8_batch_2999.pt | 0.8376666666666667 | 0.004535184807101898 |
+| Epoch_8_batch_5999.pt | 0.8376666666666666 | 0.0052422782720379785 |
+| Epoch_9_batch_5999.pt | 0.8371666666666666 | 0.005826980667867685 |
+| Epoch_6_batch_2999.pt | 0.8368333333333334 | 0.005902761437885882 |
+| Epoch_8.pt | 0.8361666666666666 | 0.00501879184714193 |
+| Epoch_5_batch_5999.pt | 0.8348333333333333 | 0.0057812489572905265 |
+| Epoch_9.pt | 0.8333333333333334 | 0.0036514837167011143 |
+| Epoch_5_batch_2999.pt | 0.8331666666666667 | 0.006769804052820161 |
+| Epoch_4_batch_5999.pt | 0.8308333333333333 | 0.004876562744139653 |
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..14a60b61c6ffe681e4fd937fc2a1145d3f638e76
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,39 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.8788333333333334 | 0.0038892856940415874 |
+| Epoch_15_batch_2999.pt | 0.8786666666666667 | 0.0034676636742949464 |
+| Epoch_17_batch_2999.pt | 0.8781666666666668 | 0.0038685978574563846 |
+| Epoch_13_batch_5999.pt | 0.8776666666666666 | 0.004053652521545486 |
+| Epoch_15.pt | 0.8776666666666666 | 0.003559026084010444 |
+| Epoch_16.pt | 0.877 | 0.003926799343793848 |
+| Epoch_16_batch_5999.pt | 0.8768333333333335 | 0.0035438174297371286 |
+| Epoch_13.pt | 0.8766666666666667 | 0.0036767538017276266 |
+| Epoch_13_batch_2999.pt | 0.8761666666666666 | 0.003685557397916 |
+| Epoch_17.pt | 0.8756666666666668 | 0.00334442598739832 |
+| Epoch_17_batch_5999.pt | 0.8755000000000001 | 0.003220785120141272 |
+| Epoch_14_batch_2999.pt | 0.8753333333333334 | 0.003572874486847457 |
+| Epoch_12_batch_5999.pt | 0.8751666666666666 | 0.004775516260631466 |
+| Epoch_14_batch_5999.pt | 0.8751666666666665 | 0.003804237403504448 |
+| Epoch_14.pt | 0.875 | 0.003379312516832349 |
+| Epoch_11_batch_2999.pt | 0.8748333333333331 | 0.002427339141661499 |
+| Epoch_15_batch_5999.pt | 0.8743333333333334 | 0.0035066075195687757 |
+| Epoch_11.pt | 0.8721666666666665 | 0.0037105954398881243 |
+| Epoch_10.pt | 0.8714999999999999 | 0.003963179294891518 |
+| Epoch_12.pt | 0.8714999999999999 | 0.0032815420945827463 |
+| Epoch_10_batch_5999.pt | 0.8713333333333333 | 0.004532461789860255 |
+| Epoch_12_batch_2999.pt | 0.8696666666666667 | 0.00409606857581484 |
+| Epoch_11_batch_5999.pt | 0.8691666666666666 | 0.003687231890232248 |
+| Epoch_10_batch_2999.pt | 0.8686666666666667 | 0.004936635531449567 |
+| Epoch_9_batch_5999.pt | 0.8473333333333333 | 0.004038395965641385 |
+| Epoch_6_batch_5999.pt | 0.8466666666666667 | 0.004520187912623998 |
+| Epoch_7_batch_5999.pt | 0.8443333333333334 | 0.003541639433446492 |
+| Epoch_8_batch_2999.pt | 0.8403333333333334 | 0.0035468643776694125 |
+| Epoch_8_batch_5999.pt | 0.8398333333333333 | 0.0042123422416149746 |
+| Epoch_5_batch_5999.pt | 0.8386666666666669 | 0.0044982849955549275 |
+| Epoch_9_batch_2999.pt | 0.8383333333333335 | 0.0035398960715719986 |
+| Epoch_4_batch_2999.pt | 0.837 | 0.0038634085890475 |
+| Epoch_8.pt | 0.8348333333333333 | 0.005002777006601213 |
+| Epoch_7_batch_2999.pt | 0.8341666666666667 | 0.0032322640461753035 |
+| Epoch_5_batch_2999.pt | 0.8321666666666667 | 0.004714372560794842 |
+| Epoch_4_batch_5999.pt | 0.8318333333333332 | 0.004241549332583645 |
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3cab9c4d0d51f26ac83fd2ffbf4a0b42595b6492
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,40 @@
++------------------------+--------------------+-----------------------+ [19/417]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.9555 | 0.003751954223311792 |
+| Epoch_13_batch_5999.pt | 0.9545 | 0.003258890972458331 |
+| Epoch_14_batch_2999.pt | 0.9543333333333333 | 0.0032791899029719442 |
+| Epoch_16_batch_5999.pt | 0.9538333333333334 | 0.0037105954398881256 |
+| Epoch_17.pt | 0.9535 | 0.003868597857456371 |
+| Epoch_15_batch_2999.pt | 0.9533333333333335 | 0.0032203059435976576 |
+| Epoch_17_batch_5999.pt | 0.9531666666666666 | 0.0036552853665768933 |
+| Epoch_16_batch_2999.pt | 0.9528333333333334 | 0.003983375948942969 |
+| Epoch_16.pt | 0.9528333333333334 | 0.0035490391674854334 |
+| Epoch_12_batch_2999.pt | 0.9526666666666668 | 0.004247003300803647 |
+| Epoch_15.pt | 0.9525 | 0.0035939764421413028 |
+| Epoch_13.pt | 0.9523333333333331 | 0.0037035185138886576 |
+| Epoch_14_batch_5999.pt | 0.9521666666666668 | 0.003905124837953329 |
+| Epoch_12_batch_5999.pt | 0.9521666666666666 | 0.0031431878132923583 |
+| Epoch_17_batch_2999.pt | 0.9516666666666668 | 0.00351364184463153 |
+| Epoch_13_batch_2999.pt | 0.9514999999999999 | 0.0037387691907536055 |
+| Epoch_11.pt | 0.9513333333333334 | 0.0036413265795942062 |
+| Epoch_10_batch_5999.pt | 0.9506666666666665 | 0.003969015799887053 |
+| Epoch_11_batch_2999.pt | 0.9506666666666665 | 0.0033259176771323956 |
+| Epoch_14.pt | 0.9503333333333336 | 0.0044430553384738215 |
+| Epoch_11_batch_5999.pt | 0.9503333333333334 | 0.00472581562625261 |
+| Epoch_12.pt | 0.9496666666666668 | 0.003422871511277636 |
+| Epoch_10.pt | 0.9481666666666667 | 0.004123479892002996 |
+| Epoch_10_batch_2999.pt | 0.9451666666666668 | 0.0030373193184081455 |
+| Epoch_9_batch_5999.pt | 0.9365 | 0.003482318654269958 |
+| Epoch_9_batch_2999.pt | 0.9288333333333334 | 0.0046946909862939975 |
+| Epoch_7_batch_5999.pt | 0.9283333333333333 | 0.005012330474951091 |
+| Epoch_6_batch_5999.pt | 0.9256666666666667 | 0.00467591675833351 |
+| Epoch_8_batch_2999.pt | 0.9233333333333335 | 0.004444444444444444 |
+| Epoch_8_batch_5999.pt | 0.9231666666666669 | 0.004085883557377088 |
+| Epoch_7_batch_2999.pt | 0.9223333333333334 | 0.004575835618216446 |
+| Epoch_6_batch_2999.pt | 0.922 | 0.003150543750835079 |
+| Epoch_9.pt | 0.9216666666666666 | 0.004260064336151288 |
+| Epoch_8.pt | 0.9215 | 0.0032815420945827433 |
+| Epoch_7.pt | 0.9198333333333333 | 0.004145874045956811 |
+| Epoch_5_batch_5999.pt | 0.9186666666666667 | 0.004917843550295721 |
+| Epoch_4_batch_5999.pt | 0.9179999999999999 | 0.003774508389214006 |
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..4edb29407f323949a4e04bae31a5fb78b1854736
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,40 @@
++------------------------+--------------------+-----------------------+ [21/481]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9118333333333333 | 0.0041309580988936265 |
+| Epoch_17_batch_5999.pt | 0.9111666666666668 | 0.00314318781329237 |
+| Epoch_16_batch_5999.pt | 0.9111666666666665 | 0.0034964708839021314 |
+| Epoch_13_batch_5999.pt | 0.9108333333333333 | 0.003344887382997867 |
+| Epoch_14_batch_2999.pt | 0.9098333333333335 | 0.0037716453494430384 |
+| Epoch_16.pt | 0.9096666666666667 | 0.0037498971179302713 |
+| Epoch_13.pt | 0.909 | 0.003212629398844659 |
+| Epoch_13_batch_2999.pt | 0.9088333333333333 | 0.003230353724468144 |
+| Epoch_15_batch_2999.pt | 0.9086666666666667 | 0.003140732005578355 |
+| Epoch_14_batch_5999.pt | 0.9081666666666667 | 0.003337497399083468 |
+| Epoch_11_batch_2999.pt | 0.908 | 0.0031407320055783557 |
+| Epoch_17.pt | 0.9078333333333333 | 0.0033245254000838207 |
+| Epoch_16_batch_2999.pt | 0.9076666666666666 | 0.0027910792637956617 |
+| Epoch_12_batch_5999.pt | 0.907 | 0.003934651379916839 |
+| Epoch_17_batch_2999.pt | 0.9066666666666666 | 0.003142696805273541 |
+| Epoch_15_batch_5999.pt | 0.9065000000000001 | 0.003697262918216626 |
+| Epoch_14.pt | 0.9065 | 0.003016927551641218 |
+| Epoch_11.pt | 0.9055 | 0.00338888888888889 |
+| Epoch_11_batch_5999.pt | 0.905 | 0.003608973726424821 |
+| Epoch_10.pt | 0.9041666666666666 | 0.004643128468533763 |
+| Epoch_12_batch_2999.pt | 0.9040000000000001 | 0.0035935470286213855 |
+| Epoch_10_batch_5999.pt | 0.9038333333333334 | 0.0033430414185178603 |
+| Epoch_12.pt | 0.9038333333333334 | 0.003751954223311792 |
+| Epoch_10_batch_2999.pt | 0.9031666666666667 | 0.003755243248026935 |
+| Epoch_9_batch_2999.pt | 0.8856666666666666 | 0.0031933573029333108 |
+| Epoch_7_batch_5999.pt | 0.8818333333333334 | 0.0034106767729268537 |
+| Epoch_9_batch_5999.pt | 0.8813333333333333 | 0.003725123247608941 |
+| Epoch_8_batch_2999.pt | 0.8773333333333333 | 0.0041290898169026756 |
+| Epoch_6_batch_5999.pt | 0.8736666666666668 | 0.005425135829214983 |
+| Epoch_8_batch_5999.pt | 0.8716666666666667 | 0.0037515428924742547 |
+| Epoch_7_batch_2999.pt | 0.8711666666666666 | 0.0017924739783224133 |
+| Epoch_4_batch_5999.pt | 0.8705 | 0.0044447916531043545 |
+| Epoch_9.pt | 0.8705 | 0.00472091485170375 |
+| Epoch_8.pt | 0.8693333333333333 | 0.00384258143684235 |
+| Epoch_5_batch_5999.pt | 0.8691666666666666 | 0.00426187520311348 |
+| Epoch_6_batch_2999.pt | 0.8688333333333332 | 0.004479376610066948 |
+| Epoch_7.pt | 0.8681666666666666 | 0.004664351277456082 |
diff --git a/bob/bio/facexzoo/models/heads/AM-Softmax/log.log b/bob/bio/facexzoo/models/heads/AM-Softmax/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..26ce560faf0fd362ff705aca72437b25807d1d51
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AM-Softmax/log.log
@@ -0,0 +1,651 @@
+INFO 2020-11-23 20:30:57 train.py: 172] Start optimization.
+INFO 2020-11-23 20:30:57 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='am-softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='am-resnet', train_file='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f828ec1aba8>)
+INFO 2020-11-23 20:31:23 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 23.431353
+INFO 2020-11-23 20:37:31 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 22.992578
+INFO 2020-11-23 20:43:32 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 22.112193
+INFO 2020-11-23 20:49:05 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 21.501875
+INFO 2020-11-23 20:54:20 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 21.017682
+INFO 2020-11-23 20:59:30 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 20.586930
+INFO 2020-11-23 21:04:16 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 20.153704
+INFO 2020-11-23 21:09:07 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 19.735421
+INFO 2020-11-23 21:13:43 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 19.322450
+INFO 2020-11-23 21:18:16 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 18.932620
+INFO 2020-11-23 21:22:12 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 18.544820
+INFO 2020-11-23 21:26:14 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 18.146690
+INFO 2020-11-23 21:30:20 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 17.753865
+INFO 2020-11-23 21:34:11 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 17.327234
+INFO 2020-11-23 21:37:29 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 16.925471
+INFO 2020-11-23 21:40:49 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-23 21:40:49 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 16.495277
+INFO 2020-11-23 21:44:05 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 16.070678
+INFO 2020-11-23 21:47:11 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 15.632312
+INFO 2020-11-23 21:50:13 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 15.213835
+INFO 2020-11-23 21:53:08 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 14.802538
+INFO 2020-11-23 21:55:48 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 14.382193
+INFO 2020-11-23 21:58:06 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 13.995301
+INFO 2020-11-23 22:00:32 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 13.619508
+INFO 2020-11-23 22:02:42 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 13.255566
+INFO 2020-11-23 22:04:50 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 12.891582
+INFO 2020-11-23 22:07:04 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 12.532326
+INFO 2020-11-23 22:08:54 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 12.189917
+INFO 2020-11-23 22:10:55 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 11.849234
+INFO 2020-11-23 22:12:52 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 11.559057
+INFO 2020-11-23 22:14:55 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 11.259521
+INFO 2020-11-23 22:16:21 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-23 22:16:21 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 10.956776
+INFO 2020-11-23 22:17:37 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 10.676456
+INFO 2020-11-23 22:18:54 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 10.420154
+INFO 2020-11-23 22:18:59 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-23 22:19:01 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 10.197703
+INFO 2020-11-23 22:20:19 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 9.851507
+INFO 2020-11-23 22:21:37 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 9.637592
+INFO 2020-11-23 22:22:54 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 9.476733
+INFO 2020-11-23 22:24:11 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 9.324735
+INFO 2020-11-23 22:25:29 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 9.205025
+INFO 2020-11-23 22:26:46 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 9.032687
+INFO 2020-11-23 22:28:04 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 8.896845
+INFO 2020-11-23 22:29:21 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 8.755357
+INFO 2020-11-23 22:30:38 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 8.651207
+INFO 2020-11-23 22:31:56 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 8.502586
+INFO 2020-11-23 22:33:13 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 8.387458
+INFO 2020-11-23 22:34:31 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 8.265476
+INFO 2020-11-23 22:35:48 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 8.182377
+INFO 2020-11-23 22:37:05 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 8.071185
+INFO 2020-11-23 22:38:22 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-23 22:38:23 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 7.976120
+INFO 2020-11-23 22:39:40 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 7.912530
+INFO 2020-11-23 22:40:58 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 7.808188
+INFO 2020-11-23 22:42:15 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 7.715766
+INFO 2020-11-23 22:43:32 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 7.646962
+INFO 2020-11-23 22:44:50 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 7.621099
+INFO 2020-11-23 22:46:07 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 7.518125
+INFO 2020-11-23 22:47:24 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 7.463833
+INFO 2020-11-23 22:48:42 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 7.397627
+INFO 2020-11-23 22:49:59 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 7.345807
+INFO 2020-11-23 22:51:17 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 7.290254
+INFO 2020-11-23 22:52:34 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 7.230198
+INFO 2020-11-23 22:53:52 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 7.203372
+INFO 2020-11-23 22:55:09 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 7.147704
+INFO 2020-11-23 22:56:26 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 7.095600
+INFO 2020-11-23 22:57:44 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-23 22:57:44 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 7.073151
+INFO 2020-11-23 22:59:01 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 7.009332
+INFO 2020-11-23 23:00:19 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 6.971970
+INFO 2020-11-23 23:00:25 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-23 23:00:26 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 6.843241
+INFO 2020-11-23 23:01:44 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 6.585406
+INFO 2020-11-23 23:03:01 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 6.568176
+INFO 2020-11-23 23:04:18 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 6.612730
+INFO 2020-11-23 23:05:36 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 6.661861
+INFO 2020-11-23 23:06:53 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 6.646130
+INFO 2020-11-23 23:08:10 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 6.662672
+INFO 2020-11-23 23:09:27 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 6.649159
+INFO 2020-11-23 23:10:45 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 6.620708
+INFO 2020-11-23 23:12:02 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 6.643027
+INFO 2020-11-23 23:13:20 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 6.600493
+INFO 2020-11-23 23:14:37 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 6.572066
+INFO 2020-11-23 23:15:54 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 6.578855
+INFO 2020-11-23 23:17:11 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 6.548282
+INFO 2020-11-23 23:18:29 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 6.510382
+INFO 2020-11-23 23:19:46 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-23 23:19:46 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 6.503403
+INFO 2020-11-23 23:21:03 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 6.485130
+INFO 2020-11-23 23:22:20 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 6.464876
+INFO 2020-11-23 23:23:38 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 6.433336
+INFO 2020-11-23 23:24:55 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 6.416342
+INFO 2020-11-23 23:26:12 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 6.420710
+INFO 2020-11-23 23:27:29 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 6.390684
+INFO 2020-11-23 23:28:47 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 6.371358
+INFO 2020-11-23 23:30:04 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 6.369870
+INFO 2020-11-23 23:31:21 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 6.312445
+INFO 2020-11-23 23:32:38 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 6.308288
+INFO 2020-11-23 23:33:55 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 6.256223
+INFO 2020-11-23 23:35:13 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 6.264229
+INFO 2020-11-23 23:36:30 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 6.231889
+INFO 2020-11-23 23:37:47 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 6.206432
+INFO 2020-11-23 23:39:04 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-23 23:39:04 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 6.197555
+INFO 2020-11-23 23:40:22 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 6.170946
+INFO 2020-11-23 23:41:39 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 6.165020
+INFO 2020-11-23 23:41:45 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-23 23:41:46 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 6.170831
+INFO 2020-11-23 23:43:04 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 5.832000
+INFO 2020-11-23 23:44:21 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 5.814424
+INFO 2020-11-23 23:45:37 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 5.887542
+INFO 2020-11-23 23:46:54 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 5.900610
+INFO 2020-11-23 23:48:10 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 5.921929
+INFO 2020-11-23 23:49:27 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 5.967691
+INFO 2020-11-23 23:50:44 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 5.963248
+INFO 2020-11-23 23:52:00 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 5.979250
+INFO 2020-11-23 23:53:17 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 5.981584
+INFO 2020-11-23 23:54:33 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 5.963605
+INFO 2020-11-23 23:55:50 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 5.967984
+INFO 2020-11-23 23:57:06 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 5.962511
+INFO 2020-11-23 23:58:23 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 5.953251
+INFO 2020-11-23 23:59:39 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 5.938224
+INFO 2020-11-24 00:00:56 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-24 00:00:56 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 5.926983
+INFO 2020-11-24 00:02:13 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 5.917728
+INFO 2020-11-24 00:03:30 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 5.908873
+INFO 2020-11-24 00:04:47 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 5.896288
+INFO 2020-11-24 00:06:04 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 5.886286
+INFO 2020-11-24 00:07:21 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 5.882384
+INFO 2020-11-24 00:08:38 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 5.863813
+INFO 2020-11-24 00:09:56 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 5.870124
+INFO 2020-11-24 00:11:13 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 5.837263
+INFO 2020-11-24 00:12:30 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 5.821667
+INFO 2020-11-24 00:13:47 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 5.850752
+INFO 2020-11-24 00:15:04 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 5.808350
+INFO 2020-11-24 00:16:21 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 5.786753
+INFO 2020-11-24 00:17:38 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 5.784242
+INFO 2020-11-24 00:18:55 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 5.775763
+INFO 2020-11-24 00:20:12 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-24 00:20:13 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 5.770726
+INFO 2020-11-24 00:21:30 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 5.774870
+INFO 2020-11-24 00:22:47 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 5.771624
+INFO 2020-11-24 00:22:53 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-24 00:22:54 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 5.707942
+INFO 2020-11-24 00:24:12 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 5.445262
+INFO 2020-11-24 00:25:30 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 5.422858
+INFO 2020-11-24 00:26:48 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 5.471771
+INFO 2020-11-24 00:28:05 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 5.524503
+INFO 2020-11-24 00:29:22 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 5.566489
+INFO 2020-11-24 00:30:40 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 5.581012
+INFO 2020-11-24 00:31:57 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 5.592611
+INFO 2020-11-24 00:33:14 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 5.607801
+INFO 2020-11-24 00:34:32 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 5.640018
+INFO 2020-11-24 00:35:49 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 5.634800
+INFO 2020-11-24 00:37:06 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 5.626814
+INFO 2020-11-24 00:38:23 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 5.608207
+INFO 2020-11-24 00:39:41 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 5.614910
+INFO 2020-11-24 00:40:58 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 5.634805
+INFO 2020-11-24 00:42:15 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-24 00:42:15 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 5.612011
+INFO 2020-11-24 00:43:32 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 5.619430
+INFO 2020-11-24 00:44:48 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 5.586245
+INFO 2020-11-24 00:46:05 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 5.595993
+INFO 2020-11-24 00:47:21 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 5.596636
+INFO 2020-11-24 00:48:38 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 5.580320
+INFO 2020-11-24 00:49:54 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 5.595710
+INFO 2020-11-24 00:51:11 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 5.580321
+INFO 2020-11-24 00:52:27 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 5.569320
+INFO 2020-11-24 00:53:44 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 5.550546
+INFO 2020-11-24 00:55:00 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 5.545074
+INFO 2020-11-24 00:56:17 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 5.546995
+INFO 2020-11-24 00:57:34 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 5.514362
+INFO 2020-11-24 00:58:50 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 5.537737
+INFO 2020-11-24 01:00:07 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 5.535563
+INFO 2020-11-24 01:01:23 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-24 01:01:23 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 5.517742
+INFO 2020-11-24 01:02:40 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 5.500482
+INFO 2020-11-24 01:03:58 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 5.494615
+INFO 2020-11-24 01:04:04 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-24 01:04:05 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 5.450038
+INFO 2020-11-24 01:05:23 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 5.192066
+INFO 2020-11-24 01:06:41 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 5.182883
+INFO 2020-11-24 01:07:58 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 5.232825
+INFO 2020-11-24 01:09:15 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 5.288545
+INFO 2020-11-24 01:10:33 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 5.323940
+INFO 2020-11-24 01:11:50 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 5.347373
+INFO 2020-11-24 01:13:07 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 5.349529
+INFO 2020-11-24 01:14:24 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 5.378345
+INFO 2020-11-24 01:15:41 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 5.395845
+INFO 2020-11-24 01:16:58 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 5.383296
+INFO 2020-11-24 01:18:15 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 5.423225
+INFO 2020-11-24 01:19:33 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 5.408470
+INFO 2020-11-24 01:20:50 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 5.410404
+INFO 2020-11-24 01:22:07 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 5.399075
+INFO 2020-11-24 01:23:24 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-24 01:23:24 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 5.398088
+INFO 2020-11-24 01:24:41 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 5.400978
+INFO 2020-11-24 01:25:58 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 5.396452
+INFO 2020-11-24 01:27:15 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 5.398574
+INFO 2020-11-24 01:28:32 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 5.378964
+INFO 2020-11-24 01:29:50 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 5.376358
+INFO 2020-11-24 01:31:07 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 5.411989
+INFO 2020-11-24 01:32:24 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 5.367356
+INFO 2020-11-24 01:33:41 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 5.371219
+INFO 2020-11-24 01:34:58 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 5.366692
+INFO 2020-11-24 01:36:15 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 5.370219
+INFO 2020-11-24 01:37:32 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 5.333378
+INFO 2020-11-24 01:38:49 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 5.357778
+INFO 2020-11-24 01:40:06 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 5.310333
+INFO 2020-11-24 01:41:23 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 5.336423
+INFO 2020-11-24 01:42:40 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-24 01:42:40 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 5.328626
+INFO 2020-11-24 01:43:56 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 5.353875
+INFO 2020-11-24 01:45:13 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 5.332091
+INFO 2020-11-24 01:45:19 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-24 01:45:20 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 5.336439
+INFO 2020-11-24 01:46:38 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 5.004231
+INFO 2020-11-24 01:47:56 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 4.990888
+INFO 2020-11-24 01:49:13 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 5.080324
+INFO 2020-11-24 01:50:31 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 5.114117
+INFO 2020-11-24 01:51:48 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 5.172408
+INFO 2020-11-24 01:53:05 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 5.192198
+INFO 2020-11-24 01:54:23 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 5.185811
+INFO 2020-11-24 01:55:40 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 5.214260
+INFO 2020-11-24 01:56:57 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 5.237810
+INFO 2020-11-24 01:58:14 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 5.241336
+INFO 2020-11-24 01:59:31 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 5.259704
+INFO 2020-11-24 02:00:49 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 5.228155
+INFO 2020-11-24 02:02:06 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 5.254398
+INFO 2020-11-24 02:03:23 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 5.267083
+INFO 2020-11-24 02:04:40 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-24 02:04:40 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 5.250467
+INFO 2020-11-24 02:05:58 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 5.233817
+INFO 2020-11-24 02:07:15 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 5.231766
+INFO 2020-11-24 02:08:32 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 5.246143
+INFO 2020-11-24 02:09:49 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 5.249186
+INFO 2020-11-24 02:11:06 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 5.240134
+INFO 2020-11-24 02:12:24 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 5.225728
+INFO 2020-11-24 02:13:41 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 5.232335
+INFO 2020-11-24 02:14:58 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 5.224760
+INFO 2020-11-24 02:16:15 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 5.223358
+INFO 2020-11-24 02:17:32 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 5.202141
+INFO 2020-11-24 02:18:49 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 5.217607
+INFO 2020-11-24 02:20:06 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 5.207081
+INFO 2020-11-24 02:21:23 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 5.208580
+INFO 2020-11-24 02:22:41 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 5.216951
+INFO 2020-11-24 02:23:58 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-24 02:23:58 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 5.199084
+INFO 2020-11-24 02:25:15 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 5.178439
+INFO 2020-11-24 02:26:32 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 5.191668
+INFO 2020-11-24 02:26:38 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-24 02:26:40 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 5.188671
+INFO 2020-11-24 02:27:58 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 4.865688
+INFO 2020-11-24 02:29:15 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 4.870638
+INFO 2020-11-24 02:30:32 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 4.937783
+INFO 2020-11-24 02:31:50 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 4.994499
+INFO 2020-11-24 02:33:07 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 5.028212
+INFO 2020-11-24 02:34:24 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 5.075426
+INFO 2020-11-24 02:35:41 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 5.085894
+INFO 2020-11-24 02:36:58 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 5.078220
+INFO 2020-11-24 02:38:16 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 5.115540
+INFO 2020-11-24 02:39:33 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 5.103897
+INFO 2020-11-24 02:40:50 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 5.109988
+INFO 2020-11-24 02:42:07 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 5.145308
+INFO 2020-11-24 02:43:24 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 5.106844
+INFO 2020-11-24 02:44:41 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 5.147266
+INFO 2020-11-24 02:45:58 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-24 02:45:59 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 5.104082
+INFO 2020-11-24 02:47:16 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 5.115193
+INFO 2020-11-24 02:48:33 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 5.124016
+INFO 2020-11-24 02:49:50 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 5.137811
+INFO 2020-11-24 02:51:07 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 5.124602
+INFO 2020-11-24 02:52:24 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 5.140168
+INFO 2020-11-24 02:53:42 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 5.124393
+INFO 2020-11-24 02:54:59 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 5.129347
+INFO 2020-11-24 02:56:16 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 5.124850
+INFO 2020-11-24 02:57:34 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 5.110467
+INFO 2020-11-24 02:58:51 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 5.117695
+INFO 2020-11-24 03:00:08 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 5.131347
+INFO 2020-11-24 03:01:25 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 5.121504
+INFO 2020-11-24 03:02:43 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 5.095711
+INFO 2020-11-24 03:04:00 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 5.101500
+INFO 2020-11-24 03:05:17 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-24 03:05:17 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 5.107354
+INFO 2020-11-24 03:06:34 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 5.112161
+INFO 2020-11-24 03:07:51 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 5.085738
+INFO 2020-11-24 03:07:57 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-24 03:07:59 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 5.051775
+INFO 2020-11-24 03:09:17 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 4.783735
+INFO 2020-11-24 03:10:35 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 4.787182
+INFO 2020-11-24 03:11:52 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 4.849713
+INFO 2020-11-24 03:13:09 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 4.898719
+INFO 2020-11-24 03:14:27 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 4.916939
+INFO 2020-11-24 03:15:44 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 4.973932
+INFO 2020-11-24 03:17:01 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 4.998854
+INFO 2020-11-24 03:18:18 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 4.986458
+INFO 2020-11-24 03:19:35 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 4.991726
+INFO 2020-11-24 03:20:53 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 5.028501
+INFO 2020-11-24 03:22:10 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 5.033619
+INFO 2020-11-24 03:23:27 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 5.029430
+INFO 2020-11-24 03:24:44 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 5.021079
+INFO 2020-11-24 03:26:01 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 5.052556
+INFO 2020-11-24 03:27:18 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-24 03:27:18 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 5.053469
+INFO 2020-11-24 03:28:36 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 5.064945
+INFO 2020-11-24 03:29:53 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 5.038971
+INFO 2020-11-24 03:31:10 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 5.010157
+INFO 2020-11-24 03:32:27 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 5.036120
+INFO 2020-11-24 03:33:44 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 5.029503
+INFO 2020-11-24 03:35:01 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 5.017757
+INFO 2020-11-24 03:36:18 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 5.034978
+INFO 2020-11-24 03:37:35 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 5.010205
+INFO 2020-11-24 03:38:52 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 5.028256
+INFO 2020-11-24 03:40:09 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 5.022256
+INFO 2020-11-24 03:41:27 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 5.036029
+INFO 2020-11-24 03:42:45 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 5.015791
+INFO 2020-11-24 03:44:02 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 5.034028
+INFO 2020-11-24 03:45:20 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 5.003483
+INFO 2020-11-24 03:46:38 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-24 03:46:39 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 5.003800
+INFO 2020-11-24 03:47:56 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 5.012013
+INFO 2020-11-24 03:49:13 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 5.004939
+INFO 2020-11-24 03:49:19 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-24 03:49:21 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 4.984909
+INFO 2020-11-24 03:50:38 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 4.706132
+INFO 2020-11-24 03:51:55 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 4.707455
+INFO 2020-11-24 03:53:12 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 4.764274
+INFO 2020-11-24 03:54:29 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 4.796268
+INFO 2020-11-24 03:55:45 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 4.833065
+INFO 2020-11-24 03:57:02 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 4.880998
+INFO 2020-11-24 03:58:19 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 4.904290
+INFO 2020-11-24 03:59:36 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 4.932658
+INFO 2020-11-24 04:00:52 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 4.924621
+INFO 2020-11-24 04:02:09 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 4.937399
+INFO 2020-11-24 04:03:26 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 4.930382
+INFO 2020-11-24 04:04:43 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 4.950013
+INFO 2020-11-24 04:06:00 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 4.974567
+INFO 2020-11-24 04:07:16 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 4.949161
+INFO 2020-11-24 04:08:33 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-24 04:08:33 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 4.946550
+INFO 2020-11-24 04:09:51 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 4.970594
+INFO 2020-11-24 04:11:08 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 4.954181
+INFO 2020-11-24 04:12:26 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 4.954727
+INFO 2020-11-24 04:13:43 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 4.970495
+INFO 2020-11-24 04:15:01 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 4.983707
+INFO 2020-11-24 04:16:18 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 4.947740
+INFO 2020-11-24 04:17:36 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 4.954028
+INFO 2020-11-24 04:18:53 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 4.971984
+INFO 2020-11-24 04:20:10 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 4.952608
+INFO 2020-11-24 04:21:28 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 4.951163
+INFO 2020-11-24 04:22:45 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 4.954459
+INFO 2020-11-24 04:24:03 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 4.940857
+INFO 2020-11-24 04:25:20 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 4.955636
+INFO 2020-11-24 04:26:38 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 4.956987
+INFO 2020-11-24 04:27:55 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-24 04:27:55 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 4.933979
+INFO 2020-11-24 04:29:13 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 4.913991
+INFO 2020-11-24 04:30:30 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 4.945502
+INFO 2020-11-24 04:30:36 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-24 04:30:38 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 5.000010
+INFO 2020-11-24 04:31:56 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 4.154199
+INFO 2020-11-24 04:33:13 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 3.994031
+INFO 2020-11-24 04:34:31 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 3.942594
+INFO 2020-11-24 04:35:48 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 3.885192
+INFO 2020-11-24 04:37:05 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 3.881911
+INFO 2020-11-24 04:38:23 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 3.834896
+INFO 2020-11-24 04:39:40 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 3.811167
+INFO 2020-11-24 04:40:57 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 3.790355
+INFO 2020-11-24 04:42:15 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 3.763783
+INFO 2020-11-24 04:43:32 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 3.754471
+INFO 2020-11-24 04:44:49 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 3.728959
+INFO 2020-11-24 04:46:06 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 3.721313
+INFO 2020-11-24 04:47:24 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 3.709909
+INFO 2020-11-24 04:48:41 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 3.693327
+INFO 2020-11-24 04:49:58 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-24 04:49:59 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 3.675561
+INFO 2020-11-24 04:51:15 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 3.657077
+INFO 2020-11-24 04:52:32 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 3.656189
+INFO 2020-11-24 04:53:48 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 3.654700
+INFO 2020-11-24 04:55:05 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 3.630731
+INFO 2020-11-24 04:56:22 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 3.592582
+INFO 2020-11-24 04:57:38 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 3.602946
+INFO 2020-11-24 04:58:55 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 3.600951
+INFO 2020-11-24 05:00:12 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 3.584942
+INFO 2020-11-24 05:01:28 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 3.583431
+INFO 2020-11-24 05:02:45 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 3.579450
+INFO 2020-11-24 05:04:01 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 3.566372
+INFO 2020-11-24 05:05:18 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 3.572648
+INFO 2020-11-24 05:06:35 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 3.566118
+INFO 2020-11-24 05:07:51 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 3.542246
+INFO 2020-11-24 05:09:08 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-24 05:09:08 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 3.563236
+INFO 2020-11-24 05:10:25 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 3.517665
+INFO 2020-11-24 05:11:43 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 3.551301
+INFO 2020-11-24 05:11:49 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-24 05:11:51 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 3.510715
+INFO 2020-11-24 05:13:08 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 3.330990
+INFO 2020-11-24 05:14:26 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 3.326119
+INFO 2020-11-24 05:15:43 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 3.319666
+INFO 2020-11-24 05:17:01 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 3.327858
+INFO 2020-11-24 05:18:18 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 3.348921
+INFO 2020-11-24 05:19:35 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 3.330388
+INFO 2020-11-24 05:20:53 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 3.327977
+INFO 2020-11-24 05:22:10 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 3.333608
+INFO 2020-11-24 05:23:28 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 3.334153
+INFO 2020-11-24 05:24:45 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 3.335261
+INFO 2020-11-24 05:26:02 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 3.324370
+INFO 2020-11-24 05:27:20 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 3.340342
+INFO 2020-11-24 05:28:37 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 3.356664
+INFO 2020-11-24 05:29:55 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 3.353062
+INFO 2020-11-24 05:31:12 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-24 05:31:12 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 3.361141
+INFO 2020-11-24 05:32:30 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 3.366452
+INFO 2020-11-24 05:33:47 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 3.348350
+INFO 2020-11-24 05:35:05 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 3.331543
+INFO 2020-11-24 05:36:22 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 3.354757
+INFO 2020-11-24 05:37:39 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 3.362968
+INFO 2020-11-24 05:38:57 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 3.367949
+INFO 2020-11-24 05:40:14 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 3.366881
+INFO 2020-11-24 05:41:32 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 3.369319
+INFO 2020-11-24 05:42:49 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 3.361776
+INFO 2020-11-24 05:44:07 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 3.360669
+INFO 2020-11-24 05:45:24 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 3.372847
+INFO 2020-11-24 05:46:42 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 3.363194
+INFO 2020-11-24 05:47:59 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 3.371010
+INFO 2020-11-24 05:49:17 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 3.373622
+INFO 2020-11-24 05:50:34 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-24 05:50:34 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 3.369140
+INFO 2020-11-24 05:51:52 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 3.387282
+INFO 2020-11-24 05:53:09 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 3.385287
+INFO 2020-11-24 05:53:15 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-24 05:53:17 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 3.363272
+INFO 2020-11-24 05:54:35 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 3.170186
+INFO 2020-11-24 05:55:52 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 3.149645
+INFO 2020-11-24 05:57:09 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 3.169130
+INFO 2020-11-24 05:58:27 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 3.189962
+INFO 2020-11-24 05:59:44 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 3.182983
+INFO 2020-11-24 06:01:01 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 3.203905
+INFO 2020-11-24 06:02:18 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 3.212434
+INFO 2020-11-24 06:03:36 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 3.239437
+INFO 2020-11-24 06:04:53 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 3.241117
+INFO 2020-11-24 06:06:11 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 3.239092
+INFO 2020-11-24 06:07:28 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 3.237745
+INFO 2020-11-24 06:08:45 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 3.234794
+INFO 2020-11-24 06:10:03 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 3.253999
+INFO 2020-11-24 06:11:20 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 3.280901
+INFO 2020-11-24 06:12:37 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-24 06:12:37 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 3.277021
+INFO 2020-11-24 06:13:54 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 3.273918
+INFO 2020-11-24 06:15:11 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 3.291123
+INFO 2020-11-24 06:16:27 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 3.281846
+INFO 2020-11-24 06:17:44 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 3.287990
+INFO 2020-11-24 06:19:01 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 3.294570
+INFO 2020-11-24 06:20:17 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 3.292180
+INFO 2020-11-24 06:21:34 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 3.297857
+INFO 2020-11-24 06:22:50 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 3.312679
+INFO 2020-11-24 06:24:07 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 3.305384
+INFO 2020-11-24 06:25:24 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 3.311811
+INFO 2020-11-24 06:26:41 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 3.312364
+INFO 2020-11-24 06:27:57 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 3.294487
+INFO 2020-11-24 06:29:14 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 3.306579
+INFO 2020-11-24 06:30:31 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 3.326816
+INFO 2020-11-24 06:31:47 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-24 06:31:48 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 3.315303
+INFO 2020-11-24 06:33:05 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 3.322655
+INFO 2020-11-24 06:34:22 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 3.328963
+INFO 2020-11-24 06:34:28 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-24 06:34:30 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 3.367159
+INFO 2020-11-24 06:35:48 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 3.064553
+INFO 2020-11-24 06:37:06 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 3.060716
+INFO 2020-11-24 06:38:23 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 3.021528
+INFO 2020-11-24 06:39:41 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 3.034977
+INFO 2020-11-24 06:40:58 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 3.023835
+INFO 2020-11-24 06:42:15 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 3.025058
+INFO 2020-11-24 06:43:33 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 3.031233
+INFO 2020-11-24 06:44:50 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 3.028460
+INFO 2020-11-24 06:46:08 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 3.014946
+INFO 2020-11-24 06:47:25 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 3.026953
+INFO 2020-11-24 06:48:43 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 3.026604
+INFO 2020-11-24 06:50:00 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 3.022412
+INFO 2020-11-24 06:51:17 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 3.025863
+INFO 2020-11-24 06:52:35 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 3.029603
+INFO 2020-11-24 06:53:52 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-24 06:53:52 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 3.039668
+INFO 2020-11-24 06:55:10 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 3.023970
+INFO 2020-11-24 06:56:27 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 3.030394
+INFO 2020-11-24 06:57:45 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 3.027165
+INFO 2020-11-24 06:59:02 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 3.026386
+INFO 2020-11-24 07:00:20 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 3.035696
+INFO 2020-11-24 07:01:37 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 3.025335
+INFO 2020-11-24 07:02:54 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 3.024760
+INFO 2020-11-24 07:04:12 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 3.038933
+INFO 2020-11-24 07:05:29 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 3.036351
+INFO 2020-11-24 07:06:47 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 3.038564
+INFO 2020-11-24 07:08:04 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 3.029024
+INFO 2020-11-24 07:09:21 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 3.036543
+INFO 2020-11-24 07:10:39 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 3.045363
+INFO 2020-11-24 07:11:56 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 3.026146
+INFO 2020-11-24 07:13:13 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-24 07:13:14 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 3.029341
+INFO 2020-11-24 07:14:31 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 3.038404
+INFO 2020-11-24 07:15:47 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 3.029541
+INFO 2020-11-24 07:15:54 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-24 07:15:55 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 2.986026
+INFO 2020-11-24 07:17:13 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 3.009964
+INFO 2020-11-24 07:18:31 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 2.985986
+INFO 2020-11-24 07:19:48 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 2.993668
+INFO 2020-11-24 07:21:06 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 2.995230
+INFO 2020-11-24 07:22:23 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 2.999048
+INFO 2020-11-24 07:23:41 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 3.011039
+INFO 2020-11-24 07:24:58 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 3.017092
+INFO 2020-11-24 07:26:16 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 3.007616
+INFO 2020-11-24 07:27:33 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 3.014714
+INFO 2020-11-24 07:28:51 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 3.003980
+INFO 2020-11-24 07:30:08 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 3.005253
+INFO 2020-11-24 07:31:26 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 3.003419
+INFO 2020-11-24 07:32:43 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 3.008633
+INFO 2020-11-24 07:34:00 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 3.009005
+INFO 2020-11-24 07:35:18 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-24 07:35:18 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 3.006640
+INFO 2020-11-24 07:36:35 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 3.017307
+INFO 2020-11-24 07:37:53 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 3.020918
+INFO 2020-11-24 07:39:10 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 3.016668
+INFO 2020-11-24 07:40:28 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 3.007162
+INFO 2020-11-24 07:41:45 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 3.013031
+INFO 2020-11-24 07:43:03 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 3.009712
+INFO 2020-11-24 07:44:20 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 3.026020
+INFO 2020-11-24 07:45:37 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 3.015756
+INFO 2020-11-24 07:46:55 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 3.014330
+INFO 2020-11-24 07:48:12 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 3.013047
+INFO 2020-11-24 07:49:30 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 3.021250
+INFO 2020-11-24 07:50:47 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 3.026684
+INFO 2020-11-24 07:52:05 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 3.020338
+INFO 2020-11-24 07:53:22 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 3.014258
+INFO 2020-11-24 07:54:39 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-24 07:54:40 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 3.006616
+INFO 2020-11-24 07:55:57 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 3.015490
+INFO 2020-11-24 07:57:15 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 3.024536
+INFO 2020-11-24 07:57:21 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-24 07:57:23 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 3.038901
+INFO 2020-11-24 07:58:40 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 2.979035
+INFO 2020-11-24 07:59:58 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 2.989299
+INFO 2020-11-24 08:01:15 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 2.998241
+INFO 2020-11-24 08:02:33 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 3.000091
+INFO 2020-11-24 08:03:50 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 2.992542
+INFO 2020-11-24 08:05:08 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 2.996305
+INFO 2020-11-24 08:06:25 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 3.011243
+INFO 2020-11-24 08:07:43 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 3.005577
+INFO 2020-11-24 08:09:00 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 2.989572
+INFO 2020-11-24 08:10:18 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 2.990408
+INFO 2020-11-24 08:11:35 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 3.002582
+INFO 2020-11-24 08:12:52 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 2.996218
+INFO 2020-11-24 08:14:10 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 2.996259
+INFO 2020-11-24 08:15:27 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 3.000973
+INFO 2020-11-24 08:16:45 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-24 08:16:45 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 3.002935
+INFO 2020-11-24 08:18:02 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 3.001099
+INFO 2020-11-24 08:19:20 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 2.985440
+INFO 2020-11-24 08:20:37 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 2.999892
+INFO 2020-11-24 08:21:55 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 2.994391
+INFO 2020-11-24 08:23:12 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 2.984912
+INFO 2020-11-24 08:24:30 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 2.994332
+INFO 2020-11-24 08:25:47 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 3.008122
+INFO 2020-11-24 08:27:05 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 3.003629
+INFO 2020-11-24 08:28:22 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 3.022021
+INFO 2020-11-24 08:29:39 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 3.009543
+INFO 2020-11-24 08:30:57 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 3.012039
+INFO 2020-11-24 08:32:14 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 3.017081
+INFO 2020-11-24 08:33:32 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 3.017457
+INFO 2020-11-24 08:34:49 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 3.001239
+INFO 2020-11-24 08:36:06 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-24 08:36:07 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 3.014142
+INFO 2020-11-24 08:37:24 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 3.007254
+INFO 2020-11-24 08:38:42 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 3.005258
+INFO 2020-11-24 08:38:48 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-24 08:38:49 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 3.071835
+INFO 2020-11-24 08:40:06 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 2.978621
+INFO 2020-11-24 08:41:23 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 2.997738
+INFO 2020-11-24 08:42:40 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 2.955296
+INFO 2020-11-24 08:43:57 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 2.974770
+INFO 2020-11-24 08:45:13 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 2.944999
+INFO 2020-11-24 08:46:30 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 2.971827
+INFO 2020-11-24 08:47:47 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 2.978880
+INFO 2020-11-24 08:49:03 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 2.963302
+INFO 2020-11-24 08:50:20 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 2.972103
+INFO 2020-11-24 08:51:36 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 2.990141
+INFO 2020-11-24 08:52:53 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 2.982921
+INFO 2020-11-24 08:54:10 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 2.977255
+INFO 2020-11-24 08:55:26 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 2.974266
+INFO 2020-11-24 08:56:43 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 2.966345
+INFO 2020-11-24 08:57:59 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-24 08:57:59 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 2.965771
+INFO 2020-11-24 08:59:17 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 2.973811
+INFO 2020-11-24 09:00:34 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 2.973932
+INFO 2020-11-24 09:01:51 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 2.981743
+INFO 2020-11-24 09:03:09 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 2.983263
+INFO 2020-11-24 09:04:26 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 2.982714
+INFO 2020-11-24 09:05:43 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 2.976597
+INFO 2020-11-24 09:07:01 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 2.969271
+INFO 2020-11-24 09:08:18 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 2.957103
+INFO 2020-11-24 09:09:35 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 2.956259
+INFO 2020-11-24 09:10:53 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 2.972634
+INFO 2020-11-24 09:12:10 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 2.956563
+INFO 2020-11-24 09:13:27 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 2.967340
+INFO 2020-11-24 09:14:45 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 2.961423
+INFO 2020-11-24 09:16:02 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 2.975215
+INFO 2020-11-24 09:17:19 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-24 09:17:20 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 2.964700
+INFO 2020-11-24 09:18:37 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 2.969290
+INFO 2020-11-24 09:19:54 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 2.965047
+INFO 2020-11-24 09:20:00 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-24 09:20:02 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 2.942841
+INFO 2020-11-24 09:21:20 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 2.968284
+INFO 2020-11-24 09:22:37 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 2.964004
+INFO 2020-11-24 09:23:55 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 2.981196
+INFO 2020-11-24 09:25:12 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 2.958751
+INFO 2020-11-24 09:26:29 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 2.977153
+INFO 2020-11-24 09:27:47 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 2.969209
+INFO 2020-11-24 09:29:04 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 2.963334
+INFO 2020-11-24 09:30:21 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 2.967177
+INFO 2020-11-24 09:31:38 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 2.970513
+INFO 2020-11-24 09:32:56 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 2.973789
+INFO 2020-11-24 09:34:13 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 2.978365
+INFO 2020-11-24 09:35:30 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 2.954196
+INFO 2020-11-24 09:36:48 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 2.978353
+INFO 2020-11-24 09:38:05 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 2.957173
+INFO 2020-11-24 09:39:22 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-24 09:39:22 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 2.982543
+INFO 2020-11-24 09:40:39 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 2.975319
+INFO 2020-11-24 09:41:55 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 2.963826
+INFO 2020-11-24 09:43:12 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 2.960883
+INFO 2020-11-24 09:44:28 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 2.975252
+INFO 2020-11-24 09:45:45 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 2.957829
+INFO 2020-11-24 09:47:01 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 2.965675
+INFO 2020-11-24 09:48:18 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 2.966600
+INFO 2020-11-24 09:49:35 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 2.956358
+INFO 2020-11-24 09:50:51 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 2.961049
+INFO 2020-11-24 09:52:08 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 2.970357
+INFO 2020-11-24 09:53:24 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 2.961374
+INFO 2020-11-24 09:54:41 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 2.957517
+INFO 2020-11-24 09:55:57 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 2.977499
+INFO 2020-11-24 09:57:14 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 2.985792
+INFO 2020-11-24 09:58:30 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-24 09:58:30 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 2.972527
+INFO 2020-11-24 09:59:48 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 2.975728
+INFO 2020-11-24 10:01:05 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 2.985188
+INFO 2020-11-24 10:01:11 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-24 10:01:11 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/.gitkeep b/bob/bio/facexzoo/models/heads/AdaCos/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b5440c3ceedce3b2bb0687309cb1cfe03e9501bb
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13.pt | 0.9538333333333334 | 0.004150338376804397 |
+| Epoch_14_batch_5999.pt | 0.9533333333333334 | 0.0039984564923214675 |
+| Epoch_15_batch_5999.pt | 0.953 | 0.004148478822798775 |
+| Epoch_14_batch_2999.pt | 0.9526666666666668 | 0.004232443767720207 |
+| Epoch_14.pt | 0.9526666666666668 | 0.0039299420408505275 |
+| Epoch_13_batch_5999.pt | 0.9526666666666666 | 0.0044804100341457685 |
+| Epoch_17.pt | 0.9525 | 0.0043336894440569795 |
+| Epoch_15_batch_2999.pt | 0.9523333333333334 | 0.004000000000000002 |
+| Epoch_15.pt | 0.9518333333333333 | 0.004495882890263381 |
+| Epoch_17_batch_2999.pt | 0.9515 | 0.004009633461281368 |
+| Epoch_11_batch_2999.pt | 0.9508333333333333 | 0.004369153942451503 |
+| Epoch_11_batch_5999.pt | 0.9508333333333333 | 0.004397319610067753 |
+| Epoch_16_batch_2999.pt | 0.9506666666666668 | 0.004562325592810622 |
+| Epoch_16_batch_5999.pt | 0.9505000000000001 | 0.004267664813801915 |
+| Epoch_13_batch_2999.pt | 0.9503333333333334 | 0.00430904876214785 |
+| Epoch_12_batch_2999.pt | 0.95 | 0.0048112522432468855 |
+| Epoch_11.pt | 0.9495000000000001 | 0.003944444444444442 |
+| Epoch_17_batch_5999.pt | 0.9491666666666667 | 0.004181455237263049 |
+| Epoch_12_batch_5999.pt | 0.9490000000000001 | 0.004247003300803639 |
+| Epoch_16.pt | 0.9490000000000001 | 0.004541985207993271 |
+| Epoch_10.pt | 0.9486666666666667 | 0.004911563578955634 |
+| Epoch_12.pt | 0.9486666666666667 | 0.004449996532111289 |
+| Epoch_10_batch_2999.pt | 0.9463333333333332 | 0.004365973934308557 |
+| Epoch_10_batch_5999.pt | 0.9445 | 0.0048499077247172034 |
+| Epoch_7_batch_5999.pt | 0.9395 | 0.004346489998681915 |
+| Epoch_8_batch_5999.pt | 0.9381666666666668 | 0.0053544143551297685 |
+| Epoch_5_batch_5999.pt | 0.9348333333333334 | 0.0056625802259053955 |
+| Epoch_7_batch_2999.pt | 0.9343333333333333 | 0.00590668171555645 |
+| Epoch_9_batch_2999.pt | 0.9339999999999999 | 0.004825344611246322 |
+| Epoch_6_batch_2999.pt | 0.9333333333333333 | 0.005810803051107949 |
+| Epoch_6_batch_5999.pt | 0.933 | 0.005504207145111479 |
+| Epoch_9_batch_5999.pt | 0.9328333333333333 | 0.005477507318092153 |
+| Epoch_5_batch_2999.pt | 0.9309999999999998 | 0.007154382236446775 |
+| Epoch_8.pt | 0.9291666666666668 | 0.005639641438562871 |
+| Epoch_8_batch_2999.pt | 0.9288333333333334 | 0.006885527853923806 |
+| Epoch_9.pt | 0.9284999999999999 | 0.00554137078018218 |
+| Epoch_4_batch_5999.pt | 0.9278333333333334 | 0.006080986355989331 |
+| Epoch_5.pt | 0.9253333333333333 | 0.005216308708580228 |
+| Epoch_3_batch_2999.pt | 0.9245000000000001 | 0.005665849614298976 |
+| Epoch_6.pt | 0.9243333333333332 | 0.006498812807062113 |
+| Epoch_3_batch_5999.pt | 0.9236666666666666 | 0.006977919319158924 |
+| Epoch_4.pt | 0.9231666666666667 | 0.006976371054238705 |
+| Epoch_7.pt | 0.9219999999999999 | 0.006028737762778017 |
+| Epoch_4_batch_2999.pt | 0.9208333333333334 | 0.0057103436579034345 |
+| Epoch_2_batch_5999.pt | 0.9115000000000002 | 0.005823801736552052 |
+| Epoch_3.pt | 0.9114999999999999 | 0.005057996968497834 |
+| Epoch_2_batch_2999.pt | 0.9083333333333332 | 0.0056108360768678135 |
+| Epoch_2.pt | 0.9039999999999999 | 0.004882571676159962 |
+| Epoch_1_batch_5999.pt | 0.9026666666666667 | 0.006868248649122486 |
+| Epoch_1.pt | 0.8986666666666666 | 0.0077650689820694 |
+| Epoch_1_batch_2999.pt | 0.8880000000000001 | 0.007792842553377317 |
+| Epoch_0_batch_5999.pt | 0.8755000000000001 | 0.005784451274166802 |
+| Epoch_0.pt | 0.8706666666666667 | 0.007597920443235 |
+| Epoch_0_batch_2999.pt | 0.8283333333333334 | 0.008734775114237137 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f584ab4635f8c27476278bc588e0214124701df5
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.9263333333333332 | 0.003996912388578883 |
+| Epoch_15_batch_2999.pt | 0.9260000000000002 | 0.0037035185138886567 |
+| Epoch_17_batch_2999.pt | 0.9260000000000002 | 0.00365317382728302 |
+| Epoch_17.pt | 0.9258333333333333 | 0.0041144881025079 |
+| Epoch_14_batch_2999.pt | 0.9256666666666666 | 0.003576328208762459 |
+| Epoch_13_batch_2999.pt | 0.925 | 0.003759760959041922 |
+| Epoch_15_batch_5999.pt | 0.9246666666666666 | 0.003632840605393737 |
+| Epoch_12_batch_5999.pt | 0.9245000000000001 | 0.0032871804872193376 |
+| Epoch_13.pt | 0.9244999999999999 | 0.0036094013046179554 |
+| Epoch_16_batch_2999.pt | 0.9244999999999999 | 0.0038171963080670876 |
+| Epoch_16.pt | 0.9243333333333335 | 0.003627739492736561 |
+| Epoch_12.pt | 0.9243333333333332 | 0.004290385461401743 |
+| Epoch_12_batch_2999.pt | 0.924 | 0.0035329140212410327 |
+| Epoch_13_batch_5999.pt | 0.924 | 0.0033444259873983166 |
+| Epoch_11_batch_5999.pt | 0.9236666666666666 | 0.003237511618740778 |
+| Epoch_16_batch_5999.pt | 0.9236666666666666 | 0.00382325567424117 |
+| Epoch_17_batch_5999.pt | 0.9236666666666666 | 0.003839367231815782 |
+| Epoch_15.pt | 0.9233333333333332 | 0.004013864859597432 |
+| Epoch_14.pt | 0.9228333333333332 | 0.0037189039935954606 |
+| Epoch_11_batch_2999.pt | 0.9221666666666666 | 0.003397984151860673 |
+| Epoch_10_batch_5999.pt | 0.9208333333333334 | 0.004225510415512539 |
+| Epoch_10.pt | 0.9200000000000002 | 0.003951870943061942 |
+| Epoch_10_batch_2999.pt | 0.9198333333333334 | 0.004108482643218736 |
+| Epoch_11.pt | 0.9189999999999999 | 0.0038425814368423556 |
+| Epoch_9_batch_5999.pt | 0.9111666666666668 | 0.004120484809064149 |
+| Epoch_8_batch_2999.pt | 0.9073333333333334 | 0.005230489909306583 |
+| Epoch_8_batch_5999.pt | 0.9071666666666666 | 0.0047534912445849386 |
+| Epoch_7_batch_2999.pt | 0.9063333333333332 | 0.004546060565661948 |
+| Epoch_5_batch_5999.pt | 0.9059999999999999 | 0.00446661138716484 |
+| Epoch_7_batch_5999.pt | 0.9055 | 0.004458657828274071 |
+| Epoch_6_batch_5999.pt | 0.905 | 0.0037101795237919977 |
+| Epoch_6_batch_2999.pt | 0.9046666666666667 | 0.004042978977480056 |
+| Epoch_9.pt | 0.9046666666666667 | 0.0034498165991688956 |
+| Epoch_9_batch_2999.pt | 0.9045 | 0.004956913117487633 |
+| Epoch_5_batch_2999.pt | 0.9023333333333333 | 0.005433094745954514 |
+| Epoch_4_batch_5999.pt | 0.9008333333333335 | 0.004649771001927981 |
+| Epoch_3_batch_5999.pt | 0.8995000000000001 | 0.004925682255181747 |
+| Epoch_5.pt | 0.8969999999999999 | 0.0037416573867739404 |
+| Epoch_8.pt | 0.8969999999999999 | 0.006406015691288698 |
+| Epoch_4_batch_2999.pt | 0.8968333333333334 | 0.005406043714397369 |
+| Epoch_6.pt | 0.8953333333333333 | 0.003950308629918039 |
+| Epoch_3_batch_2999.pt | 0.8936666666666666 | 0.005565546571719197 |
+| Epoch_2_batch_5999.pt | 0.8933333333333333 | 0.0037597609590419175 |
+| Epoch_4.pt | 0.8933333333333333 | 0.005649210586171628 |
+| Epoch_7.pt | 0.893 | 0.004758358467858628 |
+| Epoch_3.pt | 0.8925000000000001 | 0.005938210646231095 |
+| Epoch_2_batch_2999.pt | 0.8898333333333334 | 0.005440191082337609 |
+| Epoch_2.pt | 0.8876666666666667 | 0.004555555555555557 |
+| Epoch_1_batch_5999.pt | 0.8831666666666667 | 0.006253147355680327 |
+| Epoch_1.pt | 0.8746666666666668 | 0.005734883511361747 |
+| Epoch_1_batch_2999.pt | 0.8720000000000001 | 0.0060030856263289314 |
+| Epoch_0.pt | 0.8508333333333333 | 0.005335936864527369 |
+| Epoch_0_batch_5999.pt | 0.8506666666666666 | 0.0035935470286213786 |
+| Epoch_0_batch_2999.pt | 0.8084999999999999 | 0.004270556675772802 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e0695ae5ce77f43508c003b5c5b28ebb0f56951d
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.8326666666666667 | 0.0064415702402825506 |
+| Epoch_16_batch_2999.pt | 0.8320000000000001 | 0.005899361756648129 |
+| Epoch_13.pt | 0.8318333333333333 | 0.005706018048982029 |
+| Epoch_15.pt | 0.8318333333333333 | 0.005996140734147564 |
+| Epoch_17_batch_2999.pt | 0.8318333333333333 | 0.005574689273298304 |
+| Epoch_16_batch_5999.pt | 0.8315000000000001 | 0.006138574652590948 |
+| Epoch_13_batch_2999.pt | 0.8313333333333335 | 0.006343072433863145 |
+| Epoch_17.pt | 0.8313333333333333 | 0.006089862047130984 |
+| Epoch_14.pt | 0.8308333333333333 | 0.006598026266792507 |
+| Epoch_15_batch_2999.pt | 0.8305 | 0.006285639128939811 |
+| Epoch_16.pt | 0.8301666666666666 | 0.006552023530610906 |
+| Epoch_17_batch_5999.pt | 0.8299999999999998 | 0.006531027559860261 |
+| Epoch_15_batch_5999.pt | 0.8298333333333332 | 0.006389854999417455 |
+| Epoch_14_batch_5999.pt | 0.8296666666666667 | 0.0065158875161758925 |
+| Epoch_12.pt | 0.8283333333333334 | 0.005577733510227171 |
+| Epoch_14_batch_2999.pt | 0.828 | 0.006463573143221773 |
+| Epoch_12_batch_5999.pt | 0.8278333333333332 | 0.005488765180792168 |
+| Epoch_11.pt | 0.8273333333333334 | 0.007132779486979566 |
+| Epoch_11_batch_5999.pt | 0.8263333333333331 | 0.007004407783322001 |
+| Epoch_11_batch_2999.pt | 0.8261666666666665 | 0.0056549443022658 |
+| Epoch_12_batch_2999.pt | 0.8248333333333333 | 0.006887320610018472 |
+| Epoch_10.pt | 0.818 | 0.007481664830903302 |
+| Epoch_10_batch_2999.pt | 0.8166666666666667 | 0.007286042804779995 |
+| Epoch_10_batch_5999.pt | 0.8160000000000001 | 0.006517781944916645 |
+| Epoch_9_batch_2999.pt | 0.7981666666666667 | 0.007956090919292553 |
+| Epoch_9_batch_5999.pt | 0.7953333333333333 | 0.006623937138128389 |
+| Epoch_8_batch_5999.pt | 0.7941666666666667 | 0.009725396307263436 |
+| Epoch_5_batch_5999.pt | 0.7929999999999999 | 0.00676227735764575 |
+| Epoch_7_batch_2999.pt | 0.791 | 0.008517230176327465 |
+| Epoch_6_batch_2999.pt | 0.7908333333333333 | 0.008725052619318491 |
+| Epoch_5_batch_2999.pt | 0.7906666666666667 | 0.008621678104251705 |
+| Epoch_7_batch_5999.pt | 0.79 | 0.008307366952230553 |
+| Epoch_8_batch_2999.pt | 0.7898333333333333 | 0.008067806775874067 |
+| Epoch_6_batch_5999.pt | 0.7863333333333333 | 0.007139699478407707 |
+| Epoch_8.pt | 0.7853333333333333 | 0.0069557685448397915 |
+| Epoch_4.pt | 0.7846666666666666 | 0.0093847788505683 |
+| Epoch_9.pt | 0.7843333333333333 | 0.008047697316547473 |
+| Epoch_7.pt | 0.7825 | 0.007581857813308987 |
+| Epoch_4_batch_2999.pt | 0.7818333333333333 | 0.008806738441404319 |
+| Epoch_3_batch_5999.pt | 0.7816666666666667 | 0.00757269298283944 |
+| Epoch_4_batch_5999.pt | 0.7786666666666668 | 0.008141495600765505 |
+| Epoch_3_batch_2999.pt | 0.7741666666666667 | 0.008289700585815457 |
+| Epoch_5.pt | 0.774 | 0.007958999875131658 |
+| Epoch_6.pt | 0.773 | 0.008254441378381447 |
+| Epoch_2_batch_2999.pt | 0.7718333333333333 | 0.007783529619223744 |
+| Epoch_2_batch_5999.pt | 0.7701666666666667 | 0.0066223061143082795 |
+| Epoch_3.pt | 0.7689999999999999 | 0.008455402247113612 |
+| Epoch_2.pt | 0.7568333333333334 | 0.0082643440584265 |
+| Epoch_1_batch_5999.pt | 0.7521666666666667 | 0.009118729876901785 |
+| Epoch_1.pt | 0.7428333333333332 | 0.011125130045016038 |
+| Epoch_1_batch_2999.pt | 0.7394999999999999 | 0.008850786391789512 |
+| Epoch_0_batch_5999.pt | 0.7223333333333334 | 0.010089720964219106 |
+| Epoch_0.pt | 0.717 | 0.00905333987890453 |
+| Epoch_0_batch_2999.pt | 0.6669999999999999 | 0.01059291633796897 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..4e05b59ec26d8f44e65c2a6908423cee0a75cb10
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.9964999999999999 | 0.0009111788592698183 |
+| Epoch_17_batch_2999.pt | 0.9964999999999999 | 0.0007637626158259767 |
+| Epoch_15.pt | 0.9963333333333333 | 0.0008888888888888872 |
+| Epoch_11_batch_5999.pt | 0.9959999999999999 | 0.0009686442096757034 |
+| Epoch_11.pt | 0.9959999999999999 | 0.0010304020550550796 |
+| Epoch_13_batch_2999.pt | 0.9959999999999999 | 0.0009362388636862651 |
+| Epoch_13.pt | 0.9959999999999999 | 0.0008678055195451862 |
+| Epoch_14_batch_2999.pt | 0.9959999999999999 | 0.0009362388636862651 |
+| Epoch_16_batch_5999.pt | 0.9959999999999999 | 0.0008314794192830989 |
+| Epoch_17.pt | 0.9959999999999999 | 0.0008314794192830989 |
+| Epoch_10.pt | 0.9958333333333333 | 0.0011180339887498947 |
+| Epoch_15_batch_2999.pt | 0.9958333333333332 | 0.0009043789220055372 |
+| Epoch_16_batch_2999.pt | 0.9958333333333332 | 0.0009378857231185658 |
+| Epoch_17_batch_5999.pt | 0.9958333333333332 | 0.0009378857231185658 |
+| Epoch_11_batch_2999.pt | 0.9956666666666667 | 0.0011166528467912152 |
+| Epoch_10_batch_5999.pt | 0.9956666666666665 | 0.0009999999999999992 |
+| Epoch_12_batch_2999.pt | 0.9955 | 0.0009953596037316115 |
+| Epoch_14_batch_5999.pt | 0.9955 | 0.0010258991840344119 |
+| Epoch_16.pt | 0.9955 | 0.0009638528651609716 |
+| Epoch_12_batch_5999.pt | 0.9953333333333333 | 0.0012619796324000608 |
+| Epoch_14.pt | 0.9953333333333333 | 0.0010772621905369615 |
+| Epoch_15_batch_5999.pt | 0.9953333333333333 | 0.0009558139185602923 |
+| Epoch_12.pt | 0.9953333333333332 | 0.0007777777777777757 |
+| Epoch_10_batch_2999.pt | 0.9951666666666666 | 0.0011506841765115542 |
+| Epoch_7_batch_5999.pt | 0.9941666666666666 | 0.0008695819912499136 |
+| Epoch_8_batch_2999.pt | 0.994 | 0.0011967032904743344 |
+| Epoch_8.pt | 0.994 | 0.001088662107903642 |
+| Epoch_9_batch_2999.pt | 0.9936666666666667 | 0.0011055415967851315 |
+| Epoch_7.pt | 0.9935 | 0.0010957268290731164 |
+| Epoch_9_batch_5999.pt | 0.9933333333333334 | 0.0010243938285881 |
+| Epoch_8_batch_5999.pt | 0.9933333333333332 | 0.0012909944487358035 |
+| Epoch_7_batch_2999.pt | 0.9928333333333332 | 0.0014917468424552939 |
+| Epoch_4_batch_5999.pt | 0.9924999999999999 | 0.0017078251276599395 |
+| Epoch_9.pt | 0.9924999999999999 | 0.0014540280364780456 |
+| Epoch_6_batch_2999.pt | 0.9921666666666665 | 0.0018265869136415067 |
+| Epoch_6_batch_5999.pt | 0.992 | 0.001735611039090367 |
+| Epoch_5_batch_5999.pt | 0.9918333333333333 | 0.001177201116689842 |
+| Epoch_3_batch_5999.pt | 0.9916666666666666 | 0.0013146843962443628 |
+| Epoch_4_batch_2999.pt | 0.9916666666666666 | 0.00121716123890037 |
+| Epoch_4.pt | 0.9906666666666666 | 0.0015555555555555613 |
+| Epoch_5.pt | 0.9906666666666666 | 0.001688742683730079 |
+| Epoch_5_batch_2999.pt | 0.9903333333333333 | 0.0014010578014353957 |
+| Epoch_6.pt | 0.9901666666666668 | 0.001415304355872999 |
+| Epoch_3_batch_2999.pt | 0.99 | 0.0010243938285881 |
+| Epoch_2_batch_5999.pt | 0.9890000000000001 | 0.0015355861067872553 |
+| Epoch_3.pt | 0.9889999999999999 | 0.0020964402515681285 |
+| Epoch_2_batch_2999.pt | 0.9884999999999999 | 0.0012285191326386665 |
+| Epoch_1_batch_5999.pt | 0.9873333333333333 | 0.0015355861067872531 |
+| Epoch_2.pt | 0.9868333333333335 | 0.0016377114414426288 |
+| Epoch_1_batch_2999.pt | 0.9848333333333332 | 0.0018500917561496317 |
+| Epoch_1.pt | 0.9819999999999999 | 0.002148786622868184 |
+| Epoch_0_batch_5999.pt | 0.9783333333333333 | 0.003132860484415925 |
+| Epoch_0.pt | 0.9781666666666666 | 0.0019945914523351333 |
+| Epoch_0_batch_2999.pt | 0.9664999999999999 | 0.002539198862672541 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d0628b57094a636ae1ec9daa39c7b21ab5784221
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.8588333333333333 | 0.003920899997465307 |
+| Epoch_14_batch_5999.pt | 0.8575000000000002 | 0.004054033199812894 |
+| Epoch_17.pt | 0.8574999999999999 | 0.004487637339278754 |
+| Epoch_16_batch_2999.pt | 0.8573333333333334 | 0.004838120234906052 |
+| Epoch_15_batch_2999.pt | 0.8571666666666665 | 0.004187356041380279 |
+| Epoch_17_batch_5999.pt | 0.857 | 0.0042513614377131075 |
+| Epoch_14.pt | 0.8565000000000002 | 0.004475240527365853 |
+| Epoch_14_batch_2999.pt | 0.8561666666666667 | 0.004688112099953282 |
+| Epoch_13_batch_5999.pt | 0.8556666666666667 | 0.004521553322083507 |
+| Epoch_13_batch_2999.pt | 0.8554999999999999 | 0.0038972133127323535 |
+| Epoch_13.pt | 0.8553333333333333 | 0.0033591592128513264 |
+| Epoch_16_batch_5999.pt | 0.8550000000000001 | 0.004479032082388085 |
+| Epoch_15.pt | 0.8548333333333333 | 0.004570774038239256 |
+| Epoch_17_batch_2999.pt | 0.8548333333333333 | 0.004248819734444199 |
+| Epoch_12_batch_2999.pt | 0.8538333333333332 | 0.00487529676209751 |
+| Epoch_11_batch_2999.pt | 0.8536666666666669 | 0.004073400617738527 |
+| Epoch_12.pt | 0.853 | 0.005096597751635432 |
+| Epoch_15_batch_5999.pt | 0.853 | 0.004626813958590447 |
+| Epoch_11_batch_5999.pt | 0.8515 | 0.004756087716849886 |
+| Epoch_12_batch_5999.pt | 0.8515 | 0.004182931218684677 |
+| Epoch_10.pt | 0.8476666666666667 | 0.004480410034145771 |
+| Epoch_11.pt | 0.8465 | 0.004145874045956808 |
+| Epoch_10_batch_5999.pt | 0.843 | 0.004436103283928253 |
+| Epoch_10_batch_2999.pt | 0.842 | 0.004980207740225545 |
+| Epoch_8_batch_5999.pt | 0.8271666666666668 | 0.00434648999868192 |
+| Epoch_7_batch_5999.pt | 0.8244999999999999 | 0.005085988984515925 |
+| Epoch_5_batch_2999.pt | 0.8201666666666668 | 0.004852452606290584 |
+| Epoch_6_batch_5999.pt | 0.8178333333333333 | 0.0031529919163694907 |
+| Epoch_4_batch_5999.pt | 0.8170000000000002 | 0.004758358467858636 |
+| Epoch_8_batch_2999.pt | 0.8161666666666667 | 0.003735465660892049 |
+| Epoch_7_batch_2999.pt | 0.8153333333333335 | 0.004498284995554923 |
+| Epoch_5_batch_5999.pt | 0.8150000000000001 | 0.00406733449282736 |
+| Epoch_3_batch_5999.pt | 0.8086666666666668 | 0.004841946348777984 |
+| Epoch_6_batch_2999.pt | 0.808 | 0.0045119867787215395 |
+| Epoch_9_batch_5999.pt | 0.808 | 0.004351812450570249 |
+| Epoch_5.pt | 0.8078333333333333 | 0.004105476619298294 |
+| Epoch_4_batch_2999.pt | 0.8065000000000001 | 0.0059600004142845164 |
+| Epoch_9_batch_2999.pt | 0.8045 | 0.005795112883571235 |
+| Epoch_7.pt | 0.8013333333333333 | 0.005108695081143399 |
+| Epoch_4.pt | 0.8008333333333333 | 0.004995368225036444 |
+| Epoch_9.pt | 0.7955 | 0.008578072821638513 |
+| Epoch_6.pt | 0.7928333333333334 | 0.004628481339075796 |
+| Epoch_2_batch_5999.pt | 0.7916666666666667 | 0.006712803318663653 |
+| Epoch_3_batch_2999.pt | 0.7901666666666667 | 0.003482318654269964 |
+| Epoch_2.pt | 0.7890000000000001 | 0.005007401928552775 |
+| Epoch_8.pt | 0.7831666666666666 | 0.0046973199608728805 |
+| Epoch_3.pt | 0.7825 | 0.003298428357510534 |
+| Epoch_2_batch_2999.pt | 0.782 | 0.005992794026738575 |
+| Epoch_1_batch_5999.pt | 0.7638333333333334 | 0.0035228530805528177 |
+| Epoch_1_batch_2999.pt | 0.7575 | 0.005183366083021463 |
+| Epoch_1.pt | 0.7501666666666666 | 0.005602303318201882 |
+| Epoch_0_batch_5999.pt | 0.7393333333333333 | 0.003961231882216861 |
+| Epoch_0.pt | 0.7354999999999999 | 0.005931970297037756 |
+| Epoch_0_batch_2999.pt | 0.6875000000000001 | 0.006374379659965567 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6f78f613d70a035d2cad320bb39b62be17d7184b
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_2999.pt | 0.8550000000000001 | 0.003951870943061946 |
+| Epoch_13_batch_2999.pt | 0.8548333333333333 | 0.003916174121906022 |
+| Epoch_17.pt | 0.8545 | 0.004006553273800622 |
+| Epoch_13_batch_5999.pt | 0.8539999999999999 | 0.0035153982265680836 |
+| Epoch_16_batch_5999.pt | 0.8538333333333334 | 0.0041725883845834194 |
+| Epoch_15_batch_2999.pt | 0.8536666666666667 | 0.0037663225014840625 |
+| Epoch_17_batch_5999.pt | 0.8536666666666667 | 0.003981438414926418 |
+| Epoch_16_batch_2999.pt | 0.8533333333333333 | 0.004179609527512006 |
+| Epoch_15_batch_5999.pt | 0.8530000000000001 | 0.0038634085890474896 |
+| Epoch_13.pt | 0.853 | 0.003404789654676563 |
+| Epoch_14_batch_2999.pt | 0.8523333333333334 | 0.004 |
+| Epoch_12.pt | 0.852 | 0.004886362983843559 |
+| Epoch_15.pt | 0.8514999999999999 | 0.0038685978574563755 |
+| Epoch_16.pt | 0.8499999999999999 | 0.003857012212824395 |
+| Epoch_14_batch_5999.pt | 0.8496666666666666 | 0.004058218303944368 |
+| Epoch_14.pt | 0.8493333333333333 | 0.0038745768387028236 |
+| Epoch_10_batch_5999.pt | 0.849 | 0.0035676876351116286 |
+| Epoch_11_batch_2999.pt | 0.8481666666666665 | 0.0033742854753467115 |
+| Epoch_12_batch_2999.pt | 0.8479999999999999 | 0.0044430553384738275 |
+| Epoch_11_batch_5999.pt | 0.8473333333333333 | 0.004616128461742436 |
+| Epoch_12_batch_5999.pt | 0.8471666666666667 | 0.003352261075899021 |
+| Epoch_10.pt | 0.8458333333333334 | 0.004061639272540942 |
+| Epoch_10_batch_2999.pt | 0.8451666666666666 | 0.004190303327362905 |
+| Epoch_11.pt | 0.8418333333333333 | 0.004762572704477288 |
+| Epoch_8_batch_2999.pt | 0.8296666666666667 | 0.003973678831610594 |
+| Epoch_8_batch_5999.pt | 0.8185 | 0.005166666666666665 |
+| Epoch_7_batch_2999.pt | 0.8183333333333331 | 0.0059731911358736015 |
+| Epoch_6_batch_2999.pt | 0.8171666666666667 | 0.005140315117400181 |
+| Epoch_4_batch_5999.pt | 0.8166666666666667 | 0.004574486411164193 |
+| Epoch_9_batch_2999.pt | 0.8163333333333334 | 0.004653420353638204 |
+| Epoch_6_batch_5999.pt | 0.8146666666666667 | 0.004337604732431799 |
+| Epoch_7_batch_5999.pt | 0.8146666666666664 | 0.004745368112013914 |
+| Epoch_9_batch_5999.pt | 0.8143333333333335 | 0.005358735800794439 |
+| Epoch_5_batch_2999.pt | 0.8140000000000001 | 0.005329859980093982 |
+| Epoch_5_batch_5999.pt | 0.8135 | 0.003763453234319985 |
+| Epoch_4_batch_2999.pt | 0.8108333333333334 | 0.004225510415512543 |
+| Epoch_4.pt | 0.808 | 0.0048610317453838614 |
+| Epoch_8.pt | 0.8071666666666667 | 0.005049446858839772 |
+| Epoch_3_batch_2999.pt | 0.805 | 0.0042889464590264 |
+| Epoch_6.pt | 0.8026666666666668 | 0.004151453709393199 |
+| Epoch_9.pt | 0.8011666666666667 | 0.004694690986293993 |
+| Epoch_3_batch_5999.pt | 0.7996666666666666 | 0.003431876713662334 |
+| Epoch_3.pt | 0.7986666666666666 | 0.004854678283322997 |
+| Epoch_5.pt | 0.7961666666666667 | 0.0037601713908928316 |
+| Epoch_2_batch_5999.pt | 0.7956666666666665 | 0.002983493684910574 |
+| Epoch_2_batch_2999.pt | 0.7933333333333333 | 0.004707493369847746 |
+| Epoch_7.pt | 0.7905 | 0.0063003527238119335 |
+| Epoch_1_batch_5999.pt | 0.7826666666666666 | 0.004375859703892266 |
+| Epoch_1_batch_2999.pt | 0.7795 | 0.007320063751999254 |
+| Epoch_2.pt | 0.7761666666666667 | 0.0045068534916349046 |
+| Epoch_1.pt | 0.7683333333333334 | 0.007196535517649451 |
+| Epoch_0_batch_5999.pt | 0.7515 | 0.004055555555555557 |
+| Epoch_0.pt | 0.7414999999999999 | 0.00485245260629058 |
+| Epoch_0_batch_2999.pt | 0.7084999999999999 | 0.005813192743797784 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d9e0708a5c04af64708f039260164e86b94c4d79
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.9435 | 0.0036972629182166314 |
+| Epoch_17.pt | 0.9435 | 0.004182931218684679 |
+| Epoch_14_batch_2999.pt | 0.9421666666666667 | 0.004052510272185975 |
+| Epoch_16.pt | 0.9415000000000001 | 0.0037222222222222227 |
+| Epoch_16_batch_2999.pt | 0.9413333333333332 | 0.003766322501484064 |
+| Epoch_15.pt | 0.9410000000000001 | 0.003992276494049257 |
+| Epoch_17_batch_2999.pt | 0.9410000000000001 | 0.0041141130181274665 |
+| Epoch_13_batch_5999.pt | 0.9406666666666667 | 0.0038825344910347003 |
+| Epoch_11_batch_5999.pt | 0.9405000000000001 | 0.004506853491634906 |
+| Epoch_14_batch_5999.pt | 0.9403333333333335 | 0.0037989602216174953 |
+| Epoch_16_batch_5999.pt | 0.9403333333333335 | 0.0040353377327410804 |
+| Epoch_17_batch_5999.pt | 0.9400000000000001 | 0.00431763538393899 |
+| Epoch_12_batch_5999.pt | 0.9396666666666667 | 0.0036243347622889133 |
+| Epoch_15_batch_2999.pt | 0.9396666666666667 | 0.00422952584681651 |
+| Epoch_13_batch_2999.pt | 0.9395000000000001 | 0.0036349639562123495 |
+| Epoch_12_batch_2999.pt | 0.9386666666666666 | 0.003973678831610587 |
+| Epoch_14.pt | 0.9385000000000001 | 0.003908284963769233 |
+| Epoch_13.pt | 0.9376666666666666 | 0.004181086160494836 |
+| Epoch_10_batch_5999.pt | 0.9375 | 0.004076808846434961 |
+| Epoch_11_batch_2999.pt | 0.9371666666666666 | 0.003936611941790413 |
+| Epoch_12.pt | 0.9368333333333334 | 0.003931904950937892 |
+| Epoch_11.pt | 0.9365 | 0.0036805293500115363 |
+| Epoch_10_batch_2999.pt | 0.9346666666666665 | 0.0037168285963728237 |
+| Epoch_10.pt | 0.9338333333333333 | 0.003549039167485427 |
+| Epoch_7_batch_2999.pt | 0.9151666666666666 | 0.004070747800382754 |
+| Epoch_9_batch_2999.pt | 0.914 | 0.003506607519568777 |
+| Epoch_8_batch_5999.pt | 0.9131666666666666 | 0.0038526085439079595 |
+| Epoch_7_batch_5999.pt | 0.9129999999999999 | 0.0038393672318157777 |
+| Epoch_6_batch_5999.pt | 0.9126666666666667 | 0.0041440124886603395 |
+| Epoch_5_batch_5999.pt | 0.9116666666666667 | 0.004416579314300405 |
+| Epoch_9_batch_5999.pt | 0.9116666666666667 | 0.004021546904634193 |
+| Epoch_6_batch_2999.pt | 0.9106666666666667 | 0.004459696053419889 |
+| Epoch_8_batch_2999.pt | 0.9103333333333333 | 0.005257563628587098 |
+| Epoch_3_batch_5999.pt | 0.9085000000000001 | 0.005657127046456804 |
+| Epoch_5_batch_2999.pt | 0.9065 | 0.005076270138449934 |
+| Epoch_4_batch_2999.pt | 0.9049999999999999 | 0.004120110270608696 |
+| Epoch_3_batch_2999.pt | 0.9035 | 0.003713921091857391 |
+| Epoch_9.pt | 0.9021666666666667 | 0.006151633327222338 |
+| Epoch_4_batch_5999.pt | 0.9010000000000001 | 0.0043969686527576355 |
+| Epoch_4.pt | 0.901 | 0.005472715713788776 |
+| Epoch_6.pt | 0.8989999999999998 | 0.003929942040850528 |
+| Epoch_7.pt | 0.8988333333333334 | 0.005437921262239922 |
+| Epoch_2_batch_5999.pt | 0.8953333333333333 | 0.005035675197166519 |
+| Epoch_8.pt | 0.8935000000000001 | 0.00317639830198284 |
+| Epoch_5.pt | 0.8931666666666669 | 0.0053774219349672185 |
+| Epoch_2_batch_2999.pt | 0.8928333333333333 | 0.0040295972901752426 |
+| Epoch_3.pt | 0.8916666666666666 | 0.005229309611444482 |
+| Epoch_2.pt | 0.8873333333333333 | 0.005404616225340136 |
+| Epoch_1_batch_5999.pt | 0.8848333333333332 | 0.00404793805154375 |
+| Epoch_1_batch_2999.pt | 0.8768333333333335 | 0.005394613225417729 |
+| Epoch_1.pt | 0.866 | 0.004907791738968394 |
+| Epoch_0_batch_5999.pt | 0.8515 | 0.005970348543292511 |
+| Epoch_0.pt | 0.8465 | 0.006118429960797559 |
+| Epoch_0_batch_2999.pt | 0.8123333333333334 | 0.005218674918141055 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..4b2a9f6fde3ae8c2179917f9dc9d15f05f33efa1
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.8826666666666668 | 0.003627739492736556 |
+| Epoch_12_batch_5999.pt | 0.8823333333333334 | 0.0038425814368423543 |
+| Epoch_15_batch_5999.pt | 0.8821666666666668 | 0.0040601191979096635 |
+| Epoch_13.pt | 0.882 | 0.003934651379916833 |
+| Epoch_13_batch_2999.pt | 0.8818333333333334 | 0.003978724282176414 |
+| Epoch_16.pt | 0.8818333333333334 | 0.0036213529972738343 |
+| Epoch_17.pt | 0.8818333333333334 | 0.004234266447175832 |
+| Epoch_14_batch_2999.pt | 0.8816666666666666 | 0.0040292143032152755 |
+| Epoch_16_batch_5999.pt | 0.8816666666666666 | 0.003951870943061944 |
+| Epoch_17_batch_2999.pt | 0.8815 | 0.004205008771148053 |
+| Epoch_13_batch_5999.pt | 0.8813333333333333 | 0.003926799343793844 |
+| Epoch_15_batch_2999.pt | 0.881 | 0.004445833116387241 |
+| Epoch_17_batch_5999.pt | 0.8803333333333333 | 0.003349958540373635 |
+| Epoch_16_batch_2999.pt | 0.8801666666666665 | 0.0037387691907536076 |
+| Epoch_15.pt | 0.8793333333333333 | 0.00441098516472871 |
+| Epoch_10.pt | 0.8790000000000001 | 0.002973130702279918 |
+| Epoch_11_batch_5999.pt | 0.8788333333333334 | 0.004172588384583427 |
+| Epoch_12.pt | 0.8781666666666667 | 0.004495882890263376 |
+| Epoch_11_batch_2999.pt | 0.8776666666666667 | 0.0032508308529617365 |
+| Epoch_10_batch_5999.pt | 0.8775000000000001 | 0.00413693093136255 |
+| Epoch_12_batch_2999.pt | 0.8766666666666666 | 0.00364302140239 |
+| Epoch_14.pt | 0.876 | 0.004268749491621898 |
+| Epoch_10_batch_2999.pt | 0.8736666666666666 | 0.004330483393312241 |
+| Epoch_11.pt | 0.8701666666666668 | 0.003994208770670819 |
+| Epoch_9_batch_5999.pt | 0.8576666666666666 | 0.0048189440982669895 |
+| Epoch_8_batch_5999.pt | 0.8574999999999999 | 0.004964379289336685 |
+| Epoch_8_batch_2999.pt | 0.8539999999999999 | 0.004297573245736382 |
+| Epoch_7_batch_5999.pt | 0.8516666666666668 | 0.004331908597692867 |
+| Epoch_5_batch_5999.pt | 0.8506666666666666 | 0.0033536418383970173 |
+| Epoch_9.pt | 0.8501666666666667 | 0.004934446821285924 |
+| Epoch_9_batch_2999.pt | 0.8496666666666666 | 0.004484541349024565 |
+| Epoch_4_batch_5999.pt | 0.8474999999999999 | 0.004008093663428394 |
+| Epoch_8.pt | 0.8471666666666666 | 0.00444479165310436 |
+| Epoch_5_batch_2999.pt | 0.8460000000000001 | 0.004326204963095949 |
+| Epoch_4.pt | 0.8458333333333334 | 0.005742681870148161 |
+| Epoch_6_batch_2999.pt | 0.8456666666666667 | 0.003866602809178963 |
+| Epoch_6_batch_5999.pt | 0.8455 | 0.005012638348203125 |
+| Epoch_7.pt | 0.8438333333333332 | 0.003936611941790421 |
+| Epoch_4_batch_2999.pt | 0.8400000000000001 | 0.005018484351393868 |
+| Epoch_6.pt | 0.8381666666666667 | 0.004584259165750272 |
+| Epoch_3.pt | 0.8375 | 0.005495508828484462 |
+| Epoch_3_batch_5999.pt | 0.8366666666666667 | 0.004444444444444446 |
+| Epoch_7_batch_2999.pt | 0.8363333333333334 | 0.0043233503240763865 |
+| Epoch_2_batch_2999.pt | 0.8353333333333334 | 0.002287917809108222 |
+| Epoch_2_batch_5999.pt | 0.8351666666666666 | 0.004637807591875637 |
+| Epoch_3_batch_2999.pt | 0.835 | 0.003583225665910477 |
+| Epoch_5.pt | 0.8321666666666667 | 0.0050799168847093885 |
+| Epoch_1_batch_5999.pt | 0.8245000000000001 | 0.005682168339378649 |
+| Epoch_1_batch_2999.pt | 0.8171666666666667 | 0.0038252733300226725 |
+| Epoch_2.pt | 0.8168333333333335 | 0.0030976893021001438 |
+| Epoch_1.pt | 0.8133333333333335 | 0.006221230079630854 |
+| Epoch_0.pt | 0.7948333333333334 | 0.004405734198423745 |
+| Epoch_0_batch_5999.pt | 0.7893333333333333 | 0.004521553322083512 |
+| Epoch_0_batch_2999.pt | 0.7708333333333333 | 0.002293980313562507 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaCos/log.log b/bob/bio/facexzoo/models/heads/AdaCos/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..03297b4d3d8b83829773afb6ad3984ec8edde1a1
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaCos/log.log
@@ -0,0 +1,655 @@
+INFO 2020-11-24 17:57:23 train.py: 172] Start optimization.
+INFO 2020-11-24 17:57:23 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='adacos', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='arc-mobile', train_file='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f76df1891d0>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778}
+INFO 2020-11-24 17:57:45 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 11.416717
+INFO 2020-11-24 17:58:59 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 11.132950
+INFO 2020-11-24 18:00:13 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 9.655811
+INFO 2020-11-24 18:01:28 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 8.344480
+INFO 2020-11-24 18:02:43 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 7.282526
+INFO 2020-11-24 18:03:58 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 6.447086
+INFO 2020-11-24 18:05:12 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 5.743308
+INFO 2020-11-24 18:06:27 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 5.176081
+INFO 2020-11-24 18:07:42 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 4.686135
+INFO 2020-11-24 18:08:57 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 4.288819
+INFO 2020-11-24 18:10:12 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 3.973191
+INFO 2020-11-24 18:11:27 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 3.677886
+INFO 2020-11-24 18:12:41 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 3.454331
+INFO 2020-11-24 18:13:56 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 3.245951
+INFO 2020-11-24 18:15:11 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 3.073070
+INFO 2020-11-24 18:16:26 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-24 18:16:26 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 2.932677
+INFO 2020-11-24 18:17:41 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 2.805671
+INFO 2020-11-24 18:18:56 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 2.701228
+INFO 2020-11-24 18:20:11 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 2.569699
+INFO 2020-11-24 18:21:26 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 2.495928
+INFO 2020-11-24 18:22:41 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 2.437558
+INFO 2020-11-24 18:23:56 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 2.367344
+INFO 2020-11-24 18:25:10 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 2.288848
+INFO 2020-11-24 18:26:25 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 2.243524
+INFO 2020-11-24 18:27:40 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 2.185335
+INFO 2020-11-24 18:28:55 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 2.145634
+INFO 2020-11-24 18:30:10 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 2.112795
+INFO 2020-11-24 18:31:25 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 2.064050
+INFO 2020-11-24 18:32:40 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 2.030069
+INFO 2020-11-24 18:33:55 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 2.004714
+INFO 2020-11-24 18:35:09 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-24 18:35:10 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 1.977449
+INFO 2020-11-24 18:36:24 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 1.957185
+INFO 2020-11-24 18:37:38 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 1.930404
+INFO 2020-11-24 18:37:44 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-24 18:37:45 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 1.919308
+INFO 2020-11-24 18:39:01 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 1.786314
+INFO 2020-11-24 18:40:16 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 1.771297
+INFO 2020-11-24 18:41:31 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 1.776204
+INFO 2020-11-24 18:42:45 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 1.787330
+INFO 2020-11-24 18:44:01 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 1.786908
+INFO 2020-11-24 18:45:16 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 1.775076
+INFO 2020-11-24 18:46:31 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 1.776161
+INFO 2020-11-24 18:47:46 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 1.770470
+INFO 2020-11-24 18:49:01 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 1.762913
+INFO 2020-11-24 18:50:16 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 1.750178
+INFO 2020-11-24 18:51:31 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 1.743453
+INFO 2020-11-24 18:52:46 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 1.742926
+INFO 2020-11-24 18:54:01 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 1.724750
+INFO 2020-11-24 18:55:15 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 1.724824
+INFO 2020-11-24 18:56:30 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-24 18:56:30 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 1.717244
+INFO 2020-11-24 18:57:45 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 1.706369
+INFO 2020-11-24 18:59:01 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 1.695740
+INFO 2020-11-24 19:00:15 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 1.690210
+INFO 2020-11-24 19:01:30 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 1.679283
+INFO 2020-11-24 19:02:46 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 1.677418
+INFO 2020-11-24 19:04:00 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 1.675886
+INFO 2020-11-24 19:05:15 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 1.663851
+INFO 2020-11-24 19:06:30 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 1.661716
+INFO 2020-11-24 19:07:45 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 1.646499
+INFO 2020-11-24 19:09:00 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 1.647377
+INFO 2020-11-24 19:10:16 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 1.636752
+INFO 2020-11-24 19:11:30 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 1.636456
+INFO 2020-11-24 19:12:45 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 1.628192
+INFO 2020-11-24 19:14:00 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 1.616665
+INFO 2020-11-24 19:15:15 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-24 19:15:16 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 1.617624
+INFO 2020-11-24 19:16:31 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 1.615611
+INFO 2020-11-24 19:17:46 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 1.601256
+INFO 2020-11-24 19:17:52 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-24 19:17:53 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 1.625756
+INFO 2020-11-24 19:19:08 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 1.527721
+INFO 2020-11-24 19:20:22 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 1.509647
+INFO 2020-11-24 19:21:36 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 1.522225
+INFO 2020-11-24 19:22:50 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 1.522196
+INFO 2020-11-24 19:24:04 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 1.532201
+INFO 2020-11-24 19:25:18 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 1.537051
+INFO 2020-11-24 19:26:32 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 1.539282
+INFO 2020-11-24 19:27:46 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 1.534609
+INFO 2020-11-24 19:29:00 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 1.546923
+INFO 2020-11-24 19:30:14 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 1.545881
+INFO 2020-11-24 19:31:28 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 1.546089
+INFO 2020-11-24 19:32:42 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 1.543443
+INFO 2020-11-24 19:33:56 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 1.546798
+INFO 2020-11-24 19:35:10 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 1.548067
+INFO 2020-11-24 19:36:24 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-24 19:36:25 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 1.540771
+INFO 2020-11-24 19:37:39 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 1.538022
+INFO 2020-11-24 19:38:54 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 1.533676
+INFO 2020-11-24 19:40:09 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 1.532327
+INFO 2020-11-24 19:41:24 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 1.528228
+INFO 2020-11-24 19:42:39 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 1.526137
+INFO 2020-11-24 19:43:54 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 1.518209
+INFO 2020-11-24 19:45:09 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 1.521188
+INFO 2020-11-24 19:46:24 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 1.515481
+INFO 2020-11-24 19:47:38 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 1.516067
+INFO 2020-11-24 19:48:53 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 1.507745
+INFO 2020-11-24 19:50:08 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 1.507893
+INFO 2020-11-24 19:51:23 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 1.503140
+INFO 2020-11-24 19:52:38 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 1.501456
+INFO 2020-11-24 19:53:53 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 1.497328
+INFO 2020-11-24 19:55:08 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-24 19:55:08 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 1.499005
+INFO 2020-11-24 19:56:23 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 1.492542
+INFO 2020-11-24 19:57:38 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 1.493762
+INFO 2020-11-24 19:57:44 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-24 19:57:46 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 1.501257
+INFO 2020-11-24 19:59:01 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 1.415784
+INFO 2020-11-24 20:00:16 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 1.408254
+INFO 2020-11-24 20:01:31 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 1.421287
+INFO 2020-11-24 20:02:46 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 1.424631
+INFO 2020-11-24 20:04:01 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 1.432429
+INFO 2020-11-24 20:05:16 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 1.443933
+INFO 2020-11-24 20:06:31 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 1.451488
+INFO 2020-11-24 20:07:46 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 1.451521
+INFO 2020-11-24 20:09:01 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 1.454602
+INFO 2020-11-24 20:10:16 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 1.455259
+INFO 2020-11-24 20:11:31 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 1.452010
+INFO 2020-11-24 20:12:46 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 1.453682
+INFO 2020-11-24 20:14:01 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 1.447194
+INFO 2020-11-24 20:15:16 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 1.460136
+INFO 2020-11-24 20:16:31 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-24 20:16:31 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 1.449635
+INFO 2020-11-24 20:17:46 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 1.449914
+INFO 2020-11-24 20:19:00 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 1.448616
+INFO 2020-11-24 20:20:14 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 1.450514
+INFO 2020-11-24 20:21:28 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 1.453894
+INFO 2020-11-24 20:22:43 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 1.447115
+INFO 2020-11-24 20:23:57 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 1.443263
+INFO 2020-11-24 20:25:11 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 1.454202
+INFO 2020-11-24 20:26:25 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 1.443775
+INFO 2020-11-24 20:27:40 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 1.444346
+INFO 2020-11-24 20:28:54 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 1.435848
+INFO 2020-11-24 20:30:08 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 1.441113
+INFO 2020-11-24 20:31:22 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 1.441847
+INFO 2020-11-24 20:32:37 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 1.433100
+INFO 2020-11-24 20:33:51 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 1.434238
+INFO 2020-11-24 20:35:05 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-24 20:35:05 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 1.431912
+INFO 2020-11-24 20:36:20 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 1.437632
+INFO 2020-11-24 20:37:35 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 1.428208
+INFO 2020-11-24 20:37:41 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-24 20:37:43 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 1.425253
+INFO 2020-11-24 20:38:58 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 1.368639
+INFO 2020-11-24 20:40:13 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 1.356518
+INFO 2020-11-24 20:41:28 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 1.364383
+INFO 2020-11-24 20:42:43 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 1.375807
+INFO 2020-11-24 20:43:58 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 1.378041
+INFO 2020-11-24 20:45:13 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 1.386899
+INFO 2020-11-24 20:46:28 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 1.383502
+INFO 2020-11-24 20:47:43 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 1.396895
+INFO 2020-11-24 20:48:58 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 1.398383
+INFO 2020-11-24 20:50:13 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 1.396244
+INFO 2020-11-24 20:51:28 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 1.406174
+INFO 2020-11-24 20:52:43 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 1.399846
+INFO 2020-11-24 20:53:58 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 1.403289
+INFO 2020-11-24 20:55:13 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 1.403233
+INFO 2020-11-24 20:56:27 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-24 20:56:28 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 1.405860
+INFO 2020-11-24 20:57:43 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 1.402708
+INFO 2020-11-24 20:58:58 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 1.400847
+INFO 2020-11-24 21:00:13 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 1.408870
+INFO 2020-11-24 21:01:28 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 1.404592
+INFO 2020-11-24 21:02:43 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 1.402544
+INFO 2020-11-24 21:03:58 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 1.398803
+INFO 2020-11-24 21:05:13 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 1.398866
+INFO 2020-11-24 21:06:27 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 1.402501
+INFO 2020-11-24 21:07:42 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 1.399502
+INFO 2020-11-24 21:08:57 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 1.401013
+INFO 2020-11-24 21:10:12 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 1.396362
+INFO 2020-11-24 21:11:27 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 1.395240
+INFO 2020-11-24 21:12:42 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 1.400683
+INFO 2020-11-24 21:13:57 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 1.399154
+INFO 2020-11-24 21:15:12 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-24 21:15:12 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 1.389652
+INFO 2020-11-24 21:16:27 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 1.395338
+INFO 2020-11-24 21:17:41 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 1.387543
+INFO 2020-11-24 21:17:47 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-24 21:17:49 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 1.367147
+INFO 2020-11-24 21:19:04 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 1.327743
+INFO 2020-11-24 21:20:19 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 1.321898
+INFO 2020-11-24 21:21:34 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 1.325918
+INFO 2020-11-24 21:22:49 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 1.336412
+INFO 2020-11-24 21:24:04 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 1.344321
+INFO 2020-11-24 21:25:19 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 1.358709
+INFO 2020-11-24 21:26:34 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 1.358282
+INFO 2020-11-24 21:27:49 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 1.358601
+INFO 2020-11-24 21:29:04 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 1.366539
+INFO 2020-11-24 21:30:19 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 1.366163
+INFO 2020-11-24 21:31:34 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 1.366323
+INFO 2020-11-24 21:32:49 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 1.370541
+INFO 2020-11-24 21:34:04 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 1.368496
+INFO 2020-11-24 21:35:19 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 1.372332
+INFO 2020-11-24 21:36:34 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-24 21:36:35 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 1.376615
+INFO 2020-11-24 21:37:50 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 1.372517
+INFO 2020-11-24 21:39:05 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 1.376011
+INFO 2020-11-24 21:40:20 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 1.371207
+INFO 2020-11-24 21:41:35 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 1.374255
+INFO 2020-11-24 21:42:50 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 1.375235
+INFO 2020-11-24 21:44:05 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 1.377788
+INFO 2020-11-24 21:45:20 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 1.373756
+INFO 2020-11-24 21:46:35 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 1.370363
+INFO 2020-11-24 21:47:50 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 1.369006
+INFO 2020-11-24 21:49:05 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 1.371390
+INFO 2020-11-24 21:50:20 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 1.365531
+INFO 2020-11-24 21:51:35 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 1.368726
+INFO 2020-11-24 21:52:50 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 1.363939
+INFO 2020-11-24 21:54:05 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 1.369182
+INFO 2020-11-24 21:55:20 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-24 21:55:20 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 1.362293
+INFO 2020-11-24 21:56:35 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 1.363668
+INFO 2020-11-24 21:57:50 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 1.366412
+INFO 2020-11-24 21:57:56 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-24 21:57:58 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 1.373471
+INFO 2020-11-24 21:59:12 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 1.304167
+INFO 2020-11-24 22:00:26 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 1.294021
+INFO 2020-11-24 22:01:41 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 1.300894
+INFO 2020-11-24 22:02:55 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 1.315514
+INFO 2020-11-24 22:04:09 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 1.319493
+INFO 2020-11-24 22:05:23 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 1.324446
+INFO 2020-11-24 22:06:37 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 1.332261
+INFO 2020-11-24 22:07:52 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 1.337486
+INFO 2020-11-24 22:09:06 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 1.344820
+INFO 2020-11-24 22:10:20 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 1.347537
+INFO 2020-11-24 22:11:35 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 1.349921
+INFO 2020-11-24 22:12:49 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 1.348303
+INFO 2020-11-24 22:14:03 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 1.344542
+INFO 2020-11-24 22:15:18 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 1.347508
+INFO 2020-11-24 22:16:32 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-24 22:16:32 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 1.353212
+INFO 2020-11-24 22:17:47 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 1.349635
+INFO 2020-11-24 22:19:02 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 1.345761
+INFO 2020-11-24 22:20:17 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 1.348586
+INFO 2020-11-24 22:21:32 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 1.349593
+INFO 2020-11-24 22:22:47 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 1.350912
+INFO 2020-11-24 22:24:02 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 1.349537
+INFO 2020-11-24 22:25:18 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 1.352887
+INFO 2020-11-24 22:26:53 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 1.352015
+INFO 2020-11-24 22:28:32 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 1.351760
+INFO 2020-11-24 22:30:09 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 1.348667
+INFO 2020-11-24 22:31:34 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 1.351530
+INFO 2020-11-24 22:32:49 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 1.347378
+INFO 2020-11-24 22:34:04 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 1.342948
+INFO 2020-11-24 22:35:19 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 1.343790
+INFO 2020-11-24 22:36:34 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-24 22:36:34 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 1.344466
+INFO 2020-11-24 22:37:50 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 1.344174
+INFO 2020-11-24 22:39:05 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 1.337644
+INFO 2020-11-24 22:39:10 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-24 22:39:12 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 1.346722
+INFO 2020-11-24 22:40:27 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 1.281349
+INFO 2020-11-24 22:41:42 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 1.273995
+INFO 2020-11-24 22:42:57 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 1.284686
+INFO 2020-11-24 22:44:12 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 1.293039
+INFO 2020-11-24 22:45:27 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 1.295773
+INFO 2020-11-24 22:46:42 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 1.309860
+INFO 2020-11-24 22:47:57 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 1.316550
+INFO 2020-11-24 22:49:12 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 1.317121
+INFO 2020-11-24 22:50:27 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 1.322618
+INFO 2020-11-24 22:51:41 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 1.329629
+INFO 2020-11-24 22:52:56 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 1.331045
+INFO 2020-11-24 22:54:11 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 1.334171
+INFO 2020-11-24 22:55:26 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 1.327468
+INFO 2020-11-24 22:56:41 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 1.330366
+INFO 2020-11-24 22:57:56 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-24 22:57:56 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 1.331032
+INFO 2020-11-24 22:59:11 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 1.336948
+INFO 2020-11-24 23:00:26 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 1.332977
+INFO 2020-11-24 23:01:41 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 1.332363
+INFO 2020-11-24 23:02:56 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 1.336803
+INFO 2020-11-24 23:04:11 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 1.327630
+INFO 2020-11-24 23:05:26 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 1.334582
+INFO 2020-11-24 23:06:41 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 1.327688
+INFO 2020-11-24 23:08:03 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 1.333249
+INFO 2020-11-24 23:09:39 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 1.329354
+INFO 2020-11-24 23:11:16 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 1.333234
+INFO 2020-11-24 23:12:46 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 1.326928
+INFO 2020-11-24 23:14:01 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 1.334914
+INFO 2020-11-24 23:15:16 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 1.329901
+INFO 2020-11-24 23:16:32 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 1.331653
+INFO 2020-11-24 23:17:46 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-24 23:17:47 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 1.331404
+INFO 2020-11-24 23:19:01 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 1.330640
+INFO 2020-11-24 23:20:15 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 1.332471
+INFO 2020-11-24 23:20:21 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-24 23:20:23 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 1.321369
+INFO 2020-11-24 23:21:37 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 1.275679
+INFO 2020-11-24 23:22:52 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 1.264183
+INFO 2020-11-24 23:24:07 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 1.270290
+INFO 2020-11-24 23:25:22 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 1.279872
+INFO 2020-11-24 23:26:37 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 1.290014
+INFO 2020-11-24 23:27:52 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 1.294480
+INFO 2020-11-24 23:29:07 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 1.300083
+INFO 2020-11-24 23:30:22 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 1.304544
+INFO 2020-11-24 23:31:37 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 1.306816
+INFO 2020-11-24 23:32:52 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 1.314479
+INFO 2020-11-24 23:34:07 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 1.318064
+INFO 2020-11-24 23:35:22 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 1.316137
+INFO 2020-11-24 23:36:37 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 1.315622
+INFO 2020-11-24 23:37:52 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 1.318616
+INFO 2020-11-24 23:39:07 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-24 23:39:07 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 1.315701
+INFO 2020-11-24 23:40:22 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 1.317215
+INFO 2020-11-24 23:41:37 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 1.324224
+INFO 2020-11-24 23:42:52 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 1.321538
+INFO 2020-11-24 23:44:07 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 1.318713
+INFO 2020-11-24 23:45:22 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 1.321623
+INFO 2020-11-24 23:46:37 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 1.318330
+INFO 2020-11-24 23:47:52 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 1.314947
+INFO 2020-11-24 23:49:07 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 1.319453
+INFO 2020-11-24 23:50:22 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 1.320324
+INFO 2020-11-24 23:51:37 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 1.314977
+INFO 2020-11-24 23:52:53 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 1.315860
+INFO 2020-11-24 23:54:08 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 1.316496
+INFO 2020-11-24 23:55:23 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 1.324102
+INFO 2020-11-24 23:56:38 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 1.318073
+INFO 2020-11-24 23:57:53 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-24 23:57:53 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 1.320587
+INFO 2020-11-24 23:59:08 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 1.314739
+INFO 2020-11-25 00:00:23 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 1.317633
+INFO 2020-11-25 00:00:29 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-25 00:00:31 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 1.300685
+INFO 2020-11-25 00:01:46 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 1.262286
+INFO 2020-11-25 00:03:01 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 1.247746
+INFO 2020-11-25 00:04:16 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 1.263753
+INFO 2020-11-25 00:05:31 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 1.266173
+INFO 2020-11-25 00:06:46 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 1.277104
+INFO 2020-11-25 00:08:01 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 1.277577
+INFO 2020-11-25 00:09:16 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 1.285791
+INFO 2020-11-25 00:10:31 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 1.288765
+INFO 2020-11-25 00:11:46 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 1.290202
+INFO 2020-11-25 00:13:01 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 1.301500
+INFO 2020-11-25 00:14:16 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 1.301219
+INFO 2020-11-25 00:15:31 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 1.303576
+INFO 2020-11-25 00:16:46 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 1.309247
+INFO 2020-11-25 00:18:01 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 1.305372
+INFO 2020-11-25 00:19:16 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-25 00:19:16 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 1.311056
+INFO 2020-11-25 00:20:31 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 1.306957
+INFO 2020-11-25 00:21:45 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 1.312938
+INFO 2020-11-25 00:23:00 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 1.308837
+INFO 2020-11-25 00:24:15 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 1.305323
+INFO 2020-11-25 00:25:30 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 1.307396
+INFO 2020-11-25 00:26:45 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 1.311401
+INFO 2020-11-25 00:28:01 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 1.309233
+INFO 2020-11-25 00:29:16 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 1.308433
+INFO 2020-11-25 00:30:31 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 1.307102
+INFO 2020-11-25 00:31:46 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 1.314953
+INFO 2020-11-25 00:33:01 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 1.310075
+INFO 2020-11-25 00:34:16 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 1.308042
+INFO 2020-11-25 00:35:31 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 1.309480
+INFO 2020-11-25 00:36:46 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 1.309286
+INFO 2020-11-25 00:38:01 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-25 00:38:01 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 1.306205
+INFO 2020-11-25 00:39:16 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 1.306948
+INFO 2020-11-25 00:40:31 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 1.305604
+INFO 2020-11-25 00:40:37 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-25 00:40:39 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 1.304202
+INFO 2020-11-25 00:41:54 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 1.176658
+INFO 2020-11-25 00:43:08 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 1.148532
+INFO 2020-11-25 00:44:23 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 1.134307
+INFO 2020-11-25 00:45:38 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 1.126774
+INFO 2020-11-25 00:46:53 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 1.123866
+INFO 2020-11-25 00:48:08 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 1.116975
+INFO 2020-11-25 00:49:23 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 1.114827
+INFO 2020-11-25 00:50:38 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 1.110374
+INFO 2020-11-25 00:51:53 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 1.109232
+INFO 2020-11-25 00:53:08 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 1.105800
+INFO 2020-11-25 00:54:22 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 1.103239
+INFO 2020-11-25 00:55:37 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 1.100349
+INFO 2020-11-25 00:56:52 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 1.092838
+INFO 2020-11-25 00:58:07 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 1.096585
+INFO 2020-11-25 00:59:22 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-25 00:59:22 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 1.093232
+INFO 2020-11-25 01:00:36 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 1.088581
+INFO 2020-11-25 01:01:50 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 1.085622
+INFO 2020-11-25 01:03:05 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 1.083115
+INFO 2020-11-25 01:04:19 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 1.083759
+INFO 2020-11-25 01:05:33 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 1.083667
+INFO 2020-11-25 01:06:47 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 1.079579
+INFO 2020-11-25 01:08:01 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 1.078006
+INFO 2020-11-25 01:09:15 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 1.079693
+INFO 2020-11-25 01:10:30 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 1.074036
+INFO 2020-11-25 01:11:44 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 1.077814
+INFO 2020-11-25 01:12:58 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 1.074785
+INFO 2020-11-25 01:14:12 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 1.074285
+INFO 2020-11-25 01:15:27 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 1.066102
+INFO 2020-11-25 01:16:41 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 1.068799
+INFO 2020-11-25 01:17:55 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-25 01:17:55 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 1.068666
+INFO 2020-11-25 01:19:10 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 1.070428
+INFO 2020-11-25 01:20:25 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 1.068347
+INFO 2020-11-25 01:20:31 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-25 01:20:33 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 1.064234
+INFO 2020-11-25 01:21:48 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 1.025108
+INFO 2020-11-25 01:23:03 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 1.025457
+INFO 2020-11-25 01:24:18 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 1.029919
+INFO 2020-11-25 01:25:33 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 1.025832
+INFO 2020-11-25 01:26:48 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 1.028872
+INFO 2020-11-25 01:28:03 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 1.027197
+INFO 2020-11-25 01:29:18 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 1.028729
+INFO 2020-11-25 01:30:33 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 1.020073
+INFO 2020-11-25 01:31:48 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 1.027244
+INFO 2020-11-25 01:33:03 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 1.031872
+INFO 2020-11-25 01:34:18 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 1.031487
+INFO 2020-11-25 01:35:33 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 1.028213
+INFO 2020-11-25 01:36:48 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 1.026917
+INFO 2020-11-25 01:38:03 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 1.032668
+INFO 2020-11-25 01:39:18 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-25 01:39:18 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 1.031355
+INFO 2020-11-25 01:40:33 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 1.026722
+INFO 2020-11-25 01:41:48 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 1.034043
+INFO 2020-11-25 01:43:03 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 1.033295
+INFO 2020-11-25 01:44:18 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 1.030751
+INFO 2020-11-25 01:45:33 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 1.037479
+INFO 2020-11-25 01:46:47 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 1.029564
+INFO 2020-11-25 01:48:02 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 1.030864
+INFO 2020-11-25 01:49:17 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 1.032961
+INFO 2020-11-25 01:50:32 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 1.035965
+INFO 2020-11-25 01:51:47 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 1.034689
+INFO 2020-11-25 01:53:02 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 1.033343
+INFO 2020-11-25 01:54:17 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 1.033642
+INFO 2020-11-25 01:55:32 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 1.037396
+INFO 2020-11-25 01:56:47 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 1.038487
+INFO 2020-11-25 01:58:02 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-25 01:58:02 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 1.033763
+INFO 2020-11-25 01:59:17 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 1.038661
+INFO 2020-11-25 02:00:32 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 1.036838
+INFO 2020-11-25 02:00:38 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-25 02:00:40 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 1.038998
+INFO 2020-11-25 02:01:55 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 0.995148
+INFO 2020-11-25 02:03:10 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 0.998618
+INFO 2020-11-25 02:04:25 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 0.999423
+INFO 2020-11-25 02:05:40 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 0.999431
+INFO 2020-11-25 02:06:55 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.000098
+INFO 2020-11-25 02:08:10 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.004782
+INFO 2020-11-25 02:09:25 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 1.002303
+INFO 2020-11-25 02:10:40 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 1.004181
+INFO 2020-11-25 02:11:55 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 1.006173
+INFO 2020-11-25 02:13:10 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 1.006221
+INFO 2020-11-25 02:14:25 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 1.009304
+INFO 2020-11-25 02:15:40 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 1.012346
+INFO 2020-11-25 02:16:54 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 1.009386
+INFO 2020-11-25 02:18:09 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 1.011192
+INFO 2020-11-25 02:19:24 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-25 02:19:24 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 1.017104
+INFO 2020-11-25 02:20:38 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 1.015620
+INFO 2020-11-25 02:21:52 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 1.017127
+INFO 2020-11-25 02:23:07 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 1.015543
+INFO 2020-11-25 02:24:21 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 1.014963
+INFO 2020-11-25 02:25:36 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 1.019896
+INFO 2020-11-25 02:26:51 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 1.017910
+INFO 2020-11-25 02:28:06 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 1.021747
+INFO 2020-11-25 02:29:21 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 1.020708
+INFO 2020-11-25 02:30:35 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 1.020469
+INFO 2020-11-25 02:31:50 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 1.021486
+INFO 2020-11-25 02:33:05 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 1.022875
+INFO 2020-11-25 02:34:20 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 1.025001
+INFO 2020-11-25 02:35:35 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 1.023730
+INFO 2020-11-25 02:36:49 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 1.026016
+INFO 2020-11-25 02:38:04 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-25 02:38:04 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 1.026193
+INFO 2020-11-25 02:39:19 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 1.031579
+INFO 2020-11-25 02:40:34 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 1.024610
+INFO 2020-11-25 02:40:40 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-25 02:40:41 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 1.031203
+INFO 2020-11-25 02:41:56 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 0.979068
+INFO 2020-11-25 02:43:11 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 0.972301
+INFO 2020-11-25 02:44:26 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 0.976219
+INFO 2020-11-25 02:45:41 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 0.973277
+INFO 2020-11-25 02:46:56 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 0.974932
+INFO 2020-11-25 02:48:11 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 0.974172
+INFO 2020-11-25 02:49:25 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 0.975371
+INFO 2020-11-25 02:50:40 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 0.974761
+INFO 2020-11-25 02:51:55 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 0.973856
+INFO 2020-11-25 02:53:10 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 0.975677
+INFO 2020-11-25 02:54:24 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 0.971482
+INFO 2020-11-25 02:55:39 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 0.972668
+INFO 2020-11-25 02:56:54 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 0.974895
+INFO 2020-11-25 02:58:09 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 0.974955
+INFO 2020-11-25 02:59:23 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-25 02:59:23 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 0.972740
+INFO 2020-11-25 03:00:37 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 0.973307
+INFO 2020-11-25 03:01:51 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 0.976351
+INFO 2020-11-25 03:03:05 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 0.978493
+INFO 2020-11-25 03:04:19 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 0.973779
+INFO 2020-11-25 03:05:33 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 0.976580
+INFO 2020-11-25 03:06:47 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 0.971437
+INFO 2020-11-25 03:08:01 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 0.974483
+INFO 2020-11-25 03:09:15 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 0.973376
+INFO 2020-11-25 03:10:29 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 0.971711
+INFO 2020-11-25 03:11:43 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 0.974456
+INFO 2020-11-25 03:12:57 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 0.971446
+INFO 2020-11-25 03:14:11 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 0.973631
+INFO 2020-11-25 03:15:25 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 0.974200
+INFO 2020-11-25 03:16:39 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 0.973257
+INFO 2020-11-25 03:17:53 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-25 03:17:53 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 0.973297
+INFO 2020-11-25 03:19:08 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 0.977037
+INFO 2020-11-25 03:20:23 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 0.978571
+INFO 2020-11-25 03:20:28 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-25 03:20:30 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 0.975836
+INFO 2020-11-25 03:21:45 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 0.965479
+INFO 2020-11-25 03:23:00 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 0.970131
+INFO 2020-11-25 03:24:16 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 0.966709
+INFO 2020-11-25 03:25:31 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 0.969352
+INFO 2020-11-25 03:26:46 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 0.967796
+INFO 2020-11-25 03:28:01 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 0.972505
+INFO 2020-11-25 03:29:16 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 0.968585
+INFO 2020-11-25 03:30:30 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 0.970320
+INFO 2020-11-25 03:31:45 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 0.966743
+INFO 2020-11-25 03:33:00 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 0.970896
+INFO 2020-11-25 03:34:15 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 0.969638
+INFO 2020-11-25 03:35:30 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 0.970121
+INFO 2020-11-25 03:36:45 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 0.972427
+INFO 2020-11-25 03:38:00 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 0.971366
+INFO 2020-11-25 03:39:15 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-25 03:39:15 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 0.973161
+INFO 2020-11-25 03:40:30 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 0.967840
+INFO 2020-11-25 03:41:45 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 0.972479
+INFO 2020-11-25 03:43:00 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 0.970489
+INFO 2020-11-25 03:44:15 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 0.975154
+INFO 2020-11-25 03:45:30 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 0.973304
+INFO 2020-11-25 03:46:44 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 0.970431
+INFO 2020-11-25 03:48:00 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 0.973954
+INFO 2020-11-25 03:49:14 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 0.972537
+INFO 2020-11-25 03:50:29 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 0.972590
+INFO 2020-11-25 03:51:44 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 0.970980
+INFO 2020-11-25 03:52:59 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 0.970028
+INFO 2020-11-25 03:54:14 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 0.972340
+INFO 2020-11-25 03:55:29 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 0.970009
+INFO 2020-11-25 03:56:43 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 0.973213
+INFO 2020-11-25 03:57:58 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-25 03:57:59 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 0.973166
+INFO 2020-11-25 03:59:13 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 0.974608
+INFO 2020-11-25 04:00:27 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 0.975080
+INFO 2020-11-25 04:00:32 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-25 04:00:34 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 0.979018
+INFO 2020-11-25 04:01:49 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 0.963979
+INFO 2020-11-25 04:03:04 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 0.966317
+INFO 2020-11-25 04:04:19 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 0.967784
+INFO 2020-11-25 04:05:34 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 0.967673
+INFO 2020-11-25 04:06:49 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 0.969755
+INFO 2020-11-25 04:08:03 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 0.968376
+INFO 2020-11-25 04:09:18 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 0.964123
+INFO 2020-11-25 04:10:33 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 0.969228
+INFO 2020-11-25 04:11:47 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 0.967376
+INFO 2020-11-25 04:13:02 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 0.969040
+INFO 2020-11-25 04:14:17 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 0.968344
+INFO 2020-11-25 04:15:31 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 0.970091
+INFO 2020-11-25 04:16:46 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 0.970169
+INFO 2020-11-25 04:18:01 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 0.966720
+INFO 2020-11-25 04:19:15 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-25 04:19:15 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 0.971959
+INFO 2020-11-25 04:20:30 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 0.966358
+INFO 2020-11-25 04:21:45 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 0.969741
+INFO 2020-11-25 04:22:59 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 0.971326
+INFO 2020-11-25 04:24:14 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 0.964816
+INFO 2020-11-25 04:25:29 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 0.969996
+INFO 2020-11-25 04:26:44 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 0.968423
+INFO 2020-11-25 04:27:58 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 0.971637
+INFO 2020-11-25 04:29:13 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 0.971576
+INFO 2020-11-25 04:30:28 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 0.976166
+INFO 2020-11-25 04:31:43 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 0.972658
+INFO 2020-11-25 04:32:57 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 0.968336
+INFO 2020-11-25 04:34:12 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 0.972272
+INFO 2020-11-25 04:35:27 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 0.970980
+INFO 2020-11-25 04:36:42 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 0.971073
+INFO 2020-11-25 04:37:57 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-25 04:37:57 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 0.973298
+INFO 2020-11-25 04:39:12 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 0.968078
+INFO 2020-11-25 04:40:27 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 0.970595
+INFO 2020-11-25 04:40:33 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-25 04:40:34 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 0.984895
+INFO 2020-11-25 04:41:49 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 0.961366
+INFO 2020-11-25 04:43:03 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 0.966711
+INFO 2020-11-25 04:44:17 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 0.964994
+INFO 2020-11-25 04:45:32 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 0.961280
+INFO 2020-11-25 04:46:46 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 0.963230
+INFO 2020-11-25 04:48:00 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 0.964989
+INFO 2020-11-25 04:49:14 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 0.962194
+INFO 2020-11-25 04:50:28 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 0.962531
+INFO 2020-11-25 04:51:43 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 0.963279
+INFO 2020-11-25 04:52:57 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 0.962168
+INFO 2020-11-25 04:54:11 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 0.965057
+INFO 2020-11-25 04:55:25 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 0.964914
+INFO 2020-11-25 04:56:39 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 0.963014
+INFO 2020-11-25 04:57:54 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 0.966069
+INFO 2020-11-25 04:59:08 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-25 04:59:08 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 0.964854
+INFO 2020-11-25 05:00:23 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 0.966403
+INFO 2020-11-25 05:01:38 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 0.965760
+INFO 2020-11-25 05:02:53 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 0.964139
+INFO 2020-11-25 05:04:08 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 0.963679
+INFO 2020-11-25 05:05:23 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 0.965509
+INFO 2020-11-25 05:06:38 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 0.965264
+INFO 2020-11-25 05:07:53 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 0.965149
+INFO 2020-11-25 05:09:08 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 0.965419
+INFO 2020-11-25 05:10:23 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 0.962386
+INFO 2020-11-25 05:11:38 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 0.963995
+INFO 2020-11-25 05:12:53 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 0.964670
+INFO 2020-11-25 05:14:08 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 0.964155
+INFO 2020-11-25 05:15:23 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 0.961789
+INFO 2020-11-25 05:16:38 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 0.962934
+INFO 2020-11-25 05:17:52 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-25 05:17:53 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 0.965471
+INFO 2020-11-25 05:19:08 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 0.965259
+INFO 2020-11-25 05:20:23 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 0.965570
+INFO 2020-11-25 05:20:29 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-25 05:20:30 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 0.981509
+INFO 2020-11-25 05:21:45 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 0.962950
+INFO 2020-11-25 05:23:00 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 0.963583
+INFO 2020-11-25 05:24:15 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 0.963867
+INFO 2020-11-25 05:25:30 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 0.966846
+INFO 2020-11-25 05:26:45 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 0.962488
+INFO 2020-11-25 05:27:59 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 0.962950
+INFO 2020-11-25 05:29:14 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 0.963862
+INFO 2020-11-25 05:30:29 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 0.962231
+INFO 2020-11-25 05:31:43 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 0.962685
+INFO 2020-11-25 05:32:58 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 0.961242
+INFO 2020-11-25 05:34:12 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 0.961981
+INFO 2020-11-25 05:35:27 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 0.966319
+INFO 2020-11-25 05:36:42 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 0.967487
+INFO 2020-11-25 05:37:56 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 0.964566
+INFO 2020-11-25 05:39:10 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-25 05:39:11 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 0.962629
+INFO 2020-11-25 05:40:25 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 0.961897
+INFO 2020-11-25 05:41:38 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 0.963734
+INFO 2020-11-25 05:42:52 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 0.964792
+INFO 2020-11-25 05:44:06 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 0.960531
+INFO 2020-11-25 05:45:20 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 0.962284
+INFO 2020-11-25 05:46:34 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 0.964726
+INFO 2020-11-25 05:47:48 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 0.963413
+INFO 2020-11-25 05:49:02 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 0.961988
+INFO 2020-11-25 05:50:16 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 0.965384
+INFO 2020-11-25 05:51:30 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 0.962076
+INFO 2020-11-25 05:52:44 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 0.964248
+INFO 2020-11-25 05:53:58 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 0.968838
+INFO 2020-11-25 05:55:12 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 0.964142
+INFO 2020-11-25 05:56:26 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 0.962927
+INFO 2020-11-25 05:57:40 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-25 05:57:40 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 0.966201
+INFO 2020-11-25 05:58:55 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 0.963357
+INFO 2020-11-25 06:00:10 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 0.966563
+INFO 2020-11-25 06:00:16 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-25 06:00:16 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/.gitkeep b/bob/bio/facexzoo/models/heads/AdaM-Softmax/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3ea88953ce786a339658aaed31e2aaedec7c61ee
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_agedb.txt
@@ -0,0 +1,45 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.9601666666666666 | 0.003820429165989821 |
+| Epoch_16.pt | 0.9599999999999997 | 0.0038490017945975036 |
+| Epoch_17.pt | 0.9596666666666666 | 0.004316205466567778 |
+| Epoch_14_batch_2999.pt | 0.9595 | 0.00348763247015989 |
+| Epoch_13_batch_2999.pt | 0.9594999999999999 | 0.003592258480072719 |
+| Epoch_11_batch_5999.pt | 0.9591666666666667 | 0.0041518254203407505 |
+| Epoch_15_batch_5999.pt | 0.9591666666666667 | 0.003687231890232251 |
+| Epoch_16_batch_2999.pt | 0.959 | 0.003906310184721548 |
+| Epoch_12.pt | 0.9585000000000001 | 0.003705601365700504 |
+| Epoch_13_batch_5999.pt | 0.9585000000000001 | 0.003884521356980737 |
+| Epoch_15_batch_2999.pt | 0.9583333333333334 | 0.003849001794597506 |
+| Epoch_14.pt | 0.9581666666666667 | 0.0036128201084198036 |
+| Epoch_13.pt | 0.9580000000000002 | 0.0038554114606438846 |
+| Epoch_16_batch_5999.pt | 0.958 | 0.003798960221617509 |
+| Epoch_15.pt | 0.9578333333333333 | 0.0038413764251250043 |
+| Epoch_14_batch_5999.pt | 0.9576666666666667 | 0.004007708621526989 |
+| Epoch_10.pt | 0.9573333333333334 | 0.003922080576605068 |
+| Epoch_17_batch_2999.pt | 0.9573333333333334 | 0.0038022085849486 |
+| Epoch_10_batch_5999.pt | 0.9561666666666667 | 0.003952261424851649 |
+| Epoch_11.pt | 0.9559999999999998 | 0.003653173827283022 |
+| Epoch_12_batch_2999.pt | 0.9553333333333333 | 0.00354686437766942 |
+| Epoch_11_batch_2999.pt | 0.9551666666666667 | 0.003713921091857395 |
+| Epoch_12_batch_5999.pt | 0.9548333333333332 | 0.003939746811442539 |
+| Epoch_10_batch_2999.pt | 0.9543333333333333 | 0.004000000000000002 |
+| Epoch_8_batch_2999.pt | 0.9448333333333334 | 0.004604412805554102 |
+| Epoch_9.pt | 0.9445 | 0.004766459461955925 |
+| Epoch_6_batch_2999.pt | 0.9436666666666665 | 0.003208784239598587 |
+| Epoch_8_batch_5999.pt | 0.9428333333333334 | 0.003693922268862062 |
+| Epoch_9_batch_5999.pt | 0.9423333333333332 | 0.004793256580027331 |
+| Epoch_7_batch_5999.pt | 0.9404999999999999 | 0.00380099077402163 |
+| Epoch_6_batch_5999.pt | 0.9398333333333333 | 0.0039397468114425295 |
+| Epoch_9_batch_2999.pt | 0.9398333333333333 | 0.0044336674894952825 |
+| Epoch_7_batch_2999.pt | 0.9393333333333332 | 0.004319064827905956 |
+| Epoch_5_batch_5999.pt | 0.9366666666666665 | 0.004674596437325029 |
+| Epoch_5_batch_2999.pt | 0.9351666666666667 | 0.0037056013657005057 |
+| Epoch_4_batch_5999.pt | 0.9349999999999999 | 0.0043319085976928655 |
+| Epoch_5.pt | 0.9349999999999999 | 0.004388537257362552 |
+| Epoch_8.pt | 0.9331666666666667 | 0.005045778090959795 |
+| Epoch_6.pt | 0.933 | 0.004316205466567779 |
+| Epoch_3_batch_5999.pt | 0.9329999999999998 | 0.005367943257413852 |
+| Epoch_4_batch_2999.pt | 0.9301666666666668 | 0.005451526006768118 |
+| Epoch_4.pt | 0.9279999999999999 | 0.006054281211477906 |
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..4417fe0c488e0f278812eeac5ec15e9e670cd695
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_calfw.txt
@@ -0,0 +1,44 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9350000000000002 | 0.0037515428924742495 |
+| Epoch_12.pt | 0.9348333333333333 | 0.004130958098893618 |
+| Epoch_16.pt | 0.9346666666666665 | 0.0039659040661277145 |
+| Epoch_17_batch_2999.pt | 0.9346666666666665 | 0.004073400617738523 |
+| Epoch_13.pt | 0.9343333333333333 | 0.003644715437079272 |
+| Epoch_15_batch_5999.pt | 0.9343333333333333 | 0.003426476432246502 |
+| Epoch_13_batch_5999.pt | 0.9341666666666667 | 0.004121982622566562 |
+| Epoch_16_batch_2999.pt | 0.9339999999999998 | 0.003984538017120244 |
+| Epoch_15.pt | 0.9336666666666668 | 0.003871389197818742 |
+| Epoch_12_batch_2999.pt | 0.9336666666666666 | 0.00370851539295081 |
+| Epoch_14_batch_5999.pt | 0.9336666666666666 | 0.003871389197818743 |
+| Epoch_10_batch_2999.pt | 0.9333333333333333 | 0.0038086970002228064 |
+| Epoch_17.pt | 0.9333333333333333 | 0.0036514837167011065 |
+| Epoch_16_batch_5999.pt | 0.9333333333333332 | 0.003726779962499648 |
+| Epoch_11_batch_2999.pt | 0.933 | 0.0038151743807532004 |
+| Epoch_11_batch_5999.pt | 0.933 | 0.00417813237265848 |
+| Epoch_14.pt | 0.9326666666666668 | 0.003471222078180736 |
+| Epoch_12_batch_5999.pt | 0.9324999999999999 | 0.004069231128273329 |
+| Epoch_13_batch_2999.pt | 0.9324999999999999 | 0.004106979906283598 |
+| Epoch_10_batch_5999.pt | 0.9323333333333335 | 0.0038264834128272315 |
+| Epoch_14_batch_2999.pt | 0.9318333333333333 | 0.004009633461281366 |
+| Epoch_17_batch_5999.pt | 0.9313333333333335 | 0.004042978977480057 |
+| Epoch_10.pt | 0.9305 | 0.0037437190197403204 |
+| Epoch_11.pt | 0.9303333333333335 | 0.0032470309324894365 |
+| Epoch_8_batch_2999.pt | 0.9258333333333333 | 0.004210876569252709 |
+| Epoch_9_batch_2999.pt | 0.924 | 0.0039063101847215415 |
+| Epoch_9_batch_5999.pt | 0.9235 | 0.0046044128055541065 |
+| Epoch_7_batch_2999.pt | 0.9205 | 0.004224049312282305 |
+| Epoch_6_batch_5999.pt | 0.9196666666666667 | 0.0038554114606438906 |
+| Epoch_7_batch_5999.pt | 0.9191666666666667 | 0.003938179688543849 |
+| Epoch_9.pt | 0.9185000000000001 | 0.0030066798061632923 |
+| Epoch_6_batch_2999.pt | 0.9184999999999999 | 0.004313701970623148 |
+| Epoch_8_batch_5999.pt | 0.9176666666666667 | 0.004708804466759359 |
+| Epoch_5_batch_5999.pt | 0.9164999999999999 | 0.0036213529972738464 |
+| Epoch_5_batch_2999.pt | 0.9161666666666667 | 0.0038892856940415887 |
+| Epoch_6.pt | 0.915 | 0.00493788578739755 |
+| Epoch_4_batch_2999.pt | 0.9143333333333332 | 0.004876246279442591 |
+| Epoch_4_batch_5999.pt | 0.9141666666666666 | 0.004576172857824694 |
+| Epoch_3_batch_2999.pt | 0.9131666666666668 | 0.004489012649559056 |
+| Epoch_3_batch_5999.pt | 0.9123333333333333 | 0.004438885412319597 |
+| Epoch_8.pt | 0.9116666666666667 | 0.0032203059435976585 |
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..77c0c59c351833d94f6c957b264354a0f1eca699
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_cplfw.txt
@@ -0,0 +1,43 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.8385 | 0.005684340629753094 |
+| Epoch_14.pt | 0.8375 | 0.0053934688441521095 |
+| Epoch_14_batch_5999.pt | 0.8373333333333333 | 0.006000000000000003 |
+| Epoch_16_batch_5999.pt | 0.8371666666666666 | 0.0060402456821916275 |
+| Epoch_17_batch_2999.pt | 0.8363333333333335 | 0.005777777777777781 |
+| Epoch_17.pt | 0.8361666666666666 | 0.006671526006788792 |
+| Epoch_13.pt | 0.8351666666666666 | 0.0052260624178306365 |
+| Epoch_15_batch_5999.pt | 0.8351666666666666 | 0.0055413707801821735 |
+| Epoch_17_batch_5999.pt | 0.8348333333333333 | 0.0057544936815382974 |
+| Epoch_13_batch_2999.pt | 0.834 | 0.005438772555786846 |
+| Epoch_13_batch_5999.pt | 0.834 | 0.0049328828623162475 |
+| Epoch_15_batch_2999.pt | 0.8338333333333334 | 0.0059111215538754855 |
+| Epoch_16.pt | 0.8338333333333333 | 0.005900669558157607 |
+| Epoch_12_batch_2999.pt | 0.8335000000000001 | 0.005829098992273877 |
+| Epoch_14_batch_2999.pt | 0.8328333333333333 | 0.005660399584551444 |
+| Epoch_15.pt | 0.8328333333333333 | 0.005409468151215271 |
+| Epoch_12_batch_5999.pt | 0.8326666666666667 | 0.005710073403346489 |
+| Epoch_11_batch_5999.pt | 0.8311666666666667 | 0.005784451274166793 |
+| Epoch_11_batch_2999.pt | 0.829 | 0.00595870563457419 |
+| Epoch_10.pt | 0.8288333333333332 | 0.0051162413865967985 |
+| Epoch_10_batch_5999.pt | 0.8275 | 0.006296432460401289 |
+| Epoch_11.pt | 0.8275 | 0.0063452616714373905 |
+| Epoch_12.pt | 0.8273333333333334 | 0.00491407653055466 |
+| Epoch_10_batch_2999.pt | 0.8258333333333333 | 0.0064274200419034136 |
+| Epoch_9_batch_5999.pt | 0.8066666666666666 | 0.005868938953886336 |
+| Epoch_8_batch_2999.pt | 0.8063333333333332 | 0.006200358412579425 |
+| Epoch_7_batch_2999.pt | 0.8059999999999998 | 0.007462664271203106 |
+| Epoch_9_batch_2999.pt | 0.8056666666666666 | 0.007931808132255443 |
+| Epoch_7_batch_5999.pt | 0.8035 | 0.00650948974883005 |
+| Epoch_6_batch_5999.pt | 0.8005000000000001 | 0.007205322107072427 |
+| Epoch_8_batch_5999.pt | 0.8003333333333333 | 0.006642548967837832 |
+| Epoch_9.pt | 0.8003333333333333 | 0.007113714801115158 |
+| Epoch_5.pt | 0.799 | 0.006555555555555556 |
+| Epoch_6_batch_2999.pt | 0.7979999999999999 | 0.007551469483353864 |
+| Epoch_5_batch_2999.pt | 0.7963333333333333 | 0.006323579233392089 |
+| Epoch_5_batch_5999.pt | 0.7959999999999999 | 0.005283330412430556 |
+| Epoch_4_batch_2999.pt | 0.7945000000000001 | 0.006921294747925014 |
+| Epoch_3_batch_5999.pt | 0.7914999999999999 | 0.007819136070493272 |
+| Epoch_4_batch_5999.pt | 0.7898333333333334 | 0.005897530347317501 |
+| Epoch_8.pt | 0.7885 | 0.00936387199642499 |
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..279aeeac9e06f700967e10ac05520ab92cf4abee
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11_batch_2999.pt | 0.9958333333333333 | 0.0011180339887498947 |
+| Epoch_12_batch_5999.pt | 0.9958333333333333 | 0.0009702360664762748 |
+| Epoch_13_batch_2999.pt | 0.9956666666666667 | 0.0010599324460188284 |
+| Epoch_14.pt | 0.9956666666666665 | 0.0010304020550550757 |
+| Epoch_13.pt | 0.9955 | 0.0009953596037316052 |
+| Epoch_12.pt | 0.9953333333333332 | 0.0010772621905369623 |
+| Epoch_16.pt | 0.9951666666666666 | 0.001177201116689836 |
+| Epoch_17_batch_5999.pt | 0.9951666666666666 | 0.0011506841765115488 |
+| Epoch_14_batch_2999.pt | 0.9949999999999999 | 0.0011915339216404008 |
+| Epoch_15_batch_5999.pt | 0.9949999999999999 | 0.0011653431646334986 |
+| Epoch_16_batch_2999.pt | 0.9949999999999999 | 0.0011653431646334986 |
+| Epoch_8_batch_2999.pt | 0.9948333333333335 | 0.0010671873729054724 |
+| Epoch_15_batch_2999.pt | 0.9948333333333335 | 0.0010957268290731133 |
+| Epoch_11.pt | 0.9948333333333332 | 0.0012031337682059844 |
+| Epoch_12_batch_2999.pt | 0.9948333333333332 | 0.001228519132638664 |
+| Epoch_13_batch_5999.pt | 0.9948333333333332 | 0.00112354157867537 |
+| Epoch_15.pt | 0.9948333333333332 | 0.0012777777777777798 |
+| Epoch_17_batch_2999.pt | 0.9948333333333332 | 0.001228519132638664 |
+| Epoch_9_batch_5999.pt | 0.9946666666666667 | 0.001160034056545619 |
+| Epoch_14_batch_5999.pt | 0.9946666666666667 | 0.0013099806802835115 |
+| Epoch_11_batch_5999.pt | 0.9945000000000002 | 0.0011928283640879947 |
+| Epoch_10_batch_2999.pt | 0.9945 | 0.0011928283640879975 |
+| Epoch_9_batch_2999.pt | 0.9944999999999998 | 0.0011928283640879926 |
+| Epoch_17.pt | 0.9944999999999998 | 0.0013619611857923642 |
+| Epoch_8_batch_5999.pt | 0.9943333333333333 | 0.0010000000000000005 |
+| Epoch_10_batch_5999.pt | 0.9943333333333333 | 0.0012222222222222235 |
+| Epoch_9.pt | 0.9941666666666666 | 0.0010900787150193644 |
+| Epoch_16_batch_5999.pt | 0.9941666666666666 | 0.001320540480444972 |
+| Epoch_5_batch_2999.pt | 0.994 | 0.0014315665251916822 |
+| Epoch_10.pt | 0.994 | 0.0012957670877434006 |
+| Epoch_4_batch_2999.pt | 0.9933333333333334 | 0.0014054567378526145 |
+| Epoch_3_batch_2999.pt | 0.9931666666666666 | 0.0012777777777777755 |
+| Epoch_7_batch_2999.pt | 0.9931666666666666 | 0.00127777777777778 |
+| Epoch_4_batch_5999.pt | 0.993 | 0.001356283957303744 |
+| Epoch_5.pt | 0.993 | 0.001286204100310026 |
+| Epoch_6_batch_5999.pt | 0.993 | 0.0011331154474650564 |
+| Epoch_7_batch_5999.pt | 0.9926666666666668 | 0.0014740554623801788 |
+| Epoch_5_batch_5999.pt | 0.9925 | 0.0011180339887498921 |
+| Epoch_6.pt | 0.9925 | 0.0014540280364780456 |
+| Epoch_6_batch_2999.pt | 0.9921666666666666 | 0.001428328903575827 |
+| Epoch_8.pt | 0.9921666666666666 | 0.0014065543223524709 |
+| Epoch_4.pt | 0.992 | 0.0016063146994223286 |
+| Epoch_7.pt | 0.992 | 0.0016063146994223327 |
+| Epoch_3_batch_5999.pt | 0.9916666666666668 | 0.0014487116456005907 |
+| Epoch_2_batch_5999.pt | 0.9914999999999999 | 0.0012777777777777809 |
+| Epoch_2_batch_2999.pt | 0.9913333333333334 | 0.0011863420280034743 |
+| Epoch_3.pt | 0.9908333333333333 | 0.0017258027296676729 |
+| Epoch_2.pt | 0.9901666666666668 | 0.0012285191326386622 |
+| Epoch_1_batch_5999.pt | 0.9896666666666667 | 0.0013099806802835065 |
+| Epoch_1.pt | 0.9871666666666666 | 0.0020038543107634924 |
+| Epoch_1_batch_2999.pt | 0.9831666666666667 | 0.0021147629234082527 |
+| Epoch_0.pt | 0.9744999999999999 | 0.002485141027371667 |
+| Epoch_0_batch_5999.pt | 0.9718333333333333 | 0.0016377114414426305 |
+| Epoch_0_batch_2999.pt | 0.9435 | 0.003309638001912549 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ba183c6ed35f47223876b0044145132491fa6c87
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_African.txt
@@ -0,0 +1,21 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.8789999999999999 | 0.004939135726865079 |
+| Epoch_14_batch_5999.pt | 0.8783333333333333 | 0.004950370979987881 |
+| Epoch_15_batch_5999.pt | 0.8783333333333333 | 0.00481766297526195 |
+| Epoch_16_batch_2999.pt | 0.8783333333333333 | 0.004950370979987879 |
+| Epoch_17_batch_2999.pt | 0.8781666666666668 | 0.005580223014261067 |
+| Epoch_12.pt | 0.8776666666666667 | 0.0052009021059859824 |
+| Epoch_13_batch_2999.pt | 0.8773333333333333 | 0.004831736645590869 |
+| Epoch_15.pt | 0.8771666666666667 | 0.004628481339075788 |
+| Epoch_16_batch_5999.pt | 0.8770000000000001 | 0.004477653706579977 |
+| Epoch_13.pt | 0.8761666666666666 | 0.005085988984515919 |
+| Epoch_16.pt | 0.8761666666666666 | 0.00409041336315555 |
+| Epoch_13_batch_5999.pt | 0.8756666666666668 | 0.004939135726865077 |
+| Epoch_14_batch_2999.pt | 0.8755000000000001 | 0.004759979769642661 |
+| Epoch_14.pt | 0.8748333333333334 | 0.004219662967057384 |
+| Epoch_11_batch_5999.pt | 0.8743333333333334 | 0.00448729344584637 |
+| Epoch_17.pt | 0.8743333333333332 | 0.0050564712235478875 |
+| Epoch_11_batch_2999.pt | 0.874 | 0.005318265752788587 |
+| Epoch_15_batch_2999.pt | 0.874 | 0.005747785402834477 |
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7eaa98838e7405b5bc4c5fb544458c3970115f5e
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,19 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_2999.pt | 0.8836666666666666 | 0.004080970593823769 |
+| Epoch_15_batch_2999.pt | 0.883 | 0.00402768199119819 |
+| Epoch_16_batch_2999.pt | 0.883 | 0.0038232556742411714 |
+| Epoch_14_batch_5999.pt | 0.8818333333333334 | 0.004794866081627384 |
+| Epoch_17_batch_5999.pt | 0.8816666666666666 | 0.004135065347984088 |
+| Epoch_13_batch_5999.pt | 0.8815000000000002 | 0.004461425891758637 |
+| Epoch_16.pt | 0.881 | 0.004466611387164843 |
+| Epoch_13.pt | 0.8805 | 0.0030434102055116362 |
+| Epoch_15_batch_5999.pt | 0.8798333333333334 | 0.003916174121906021 |
+| Epoch_14.pt | 0.8785000000000001 | 0.004617799700514747 |
+| Epoch_15.pt | 0.8781666666666667 | 0.0038123418809550592 |
+| Epoch_17.pt | 0.8781666666666667 | 0.004123479892003004 |
+| Epoch_14_batch_2999.pt | 0.8780000000000001 | 0.003564225540521213 |
+| Epoch_16_batch_5999.pt | 0.8768333333333335 | 0.0038365525934710176 |
+| Epoch_11_batch_2999.pt | 0.8756666666666668 | 0.0039534326388127065 |
+| Epoch_13_batch_2999.pt | 0.875 | 0.004492792582034612 |
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..5f9a2b51c0624ab9ed9d387cb0928a3d06f5dc88
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,21 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.9531666666666666 | 0.0032054159415003686 |
+| Epoch_15.pt | 0.9526666666666666 | 0.004129089816902677 |
+| Epoch_12.pt | 0.9514999999999999 | 0.0022696331350906184 |
+| Epoch_15_batch_2999.pt | 0.9511666666666667 | 0.003043410205511633 |
+| Epoch_16_batch_2999.pt | 0.9511666666666667 | 0.003600840094021955 |
+| Epoch_14_batch_2999.pt | 0.951 | 0.0036021255727660753 |
+| Epoch_14_batch_5999.pt | 0.951 | 0.0032317865716108844 |
+| Epoch_16_batch_5999.pt | 0.9508333333333333 | 0.003745367509040707 |
+| Epoch_13_batch_5999.pt | 0.9506666666666665 | 0.0033993463423951913 |
+| Epoch_16.pt | 0.9506666666666665 | 0.0029938207967349904 |
+| Epoch_17_batch_5999.pt | 0.9506666666666665 | 0.003307305792474936 |
+| Epoch_11_batch_5999.pt | 0.9504999999999999 | 0.0038733817807896378 |
+| Epoch_11.pt | 0.9504999999999999 | 0.0034787716009896078 |
+| Epoch_17_batch_2999.pt | 0.9503333333333334 | 0.0037745083892139998 |
+| Epoch_13_batch_2999.pt | 0.9501666666666665 | 0.0029860788111948223 |
+| Epoch_17.pt | 0.9498333333333333 | 0.0034645470728152973 |
+| Epoch_10_batch_5999.pt | 0.9494999999999999 | 0.0032015621187164224 |
+| Epoch_11_batch_2999.pt | 0.9493333333333334 | 0.0028458329944145992 |
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7be65d66610d828241e96dec9ffb086f40685296
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,21 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.9113333333333333 | 0.0036750745352313874 |
+| Epoch_16.pt | 0.9108333333333334 | 0.0029000851413640387 |
+| Epoch_17.pt | 0.9105000000000001 | 0.0031822229981377115 |
+| Epoch_16_batch_2999.pt | 0.9099999999999999 | 0.0033701668640229083 |
+| Epoch_16_batch_5999.pt | 0.9099999999999999 | 0.0034426518632954812 |
+| Epoch_13.pt | 0.9098333333333333 | 0.003137290645221223 |
+| Epoch_14.pt | 0.9096666666666667 | 0.0030510067150546676 |
+| Epoch_15_batch_2999.pt | 0.9096666666666667 | 0.003111111111111112 |
+| Epoch_17_batch_2999.pt | 0.9096666666666666 | 0.003440858348267116 |
+| Epoch_14_batch_5999.pt | 0.9094999999999999 | 0.0036179422567033156 |
+| Epoch_13_batch_2999.pt | 0.9093333333333333 | 0.003212629398844659 |
+| Epoch_12_batch_5999.pt | 0.9091666666666667 | 0.0028894230275435186 |
+| Epoch_14_batch_2999.pt | 0.9091666666666667 | 0.0029528182813151806 |
+| Epoch_12.pt | 0.9088333333333333 | 0.004332264825529307 |
+| Epoch_17_batch_5999.pt | 0.9083333333333334 | 0.003063121944908938 |
+| Epoch_12_batch_2999.pt | 0.908 | 0.0031111111111111075 |
+| Epoch_15_batch_5999.pt | 0.9076666666666666 | 0.0031836775070876442 |
+| Epoch_11_batch_5999.pt | 0.9075000000000001 | 0.002620550314460163 |
diff --git a/bob/bio/facexzoo/models/heads/AdaM-Softmax/log.log b/bob/bio/facexzoo/models/heads/AdaM-Softmax/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..84d56aef98f465a5a4a11bae22f3e8d8ed268ab3
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/AdaM-Softmax/log.log
@@ -0,0 +1,655 @@
+INFO 2020-11-27 11:25:25 train.py: 177] Start optimization.
+INFO 2020-11-27 11:25:25 train.py: 178] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='Adam_Softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='adam-mobile', train_file='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7fb4996eeef0>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'scale': 32, 'lamda': 70.0}
+INFO 2020-11-27 11:25:45 train.py: 79] Epoch 0, iter 0/6416, lr 0.100000, loss 68.895447
+INFO 2020-11-27 11:27:02 train.py: 79] Epoch 0, iter 200/6416, lr 0.100000, loss 66.453452
+INFO 2020-11-27 11:28:18 train.py: 79] Epoch 0, iter 400/6416, lr 0.100000, loss 61.217521
+INFO 2020-11-27 11:29:35 train.py: 79] Epoch 0, iter 600/6416, lr 0.100000, loss 56.275663
+INFO 2020-11-27 11:30:51 train.py: 79] Epoch 0, iter 800/6416, lr 0.100000, loss 51.401999
+INFO 2020-11-27 11:32:07 train.py: 79] Epoch 0, iter 1000/6416, lr 0.100000, loss 47.736430
+INFO 2020-11-27 11:33:24 train.py: 79] Epoch 0, iter 1200/6416, lr 0.100000, loss 45.576048
+INFO 2020-11-27 11:34:40 train.py: 79] Epoch 0, iter 1400/6416, lr 0.100000, loss 44.220327
+INFO 2020-11-27 11:35:57 train.py: 79] Epoch 0, iter 1600/6416, lr 0.100000, loss 43.221189
+INFO 2020-11-27 11:37:14 train.py: 79] Epoch 0, iter 1800/6416, lr 0.100000, loss 42.476572
+INFO 2020-11-27 11:38:30 train.py: 79] Epoch 0, iter 2000/6416, lr 0.100000, loss 41.735169
+INFO 2020-11-27 11:39:47 train.py: 79] Epoch 0, iter 2200/6416, lr 0.100000, loss 41.068170
+INFO 2020-11-27 11:41:03 train.py: 79] Epoch 0, iter 2400/6416, lr 0.100000, loss 40.607779
+INFO 2020-11-27 11:42:20 train.py: 79] Epoch 0, iter 2600/6416, lr 0.100000, loss 40.068805
+INFO 2020-11-27 11:43:36 train.py: 79] Epoch 0, iter 2800/6416, lr 0.100000, loss 39.633740
+INFO 2020-11-27 11:44:53 train.py: 92] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-27 11:44:53 train.py: 79] Epoch 0, iter 3000/6416, lr 0.100000, loss 39.079019
+INFO 2020-11-27 11:46:10 train.py: 79] Epoch 0, iter 3200/6416, lr 0.100000, loss 38.623620
+INFO 2020-11-27 11:47:28 train.py: 79] Epoch 0, iter 3400/6416, lr 0.100000, loss 38.154093
+INFO 2020-11-27 11:48:45 train.py: 79] Epoch 0, iter 3600/6416, lr 0.100000, loss 37.649786
+INFO 2020-11-27 11:50:02 train.py: 79] Epoch 0, iter 3800/6416, lr 0.100000, loss 37.255512
+INFO 2020-11-27 11:51:19 train.py: 79] Epoch 0, iter 4000/6416, lr 0.100000, loss 36.827047
+INFO 2020-11-27 11:52:37 train.py: 79] Epoch 0, iter 4200/6416, lr 0.100000, loss 36.372875
+INFO 2020-11-27 11:53:54 train.py: 79] Epoch 0, iter 4400/6416, lr 0.100000, loss 35.929390
+INFO 2020-11-27 11:55:11 train.py: 79] Epoch 0, iter 4600/6416, lr 0.100000, loss 35.571391
+INFO 2020-11-27 11:56:28 train.py: 79] Epoch 0, iter 4800/6416, lr 0.100000, loss 35.173256
+INFO 2020-11-27 11:57:45 train.py: 79] Epoch 0, iter 5000/6416, lr 0.100000, loss 34.815217
+INFO 2020-11-27 11:59:03 train.py: 79] Epoch 0, iter 5200/6416, lr 0.100000, loss 34.404765
+INFO 2020-11-27 12:00:20 train.py: 79] Epoch 0, iter 5400/6416, lr 0.100000, loss 34.022353
+INFO 2020-11-27 12:01:37 train.py: 79] Epoch 0, iter 5600/6416, lr 0.100000, loss 33.711045
+INFO 2020-11-27 12:02:55 train.py: 79] Epoch 0, iter 5800/6416, lr 0.100000, loss 33.412540
+INFO 2020-11-27 12:04:11 train.py: 92] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-27 12:04:12 train.py: 79] Epoch 0, iter 6000/6416, lr 0.100000, loss 33.060440
+INFO 2020-11-27 12:05:29 train.py: 79] Epoch 0, iter 6200/6416, lr 0.100000, loss 32.789185
+INFO 2020-11-27 12:06:46 train.py: 79] Epoch 0, iter 6400/6416, lr 0.100000, loss 32.510369
+INFO 2020-11-27 12:06:52 train.py: 97] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-27 12:06:54 train.py: 79] Epoch 1, iter 0/6416, lr 0.100000, loss 32.480811
+INFO 2020-11-27 12:08:11 train.py: 79] Epoch 1, iter 200/6416, lr 0.100000, loss 31.700433
+INFO 2020-11-27 12:09:29 train.py: 79] Epoch 1, iter 400/6416, lr 0.100000, loss 31.576547
+INFO 2020-11-27 12:10:46 train.py: 79] Epoch 1, iter 600/6416, lr 0.100000, loss 31.354598
+INFO 2020-11-27 12:12:03 train.py: 79] Epoch 1, iter 800/6416, lr 0.100000, loss 31.224222
+INFO 2020-11-27 12:13:21 train.py: 79] Epoch 1, iter 1000/6416, lr 0.100000, loss 31.056949
+INFO 2020-11-27 12:14:38 train.py: 79] Epoch 1, iter 1200/6416, lr 0.100000, loss 30.974754
+INFO 2020-11-27 12:15:55 train.py: 79] Epoch 1, iter 1400/6416, lr 0.100000, loss 30.816703
+INFO 2020-11-27 12:17:12 train.py: 79] Epoch 1, iter 1600/6416, lr 0.100000, loss 30.637983
+INFO 2020-11-27 12:18:29 train.py: 79] Epoch 1, iter 1800/6416, lr 0.100000, loss 30.525500
+INFO 2020-11-27 12:19:47 train.py: 79] Epoch 1, iter 2000/6416, lr 0.100000, loss 30.331987
+INFO 2020-11-27 12:21:04 train.py: 79] Epoch 1, iter 2200/6416, lr 0.100000, loss 30.243652
+INFO 2020-11-27 12:22:21 train.py: 79] Epoch 1, iter 2400/6416, lr 0.100000, loss 30.118532
+INFO 2020-11-27 12:23:39 train.py: 79] Epoch 1, iter 2600/6416, lr 0.100000, loss 30.025257
+INFO 2020-11-27 12:24:56 train.py: 79] Epoch 1, iter 2800/6416, lr 0.100000, loss 29.925623
+INFO 2020-11-27 12:26:13 train.py: 92] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-27 12:26:13 train.py: 79] Epoch 1, iter 3000/6416, lr 0.100000, loss 29.799252
+INFO 2020-11-27 12:27:30 train.py: 79] Epoch 1, iter 3200/6416, lr 0.100000, loss 29.739019
+INFO 2020-11-27 12:28:47 train.py: 79] Epoch 1, iter 3400/6416, lr 0.100000, loss 29.618822
+INFO 2020-11-27 12:30:03 train.py: 79] Epoch 1, iter 3600/6416, lr 0.100000, loss 29.564026
+INFO 2020-11-27 12:31:20 train.py: 79] Epoch 1, iter 3800/6416, lr 0.100000, loss 29.499074
+INFO 2020-11-27 12:32:36 train.py: 79] Epoch 1, iter 4000/6416, lr 0.100000, loss 29.360484
+INFO 2020-11-27 12:33:53 train.py: 79] Epoch 1, iter 4200/6416, lr 0.100000, loss 29.312003
+INFO 2020-11-27 12:35:10 train.py: 79] Epoch 1, iter 4400/6416, lr 0.100000, loss 29.241322
+INFO 2020-11-27 12:36:26 train.py: 79] Epoch 1, iter 4600/6416, lr 0.100000, loss 29.153561
+INFO 2020-11-27 12:37:43 train.py: 79] Epoch 1, iter 4800/6416, lr 0.100000, loss 29.087117
+INFO 2020-11-27 12:38:59 train.py: 79] Epoch 1, iter 5000/6416, lr 0.100000, loss 29.070027
+INFO 2020-11-27 12:40:16 train.py: 79] Epoch 1, iter 5200/6416, lr 0.100000, loss 29.030273
+INFO 2020-11-27 12:41:32 train.py: 79] Epoch 1, iter 5400/6416, lr 0.100000, loss 28.929905
+INFO 2020-11-27 12:42:49 train.py: 79] Epoch 1, iter 5600/6416, lr 0.100000, loss 28.869163
+INFO 2020-11-27 12:44:06 train.py: 79] Epoch 1, iter 5800/6416, lr 0.100000, loss 28.823680
+INFO 2020-11-27 12:45:22 train.py: 92] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-27 12:45:23 train.py: 79] Epoch 1, iter 6000/6416, lr 0.100000, loss 28.832928
+INFO 2020-11-27 12:46:40 train.py: 79] Epoch 1, iter 6200/6416, lr 0.100000, loss 28.736330
+INFO 2020-11-27 12:47:57 train.py: 79] Epoch 1, iter 6400/6416, lr 0.100000, loss 28.747449
+INFO 2020-11-27 12:48:03 train.py: 97] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-27 12:48:05 train.py: 79] Epoch 2, iter 0/6416, lr 0.100000, loss 28.593937
+INFO 2020-11-27 12:49:22 train.py: 79] Epoch 2, iter 200/6416, lr 0.100000, loss 28.173371
+INFO 2020-11-27 12:50:39 train.py: 79] Epoch 2, iter 400/6416, lr 0.100000, loss 28.174336
+INFO 2020-11-27 12:51:56 train.py: 79] Epoch 2, iter 600/6416, lr 0.100000, loss 28.204712
+INFO 2020-11-27 12:53:14 train.py: 79] Epoch 2, iter 800/6416, lr 0.100000, loss 28.273016
+INFO 2020-11-27 12:54:31 train.py: 79] Epoch 2, iter 1000/6416, lr 0.100000, loss 28.321223
+INFO 2020-11-27 12:55:48 train.py: 79] Epoch 2, iter 1200/6416, lr 0.100000, loss 28.311304
+INFO 2020-11-27 12:57:05 train.py: 79] Epoch 2, iter 1400/6416, lr 0.100000, loss 28.303212
+INFO 2020-11-27 12:58:22 train.py: 79] Epoch 2, iter 1600/6416, lr 0.100000, loss 28.308795
+INFO 2020-11-27 12:59:40 train.py: 79] Epoch 2, iter 1800/6416, lr 0.100000, loss 28.303300
+INFO 2020-11-27 13:00:57 train.py: 79] Epoch 2, iter 2000/6416, lr 0.100000, loss 28.267853
+INFO 2020-11-27 13:02:14 train.py: 79] Epoch 2, iter 2200/6416, lr 0.100000, loss 28.245044
+INFO 2020-11-27 13:03:31 train.py: 79] Epoch 2, iter 2400/6416, lr 0.100000, loss 28.236938
+INFO 2020-11-27 13:04:48 train.py: 79] Epoch 2, iter 2600/6416, lr 0.100000, loss 28.230038
+INFO 2020-11-27 13:06:05 train.py: 79] Epoch 2, iter 2800/6416, lr 0.100000, loss 28.171071
+INFO 2020-11-27 13:07:23 train.py: 92] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-27 13:07:23 train.py: 79] Epoch 2, iter 3000/6416, lr 0.100000, loss 28.178500
+INFO 2020-11-27 13:08:40 train.py: 79] Epoch 2, iter 3200/6416, lr 0.100000, loss 28.163017
+INFO 2020-11-27 13:09:57 train.py: 79] Epoch 2, iter 3400/6416, lr 0.100000, loss 28.155609
+INFO 2020-11-27 13:11:15 train.py: 79] Epoch 2, iter 3600/6416, lr 0.100000, loss 28.132254
+INFO 2020-11-27 13:12:32 train.py: 79] Epoch 2, iter 3800/6416, lr 0.100000, loss 28.093806
+INFO 2020-11-27 13:13:49 train.py: 79] Epoch 2, iter 4000/6416, lr 0.100000, loss 28.087449
+INFO 2020-11-27 13:15:06 train.py: 79] Epoch 2, iter 4200/6416, lr 0.100000, loss 28.073048
+INFO 2020-11-27 13:16:23 train.py: 79] Epoch 2, iter 4400/6416, lr 0.100000, loss 28.027729
+INFO 2020-11-27 13:17:41 train.py: 79] Epoch 2, iter 4600/6416, lr 0.100000, loss 28.014697
+INFO 2020-11-27 13:18:58 train.py: 79] Epoch 2, iter 4800/6416, lr 0.100000, loss 27.983244
+INFO 2020-11-27 13:20:15 train.py: 79] Epoch 2, iter 5000/6416, lr 0.100000, loss 27.917437
+INFO 2020-11-27 13:21:32 train.py: 79] Epoch 2, iter 5200/6416, lr 0.100000, loss 27.934515
+INFO 2020-11-27 13:22:49 train.py: 79] Epoch 2, iter 5400/6416, lr 0.100000, loss 27.890959
+INFO 2020-11-27 13:24:07 train.py: 79] Epoch 2, iter 5600/6416, lr 0.100000, loss 27.905500
+INFO 2020-11-27 13:25:24 train.py: 79] Epoch 2, iter 5800/6416, lr 0.100000, loss 27.885338
+INFO 2020-11-27 13:26:41 train.py: 92] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-27 13:26:42 train.py: 79] Epoch 2, iter 6000/6416, lr 0.100000, loss 27.911852
+INFO 2020-11-27 13:27:58 train.py: 79] Epoch 2, iter 6200/6416, lr 0.100000, loss 27.845165
+INFO 2020-11-27 13:29:15 train.py: 79] Epoch 2, iter 6400/6416, lr 0.100000, loss 27.785669
+INFO 2020-11-27 13:29:21 train.py: 97] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-27 13:29:22 train.py: 79] Epoch 3, iter 0/6416, lr 0.100000, loss 27.794654
+INFO 2020-11-27 13:30:40 train.py: 79] Epoch 3, iter 200/6416, lr 0.100000, loss 27.331309
+INFO 2020-11-27 13:31:57 train.py: 79] Epoch 3, iter 400/6416, lr 0.100000, loss 27.304826
+INFO 2020-11-27 13:33:14 train.py: 79] Epoch 3, iter 600/6416, lr 0.100000, loss 27.456631
+INFO 2020-11-27 13:34:32 train.py: 79] Epoch 3, iter 800/6416, lr 0.100000, loss 27.456618
+INFO 2020-11-27 13:35:49 train.py: 79] Epoch 3, iter 1000/6416, lr 0.100000, loss 27.560082
+INFO 2020-11-27 13:37:06 train.py: 79] Epoch 3, iter 1200/6416, lr 0.100000, loss 27.578836
+INFO 2020-11-27 13:38:23 train.py: 79] Epoch 3, iter 1400/6416, lr 0.100000, loss 27.584422
+INFO 2020-11-27 13:39:41 train.py: 79] Epoch 3, iter 1600/6416, lr 0.100000, loss 27.597026
+INFO 2020-11-27 13:40:58 train.py: 79] Epoch 3, iter 1800/6416, lr 0.100000, loss 27.625970
+INFO 2020-11-27 13:42:15 train.py: 79] Epoch 3, iter 2000/6416, lr 0.100000, loss 27.590219
+INFO 2020-11-27 13:43:32 train.py: 79] Epoch 3, iter 2200/6416, lr 0.100000, loss 27.541043
+INFO 2020-11-27 13:44:50 train.py: 79] Epoch 3, iter 2400/6416, lr 0.100000, loss 27.565358
+INFO 2020-11-27 13:46:07 train.py: 79] Epoch 3, iter 2600/6416, lr 0.100000, loss 27.567240
+INFO 2020-11-27 13:47:24 train.py: 79] Epoch 3, iter 2800/6416, lr 0.100000, loss 27.575654
+INFO 2020-11-27 13:48:41 train.py: 92] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-27 13:48:41 train.py: 79] Epoch 3, iter 3000/6416, lr 0.100000, loss 27.545376
+INFO 2020-11-27 13:49:59 train.py: 79] Epoch 3, iter 3200/6416, lr 0.100000, loss 27.524548
+INFO 2020-11-27 13:51:16 train.py: 79] Epoch 3, iter 3400/6416, lr 0.100000, loss 27.543726
+INFO 2020-11-27 13:52:33 train.py: 79] Epoch 3, iter 3600/6416, lr 0.100000, loss 27.506469
+INFO 2020-11-27 13:53:50 train.py: 79] Epoch 3, iter 3800/6416, lr 0.100000, loss 27.515937
+INFO 2020-11-27 13:55:07 train.py: 79] Epoch 3, iter 4000/6416, lr 0.100000, loss 27.512061
+INFO 2020-11-27 13:56:25 train.py: 79] Epoch 3, iter 4200/6416, lr 0.100000, loss 27.492437
+INFO 2020-11-27 13:57:42 train.py: 79] Epoch 3, iter 4400/6416, lr 0.100000, loss 27.499637
+INFO 2020-11-27 13:58:59 train.py: 79] Epoch 3, iter 4600/6416, lr 0.100000, loss 27.467361
+INFO 2020-11-27 14:00:16 train.py: 79] Epoch 3, iter 4800/6416, lr 0.100000, loss 27.465655
+INFO 2020-11-27 14:01:34 train.py: 79] Epoch 3, iter 5000/6416, lr 0.100000, loss 27.470464
+INFO 2020-11-27 14:02:51 train.py: 79] Epoch 3, iter 5200/6416, lr 0.100000, loss 27.462713
+INFO 2020-11-27 14:04:08 train.py: 79] Epoch 3, iter 5400/6416, lr 0.100000, loss 27.437865
+INFO 2020-11-27 14:05:25 train.py: 79] Epoch 3, iter 5600/6416, lr 0.100000, loss 27.378889
+INFO 2020-11-27 14:06:43 train.py: 79] Epoch 3, iter 5800/6416, lr 0.100000, loss 27.368946
+INFO 2020-11-27 14:08:00 train.py: 92] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-27 14:08:00 train.py: 79] Epoch 3, iter 6000/6416, lr 0.100000, loss 27.376025
+INFO 2020-11-27 14:09:17 train.py: 79] Epoch 3, iter 6200/6416, lr 0.100000, loss 27.379211
+INFO 2020-11-27 14:10:35 train.py: 79] Epoch 3, iter 6400/6416, lr 0.100000, loss 27.362750
+INFO 2020-11-27 14:10:41 train.py: 97] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-27 14:10:43 train.py: 79] Epoch 4, iter 0/6416, lr 0.100000, loss 27.380721
+INFO 2020-11-27 14:11:59 train.py: 79] Epoch 4, iter 200/6416, lr 0.100000, loss 26.836015
+INFO 2020-11-27 14:13:16 train.py: 79] Epoch 4, iter 400/6416, lr 0.100000, loss 26.903098
+INFO 2020-11-27 14:14:32 train.py: 79] Epoch 4, iter 600/6416, lr 0.100000, loss 27.023604
+INFO 2020-11-27 14:15:49 train.py: 79] Epoch 4, iter 800/6416, lr 0.100000, loss 27.070077
+INFO 2020-11-27 14:17:06 train.py: 79] Epoch 4, iter 1000/6416, lr 0.100000, loss 27.093762
+INFO 2020-11-27 14:18:22 train.py: 79] Epoch 4, iter 1200/6416, lr 0.100000, loss 27.167080
+INFO 2020-11-27 14:19:39 train.py: 79] Epoch 4, iter 1400/6416, lr 0.100000, loss 27.156359
+INFO 2020-11-27 14:20:56 train.py: 79] Epoch 4, iter 1600/6416, lr 0.100000, loss 27.208953
+INFO 2020-11-27 14:22:12 train.py: 79] Epoch 4, iter 1800/6416, lr 0.100000, loss 27.183998
+INFO 2020-11-27 14:23:29 train.py: 79] Epoch 4, iter 2000/6416, lr 0.100000, loss 27.181064
+INFO 2020-11-27 14:24:45 train.py: 79] Epoch 4, iter 2200/6416, lr 0.100000, loss 27.214739
+INFO 2020-11-27 14:26:02 train.py: 79] Epoch 4, iter 2400/6416, lr 0.100000, loss 27.206284
+INFO 2020-11-27 14:27:19 train.py: 79] Epoch 4, iter 2600/6416, lr 0.100000, loss 27.197754
+INFO 2020-11-27 14:28:35 train.py: 79] Epoch 4, iter 2800/6416, lr 0.100000, loss 27.188528
+INFO 2020-11-27 14:29:51 train.py: 92] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-27 14:29:52 train.py: 79] Epoch 4, iter 3000/6416, lr 0.100000, loss 27.211198
+INFO 2020-11-27 14:31:09 train.py: 79] Epoch 4, iter 3200/6416, lr 0.100000, loss 27.214330
+INFO 2020-11-27 14:32:26 train.py: 79] Epoch 4, iter 3400/6416, lr 0.100000, loss 27.167536
+INFO 2020-11-27 14:33:44 train.py: 79] Epoch 4, iter 3600/6416, lr 0.100000, loss 27.204409
+INFO 2020-11-27 14:35:01 train.py: 79] Epoch 4, iter 3800/6416, lr 0.100000, loss 27.183722
+INFO 2020-11-27 14:36:18 train.py: 79] Epoch 4, iter 4000/6416, lr 0.100000, loss 27.201032
+INFO 2020-11-27 14:37:35 train.py: 79] Epoch 4, iter 4200/6416, lr 0.100000, loss 27.125988
+INFO 2020-11-27 14:38:52 train.py: 79] Epoch 4, iter 4400/6416, lr 0.100000, loss 27.148296
+INFO 2020-11-27 14:40:10 train.py: 79] Epoch 4, iter 4600/6416, lr 0.100000, loss 27.100156
+INFO 2020-11-27 14:41:27 train.py: 79] Epoch 4, iter 4800/6416, lr 0.100000, loss 27.081549
+INFO 2020-11-27 14:42:44 train.py: 79] Epoch 4, iter 5000/6416, lr 0.100000, loss 27.113391
+INFO 2020-11-27 14:44:01 train.py: 79] Epoch 4, iter 5200/6416, lr 0.100000, loss 27.147206
+INFO 2020-11-27 14:45:18 train.py: 79] Epoch 4, iter 5400/6416, lr 0.100000, loss 27.114991
+INFO 2020-11-27 14:46:36 train.py: 79] Epoch 4, iter 5600/6416, lr 0.100000, loss 27.140727
+INFO 2020-11-27 14:47:53 train.py: 79] Epoch 4, iter 5800/6416, lr 0.100000, loss 27.129555
+INFO 2020-11-27 14:49:10 train.py: 92] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-27 14:49:10 train.py: 79] Epoch 4, iter 6000/6416, lr 0.100000, loss 27.096793
+INFO 2020-11-27 14:50:28 train.py: 79] Epoch 4, iter 6200/6416, lr 0.100000, loss 27.101055
+INFO 2020-11-27 14:51:45 train.py: 79] Epoch 4, iter 6400/6416, lr 0.100000, loss 27.090219
+INFO 2020-11-27 14:51:51 train.py: 97] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-27 14:51:53 train.py: 79] Epoch 5, iter 0/6416, lr 0.100000, loss 27.080042
+INFO 2020-11-27 14:53:10 train.py: 79] Epoch 5, iter 200/6416, lr 0.100000, loss 26.569350
+INFO 2020-11-27 14:54:27 train.py: 79] Epoch 5, iter 400/6416, lr 0.100000, loss 26.637917
+INFO 2020-11-27 14:55:44 train.py: 79] Epoch 5, iter 600/6416, lr 0.100000, loss 26.794215
+INFO 2020-11-27 14:57:02 train.py: 79] Epoch 5, iter 800/6416, lr 0.100000, loss 26.809539
+INFO 2020-11-27 14:58:19 train.py: 79] Epoch 5, iter 1000/6416, lr 0.100000, loss 26.856054
+INFO 2020-11-27 14:59:36 train.py: 79] Epoch 5, iter 1200/6416, lr 0.100000, loss 26.829732
+INFO 2020-11-27 15:00:53 train.py: 79] Epoch 5, iter 1400/6416, lr 0.100000, loss 26.879274
+INFO 2020-11-27 15:02:11 train.py: 79] Epoch 5, iter 1600/6416, lr 0.100000, loss 26.933818
+INFO 2020-11-27 15:03:28 train.py: 79] Epoch 5, iter 1800/6416, lr 0.100000, loss 26.909025
+INFO 2020-11-27 15:04:45 train.py: 79] Epoch 5, iter 2000/6416, lr 0.100000, loss 26.920821
+INFO 2020-11-27 15:06:02 train.py: 79] Epoch 5, iter 2200/6416, lr 0.100000, loss 27.009365
+INFO 2020-11-27 15:07:19 train.py: 79] Epoch 5, iter 2400/6416, lr 0.100000, loss 26.977531
+INFO 2020-11-27 15:08:37 train.py: 79] Epoch 5, iter 2600/6416, lr 0.100000, loss 26.971922
+INFO 2020-11-27 15:09:54 train.py: 79] Epoch 5, iter 2800/6416, lr 0.100000, loss 26.990467
+INFO 2020-11-27 15:11:11 train.py: 92] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-27 15:11:11 train.py: 79] Epoch 5, iter 3000/6416, lr 0.100000, loss 26.980199
+INFO 2020-11-27 15:12:28 train.py: 79] Epoch 5, iter 3200/6416, lr 0.100000, loss 26.933842
+INFO 2020-11-27 15:13:44 train.py: 79] Epoch 5, iter 3400/6416, lr 0.100000, loss 26.883500
+INFO 2020-11-27 15:15:01 train.py: 79] Epoch 5, iter 3600/6416, lr 0.100000, loss 26.962715
+INFO 2020-11-27 15:16:18 train.py: 79] Epoch 5, iter 3800/6416, lr 0.100000, loss 26.941409
+INFO 2020-11-27 15:17:34 train.py: 79] Epoch 5, iter 4000/6416, lr 0.100000, loss 26.970323
+INFO 2020-11-27 15:18:51 train.py: 79] Epoch 5, iter 4200/6416, lr 0.100000, loss 26.955577
+INFO 2020-11-27 15:20:08 train.py: 79] Epoch 5, iter 4400/6416, lr 0.100000, loss 26.907030
+INFO 2020-11-27 15:21:24 train.py: 79] Epoch 5, iter 4600/6416, lr 0.100000, loss 26.959151
+INFO 2020-11-27 15:22:41 train.py: 79] Epoch 5, iter 4800/6416, lr 0.100000, loss 26.890113
+INFO 2020-11-27 15:23:57 train.py: 79] Epoch 5, iter 5000/6416, lr 0.100000, loss 26.946051
+INFO 2020-11-27 15:25:14 train.py: 79] Epoch 5, iter 5200/6416, lr 0.100000, loss 26.908221
+INFO 2020-11-27 15:26:30 train.py: 79] Epoch 5, iter 5400/6416, lr 0.100000, loss 26.885252
+INFO 2020-11-27 15:27:47 train.py: 79] Epoch 5, iter 5600/6416, lr 0.100000, loss 26.921996
+INFO 2020-11-27 15:29:04 train.py: 79] Epoch 5, iter 5800/6416, lr 0.100000, loss 26.899996
+INFO 2020-11-27 15:30:20 train.py: 92] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-27 15:30:21 train.py: 79] Epoch 5, iter 6000/6416, lr 0.100000, loss 26.892551
+INFO 2020-11-27 15:31:38 train.py: 79] Epoch 5, iter 6200/6416, lr 0.100000, loss 26.925572
+INFO 2020-11-27 15:32:55 train.py: 79] Epoch 5, iter 6400/6416, lr 0.100000, loss 26.925721
+INFO 2020-11-27 15:33:01 train.py: 97] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-27 15:33:03 train.py: 79] Epoch 6, iter 0/6416, lr 0.100000, loss 26.843349
+INFO 2020-11-27 15:34:20 train.py: 79] Epoch 6, iter 200/6416, lr 0.100000, loss 26.333796
+INFO 2020-11-27 15:35:37 train.py: 79] Epoch 6, iter 400/6416, lr 0.100000, loss 26.494543
+INFO 2020-11-27 15:36:55 train.py: 79] Epoch 6, iter 600/6416, lr 0.100000, loss 26.552440
+INFO 2020-11-27 15:38:12 train.py: 79] Epoch 6, iter 800/6416, lr 0.100000, loss 26.582460
+INFO 2020-11-27 15:39:29 train.py: 79] Epoch 6, iter 1000/6416, lr 0.100000, loss 26.632009
+INFO 2020-11-27 15:40:46 train.py: 79] Epoch 6, iter 1200/6416, lr 0.100000, loss 26.644750
+INFO 2020-11-27 15:42:04 train.py: 79] Epoch 6, iter 1400/6416, lr 0.100000, loss 26.731838
+INFO 2020-11-27 15:43:21 train.py: 79] Epoch 6, iter 1600/6416, lr 0.100000, loss 26.737685
+INFO 2020-11-27 15:44:38 train.py: 79] Epoch 6, iter 1800/6416, lr 0.100000, loss 26.744759
+INFO 2020-11-27 15:45:55 train.py: 79] Epoch 6, iter 2000/6416, lr 0.100000, loss 26.746034
+INFO 2020-11-27 15:47:13 train.py: 79] Epoch 6, iter 2200/6416, lr 0.100000, loss 26.764201
+INFO 2020-11-27 15:48:30 train.py: 79] Epoch 6, iter 2400/6416, lr 0.100000, loss 26.812925
+INFO 2020-11-27 15:49:47 train.py: 79] Epoch 6, iter 2600/6416, lr 0.100000, loss 26.820769
+INFO 2020-11-27 15:51:04 train.py: 79] Epoch 6, iter 2800/6416, lr 0.100000, loss 26.836431
+INFO 2020-11-27 15:52:21 train.py: 92] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-27 15:52:22 train.py: 79] Epoch 6, iter 3000/6416, lr 0.100000, loss 26.786902
+INFO 2020-11-27 15:53:39 train.py: 79] Epoch 6, iter 3200/6416, lr 0.100000, loss 26.804522
+INFO 2020-11-27 15:54:56 train.py: 79] Epoch 6, iter 3400/6416, lr 0.100000, loss 26.826138
+INFO 2020-11-27 15:56:13 train.py: 79] Epoch 6, iter 3600/6416, lr 0.100000, loss 26.804703
+INFO 2020-11-27 15:57:30 train.py: 79] Epoch 6, iter 3800/6416, lr 0.100000, loss 26.801648
+INFO 2020-11-27 15:58:48 train.py: 79] Epoch 6, iter 4000/6416, lr 0.100000, loss 26.793580
+INFO 2020-11-27 16:00:05 train.py: 79] Epoch 6, iter 4200/6416, lr 0.100000, loss 26.763612
+INFO 2020-11-27 16:01:22 train.py: 79] Epoch 6, iter 4400/6416, lr 0.100000, loss 26.796216
+INFO 2020-11-27 16:02:40 train.py: 79] Epoch 6, iter 4600/6416, lr 0.100000, loss 26.796974
+INFO 2020-11-27 16:03:57 train.py: 79] Epoch 6, iter 4800/6416, lr 0.100000, loss 26.774140
+INFO 2020-11-27 16:05:14 train.py: 79] Epoch 6, iter 5000/6416, lr 0.100000, loss 26.746327
+INFO 2020-11-27 16:06:31 train.py: 79] Epoch 6, iter 5200/6416, lr 0.100000, loss 26.750737
+INFO 2020-11-27 16:07:48 train.py: 79] Epoch 6, iter 5400/6416, lr 0.100000, loss 26.747084
+INFO 2020-11-27 16:09:06 train.py: 79] Epoch 6, iter 5600/6416, lr 0.100000, loss 26.776214
+INFO 2020-11-27 16:10:23 train.py: 79] Epoch 6, iter 5800/6416, lr 0.100000, loss 26.715030
+INFO 2020-11-27 16:11:40 train.py: 92] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-27 16:11:40 train.py: 79] Epoch 6, iter 6000/6416, lr 0.100000, loss 26.787295
+INFO 2020-11-27 16:12:58 train.py: 79] Epoch 6, iter 6200/6416, lr 0.100000, loss 26.734481
+INFO 2020-11-27 16:14:15 train.py: 79] Epoch 6, iter 6400/6416, lr 0.100000, loss 26.775375
+INFO 2020-11-27 16:14:21 train.py: 97] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-27 16:14:23 train.py: 79] Epoch 7, iter 0/6416, lr 0.100000, loss 26.625604
+INFO 2020-11-27 16:15:40 train.py: 79] Epoch 7, iter 200/6416, lr 0.100000, loss 26.281002
+INFO 2020-11-27 16:16:57 train.py: 79] Epoch 7, iter 400/6416, lr 0.100000, loss 26.348980
+INFO 2020-11-27 16:18:14 train.py: 79] Epoch 7, iter 600/6416, lr 0.100000, loss 26.400044
+INFO 2020-11-27 16:19:31 train.py: 79] Epoch 7, iter 800/6416, lr 0.100000, loss 26.485080
+INFO 2020-11-27 16:20:48 train.py: 79] Epoch 7, iter 1000/6416, lr 0.100000, loss 26.513556
+INFO 2020-11-27 16:22:06 train.py: 79] Epoch 7, iter 1200/6416, lr 0.100000, loss 26.563984
+INFO 2020-11-27 16:23:23 train.py: 79] Epoch 7, iter 1400/6416, lr 0.100000, loss 26.551700
+INFO 2020-11-27 16:24:40 train.py: 79] Epoch 7, iter 1600/6416, lr 0.100000, loss 26.599454
+INFO 2020-11-27 16:25:57 train.py: 79] Epoch 7, iter 1800/6416, lr 0.100000, loss 26.622616
+INFO 2020-11-27 16:27:15 train.py: 79] Epoch 7, iter 2000/6416, lr 0.100000, loss 26.617826
+INFO 2020-11-27 16:28:32 train.py: 79] Epoch 7, iter 2200/6416, lr 0.100000, loss 26.667981
+INFO 2020-11-27 16:29:49 train.py: 79] Epoch 7, iter 2400/6416, lr 0.100000, loss 26.647871
+INFO 2020-11-27 16:31:06 train.py: 79] Epoch 7, iter 2600/6416, lr 0.100000, loss 26.653957
+INFO 2020-11-27 16:32:23 train.py: 79] Epoch 7, iter 2800/6416, lr 0.100000, loss 26.670814
+INFO 2020-11-27 16:33:40 train.py: 92] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-27 16:33:41 train.py: 79] Epoch 7, iter 3000/6416, lr 0.100000, loss 26.699703
+INFO 2020-11-27 16:34:58 train.py: 79] Epoch 7, iter 3200/6416, lr 0.100000, loss 26.685563
+INFO 2020-11-27 16:36:15 train.py: 79] Epoch 7, iter 3400/6416, lr 0.100000, loss 26.624901
+INFO 2020-11-27 16:37:32 train.py: 79] Epoch 7, iter 3600/6416, lr 0.100000, loss 26.667817
+INFO 2020-11-27 16:38:50 train.py: 79] Epoch 7, iter 3800/6416, lr 0.100000, loss 26.672129
+INFO 2020-11-27 16:40:07 train.py: 79] Epoch 7, iter 4000/6416, lr 0.100000, loss 26.666484
+INFO 2020-11-27 16:41:24 train.py: 79] Epoch 7, iter 4200/6416, lr 0.100000, loss 26.624046
+INFO 2020-11-27 16:42:41 train.py: 79] Epoch 7, iter 4400/6416, lr 0.100000, loss 26.669233
+INFO 2020-11-27 16:43:59 train.py: 79] Epoch 7, iter 4600/6416, lr 0.100000, loss 26.643160
+INFO 2020-11-27 16:45:16 train.py: 79] Epoch 7, iter 4800/6416, lr 0.100000, loss 26.628013
+INFO 2020-11-27 16:46:33 train.py: 79] Epoch 7, iter 5000/6416, lr 0.100000, loss 26.649422
+INFO 2020-11-27 16:47:50 train.py: 79] Epoch 7, iter 5200/6416, lr 0.100000, loss 26.663376
+INFO 2020-11-27 16:49:08 train.py: 79] Epoch 7, iter 5400/6416, lr 0.100000, loss 26.617261
+INFO 2020-11-27 16:50:25 train.py: 79] Epoch 7, iter 5600/6416, lr 0.100000, loss 26.638865
+INFO 2020-11-27 16:51:42 train.py: 79] Epoch 7, iter 5800/6416, lr 0.100000, loss 26.626876
+INFO 2020-11-27 16:52:59 train.py: 92] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-27 16:53:00 train.py: 79] Epoch 7, iter 6000/6416, lr 0.100000, loss 26.623282
+INFO 2020-11-27 16:54:16 train.py: 79] Epoch 7, iter 6200/6416, lr 0.100000, loss 26.608983
+INFO 2020-11-27 16:55:34 train.py: 79] Epoch 7, iter 6400/6416, lr 0.100000, loss 26.611998
+INFO 2020-11-27 16:55:40 train.py: 97] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-27 16:55:42 train.py: 79] Epoch 8, iter 0/6416, lr 0.100000, loss 26.785995
+INFO 2020-11-27 16:56:59 train.py: 79] Epoch 8, iter 200/6416, lr 0.100000, loss 26.144065
+INFO 2020-11-27 16:58:16 train.py: 79] Epoch 8, iter 400/6416, lr 0.100000, loss 26.240759
+INFO 2020-11-27 16:59:33 train.py: 79] Epoch 8, iter 600/6416, lr 0.100000, loss 26.304228
+INFO 2020-11-27 17:00:50 train.py: 79] Epoch 8, iter 800/6416, lr 0.100000, loss 26.356779
+INFO 2020-11-27 17:02:08 train.py: 79] Epoch 8, iter 1000/6416, lr 0.100000, loss 26.448754
+INFO 2020-11-27 17:03:25 train.py: 79] Epoch 8, iter 1200/6416, lr 0.100000, loss 26.423890
+INFO 2020-11-27 17:04:42 train.py: 79] Epoch 8, iter 1400/6416, lr 0.100000, loss 26.466591
+INFO 2020-11-27 17:06:00 train.py: 79] Epoch 8, iter 1600/6416, lr 0.100000, loss 26.511736
+INFO 2020-11-27 17:07:17 train.py: 79] Epoch 8, iter 1800/6416, lr 0.100000, loss 26.578622
+INFO 2020-11-27 17:08:34 train.py: 79] Epoch 8, iter 2000/6416, lr 0.100000, loss 26.522598
+INFO 2020-11-27 17:09:51 train.py: 79] Epoch 8, iter 2200/6416, lr 0.100000, loss 26.499182
+INFO 2020-11-27 17:11:09 train.py: 79] Epoch 8, iter 2400/6416, lr 0.100000, loss 26.545229
+INFO 2020-11-27 17:12:26 train.py: 79] Epoch 8, iter 2600/6416, lr 0.100000, loss 26.531583
+INFO 2020-11-27 17:13:43 train.py: 79] Epoch 8, iter 2800/6416, lr 0.100000, loss 26.538433
+INFO 2020-11-27 17:15:00 train.py: 92] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-27 17:15:01 train.py: 79] Epoch 8, iter 3000/6416, lr 0.100000, loss 26.486599
+INFO 2020-11-27 17:16:17 train.py: 79] Epoch 8, iter 3200/6416, lr 0.100000, loss 26.570876
+INFO 2020-11-27 17:17:34 train.py: 79] Epoch 8, iter 3400/6416, lr 0.100000, loss 26.513811
+INFO 2020-11-27 17:18:50 train.py: 79] Epoch 8, iter 3600/6416, lr 0.100000, loss 26.543802
+INFO 2020-11-27 17:20:07 train.py: 79] Epoch 8, iter 3800/6416, lr 0.100000, loss 26.545988
+INFO 2020-11-27 17:21:24 train.py: 79] Epoch 8, iter 4000/6416, lr 0.100000, loss 26.559745
+INFO 2020-11-27 17:22:40 train.py: 79] Epoch 8, iter 4200/6416, lr 0.100000, loss 26.556153
+INFO 2020-11-27 17:23:57 train.py: 79] Epoch 8, iter 4400/6416, lr 0.100000, loss 26.573778
+INFO 2020-11-27 17:25:14 train.py: 79] Epoch 8, iter 4600/6416, lr 0.100000, loss 26.542522
+INFO 2020-11-27 17:26:30 train.py: 79] Epoch 8, iter 4800/6416, lr 0.100000, loss 26.533074
+INFO 2020-11-27 17:27:47 train.py: 79] Epoch 8, iter 5000/6416, lr 0.100000, loss 26.559845
+INFO 2020-11-27 17:29:04 train.py: 79] Epoch 8, iter 5200/6416, lr 0.100000, loss 26.554198
+INFO 2020-11-27 17:30:20 train.py: 79] Epoch 8, iter 5400/6416, lr 0.100000, loss 26.485994
+INFO 2020-11-27 17:31:37 train.py: 79] Epoch 8, iter 5600/6416, lr 0.100000, loss 26.550967
+INFO 2020-11-27 17:32:53 train.py: 79] Epoch 8, iter 5800/6416, lr 0.100000, loss 26.500807
+INFO 2020-11-27 17:34:10 train.py: 92] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-27 17:34:10 train.py: 79] Epoch 8, iter 6000/6416, lr 0.100000, loss 26.552850
+INFO 2020-11-27 17:35:27 train.py: 79] Epoch 8, iter 6200/6416, lr 0.100000, loss 26.520036
+INFO 2020-11-27 17:36:44 train.py: 79] Epoch 8, iter 6400/6416, lr 0.100000, loss 26.478472
+INFO 2020-11-27 17:36:51 train.py: 97] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-27 17:36:52 train.py: 79] Epoch 9, iter 0/6416, lr 0.100000, loss 26.624666
+INFO 2020-11-27 17:38:10 train.py: 79] Epoch 9, iter 200/6416, lr 0.100000, loss 26.047869
+INFO 2020-11-27 17:39:27 train.py: 79] Epoch 9, iter 400/6416, lr 0.100000, loss 26.122794
+INFO 2020-11-27 17:40:44 train.py: 79] Epoch 9, iter 600/6416, lr 0.100000, loss 26.241400
+INFO 2020-11-27 17:42:01 train.py: 79] Epoch 9, iter 800/6416, lr 0.100000, loss 26.252919
+INFO 2020-11-27 17:43:18 train.py: 79] Epoch 9, iter 1000/6416, lr 0.100000, loss 26.282740
+INFO 2020-11-27 17:44:35 train.py: 79] Epoch 9, iter 1200/6416, lr 0.100000, loss 26.356872
+INFO 2020-11-27 17:45:53 train.py: 79] Epoch 9, iter 1400/6416, lr 0.100000, loss 26.374546
+INFO 2020-11-27 17:47:10 train.py: 79] Epoch 9, iter 1600/6416, lr 0.100000, loss 26.390618
+INFO 2020-11-27 17:48:27 train.py: 79] Epoch 9, iter 1800/6416, lr 0.100000, loss 26.420870
+INFO 2020-11-27 17:49:45 train.py: 79] Epoch 9, iter 2000/6416, lr 0.100000, loss 26.451176
+INFO 2020-11-27 17:51:02 train.py: 79] Epoch 9, iter 2200/6416, lr 0.100000, loss 26.424836
+INFO 2020-11-27 17:52:19 train.py: 79] Epoch 9, iter 2400/6416, lr 0.100000, loss 26.472089
+INFO 2020-11-27 17:53:36 train.py: 79] Epoch 9, iter 2600/6416, lr 0.100000, loss 26.453868
+INFO 2020-11-27 17:54:53 train.py: 79] Epoch 9, iter 2800/6416, lr 0.100000, loss 26.456881
+INFO 2020-11-27 17:56:10 train.py: 92] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-27 17:56:11 train.py: 79] Epoch 9, iter 3000/6416, lr 0.100000, loss 26.489091
+INFO 2020-11-27 17:57:27 train.py: 79] Epoch 9, iter 3200/6416, lr 0.100000, loss 26.469661
+INFO 2020-11-27 17:58:44 train.py: 79] Epoch 9, iter 3400/6416, lr 0.100000, loss 26.471517
+INFO 2020-11-27 18:00:01 train.py: 79] Epoch 9, iter 3600/6416, lr 0.100000, loss 26.453394
+INFO 2020-11-27 18:01:18 train.py: 79] Epoch 9, iter 3800/6416, lr 0.100000, loss 26.463491
+INFO 2020-11-27 18:02:36 train.py: 79] Epoch 9, iter 4000/6416, lr 0.100000, loss 26.441960
+INFO 2020-11-27 18:03:53 train.py: 79] Epoch 9, iter 4200/6416, lr 0.100000, loss 26.483037
+INFO 2020-11-27 18:05:10 train.py: 79] Epoch 9, iter 4400/6416, lr 0.100000, loss 26.403237
+INFO 2020-11-27 18:06:27 train.py: 79] Epoch 9, iter 4600/6416, lr 0.100000, loss 26.486957
+INFO 2020-11-27 18:07:45 train.py: 79] Epoch 9, iter 4800/6416, lr 0.100000, loss 26.482731
+INFO 2020-11-27 18:09:02 train.py: 79] Epoch 9, iter 5000/6416, lr 0.100000, loss 26.440016
+INFO 2020-11-27 18:10:19 train.py: 79] Epoch 9, iter 5200/6416, lr 0.100000, loss 26.405595
+INFO 2020-11-27 18:11:36 train.py: 79] Epoch 9, iter 5400/6416, lr 0.100000, loss 26.428556
+INFO 2020-11-27 18:12:54 train.py: 79] Epoch 9, iter 5600/6416, lr 0.100000, loss 26.379415
+INFO 2020-11-27 18:14:11 train.py: 79] Epoch 9, iter 5800/6416, lr 0.100000, loss 26.450669
+INFO 2020-11-27 18:15:28 train.py: 92] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-27 18:15:28 train.py: 79] Epoch 9, iter 6000/6416, lr 0.100000, loss 26.445732
+INFO 2020-11-27 18:16:45 train.py: 79] Epoch 9, iter 6200/6416, lr 0.100000, loss 26.428099
+INFO 2020-11-27 18:18:02 train.py: 79] Epoch 9, iter 6400/6416, lr 0.100000, loss 26.435389
+INFO 2020-11-27 18:18:09 train.py: 97] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-27 18:18:10 train.py: 79] Epoch 10, iter 0/6416, lr 0.010000, loss 26.353246
+INFO 2020-11-27 18:19:27 train.py: 79] Epoch 10, iter 200/6416, lr 0.010000, loss 25.366233
+INFO 2020-11-27 18:20:43 train.py: 79] Epoch 10, iter 400/6416, lr 0.010000, loss 25.134344
+INFO 2020-11-27 18:22:00 train.py: 79] Epoch 10, iter 600/6416, lr 0.010000, loss 24.939041
+INFO 2020-11-27 18:23:16 train.py: 79] Epoch 10, iter 800/6416, lr 0.010000, loss 24.881675
+INFO 2020-11-27 18:24:33 train.py: 79] Epoch 10, iter 1000/6416, lr 0.010000, loss 24.780506
+INFO 2020-11-27 18:25:50 train.py: 79] Epoch 10, iter 1200/6416, lr 0.010000, loss 24.735566
+INFO 2020-11-27 18:27:06 train.py: 79] Epoch 10, iter 1400/6416, lr 0.010000, loss 24.651775
+INFO 2020-11-27 18:28:23 train.py: 79] Epoch 10, iter 1600/6416, lr 0.010000, loss 24.613847
+INFO 2020-11-27 18:29:39 train.py: 79] Epoch 10, iter 1800/6416, lr 0.010000, loss 24.611395
+INFO 2020-11-27 18:30:56 train.py: 79] Epoch 10, iter 2000/6416, lr 0.010000, loss 24.542444
+INFO 2020-11-27 18:32:13 train.py: 79] Epoch 10, iter 2200/6416, lr 0.010000, loss 24.483578
+INFO 2020-11-27 18:33:29 train.py: 79] Epoch 10, iter 2400/6416, lr 0.010000, loss 24.467519
+INFO 2020-11-27 18:34:46 train.py: 79] Epoch 10, iter 2600/6416, lr 0.010000, loss 24.449910
+INFO 2020-11-27 18:36:03 train.py: 79] Epoch 10, iter 2800/6416, lr 0.010000, loss 24.407115
+INFO 2020-11-27 18:37:19 train.py: 92] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-27 18:37:19 train.py: 79] Epoch 10, iter 3000/6416, lr 0.010000, loss 24.378046
+INFO 2020-11-27 18:38:37 train.py: 79] Epoch 10, iter 3200/6416, lr 0.010000, loss 24.332232
+INFO 2020-11-27 18:39:54 train.py: 79] Epoch 10, iter 3400/6416, lr 0.010000, loss 24.346177
+INFO 2020-11-27 18:41:11 train.py: 79] Epoch 10, iter 3600/6416, lr 0.010000, loss 24.290777
+INFO 2020-11-27 18:42:28 train.py: 79] Epoch 10, iter 3800/6416, lr 0.010000, loss 24.300226
+INFO 2020-11-27 18:43:46 train.py: 79] Epoch 10, iter 4000/6416, lr 0.010000, loss 24.253135
+INFO 2020-11-27 18:45:03 train.py: 79] Epoch 10, iter 4200/6416, lr 0.010000, loss 24.290047
+INFO 2020-11-27 18:46:20 train.py: 79] Epoch 10, iter 4400/6416, lr 0.010000, loss 24.246174
+INFO 2020-11-27 18:47:37 train.py: 79] Epoch 10, iter 4600/6416, lr 0.010000, loss 24.223171
+INFO 2020-11-27 18:48:55 train.py: 79] Epoch 10, iter 4800/6416, lr 0.010000, loss 24.206740
+INFO 2020-11-27 18:50:12 train.py: 79] Epoch 10, iter 5000/6416, lr 0.010000, loss 24.153028
+INFO 2020-11-27 18:51:29 train.py: 79] Epoch 10, iter 5200/6416, lr 0.010000, loss 24.147221
+INFO 2020-11-27 18:52:46 train.py: 79] Epoch 10, iter 5400/6416, lr 0.010000, loss 24.178179
+INFO 2020-11-27 18:54:04 train.py: 79] Epoch 10, iter 5600/6416, lr 0.010000, loss 24.161250
+INFO 2020-11-27 18:55:21 train.py: 79] Epoch 10, iter 5800/6416, lr 0.010000, loss 24.119541
+INFO 2020-11-27 18:56:38 train.py: 92] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-27 18:56:38 train.py: 79] Epoch 10, iter 6000/6416, lr 0.010000, loss 24.130006
+INFO 2020-11-27 18:57:55 train.py: 79] Epoch 10, iter 6200/6416, lr 0.010000, loss 24.113544
+INFO 2020-11-27 18:59:13 train.py: 79] Epoch 10, iter 6400/6416, lr 0.010000, loss 24.050037
+INFO 2020-11-27 18:59:19 train.py: 97] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-27 18:59:20 train.py: 79] Epoch 11, iter 0/6416, lr 0.010000, loss 24.293459
+INFO 2020-11-27 19:00:37 train.py: 79] Epoch 11, iter 200/6416, lr 0.010000, loss 23.817109
+INFO 2020-11-27 19:01:54 train.py: 79] Epoch 11, iter 400/6416, lr 0.010000, loss 23.813827
+INFO 2020-11-27 19:03:10 train.py: 79] Epoch 11, iter 600/6416, lr 0.010000, loss 23.813719
+INFO 2020-11-27 19:04:27 train.py: 79] Epoch 11, iter 800/6416, lr 0.010000, loss 23.829766
+INFO 2020-11-27 19:05:45 train.py: 79] Epoch 11, iter 1000/6416, lr 0.010000, loss 23.802465
+INFO 2020-11-27 19:07:02 train.py: 79] Epoch 11, iter 1200/6416, lr 0.010000, loss 23.808603
+INFO 2020-11-27 19:08:19 train.py: 79] Epoch 11, iter 1400/6416, lr 0.010000, loss 23.804885
+INFO 2020-11-27 19:09:36 train.py: 79] Epoch 11, iter 1600/6416, lr 0.010000, loss 23.804163
+INFO 2020-11-27 19:10:53 train.py: 79] Epoch 11, iter 1800/6416, lr 0.010000, loss 23.795932
+INFO 2020-11-27 19:12:11 train.py: 79] Epoch 11, iter 2000/6416, lr 0.010000, loss 23.839948
+INFO 2020-11-27 19:13:28 train.py: 79] Epoch 11, iter 2200/6416, lr 0.010000, loss 23.831888
+INFO 2020-11-27 19:14:45 train.py: 79] Epoch 11, iter 2400/6416, lr 0.010000, loss 23.841614
+INFO 2020-11-27 19:16:02 train.py: 79] Epoch 11, iter 2600/6416, lr 0.010000, loss 23.799354
+INFO 2020-11-27 19:17:20 train.py: 79] Epoch 11, iter 2800/6416, lr 0.010000, loss 23.865570
+INFO 2020-11-27 19:18:37 train.py: 92] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-27 19:18:37 train.py: 79] Epoch 11, iter 3000/6416, lr 0.010000, loss 23.832731
+INFO 2020-11-27 19:19:54 train.py: 79] Epoch 11, iter 3200/6416, lr 0.010000, loss 23.788771
+INFO 2020-11-27 19:21:12 train.py: 79] Epoch 11, iter 3400/6416, lr 0.010000, loss 23.833594
+INFO 2020-11-27 19:22:29 train.py: 79] Epoch 11, iter 3600/6416, lr 0.010000, loss 23.824566
+INFO 2020-11-27 19:23:46 train.py: 79] Epoch 11, iter 3800/6416, lr 0.010000, loss 23.839771
+INFO 2020-11-27 19:25:03 train.py: 79] Epoch 11, iter 4000/6416, lr 0.010000, loss 23.850836
+INFO 2020-11-27 19:26:20 train.py: 79] Epoch 11, iter 4200/6416, lr 0.010000, loss 23.835896
+INFO 2020-11-27 19:27:38 train.py: 79] Epoch 11, iter 4400/6416, lr 0.010000, loss 23.866312
+INFO 2020-11-27 19:28:55 train.py: 79] Epoch 11, iter 4600/6416, lr 0.010000, loss 23.831772
+INFO 2020-11-27 19:30:12 train.py: 79] Epoch 11, iter 4800/6416, lr 0.010000, loss 23.795159
+INFO 2020-11-27 19:31:29 train.py: 79] Epoch 11, iter 5000/6416, lr 0.010000, loss 23.882771
+INFO 2020-11-27 19:32:46 train.py: 79] Epoch 11, iter 5200/6416, lr 0.010000, loss 23.834567
+INFO 2020-11-27 19:34:04 train.py: 79] Epoch 11, iter 5400/6416, lr 0.010000, loss 23.880144
+INFO 2020-11-27 19:35:21 train.py: 79] Epoch 11, iter 5600/6416, lr 0.010000, loss 23.829198
+INFO 2020-11-27 19:36:38 train.py: 79] Epoch 11, iter 5800/6416, lr 0.010000, loss 23.838647
+INFO 2020-11-27 19:37:55 train.py: 92] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-27 19:37:56 train.py: 79] Epoch 11, iter 6000/6416, lr 0.010000, loss 23.820171
+INFO 2020-11-27 19:39:12 train.py: 79] Epoch 11, iter 6200/6416, lr 0.010000, loss 23.816880
+INFO 2020-11-27 19:40:29 train.py: 79] Epoch 11, iter 6400/6416, lr 0.010000, loss 23.843305
+INFO 2020-11-27 19:40:35 train.py: 97] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-27 19:40:36 train.py: 79] Epoch 12, iter 0/6416, lr 0.010000, loss 23.819377
+INFO 2020-11-27 19:41:54 train.py: 79] Epoch 12, iter 200/6416, lr 0.010000, loss 23.548012
+INFO 2020-11-27 19:43:11 train.py: 79] Epoch 12, iter 400/6416, lr 0.010000, loss 23.577025
+INFO 2020-11-27 19:44:28 train.py: 79] Epoch 12, iter 600/6416, lr 0.010000, loss 23.591395
+INFO 2020-11-27 19:45:45 train.py: 79] Epoch 12, iter 800/6416, lr 0.010000, loss 23.605145
+INFO 2020-11-27 19:47:02 train.py: 79] Epoch 12, iter 1000/6416, lr 0.010000, loss 23.642389
+INFO 2020-11-27 19:48:20 train.py: 79] Epoch 12, iter 1200/6416, lr 0.010000, loss 23.618043
+INFO 2020-11-27 19:49:37 train.py: 79] Epoch 12, iter 1400/6416, lr 0.010000, loss 23.607647
+INFO 2020-11-27 19:50:54 train.py: 79] Epoch 12, iter 1600/6416, lr 0.010000, loss 23.656293
+INFO 2020-11-27 19:52:11 train.py: 79] Epoch 12, iter 1800/6416, lr 0.010000, loss 23.643595
+INFO 2020-11-27 19:53:28 train.py: 79] Epoch 12, iter 2000/6416, lr 0.010000, loss 23.667570
+INFO 2020-11-27 19:54:46 train.py: 79] Epoch 12, iter 2200/6416, lr 0.010000, loss 23.686193
+INFO 2020-11-27 19:56:03 train.py: 79] Epoch 12, iter 2400/6416, lr 0.010000, loss 23.706958
+INFO 2020-11-27 19:57:20 train.py: 79] Epoch 12, iter 2600/6416, lr 0.010000, loss 23.672709
+INFO 2020-11-27 19:58:37 train.py: 79] Epoch 12, iter 2800/6416, lr 0.010000, loss 23.704939
+INFO 2020-11-27 19:59:54 train.py: 92] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-27 19:59:55 train.py: 79] Epoch 12, iter 3000/6416, lr 0.010000, loss 23.678014
+INFO 2020-11-27 20:01:12 train.py: 79] Epoch 12, iter 3200/6416, lr 0.010000, loss 23.665280
+INFO 2020-11-27 20:02:29 train.py: 79] Epoch 12, iter 3400/6416, lr 0.010000, loss 23.720366
+INFO 2020-11-27 20:03:47 train.py: 79] Epoch 12, iter 3600/6416, lr 0.010000, loss 23.705966
+INFO 2020-11-27 20:05:04 train.py: 79] Epoch 12, iter 3800/6416, lr 0.010000, loss 23.704745
+INFO 2020-11-27 20:06:21 train.py: 79] Epoch 12, iter 4000/6416, lr 0.010000, loss 23.734378
+INFO 2020-11-27 20:07:38 train.py: 79] Epoch 12, iter 4200/6416, lr 0.010000, loss 23.764667
+INFO 2020-11-27 20:08:55 train.py: 79] Epoch 12, iter 4400/6416, lr 0.010000, loss 23.712801
+INFO 2020-11-27 20:10:13 train.py: 79] Epoch 12, iter 4600/6416, lr 0.010000, loss 23.674852
+INFO 2020-11-27 20:11:30 train.py: 79] Epoch 12, iter 4800/6416, lr 0.010000, loss 23.730592
+INFO 2020-11-27 20:12:47 train.py: 79] Epoch 12, iter 5000/6416, lr 0.010000, loss 23.726943
+INFO 2020-11-27 20:14:04 train.py: 79] Epoch 12, iter 5200/6416, lr 0.010000, loss 23.703507
+INFO 2020-11-27 20:15:21 train.py: 79] Epoch 12, iter 5400/6416, lr 0.010000, loss 23.691990
+INFO 2020-11-27 20:16:39 train.py: 79] Epoch 12, iter 5600/6416, lr 0.010000, loss 23.770929
+INFO 2020-11-27 20:17:56 train.py: 79] Epoch 12, iter 5800/6416, lr 0.010000, loss 23.774558
+INFO 2020-11-27 20:19:13 train.py: 92] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-27 20:19:13 train.py: 79] Epoch 12, iter 6000/6416, lr 0.010000, loss 23.726325
+INFO 2020-11-27 20:20:30 train.py: 79] Epoch 12, iter 6200/6416, lr 0.010000, loss 23.754166
+INFO 2020-11-27 20:21:47 train.py: 79] Epoch 12, iter 6400/6416, lr 0.010000, loss 23.747228
+INFO 2020-11-27 20:21:53 train.py: 97] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-27 20:21:55 train.py: 79] Epoch 13, iter 0/6416, lr 0.001000, loss 23.727625
+INFO 2020-11-27 20:23:12 train.py: 79] Epoch 13, iter 200/6416, lr 0.001000, loss 23.384237
+INFO 2020-11-27 20:24:28 train.py: 79] Epoch 13, iter 400/6416, lr 0.001000, loss 23.371738
+INFO 2020-11-27 20:25:45 train.py: 79] Epoch 13, iter 600/6416, lr 0.001000, loss 23.337761
+INFO 2020-11-27 20:27:02 train.py: 79] Epoch 13, iter 800/6416, lr 0.001000, loss 23.329488
+INFO 2020-11-27 20:28:18 train.py: 79] Epoch 13, iter 1000/6416, lr 0.001000, loss 23.322887
+INFO 2020-11-27 20:29:35 train.py: 79] Epoch 13, iter 1200/6416, lr 0.001000, loss 23.322364
+INFO 2020-11-27 20:30:52 train.py: 79] Epoch 13, iter 1400/6416, lr 0.001000, loss 23.319793
+INFO 2020-11-27 20:32:08 train.py: 79] Epoch 13, iter 1600/6416, lr 0.001000, loss 23.341679
+INFO 2020-11-27 20:33:25 train.py: 79] Epoch 13, iter 1800/6416, lr 0.001000, loss 23.336878
+INFO 2020-11-27 20:34:41 train.py: 79] Epoch 13, iter 2000/6416, lr 0.001000, loss 23.331237
+INFO 2020-11-27 20:35:58 train.py: 79] Epoch 13, iter 2200/6416, lr 0.001000, loss 23.311621
+INFO 2020-11-27 20:37:14 train.py: 79] Epoch 13, iter 2400/6416, lr 0.001000, loss 23.313848
+INFO 2020-11-27 20:38:31 train.py: 79] Epoch 13, iter 2600/6416, lr 0.001000, loss 23.326241
+INFO 2020-11-27 20:39:48 train.py: 79] Epoch 13, iter 2800/6416, lr 0.001000, loss 23.317299
+INFO 2020-11-27 20:41:04 train.py: 92] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-27 20:41:04 train.py: 79] Epoch 13, iter 3000/6416, lr 0.001000, loss 23.276657
+INFO 2020-11-27 20:42:21 train.py: 79] Epoch 13, iter 3200/6416, lr 0.001000, loss 23.272422
+INFO 2020-11-27 20:43:39 train.py: 79] Epoch 13, iter 3400/6416, lr 0.001000, loss 23.287074
+INFO 2020-11-27 20:44:56 train.py: 79] Epoch 13, iter 3600/6416, lr 0.001000, loss 23.300940
+INFO 2020-11-27 20:46:13 train.py: 79] Epoch 13, iter 3800/6416, lr 0.001000, loss 23.286703
+INFO 2020-11-27 20:47:30 train.py: 79] Epoch 13, iter 4000/6416, lr 0.001000, loss 23.303239
+INFO 2020-11-27 20:48:48 train.py: 79] Epoch 13, iter 4200/6416, lr 0.001000, loss 23.302035
+INFO 2020-11-27 20:50:05 train.py: 79] Epoch 13, iter 4400/6416, lr 0.001000, loss 23.348301
+INFO 2020-11-27 20:51:22 train.py: 79] Epoch 13, iter 4600/6416, lr 0.001000, loss 23.308829
+INFO 2020-11-27 20:52:39 train.py: 79] Epoch 13, iter 4800/6416, lr 0.001000, loss 23.321386
+INFO 2020-11-27 20:53:56 train.py: 79] Epoch 13, iter 5000/6416, lr 0.001000, loss 23.319429
+INFO 2020-11-27 20:55:13 train.py: 79] Epoch 13, iter 5200/6416, lr 0.001000, loss 23.266885
+INFO 2020-11-27 20:56:31 train.py: 79] Epoch 13, iter 5400/6416, lr 0.001000, loss 23.317021
+INFO 2020-11-27 20:57:48 train.py: 79] Epoch 13, iter 5600/6416, lr 0.001000, loss 23.307620
+INFO 2020-11-27 20:59:05 train.py: 79] Epoch 13, iter 5800/6416, lr 0.001000, loss 23.287010
+INFO 2020-11-27 21:00:22 train.py: 92] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-27 21:00:22 train.py: 79] Epoch 13, iter 6000/6416, lr 0.001000, loss 23.291159
+INFO 2020-11-27 21:01:40 train.py: 79] Epoch 13, iter 6200/6416, lr 0.001000, loss 23.313620
+INFO 2020-11-27 21:02:57 train.py: 79] Epoch 13, iter 6400/6416, lr 0.001000, loss 23.312954
+INFO 2020-11-27 21:03:03 train.py: 97] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-27 21:03:04 train.py: 79] Epoch 14, iter 0/6416, lr 0.001000, loss 23.313574
+INFO 2020-11-27 21:04:22 train.py: 79] Epoch 14, iter 200/6416, lr 0.001000, loss 23.204394
+INFO 2020-11-27 21:05:39 train.py: 79] Epoch 14, iter 400/6416, lr 0.001000, loss 23.219519
+INFO 2020-11-27 21:06:56 train.py: 79] Epoch 14, iter 600/6416, lr 0.001000, loss 23.221956
+INFO 2020-11-27 21:08:13 train.py: 79] Epoch 14, iter 800/6416, lr 0.001000, loss 23.224297
+INFO 2020-11-27 21:09:31 train.py: 79] Epoch 14, iter 1000/6416, lr 0.001000, loss 23.252360
+INFO 2020-11-27 21:10:48 train.py: 79] Epoch 14, iter 1200/6416, lr 0.001000, loss 23.261930
+INFO 2020-11-27 21:12:05 train.py: 79] Epoch 14, iter 1400/6416, lr 0.001000, loss 23.255049
+INFO 2020-11-27 21:13:22 train.py: 79] Epoch 14, iter 1600/6416, lr 0.001000, loss 23.246408
+INFO 2020-11-27 21:14:39 train.py: 79] Epoch 14, iter 1800/6416, lr 0.001000, loss 23.304769
+INFO 2020-11-27 21:15:56 train.py: 79] Epoch 14, iter 2000/6416, lr 0.001000, loss 23.313794
+INFO 2020-11-27 21:17:13 train.py: 79] Epoch 14, iter 2200/6416, lr 0.001000, loss 23.291044
+INFO 2020-11-27 21:18:31 train.py: 79] Epoch 14, iter 2400/6416, lr 0.001000, loss 23.266085
+INFO 2020-11-27 21:19:48 train.py: 79] Epoch 14, iter 2600/6416, lr 0.001000, loss 23.236668
+INFO 2020-11-27 21:21:05 train.py: 79] Epoch 14, iter 2800/6416, lr 0.001000, loss 23.249743
+INFO 2020-11-27 21:22:22 train.py: 92] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-27 21:22:22 train.py: 79] Epoch 14, iter 3000/6416, lr 0.001000, loss 23.241882
+INFO 2020-11-27 21:23:38 train.py: 79] Epoch 14, iter 3200/6416, lr 0.001000, loss 23.271876
+INFO 2020-11-27 21:24:55 train.py: 79] Epoch 14, iter 3400/6416, lr 0.001000, loss 23.281327
+INFO 2020-11-27 21:26:11 train.py: 79] Epoch 14, iter 3600/6416, lr 0.001000, loss 23.270969
+INFO 2020-11-27 21:27:27 train.py: 79] Epoch 14, iter 3800/6416, lr 0.001000, loss 23.256917
+INFO 2020-11-27 21:28:44 train.py: 79] Epoch 14, iter 4000/6416, lr 0.001000, loss 23.272061
+INFO 2020-11-27 21:30:00 train.py: 79] Epoch 14, iter 4200/6416, lr 0.001000, loss 23.313885
+INFO 2020-11-27 21:31:17 train.py: 79] Epoch 14, iter 4400/6416, lr 0.001000, loss 23.296111
+INFO 2020-11-27 21:32:33 train.py: 79] Epoch 14, iter 4600/6416, lr 0.001000, loss 23.256415
+INFO 2020-11-27 21:33:49 train.py: 79] Epoch 14, iter 4800/6416, lr 0.001000, loss 23.295370
+INFO 2020-11-27 21:35:06 train.py: 79] Epoch 14, iter 5000/6416, lr 0.001000, loss 23.290063
+INFO 2020-11-27 21:36:22 train.py: 79] Epoch 14, iter 5200/6416, lr 0.001000, loss 23.301908
+INFO 2020-11-27 21:37:39 train.py: 79] Epoch 14, iter 5400/6416, lr 0.001000, loss 23.275354
+INFO 2020-11-27 21:38:55 train.py: 79] Epoch 14, iter 5600/6416, lr 0.001000, loss 23.261909
+INFO 2020-11-27 21:40:11 train.py: 79] Epoch 14, iter 5800/6416, lr 0.001000, loss 23.292931
+INFO 2020-11-27 21:41:28 train.py: 92] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-27 21:41:28 train.py: 79] Epoch 14, iter 6000/6416, lr 0.001000, loss 23.275218
+INFO 2020-11-27 21:42:45 train.py: 79] Epoch 14, iter 6200/6416, lr 0.001000, loss 23.277818
+INFO 2020-11-27 21:44:02 train.py: 79] Epoch 14, iter 6400/6416, lr 0.001000, loss 23.256935
+INFO 2020-11-27 21:44:08 train.py: 97] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-27 21:44:10 train.py: 79] Epoch 15, iter 0/6416, lr 0.001000, loss 23.359136
+INFO 2020-11-27 21:45:27 train.py: 79] Epoch 15, iter 200/6416, lr 0.001000, loss 23.198156
+INFO 2020-11-27 21:46:44 train.py: 79] Epoch 15, iter 400/6416, lr 0.001000, loss 23.221968
+INFO 2020-11-27 21:48:02 train.py: 79] Epoch 15, iter 600/6416, lr 0.001000, loss 23.239415
+INFO 2020-11-27 21:49:19 train.py: 79] Epoch 15, iter 800/6416, lr 0.001000, loss 23.234750
+INFO 2020-11-27 21:50:36 train.py: 79] Epoch 15, iter 1000/6416, lr 0.001000, loss 23.216742
+INFO 2020-11-27 21:51:54 train.py: 79] Epoch 15, iter 1200/6416, lr 0.001000, loss 23.201826
+INFO 2020-11-27 21:53:11 train.py: 79] Epoch 15, iter 1400/6416, lr 0.001000, loss 23.254077
+INFO 2020-11-27 21:54:28 train.py: 79] Epoch 15, iter 1600/6416, lr 0.001000, loss 23.243932
+INFO 2020-11-27 21:55:45 train.py: 79] Epoch 15, iter 1800/6416, lr 0.001000, loss 23.239057
+INFO 2020-11-27 21:57:02 train.py: 79] Epoch 15, iter 2000/6416, lr 0.001000, loss 23.274510
+INFO 2020-11-27 21:58:20 train.py: 79] Epoch 15, iter 2200/6416, lr 0.001000, loss 23.218772
+INFO 2020-11-27 21:59:37 train.py: 79] Epoch 15, iter 2400/6416, lr 0.001000, loss 23.261813
+INFO 2020-11-27 22:00:54 train.py: 79] Epoch 15, iter 2600/6416, lr 0.001000, loss 23.246546
+INFO 2020-11-27 22:02:11 train.py: 79] Epoch 15, iter 2800/6416, lr 0.001000, loss 23.248505
+INFO 2020-11-27 22:03:28 train.py: 92] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-27 22:03:28 train.py: 79] Epoch 15, iter 3000/6416, lr 0.001000, loss 23.235488
+INFO 2020-11-27 22:04:45 train.py: 79] Epoch 15, iter 3200/6416, lr 0.001000, loss 23.244191
+INFO 2020-11-27 22:06:02 train.py: 79] Epoch 15, iter 3400/6416, lr 0.001000, loss 23.217826
+INFO 2020-11-27 22:07:19 train.py: 79] Epoch 15, iter 3600/6416, lr 0.001000, loss 23.239476
+INFO 2020-11-27 22:08:36 train.py: 79] Epoch 15, iter 3800/6416, lr 0.001000, loss 23.245796
+INFO 2020-11-27 22:09:53 train.py: 79] Epoch 15, iter 4000/6416, lr 0.001000, loss 23.277841
+INFO 2020-11-27 22:11:10 train.py: 79] Epoch 15, iter 4200/6416, lr 0.001000, loss 23.231695
+INFO 2020-11-27 22:12:28 train.py: 79] Epoch 15, iter 4400/6416, lr 0.001000, loss 23.254363
+INFO 2020-11-27 22:13:45 train.py: 79] Epoch 15, iter 4600/6416, lr 0.001000, loss 23.255496
+INFO 2020-11-27 22:15:02 train.py: 79] Epoch 15, iter 4800/6416, lr 0.001000, loss 23.264644
+INFO 2020-11-27 22:16:19 train.py: 79] Epoch 15, iter 5000/6416, lr 0.001000, loss 23.254548
+INFO 2020-11-27 22:17:36 train.py: 79] Epoch 15, iter 5200/6416, lr 0.001000, loss 23.235509
+INFO 2020-11-27 22:18:53 train.py: 79] Epoch 15, iter 5400/6416, lr 0.001000, loss 23.286478
+INFO 2020-11-27 22:20:10 train.py: 79] Epoch 15, iter 5600/6416, lr 0.001000, loss 23.261007
+INFO 2020-11-27 22:21:27 train.py: 79] Epoch 15, iter 5800/6416, lr 0.001000, loss 23.251082
+INFO 2020-11-27 22:22:43 train.py: 92] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-27 22:22:44 train.py: 79] Epoch 15, iter 6000/6416, lr 0.001000, loss 23.293034
+INFO 2020-11-27 22:24:00 train.py: 79] Epoch 15, iter 6200/6416, lr 0.001000, loss 23.270165
+INFO 2020-11-27 22:25:16 train.py: 79] Epoch 15, iter 6400/6416, lr 0.001000, loss 23.272867
+INFO 2020-11-27 22:25:22 train.py: 97] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-27 22:25:24 train.py: 79] Epoch 16, iter 0/6416, lr 0.000100, loss 23.244636
+INFO 2020-11-27 22:26:41 train.py: 79] Epoch 16, iter 200/6416, lr 0.000100, loss 23.209590
+INFO 2020-11-27 22:27:58 train.py: 79] Epoch 16, iter 400/6416, lr 0.000100, loss 23.215623
+INFO 2020-11-27 22:29:16 train.py: 79] Epoch 16, iter 600/6416, lr 0.000100, loss 23.225994
+INFO 2020-11-27 22:30:33 train.py: 79] Epoch 16, iter 800/6416, lr 0.000100, loss 23.180587
+INFO 2020-11-27 22:31:50 train.py: 79] Epoch 16, iter 1000/6416, lr 0.000100, loss 23.185012
+INFO 2020-11-27 22:33:07 train.py: 79] Epoch 16, iter 1200/6416, lr 0.000100, loss 23.203271
+INFO 2020-11-27 22:34:24 train.py: 79] Epoch 16, iter 1400/6416, lr 0.000100, loss 23.207087
+INFO 2020-11-27 22:35:41 train.py: 79] Epoch 16, iter 1600/6416, lr 0.000100, loss 23.201103
+INFO 2020-11-27 22:36:58 train.py: 79] Epoch 16, iter 1800/6416, lr 0.000100, loss 23.243564
+INFO 2020-11-27 22:38:15 train.py: 79] Epoch 16, iter 2000/6416, lr 0.000100, loss 23.178899
+INFO 2020-11-27 22:39:32 train.py: 79] Epoch 16, iter 2200/6416, lr 0.000100, loss 23.228556
+INFO 2020-11-27 22:40:49 train.py: 79] Epoch 16, iter 2400/6416, lr 0.000100, loss 23.187529
+INFO 2020-11-27 22:42:06 train.py: 79] Epoch 16, iter 2600/6416, lr 0.000100, loss 23.218757
+INFO 2020-11-27 22:43:23 train.py: 79] Epoch 16, iter 2800/6416, lr 0.000100, loss 23.171763
+INFO 2020-11-27 22:44:40 train.py: 92] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-27 22:44:40 train.py: 79] Epoch 16, iter 3000/6416, lr 0.000100, loss 23.194116
+INFO 2020-11-27 22:45:58 train.py: 79] Epoch 16, iter 3200/6416, lr 0.000100, loss 23.225780
+INFO 2020-11-27 22:47:15 train.py: 79] Epoch 16, iter 3400/6416, lr 0.000100, loss 23.225334
+INFO 2020-11-27 22:48:32 train.py: 79] Epoch 16, iter 3600/6416, lr 0.000100, loss 23.207852
+INFO 2020-11-27 22:49:48 train.py: 79] Epoch 16, iter 3800/6416, lr 0.000100, loss 23.210509
+INFO 2020-11-27 22:51:05 train.py: 79] Epoch 16, iter 4000/6416, lr 0.000100, loss 23.189287
+INFO 2020-11-27 22:52:22 train.py: 79] Epoch 16, iter 4200/6416, lr 0.000100, loss 23.241467
+INFO 2020-11-27 22:53:40 train.py: 79] Epoch 16, iter 4400/6416, lr 0.000100, loss 23.230613
+INFO 2020-11-27 22:54:56 train.py: 79] Epoch 16, iter 4600/6416, lr 0.000100, loss 23.178476
+INFO 2020-11-27 22:56:13 train.py: 79] Epoch 16, iter 4800/6416, lr 0.000100, loss 23.184672
+INFO 2020-11-27 22:57:30 train.py: 79] Epoch 16, iter 5000/6416, lr 0.000100, loss 23.179677
+INFO 2020-11-27 22:58:47 train.py: 79] Epoch 16, iter 5200/6416, lr 0.000100, loss 23.210603
+INFO 2020-11-27 23:00:04 train.py: 79] Epoch 16, iter 5400/6416, lr 0.000100, loss 23.186464
+INFO 2020-11-27 23:01:21 train.py: 79] Epoch 16, iter 5600/6416, lr 0.000100, loss 23.143685
+INFO 2020-11-27 23:02:38 train.py: 79] Epoch 16, iter 5800/6416, lr 0.000100, loss 23.192623
+INFO 2020-11-27 23:03:55 train.py: 92] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-27 23:03:55 train.py: 79] Epoch 16, iter 6000/6416, lr 0.000100, loss 23.177588
+INFO 2020-11-27 23:05:12 train.py: 79] Epoch 16, iter 6200/6416, lr 0.000100, loss 23.164168
+INFO 2020-11-27 23:06:29 train.py: 79] Epoch 16, iter 6400/6416, lr 0.000100, loss 23.186432
+INFO 2020-11-27 23:06:35 train.py: 97] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-27 23:06:37 train.py: 79] Epoch 17, iter 0/6416, lr 0.000100, loss 23.225369
+INFO 2020-11-27 23:07:54 train.py: 79] Epoch 17, iter 200/6416, lr 0.000100, loss 23.172257
+INFO 2020-11-27 23:09:11 train.py: 79] Epoch 17, iter 400/6416, lr 0.000100, loss 23.177428
+INFO 2020-11-27 23:10:28 train.py: 79] Epoch 17, iter 600/6416, lr 0.000100, loss 23.180162
+INFO 2020-11-27 23:11:46 train.py: 79] Epoch 17, iter 800/6416, lr 0.000100, loss 23.203078
+INFO 2020-11-27 23:13:03 train.py: 79] Epoch 17, iter 1000/6416, lr 0.000100, loss 23.218781
+INFO 2020-11-27 23:14:20 train.py: 79] Epoch 17, iter 1200/6416, lr 0.000100, loss 23.172844
+INFO 2020-11-27 23:15:37 train.py: 79] Epoch 17, iter 1400/6416, lr 0.000100, loss 23.177171
+INFO 2020-11-27 23:16:54 train.py: 79] Epoch 17, iter 1600/6416, lr 0.000100, loss 23.216808
+INFO 2020-11-27 23:18:11 train.py: 79] Epoch 17, iter 1800/6416, lr 0.000100, loss 23.178026
+INFO 2020-11-27 23:19:28 train.py: 79] Epoch 17, iter 2000/6416, lr 0.000100, loss 23.185704
+INFO 2020-11-27 23:20:45 train.py: 79] Epoch 17, iter 2200/6416, lr 0.000100, loss 23.226472
+INFO 2020-11-27 23:22:02 train.py: 79] Epoch 17, iter 2400/6416, lr 0.000100, loss 23.226721
+INFO 2020-11-27 23:23:19 train.py: 79] Epoch 17, iter 2600/6416, lr 0.000100, loss 23.176181
+INFO 2020-11-27 23:24:36 train.py: 79] Epoch 17, iter 2800/6416, lr 0.000100, loss 23.179059
+INFO 2020-11-27 23:25:53 train.py: 92] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-27 23:25:53 train.py: 79] Epoch 17, iter 3000/6416, lr 0.000100, loss 23.172584
+INFO 2020-11-27 23:27:10 train.py: 79] Epoch 17, iter 3200/6416, lr 0.000100, loss 23.235901
+INFO 2020-11-27 23:28:27 train.py: 79] Epoch 17, iter 3400/6416, lr 0.000100, loss 23.165605
+INFO 2020-11-27 23:29:44 train.py: 79] Epoch 17, iter 3600/6416, lr 0.000100, loss 23.205151
+INFO 2020-11-27 23:31:01 train.py: 79] Epoch 17, iter 3800/6416, lr 0.000100, loss 23.167248
+INFO 2020-11-27 23:32:18 train.py: 79] Epoch 17, iter 4000/6416, lr 0.000100, loss 23.207441
+INFO 2020-11-27 23:33:35 train.py: 79] Epoch 17, iter 4200/6416, lr 0.000100, loss 23.243991
+INFO 2020-11-27 23:34:52 train.py: 79] Epoch 17, iter 4400/6416, lr 0.000100, loss 23.175317
+INFO 2020-11-27 23:36:09 train.py: 79] Epoch 17, iter 4600/6416, lr 0.000100, loss 23.201301
+INFO 2020-11-27 23:37:26 train.py: 79] Epoch 17, iter 4800/6416, lr 0.000100, loss 23.210275
+INFO 2020-11-27 23:38:43 train.py: 79] Epoch 17, iter 5000/6416, lr 0.000100, loss 23.202372
+INFO 2020-11-27 23:40:00 train.py: 79] Epoch 17, iter 5200/6416, lr 0.000100, loss 23.197515
+INFO 2020-11-27 23:41:17 train.py: 79] Epoch 17, iter 5400/6416, lr 0.000100, loss 23.156143
+INFO 2020-11-27 23:42:34 train.py: 79] Epoch 17, iter 5600/6416, lr 0.000100, loss 23.183499
+INFO 2020-11-27 23:43:51 train.py: 79] Epoch 17, iter 5800/6416, lr 0.000100, loss 23.181795
+INFO 2020-11-27 23:45:08 train.py: 92] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-27 23:45:08 train.py: 79] Epoch 17, iter 6000/6416, lr 0.000100, loss 23.236270
+INFO 2020-11-27 23:46:25 train.py: 79] Epoch 17, iter 6200/6416, lr 0.000100, loss 23.180215
+INFO 2020-11-27 23:47:41 train.py: 79] Epoch 17, iter 6400/6416, lr 0.000100, loss 23.176069
+INFO 2020-11-27 23:47:47 train.py: 97] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-27 23:47:48 train.py: 180] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/.gitkeep b/bob/bio/facexzoo/models/heads/ArcFace/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..1a8744b81fef9893cd0d6ea20cffffcb617c6afb
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_agedb.txt
@@ -0,0 +1,50 @@
++------------------------+--------------------+-----------------------+ [9/1472]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9623333333333333 | 0.0026199613605670277 |
+| Epoch_13.pt | 0.9620000000000001 | 0.0028846122190549313 |
+| Epoch_14_batch_2999.pt | 0.9618333333333332 | 0.002551325000712231 |
+| Epoch_14_batch_5999.pt | 0.9618333333333332 | 0.0025873624493766745 |
+| Epoch_15_batch_5999.pt | 0.961 | 0.003044931023565495 |
+| Epoch_16.pt | 0.9608333333333332 | 0.002747052292239141 |
+| Epoch_17_batch_2999.pt | 0.9608333333333332 | 0.0027694319068032277 |
+| Epoch_13_batch_5999.pt | 0.9606666666666666 | 0.0024870032539554866 |
+| Epoch_16_batch_2999.pt | 0.9604999999999999 | 0.002594509873074581 |
+| Epoch_12_batch_5999.pt | 0.9603333333333334 | 0.00255555555555556 |
+| Epoch_12.pt | 0.9601666666666666 | 0.002158102733283899 |
+| Epoch_17_batch_5999.pt | 0.9598333333333333 | 0.0026579719364234855 |
+| Epoch_17.pt | 0.9596666666666668 | 0.0025190631219454756 |
+| Epoch_15.pt | 0.9594999999999999 | 0.0030025709148603684 |
+| Epoch_13_batch_2999.pt | 0.9593333333333334 | 0.0025361582690029633 |
+| Epoch_15_batch_2999.pt | 0.9591666666666667 | 0.0028463752127665543 |
+| Epoch_14.pt | 0.959 | 0.00286744175568088 |
+| Epoch_11.pt | 0.9586666666666666 | 0.0023934065809486744 |
+| Epoch_11_batch_2999.pt | 0.9576666666666668 | 0.0029627314724385324 |
+| Epoch_12_batch_2999.pt | 0.9571666666666667 | 0.002447598974656749 |
+| Epoch_10.pt | 0.9570000000000001 | 0.002393406580948667 |
+| Epoch_10_batch_2999.pt | 0.9561666666666667 | 0.0024349563334234606 |
+| Epoch_10_batch_5999.pt | 0.9561666666666666 | 0.0025464814747479646 |
+| Epoch_11_batch_5999.pt | 0.9555 | 0.0027222222222222266 |
+| Epoch_9_batch_5999.pt | 0.9463333333333332 | 0.0028631330503833662 |
+| Epoch_8_batch_5999.pt | 0.945 | 0.003397529966982365 |
+| Epoch_9_batch_2999.pt | 0.944 | 0.0032508308529617296 |
+| Epoch_6_batch_5999.pt | 0.9410000000000001 | 0.004158881616047304 |
+| Epoch_8.pt | 0.9406666666666667 | 0.003381138678822877 |
+| Epoch_5_batch_2999.pt | 0.9398333333333333 | 0.003437717445423328 |
+| Epoch_7_batch_2999.pt | 0.9391666666666666 | 0.005242572641483812 |
+| Epoch_9.pt | 0.9365 | 0.004040306185683713 |
+| Epoch_7_batch_5999.pt | 0.9361666666666666 | 0.005317395166622909 |
+| Epoch_6_batch_2999.pt | 0.9356666666666665 | 0.005254040185571948 |
+| Epoch_7.pt | 0.9354999999999999 | 0.004112987559751023 |
+| Epoch_5_batch_5999.pt | 0.9351666666666667 | 0.0040631587784947885 |
+| Epoch_8_batch_2999.pt | 0.9339999999999999 | 0.005838357624630958 |
+| Epoch_4_batch_5999.pt | 0.932 | 0.004477653706579977 |
+| Epoch_6.pt | 0.9315000000000001 | 0.005226062417830629 |
+| Epoch_3_batch_5999.pt | 0.9308333333333334 | 0.00506287626249967 |
+| Epoch_4_batch_2999.pt | 0.9301666666666666 | 0.004047938051543747 |
+| Epoch_3_batch_2999.pt | 0.9266666666666665 | 0.005049141230000782 |
+| Epoch_4.pt | 0.9256666666666666 | 0.004850862211765129 |
+| Epoch_3.pt | 0.924 | 0.005318265752788589 |
+| Epoch_5.pt | 0.9236666666666666 | 0.004316205466567784 |
+| Epoch_2_batch_2999.pt | 0.917 | 0.006304025756668571 |
+| Epoch_2_batch_5999.pt | 0.9143333333333334 | 0.00568841143829787 |
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..392315346db5c0fcf7a86adee4363ac6772bb0f6
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_calfw.txt
@@ -0,0 +1,50 @@
++------------------------+--------------------+-----------------------+ [10/1410]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.9398333333333333 | 0.0027605018330677327 |
+| Epoch_15_batch_5999.pt | 0.9396666666666667 | 0.0027755546659548433 |
+| Epoch_16_batch_2999.pt | 0.9396666666666667 | 0.0028087165910587767 |
+| Epoch_17.pt | 0.9395000000000001 | 0.002940248582755381 |
+| Epoch_12.pt | 0.9395 | 0.002710860644167964 |
+| Epoch_14.pt | 0.9393333333333332 | 0.0025361582690029633 |
+| Epoch_13_batch_5999.pt | 0.9388333333333334 | 0.0027894200468073682 |
+| Epoch_16.pt | 0.9388333333333334 | 0.0022641870969238686 |
+| Epoch_16_batch_5999.pt | 0.9386666666666666 | 0.0027307123838765488 |
+| Epoch_17_batch_2999.pt | 0.9386666666666666 | 0.0026736020923368766 |
+| Epoch_12_batch_2999.pt | 0.9381666666666668 | 0.002669558617051987 |
+| Epoch_13_batch_2999.pt | 0.9381666666666668 | 0.0029860788111948098 |
+| Epoch_17_batch_5999.pt | 0.938 | 0.00268512132746546 |
+| Epoch_14_batch_2999.pt | 0.937 | 0.002905932629027115 |
+| Epoch_10_batch_2999.pt | 0.9369999999999999 | 0.002567604446286963 |
+| Epoch_13.pt | 0.9369999999999999 | 0.0030102704853653467 |
+| Epoch_15.pt | 0.9369999999999999 | 0.0034587516480607504 |
+| Epoch_11_batch_2999.pt | 0.9368333333333332 | 0.0028267898296018947 |
+| Epoch_15_batch_2999.pt | 0.9366666666666665 | 0.0032203059435976502 |
+| Epoch_11_batch_5999.pt | 0.9353333333333333 | 0.0030912061651652265 |
+| Epoch_11.pt | 0.9353333333333333 | 0.0032848323331321023 |
+| Epoch_10_batch_5999.pt | 0.9351666666666668 | 0.0020253029037286814 |
+| Epoch_10.pt | 0.9348333333333333 | 0.0028158501994387073 |
+| Epoch_12_batch_5999.pt | 0.9348333333333333 | 0.0032150302880511674 |
+| Epoch_9_batch_2999.pt | 0.9235000000000001 | 0.0036383587400073197 |
+| Epoch_9_batch_5999.pt | 0.9216666666666666 | 0.004216370213557845 |
+| Epoch_8_batch_5999.pt | 0.9211666666666666 | 0.004374801582802229 |
+| Epoch_7_batch_5999.pt | 0.9209999999999999 | 0.0040230815533848115 |
+| Epoch_8_batch_2999.pt | 0.9200000000000002 | 0.003912625969257564 |
+| Epoch_7_batch_2999.pt | 0.9199999999999999 | 0.003557291243018254 |
+| Epoch_6_batch_5999.pt | 0.9198333333333333 | 0.0030776975521032285 |
+| Epoch_4_batch_5999.pt | 0.9186666666666665 | 0.004148478822798773 |
+| Epoch_9.pt | 0.9181666666666667 | 0.0045027426484338 |
+| Epoch_6_batch_2999.pt | 0.9171666666666667 | 0.003531603350069511 |
+| Epoch_5_batch_2999.pt | 0.9171666666666665 | 0.004127968441835539 |
+| Epoch_8.pt | 0.916 | 0.004535184807101904 |
+| Epoch_6.pt | 0.9151666666666666 | 0.0028267898296019055 |
+| Epoch_5_batch_5999.pt | 0.9146666666666668 | 0.003782676562344272 |
+| Epoch_7.pt | 0.9123333333333333 | 0.003744955454745049 |
+| Epoch_4_batch_2999.pt | 0.9116666666666667 | 0.004172218523448578 |
+| Epoch_3.pt | 0.9108333333333333 | 0.004958158260214507 |
+| Epoch_4.pt | 0.9099999999999999 | 0.004881307252553064 |
+| Epoch_3_batch_2999.pt | 0.9095000000000001 | 0.005158296685084918 |
+| Epoch_3_batch_5999.pt | 0.9063333333333334 | 0.004573136806057045 |
+| Epoch_5.pt | 0.9038333333333334 | 0.004075294430020738 |
+| Epoch_2.pt | 0.9028333333333334 | 0.004097951912424451 |
+| Epoch_2_batch_5999.pt | 0.9016666666666667 | 0.004956601782932328 |
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3164282c4a311096127fe72f9ac1e8e8ad86af99
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_cplfw.txt
@@ -0,0 +1,50 @@
++------------------------+--------------------+-----------------------+ [13/1347]
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8368333333333334 | 0.006801642866176702 |
+| Epoch_14.pt | 0.8361666666666666 | 0.006354011078848025 |
+| Epoch_13_batch_5999.pt | 0.8356666666666668 | 0.005989703098641585 |
+| Epoch_14_batch_5999.pt | 0.8348333333333333 | 0.0065989617589660475 |
+| Epoch_15.pt | 0.8348333333333333 | 0.00721987141228128 |
+| Epoch_15_batch_5999.pt | 0.8346666666666668 | 0.006230153675056078 |
+| Epoch_11.pt | 0.8345 | 0.007031893831977101 |
+| Epoch_16_batch_2999.pt | 0.8335000000000001 | 0.0071855908629271876 |
+| Epoch_17_batch_5999.pt | 0.833 | 0.006572482853853442 |
+| Epoch_13_batch_2999.pt | 0.8328333333333333 | 0.00680436497522358 |
+| Epoch_13.pt | 0.8321666666666665 | 0.006378252014888721 |
+| Epoch_16.pt | 0.8321666666666665 | 0.006383089154985522 |
+| Epoch_17_batch_2999.pt | 0.8321666666666665 | 0.006772538958958984 |
+| Epoch_14_batch_2999.pt | 0.8319999999999999 | 0.00630402575666857 |
+| Epoch_16_batch_5999.pt | 0.8316666666666667 | 0.006961091158967592 |
+| Epoch_12_batch_5999.pt | 0.8313333333333333 | 0.006338204754200736 |
+| Epoch_15_batch_2999.pt | 0.8311666666666667 | 0.006241290227485316 |
+| Epoch_10_batch_5999.pt | 0.8305 | 0.006781647347123022 |
+| Epoch_11_batch_5999.pt | 0.8300000000000001 | 0.006039990188856397 |
+| Epoch_11_batch_2999.pt | 0.829 | 0.006890680770539397 |
+| Epoch_10_batch_2999.pt | 0.8288333333333334 | 0.006992279693250402 |
+| Epoch_12_batch_2999.pt | 0.8285 | 0.008060151945386394 |
+| Epoch_10.pt | 0.8271666666666666 | 0.0065783502003180575 |
+| Epoch_12.pt | 0.8265 | 0.007283712592888499 |
+| Epoch_9_batch_2999.pt | 0.8091666666666667 | 0.008237410896125367 |
+| Epoch_8_batch_5999.pt | 0.8056666666666666 | 0.0065508457657705195 |
+| Epoch_7_batch_5999.pt | 0.8045 | 0.006116411842507053 |
+| Epoch_8_batch_2999.pt | 0.8031666666666666 | 0.007885954070684046 |
+| Epoch_6_batch_2999.pt | 0.8029999999999999 | 0.007039570693980959 |
+| Epoch_9_batch_5999.pt | 0.8028333333333333 | 0.006943555498659381 |
+| Epoch_8.pt | 0.8025 | 0.007487644143174462 |
+| Epoch_6_batch_5999.pt | 0.8015000000000001 | 0.00635595375597937 |
+| Epoch_5_batch_5999.pt | 0.7958333333333334 | 0.008510160956537362 |
+| Epoch_6.pt | 0.7955000000000001 | 0.008054789235580883 |
+| Epoch_7_batch_2999.pt | 0.7938333333333334 | 0.008311267081127116 |
+| Epoch_5_batch_2999.pt | 0.7921666666666666 | 0.007495060101555291 |
+| Epoch_9.pt | 0.7916666666666666 | 0.0071015560805665736 |
+| Epoch_4_batch_5999.pt | 0.791 | 0.0076827528502500155 |
+| Epoch_3.pt | 0.7893333333333332 | 0.0068231631634862394 |
+| Epoch_5.pt | 0.7891666666666667 | 0.007561476438231221 |
+| Epoch_4_batch_2999.pt | 0.7821666666666667 | 0.008692447409982337 |
+| Epoch_4.pt | 0.7819999999999998 | 0.00758002573369358 |
+| Epoch_3_batch_2999.pt | 0.7816666666666666 | 0.008104258901693542 |
+| Epoch_7.pt | 0.781 | 0.009382805392687044 |
+| Epoch_3_batch_5999.pt | 0.7798333333333334 | 0.007545949776183998 |
+| Epoch_2_batch_5999.pt | 0.7796666666666666 | 0.007984552988098532 |
+| Epoch_2_batch_2999.pt | 0.7673333333333334 | 0.00808214004174148 |
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3f53df0b96b9f2386c951a3934a12cd59fb2f783
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13.pt | 0.9956666666666665 | 0.0010599324460188284 |
+| Epoch_16.pt | 0.9953333333333335 | 0.001160034056545619 |
+| Epoch_17_batch_2999.pt | 0.9953333333333335 | 0.001160034056545619 |
+| Epoch_15_batch_2999.pt | 0.9953333333333333 | 0.0010772621905369623 |
+| Epoch_12_batch_2999.pt | 0.9951666666666668 | 0.0011772011166898413 |
+| Epoch_13_batch_2999.pt | 0.9951666666666668 | 0.001277777777777775 |
+| Epoch_15_batch_5999.pt | 0.9951666666666668 | 0.0012031337682059844 |
+| Epoch_15.pt | 0.9951666666666666 | 0.0011772011166898376 |
+| Epoch_13_batch_5999.pt | 0.9950000000000001 | 0.0013146843962443548 |
+| Epoch_14_batch_2999.pt | 0.9950000000000001 | 0.0011385500851066213 |
+| Epoch_14_batch_5999.pt | 0.9950000000000001 | 0.0011915339216404008 |
+| Epoch_11_batch_2999.pt | 0.9949999999999999 | 0.0012668615834434884 |
+| Epoch_17_batch_5999.pt | 0.9949999999999999 | 0.0012668615834434884 |
+| Epoch_10_batch_2999.pt | 0.9948333333333335 | 0.0010957268290731129 |
+| Epoch_10_batch_5999.pt | 0.9948333333333335 | 0.0011772011166898408 |
+| Epoch_16_batch_2999.pt | 0.9948333333333335 | 0.0012031337682059844 |
+| Epoch_16_batch_5999.pt | 0.9948333333333335 | 0.0013709958532503379 |
+| Epoch_10.pt | 0.9946666666666667 | 0.0009229582069908961 |
+| Epoch_14.pt | 0.9946666666666667 | 0.0011331154474650623 |
+| Epoch_17.pt | 0.9946666666666667 | 0.0011863420280034786 |
+| Epoch_12_batch_5999.pt | 0.9943333333333333 | 0.0012222222222222259 |
+| Epoch_11.pt | 0.9941666666666666 | 0.00114530711822713 |
+| Epoch_11_batch_5999.pt | 0.9940000000000001 | 0.0012472191289246398 |
+| Epoch_12.pt | 0.994 | 0.0014740554623801777 |
+| Epoch_6_batch_2999.pt | 0.9936666666666667 | 0.0013333333333333326 |
+| Epoch_9_batch_5999.pt | 0.9936666666666666 | 0.0013788526273323211 |
+| Epoch_5_batch_5999.pt | 0.9935 | 0.001393326244887162 |
+| Epoch_9.pt | 0.9935 | 0.0011772011166898408 |
+| Epoch_4_batch_2999.pt | 0.9933333333333334 | 0.0010243938285881 |
+| Epoch_8_batch_5999.pt | 0.9933333333333334 | 0.0012909944487358082 |
+| Epoch_9_batch_2999.pt | 0.9931666666666666 | 0.0014792807728549245 |
+| Epoch_8.pt | 0.9926666666666668 | 0.0014948471163415214 |
+| Epoch_4_batch_5999.pt | 0.9925 | 0.0012484558363469022 |
+| Epoch_6_batch_5999.pt | 0.9924999999999999 | 0.0014109361221333657 |
+| Epoch_8_batch_2999.pt | 0.9923333333333332 | 0.0015355861067872492 |
+| Epoch_2_batch_5999.pt | 0.9921666666666666 | 0.0012921892610681092 |
+| Epoch_4.pt | 0.992 | 0.0012862041003100224 |
+| Epoch_3_batch_5999.pt | 0.9916666666666668 | 0.0013146843962443676 |
+| Epoch_3_batch_2999.pt | 0.9913333333333332 | 0.0013562839573037495 |
+| Epoch_7_batch_5999.pt | 0.9911666666666668 | 0.0016301556390134685 |
+| Epoch_5_batch_2999.pt | 0.991 | 0.001515353521887317 |
+| Epoch_3.pt | 0.9908333333333333 | 0.0015565473029024329 |
+| Epoch_7_batch_2999.pt | 0.9908333333333333 | 0.0016149379837498495 |
+| Epoch_2.pt | 0.9906666666666668 | 0.0015947444549341545 |
+| Epoch_6.pt | 0.9903333333333333 | 0.002045772515502441 |
+| Epoch_7.pt | 0.9901666666666665 | 0.0022284634577924 |
+| Epoch_2_batch_2999.pt | 0.9896666666666667 | 0.0014865653511399585 |
+| Epoch_5.pt | 0.9894999999999999 | 0.0016489802310728709 |
+| Epoch_1_batch_5999.pt | 0.9879999999999999 | 0.0013562839573037482 |
+| Epoch_1.pt | 0.9863333333333333 | 0.0017177360926378049 |
+| Epoch_1_batch_2999.pt | 0.9821666666666667 | 0.00149174684245528 |
+| Epoch_0_batch_5999.pt | 0.9768333333333334 | 0.0034556270089075476 |
+| Epoch_0.pt | 0.9713333333333333 | 0.002 |
+| Epoch_0_batch_2999.pt | 0.9408333333333333 | 0.004397319610067753 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0424e0027d89a5c60778bd24d6465b1713e11f0e
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.8821666666666668 | 0.004534163867275719 |
+| Epoch_13.pt | 0.8821666666666668 | 0.005006477285958144 |
+| Epoch_14_batch_2999.pt | 0.8818333333333334 | 0.005231964907195338 |
+| Epoch_14_batch_5999.pt | 0.8818333333333334 | 0.004984234403857082 |
+| Epoch_15_batch_2999.pt | 0.8818333333333334 | 0.005552499159259757 |
+| Epoch_16_batch_2999.pt | 0.8815 | 0.005569150033781232 |
+| Epoch_16_batch_5999.pt | 0.881 | 0.0051111111111111045 |
+| Epoch_17.pt | 0.8805 | 0.005146315954275886 |
+| Epoch_17_batch_2999.pt | 0.8803333333333334 | 0.005521003666135487 |
+| Epoch_17_batch_5999.pt | 0.8803333333333333 | 0.005251689910265101 |
+| Epoch_15_batch_5999.pt | 0.8791666666666667 | 0.004649771001927978 |
+| Epoch_16.pt | 0.8785000000000001 | 0.0053313074856148395 |
+| Epoch_12_batch_5999.pt | 0.8778333333333335 | 0.00504944685883978 |
+| Epoch_15.pt | 0.8768333333333335 | 0.0050151006538155605 |
+| Epoch_11_batch_5999.pt | 0.8765000000000001 | 0.004611111111111107 |
+| Epoch_13_batch_5999.pt | 0.8765000000000001 | 0.00506409535143198 |
+| Epoch_14.pt | 0.8765000000000001 | 0.005319716413122925 |
+| Epoch_12_batch_2999.pt | 0.8753333333333334 | 0.004546060565661954 |
+| Epoch_10_batch_5999.pt | 0.8741666666666668 | 0.005242572641483808 |
+| Epoch_11_batch_2999.pt | 0.873 | 0.0051147330162150725 |
+| Epoch_11.pt | 0.8718333333333333 | 0.004730058856940733 |
+| Epoch_12.pt | 0.8714999999999999 | 0.004536885862938736 |
+| Epoch_10_batch_2999.pt | 0.8698333333333335 | 0.004637807591875637 |
+| Epoch_10.pt | 0.8696666666666667 | 0.00471273556898855 |
+| Epoch_5_batch_2999.pt | 0.8458333333333334 | 0.004439233055771769 |
+| Epoch_9_batch_2999.pt | 0.841 | 0.006374137559672649 |
+| Epoch_6_batch_5999.pt | 0.8401666666666667 | 0.004934446821285926 |
+| Epoch_7_batch_2999.pt | 0.8400000000000001 | 0.00447213595499958 |
+| Epoch_7_batch_5999.pt | 0.8394999999999999 | 0.0056111111111111145 |
+| Epoch_6.pt | 0.8371666666666666 | 0.004416928713400213 |
+| Epoch_8_batch_5999.pt | 0.8365 | 0.0036972629182166275 |
+| Epoch_9_batch_5999.pt | 0.8358333333333332 | 0.0045423249607863545 |
+| Epoch_6_batch_2999.pt | 0.8351666666666666 | 0.0049966037848438675 |
+| Epoch_8_batch_2999.pt | 0.8351666666666666 | 0.005394613225417728 |
+| Epoch_5_batch_5999.pt | 0.8326666666666667 | 0.004850862211765134 |
+| Epoch_7.pt | 0.8324999999999999 | 0.004397319610067748 |
+| Epoch_8.pt | 0.8316666666666667 | 0.0050184843513938785 |
+| Epoch_3_batch_5999.pt | 0.8306666666666667 | 0.005639367795755346 |
+| Epoch_4_batch_5999.pt | 0.8276666666666668 | 0.005265775827143459 |
+| Epoch_9.pt | 0.826 | 0.005212757380328837 |
+| Epoch_4_batch_2999.pt | 0.8233333333333333 | 0.006790516262487791 |
+| Epoch_3_batch_2999.pt | 0.8151666666666666 | 0.005045778090959793 |
+| Epoch_4.pt | 0.8116666666666668 | 0.00592650460137537 |
+| Epoch_5.pt | 0.8115 | 0.005684340629753092 |
+| Epoch_2_batch_5999.pt | 0.8108333333333333 | 0.006046374283428758 |
+| Epoch_3.pt | 0.8094999999999999 | 0.004913134324327886 |
+| Epoch_2.pt | 0.8015000000000001 | 0.006052496679006334 |
+| Epoch_2_batch_2999.pt | 0.7971666666666666 | 0.004688112099953285 |
+| Epoch_1_batch_5999.pt | 0.7835 | 0.005008942620238278 |
+| Epoch_1.pt | 0.7745 | 0.00441692871340021 |
+| Epoch_1_batch_2999.pt | 0.7651666666666667 | 0.005308100029758853 |
+| Epoch_0.pt | 0.733 | 0.0036243347622889064 |
+| Epoch_0_batch_5999.pt | 0.7166666666666666 | 0.0051520102752753896 |
+| Epoch_0_batch_2999.pt | 0.6196666666666666 | 0.007234178138070234 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..c001ee5bdf1ceca8c242d0de7620d1426a07eff4
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11_batch_5999.pt | 0.8800000000000001 | 0.00372677996249965 |
+| Epoch_17.pt | 0.8799999999999999 | 0.004059739090321136 |
+| Epoch_16_batch_5999.pt | 0.8783333333333335 | 0.004238273582835694 |
+| Epoch_14_batch_5999.pt | 0.8776666666666666 | 0.004452769979891972 |
+| Epoch_13_batch_2999.pt | 0.8771666666666667 | 0.004037249399783156 |
+| Epoch_16_batch_2999.pt | 0.8766666666666667 | 0.0041869874847592845 |
+| Epoch_17_batch_5999.pt | 0.8766666666666667 | 0.004339027597725921 |
+| Epoch_15_batch_5999.pt | 0.8765000000000001 | 0.003844588950399093 |
+| Epoch_12_batch_5999.pt | 0.8763333333333334 | 0.003911047979288458 |
+| Epoch_16.pt | 0.8760000000000001 | 0.0037614024177357324 |
+| Epoch_11_batch_2999.pt | 0.876 | 0.0032030078456443583 |
+| Epoch_12_batch_2999.pt | 0.876 | 0.003753187945345449 |
+| Epoch_13_batch_5999.pt | 0.876 | 0.004232443767720206 |
+| Epoch_15_batch_2999.pt | 0.876 | 0.004188461516660172 |
+| Epoch_14.pt | 0.875 | 0.004245549594342845 |
+| Epoch_11.pt | 0.8743333333333334 | 0.003961231882216867 |
+| Epoch_14_batch_2999.pt | 0.874 | 0.00458257569495584 |
+| Epoch_13.pt | 0.8736666666666666 | 0.0034587516480607526 |
+| Epoch_17_batch_2999.pt | 0.8736666666666666 | 0.0040960685758148355 |
+| Epoch_12.pt | 0.8728333333333331 | 0.003685557397915998 |
+| Epoch_15.pt | 0.8711666666666666 | 0.00483077837962853 |
+| Epoch_10_batch_5999.pt | 0.8708333333333333 | 0.0029423472615256402 |
+| Epoch_10.pt | 0.8693333333333333 | 0.003602125572766073 |
+| Epoch_10_batch_2999.pt | 0.869 | 0.004061259307221572 |
+| Epoch_9_batch_5999.pt | 0.8438333333333332 | 0.005334779896416515 |
+| Epoch_6_batch_2999.pt | 0.8423333333333334 | 0.0048253446112463275 |
+| Epoch_6_batch_5999.pt | 0.8413333333333334 | 0.005047918529600243 |
+| Epoch_9_batch_2999.pt | 0.8388333333333333 | 0.00537512560863855 |
+| Epoch_7_batch_5999.pt | 0.8386666666666667 | 0.004633479880442895 |
+| Epoch_7_batch_2999.pt | 0.8378333333333334 | 0.004296495844663398 |
+| Epoch_8_batch_5999.pt | 0.8375 | 0.003867001901176638 |
+| Epoch_5_batch_5999.pt | 0.8353333333333334 | 0.006274581290442445 |
+| Epoch_4_batch_5999.pt | 0.8346666666666668 | 0.005011098792790973 |
+| Epoch_6.pt | 0.8344999999999999 | 0.006795287037787724 |
+| Epoch_8_batch_2999.pt | 0.8331666666666665 | 0.003939746811442541 |
+| Epoch_8.pt | 0.8323333333333333 | 0.0038103173776627246 |
+| Epoch_7.pt | 0.8291666666666666 | 0.007650748369753379 |
+| Epoch_5_batch_2999.pt | 0.8273333333333334 | 0.007201680188043912 |
+| Epoch_9.pt | 0.8268333333333333 | 0.004405734198423743 |
+| Epoch_4_batch_2999.pt | 0.8216666666666667 | 0.0042600643361512935 |
+| Epoch_4.pt | 0.8183333333333334 | 0.004296136650929157 |
+| Epoch_3.pt | 0.8165000000000001 | 0.005220153254455277 |
+| Epoch_3_batch_5999.pt | 0.8160000000000001 | 0.004656072631057105 |
+| Epoch_2_batch_5999.pt | 0.812 | 0.0055876848714134005 |
+| Epoch_5.pt | 0.8118333333333332 | 0.0049842344038570855 |
+| Epoch_3_batch_2999.pt | 0.8073333333333332 | 0.003969015799887058 |
+| Epoch_2.pt | 0.8061666666666666 | 0.005253158955556715 |
+| Epoch_2_batch_2999.pt | 0.7991666666666667 | 0.006112373606964084 |
+| Epoch_1.pt | 0.7878333333333334 | 0.004887941827460916 |
+| Epoch_1_batch_5999.pt | 0.7828333333333333 | 0.00559458512456906 |
+| Epoch_1_batch_2999.pt | 0.7671666666666667 | 0.004655078204114274 |
+| Epoch_0.pt | 0.7381666666666666 | 0.004197662488858576 |
+| Epoch_0_batch_5999.pt | 0.7333333333333335 | 0.004310481053615736 |
+| Epoch_0_batch_2999.pt | 0.6751666666666667 | 0.006958652132861134 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..684254e661d3f17ecb26d8dd792d4ff9e92dc9f2
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.9513333333333334 | 0.003331480966792213 |
+| Epoch_13_batch_2999.pt | 0.951 | 0.0027011657291429376 |
+| Epoch_17_batch_5999.pt | 0.9506666666666668 | 0.002802115602870778 |
+| Epoch_17.pt | 0.9506666666666668 | 0.0025603819159562015 |
+| Epoch_14_batch_5999.pt | 0.9505000000000001 | 0.0027448042948968092 |
+| Epoch_15.pt | 0.9503333333333333 | 0.003091206165165229 |
+| Epoch_12.pt | 0.95 | 0.0028109134757052273 |
+| Epoch_13_batch_5999.pt | 0.95 | 0.0026874192494328528 |
+| Epoch_16.pt | 0.9499999999999998 | 0.002545875386086581 |
+| Epoch_14_batch_2999.pt | 0.9496666666666668 | 0.002819683897877674 |
+| Epoch_15_batch_2999.pt | 0.9490000000000001 | 0.0027688746209726914 |
+| Epoch_15_batch_5999.pt | 0.9485000000000001 | 0.0025633937766798556 |
+| Epoch_17_batch_2999.pt | 0.9484999999999999 | 0.003437717445423324 |
+| Epoch_16_batch_2999.pt | 0.9481666666666667 | 0.0032341732395173104 |
+| Epoch_12_batch_5999.pt | 0.9478333333333333 | 0.002509857110683667 |
+| Epoch_13.pt | 0.9475 | 0.002824605293085815 |
+| Epoch_16_batch_5999.pt | 0.9473333333333332 | 0.002834966849371794 |
+| Epoch_12_batch_2999.pt | 0.9469999999999998 | 0.0024062675364119645 |
+| Epoch_11_batch_5999.pt | 0.9468333333333334 | 0.003057575090181551 |
+| Epoch_11.pt | 0.9465 | 0.0025633937766798495 |
+| Epoch_11_batch_2999.pt | 0.9458333333333332 | 0.0029423472615256307 |
+| Epoch_10_batch_5999.pt | 0.9456666666666667 | 0.0021401511426953597 |
+| Epoch_10_batch_2999.pt | 0.9446666666666668 | 0.0027532248207475275 |
+| Epoch_10.pt | 0.9445 | 0.002409472049133492 |
+| Epoch_9_batch_5999.pt | 0.9258333333333335 | 0.0029212926278077565 |
+| Epoch_9_batch_2999.pt | 0.9246666666666666 | 0.004020011670027897 |
+| Epoch_7_batch_5999.pt | 0.9243333333333332 | 0.003034777840832814 |
+| Epoch_8_batch_5999.pt | 0.9236666666666669 | 0.003294214906749693 |
+| Epoch_8_batch_2999.pt | 0.9206666666666667 | 0.0036868133384526836 |
+| Epoch_6_batch_2999.pt | 0.9193333333333333 | 0.002462057756240044 |
+| Epoch_7_batch_2999.pt | 0.9190000000000002 | 0.004247003300803639 |
+| Epoch_5_batch_5999.pt | 0.9185000000000001 | 0.003224615969095879 |
+| Epoch_6_batch_5999.pt | 0.9179999999999999 | 0.0031308895119123033 |
+| Epoch_5_batch_2999.pt | 0.9163333333333334 | 0.0040734006177385205 |
+| Epoch_4_batch_5999.pt | 0.9161666666666666 | 0.003643444984904356 |
+| Epoch_8.pt | 0.9148333333333334 | 0.004820545023230016 |
+| Epoch_6.pt | 0.9123333333333334 | 0.004083994655442028 |
+| Epoch_9.pt | 0.9099999999999999 | 0.003122993182790048 |
+| Epoch_4.pt | 0.9088333333333335 | 0.00331522861063015 |
+| Epoch_3_batch_5999.pt | 0.9086666666666667 | 0.0031603250340728668 |
+| Epoch_4_batch_2999.pt | 0.908 | 0.004673275743291798 |
+| Epoch_3_batch_2999.pt | 0.9063333333333334 | 0.004323350324076378 |
+| Epoch_3.pt | 0.9045 | 0.004668319813010154 |
+| Epoch_7.pt | 0.9028333333333333 | 0.003685557397916 |
+| Epoch_5.pt | 0.9023333333333333 | 0.0029418227321941536 |
+| Epoch_2_batch_5999.pt | 0.8991666666666667 | 0.0034805455794837954 |
+| Epoch_2.pt | 0.8958333333333334 | 0.004166666666666659 |
+| Epoch_2_batch_2999.pt | 0.8939999999999999 | 0.003984538017120241 |
+| Epoch_1.pt | 0.8891666666666665 | 0.0037039351779518397 |
+| Epoch_1_batch_5999.pt | 0.8881666666666668 | 0.004009633461281367 |
+| Epoch_1_batch_2999.pt | 0.8751666666666666 | 0.003552516061944044 |
+| Epoch_0.pt | 0.8418333333333334 | 0.004327988156347185 |
+| Epoch_0_batch_5999.pt | 0.8383333333333333 | 0.006564024661218636 |
+| Epoch_0_batch_2999.pt | 0.7586666666666667 | 0.008387969044492119 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..a66268adcacef92e54ff314cbfcbc48572f5db46
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.9069999999999998 | 0.003422871511277636 |
+| Epoch_15_batch_2999.pt | 0.9056666666666666 | 0.0037118429085533523 |
+| Epoch_13_batch_5999.pt | 0.9055 | 0.0041279684418355405 |
+| Epoch_16.pt | 0.9035 | 0.004032659876435444 |
+| Epoch_15_batch_5999.pt | 0.9033333333333335 | 0.00333333333333333 |
+| Epoch_13_batch_2999.pt | 0.9033333333333333 | 0.004628147911035227 |
+| Epoch_14_batch_2999.pt | 0.9033333333333331 | 0.004388537257362553 |
+| Epoch_14_batch_5999.pt | 0.9028333333333333 | 0.004628481339075796 |
+| Epoch_16_batch_2999.pt | 0.9026666666666667 | 0.004136557881996948 |
+| Epoch_13.pt | 0.9023333333333333 | 0.003661612678507565 |
+| Epoch_16_batch_5999.pt | 0.9021666666666667 | 0.0037765520887460737 |
+| Epoch_12_batch_2999.pt | 0.9016666666666666 | 0.004785523437190674 |
+| Epoch_12.pt | 0.9005000000000001 | 0.004296495844663402 |
+| Epoch_17_batch_2999.pt | 0.9004999999999999 | 0.0036434449849043517 |
+| Epoch_11_batch_5999.pt | 0.9001666666666666 | 0.003994208770670818 |
+| Epoch_11_batch_2999.pt | 0.9 | 0.0038086970002228038 |
+| Epoch_17.pt | 0.8996666666666668 | 0.0035986966090448074 |
+| Epoch_12_batch_5999.pt | 0.8995 | 0.003296556377588173 |
+| Epoch_17_batch_5999.pt | 0.8993333333333332 | 0.004326204963095945 |
+| Epoch_15.pt | 0.8991666666666667 | 0.003794489181450923 |
+| Epoch_10_batch_5999.pt | 0.8978333333333334 | 0.0036603480911175765 |
+| Epoch_10_batch_2999.pt | 0.8968333333333334 | 0.00395538389140612 |
+| Epoch_11.pt | 0.8959999999999999 | 0.0042105100714436424 |
+| Epoch_10.pt | 0.8943333333333333 | 0.0032030078456443496 |
+| Epoch_9_batch_2999.pt | 0.8761666666666666 | 0.004388888888888896 |
+| Epoch_6_batch_2999.pt | 0.8758333333333332 | 0.00430510750356875 |
+| Epoch_7_batch_5999.pt | 0.8753333333333334 | 0.0062202377787904 |
+| Epoch_6_batch_5999.pt | 0.875 | 0.0033425797681091883 |
+| Epoch_5_batch_2999.pt | 0.8728333333333333 | 0.004931944248881379 |
+| Epoch_9.pt | 0.8711666666666668 | 0.004944444444444441 |
+| Epoch_8_batch_2999.pt | 0.868 | 0.004484541349024564 |
+| Epoch_9_batch_5999.pt | 0.8673333333333334 | 0.004053652521545493 |
+| Epoch_8_batch_5999.pt | 0.8671666666666666 | 0.0046484432546131 |
+| Epoch_8.pt | 0.8671666666666665 | 0.005916340627427828 |
+| Epoch_5_batch_5999.pt | 0.8663333333333334 | 0.004155912046503345 |
+| Epoch_4_batch_5999.pt | 0.8633333333333333 | 0.006186404847588914 |
+| Epoch_7_batch_2999.pt | 0.8633333333333333 | 0.004740162001711448 |
+| Epoch_4_batch_2999.pt | 0.8616666666666667 | 0.005299662230094137 |
+| Epoch_3_batch_5999.pt | 0.8585 | 0.004212342241614968 |
+| Epoch_6.pt | 0.8583333333333332 | 0.005351819812796581 |
+| Epoch_7.pt | 0.8571666666666667 | 0.005098111487672739 |
+| Epoch_3.pt | 0.8546666666666667 | 0.004706181907677196 |
+| Epoch_3_batch_2999.pt | 0.8530000000000001 | 0.004738859580474691 |
+| Epoch_2_batch_5999.pt | 0.85 | 0.005621826951410453 |
+| Epoch_4.pt | 0.8488333333333333 | 0.004044887033170529 |
+| Epoch_5.pt | 0.8479999999999999 | 0.003895232920657351 |
+| Epoch_2.pt | 0.8423333333333334 | 0.003014368881389011 |
+| Epoch_2_batch_2999.pt | 0.8358333333333332 | 0.0025968879761156124 |
+| Epoch_1_batch_5999.pt | 0.8291666666666666 | 0.004976798018658033 |
+| Epoch_1.pt | 0.8223333333333332 | 0.005123174170399335 |
+| Epoch_1_batch_2999.pt | 0.8073333333333335 | 0.00524227827203798 |
+| Epoch_0_batch_5999.pt | 0.7898333333333334 | 0.003970959395224095 |
+| Epoch_0.pt | 0.7803333333333333 | 0.0053736899063750055 |
+| Epoch_0_batch_2999.pt | 0.7038333333333334 | 0.004714372560794844 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/ArcFace/log.log b/bob/bio/facexzoo/models/heads/ArcFace/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..cbd893a93d786db9f0f6edc2e0bbbba7d9285733
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/ArcFace/log.log
@@ -0,0 +1,651 @@
+INFO 2020-11-24 00:29:12 train.py: 172] Start optimization.
+INFO 2020-11-24 00:29:12 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='arcface', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='arc-mobile', train_file='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f8ee40187b8>)
+INFO 2020-11-24 00:29:53 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 23.359434
+INFO 2020-11-24 00:38:17 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 22.847342
+INFO 2020-11-24 00:45:57 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 22.217041
+INFO 2020-11-24 00:52:58 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 21.791875
+INFO 2020-11-24 00:59:50 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 21.443408
+INFO 2020-11-24 01:06:58 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 21.076518
+INFO 2020-11-24 01:13:09 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 20.673717
+INFO 2020-11-24 01:19:36 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 20.258046
+INFO 2020-11-24 01:26:01 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 19.830074
+INFO 2020-11-24 01:32:16 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 19.418279
+INFO 2020-11-24 01:38:42 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 18.993834
+INFO 2020-11-24 01:45:11 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 18.566143
+INFO 2020-11-24 01:51:23 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 18.108684
+INFO 2020-11-24 01:57:49 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 17.637085
+INFO 2020-11-24 02:04:15 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 17.140226
+INFO 2020-11-24 02:10:27 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-24 02:10:27 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 16.658029
+INFO 2020-11-24 02:16:59 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 16.150862
+INFO 2020-11-24 02:23:10 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 15.592323
+INFO 2020-11-24 02:29:54 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 15.077606
+INFO 2020-11-24 02:36:10 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 14.570124
+INFO 2020-11-24 02:42:41 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 14.016759
+INFO 2020-11-24 02:56:07 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 13.490575
+INFO 2020-11-24 03:02:48 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 12.965636
+INFO 2020-11-24 03:09:18 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 12.419055
+INFO 2020-11-24 03:15:29 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 11.930718
+INFO 2020-11-24 03:22:05 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 11.403103
+INFO 2020-11-24 03:28:08 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 10.970842
+INFO 2020-11-24 03:34:42 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 10.545405
+INFO 2020-11-24 03:40:57 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 10.064106
+INFO 2020-11-24 03:47:24 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 9.671036
+INFO 2020-11-24 03:53:34 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-24 03:53:34 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 9.325449
+INFO 2020-11-24 04:00:04 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 9.009621
+INFO 2020-11-24 04:06:16 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 8.697339
+INFO 2020-11-24 04:06:44 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-24 04:06:46 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 8.564224
+INFO 2020-11-24 04:08:04 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 8.043653
+INFO 2020-11-24 04:09:23 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 7.813574
+INFO 2020-11-24 04:10:41 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 7.658158
+INFO 2020-11-24 04:11:59 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 7.518124
+INFO 2020-11-24 04:13:17 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 7.372449
+INFO 2020-11-24 04:14:35 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 7.258654
+INFO 2020-11-24 04:15:53 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 7.121700
+INFO 2020-11-24 04:17:11 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 6.975845
+INFO 2020-11-24 04:18:29 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 6.874810
+INFO 2020-11-24 04:19:47 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 6.779687
+INFO 2020-11-24 04:21:05 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 6.673733
+INFO 2020-11-24 04:22:23 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 6.561958
+INFO 2020-11-24 04:23:41 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 6.476950
+INFO 2020-11-24 04:24:59 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 6.400562
+INFO 2020-11-24 04:26:17 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-24 04:26:17 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 6.286831
+INFO 2020-11-24 04:27:35 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 6.194930
+INFO 2020-11-24 04:28:52 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 6.154781
+INFO 2020-11-24 04:30:10 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 6.086702
+INFO 2020-11-24 04:31:27 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 6.029971
+INFO 2020-11-24 04:32:45 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 5.942696
+INFO 2020-11-24 04:34:02 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 5.885881
+INFO 2020-11-24 04:35:20 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 5.837472
+INFO 2020-11-24 04:36:37 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 5.788877
+INFO 2020-11-24 04:37:55 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 5.757636
+INFO 2020-11-24 04:39:12 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 5.714747
+INFO 2020-11-24 04:40:30 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 5.616349
+INFO 2020-11-24 04:41:47 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 5.610363
+INFO 2020-11-24 04:43:04 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 5.578490
+INFO 2020-11-24 04:44:22 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 5.508913
+INFO 2020-11-24 04:45:39 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-24 04:45:40 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 5.498577
+INFO 2020-11-24 04:46:58 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 5.483171
+INFO 2020-11-24 04:48:16 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 5.415438
+INFO 2020-11-24 04:48:22 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-24 04:48:24 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 5.440089
+INFO 2020-11-24 04:49:42 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 5.026180
+INFO 2020-11-24 04:51:01 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 5.053122
+INFO 2020-11-24 04:52:19 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 5.106497
+INFO 2020-11-24 04:53:37 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 5.136229
+INFO 2020-11-24 04:54:56 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 5.132768
+INFO 2020-11-24 04:56:14 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 5.164696
+INFO 2020-11-24 04:57:32 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 5.157776
+INFO 2020-11-24 04:58:50 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 5.135490
+INFO 2020-11-24 05:00:08 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 5.116055
+INFO 2020-11-24 05:01:26 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 5.122944
+INFO 2020-11-24 05:02:44 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 5.096611
+INFO 2020-11-24 05:04:02 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 5.097762
+INFO 2020-11-24 05:05:20 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 5.086466
+INFO 2020-11-24 05:06:38 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 5.045042
+INFO 2020-11-24 05:07:56 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-24 05:07:57 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 5.059261
+INFO 2020-11-24 05:09:15 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 5.032098
+INFO 2020-11-24 05:10:33 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 5.027476
+INFO 2020-11-24 05:11:50 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 4.993284
+INFO 2020-11-24 05:13:08 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 4.983350
+INFO 2020-11-24 05:14:26 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 4.958266
+INFO 2020-11-24 05:15:44 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 4.936499
+INFO 2020-11-24 05:17:02 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 4.933462
+INFO 2020-11-24 05:18:20 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 4.937922
+INFO 2020-11-24 05:19:38 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 4.897083
+INFO 2020-11-24 05:20:56 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 4.866741
+INFO 2020-11-24 05:22:14 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 4.846557
+INFO 2020-11-24 05:23:32 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 4.844702
+INFO 2020-11-24 05:24:50 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 4.827470
+INFO 2020-11-24 05:26:08 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 4.834561
+INFO 2020-11-24 05:27:26 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-24 05:27:26 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 4.803743
+INFO 2020-11-24 05:28:44 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 4.787435
+INFO 2020-11-24 05:30:02 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 4.771263
+INFO 2020-11-24 05:30:08 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-24 05:30:09 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 4.665100
+INFO 2020-11-24 05:31:28 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 4.450923
+INFO 2020-11-24 05:32:46 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 4.442122
+INFO 2020-11-24 05:34:05 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 4.482682
+INFO 2020-11-24 05:35:23 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 4.523028
+INFO 2020-11-24 05:36:41 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 4.545702
+INFO 2020-11-24 05:38:00 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 4.545225
+INFO 2020-11-24 05:39:18 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 4.600934
+INFO 2020-11-24 05:40:36 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 4.616325
+INFO 2020-11-24 05:41:54 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 4.617960
+INFO 2020-11-24 05:43:12 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 4.597656
+INFO 2020-11-24 05:44:29 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 4.606315
+INFO 2020-11-24 05:45:47 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 4.568970
+INFO 2020-11-24 05:47:05 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 4.596194
+INFO 2020-11-24 05:48:23 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 4.570230
+INFO 2020-11-24 05:49:41 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-24 05:49:41 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 4.574229
+INFO 2020-11-24 05:50:58 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 4.572266
+INFO 2020-11-24 05:52:15 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 4.580258
+INFO 2020-11-24 05:53:33 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 4.533460
+INFO 2020-11-24 05:54:50 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 4.536240
+INFO 2020-11-24 05:56:07 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 4.557130
+INFO 2020-11-24 05:57:24 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 4.523213
+INFO 2020-11-24 05:58:41 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 4.503493
+INFO 2020-11-24 05:59:59 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 4.518086
+INFO 2020-11-24 06:01:16 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 4.502888
+INFO 2020-11-24 06:02:33 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 4.505856
+INFO 2020-11-24 06:03:50 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 4.486231
+INFO 2020-11-24 06:05:07 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 4.482046
+INFO 2020-11-24 06:06:25 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 4.460350
+INFO 2020-11-24 06:07:42 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 4.460023
+INFO 2020-11-24 06:08:59 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-24 06:08:59 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 4.455923
+INFO 2020-11-24 06:10:17 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 4.439720
+INFO 2020-11-24 06:11:35 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 4.420194
+INFO 2020-11-24 06:11:41 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-24 06:11:43 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 4.540029
+INFO 2020-11-24 06:13:01 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 4.096731
+INFO 2020-11-24 06:14:20 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 4.102299
+INFO 2020-11-24 06:15:38 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 4.138867
+INFO 2020-11-24 06:16:56 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 4.192034
+INFO 2020-11-24 06:18:15 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 4.244839
+INFO 2020-11-24 06:19:33 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 4.268064
+INFO 2020-11-24 06:20:51 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 4.274738
+INFO 2020-11-24 06:22:09 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 4.272569
+INFO 2020-11-24 06:23:27 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 4.322196
+INFO 2020-11-24 06:24:45 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 4.292552
+INFO 2020-11-24 06:26:03 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 4.308279
+INFO 2020-11-24 06:27:21 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 4.308598
+INFO 2020-11-24 06:28:39 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 4.303280
+INFO 2020-11-24 06:29:57 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 4.319410
+INFO 2020-11-24 06:31:15 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-24 06:31:15 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 4.301726
+INFO 2020-11-24 06:32:33 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 4.317716
+INFO 2020-11-24 06:33:50 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 4.312454
+INFO 2020-11-24 06:35:08 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 4.313283
+INFO 2020-11-24 06:36:26 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 4.314122
+INFO 2020-11-24 06:37:44 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 4.286486
+INFO 2020-11-24 06:39:02 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 4.249253
+INFO 2020-11-24 06:40:20 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 4.289481
+INFO 2020-11-24 06:41:38 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 4.246100
+INFO 2020-11-24 06:42:56 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 4.264654
+INFO 2020-11-24 06:44:14 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 4.257778
+INFO 2020-11-24 06:45:32 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 4.293045
+INFO 2020-11-24 06:46:50 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 4.253713
+INFO 2020-11-24 06:48:08 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 4.249499
+INFO 2020-11-24 06:49:26 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 4.244498
+INFO 2020-11-24 06:50:44 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-24 06:50:44 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 4.244528
+INFO 2020-11-24 06:52:02 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 4.216490
+INFO 2020-11-24 06:53:20 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 4.204208
+INFO 2020-11-24 06:53:26 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-24 06:53:28 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 4.202881
+INFO 2020-11-24 06:54:46 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 3.885112
+INFO 2020-11-24 06:56:05 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 3.881414
+INFO 2020-11-24 06:57:23 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 3.922434
+INFO 2020-11-24 06:58:41 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 3.989189
+INFO 2020-11-24 07:00:00 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 4.039934
+INFO 2020-11-24 07:01:18 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 4.076166
+INFO 2020-11-24 07:02:36 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 4.098873
+INFO 2020-11-24 07:03:54 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 4.107318
+INFO 2020-11-24 07:05:12 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 4.110902
+INFO 2020-11-24 07:06:30 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 4.117602
+INFO 2020-11-24 07:07:48 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 4.134322
+INFO 2020-11-24 07:09:06 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 4.126344
+INFO 2020-11-24 07:10:24 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 4.119838
+INFO 2020-11-24 07:11:42 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 4.137152
+INFO 2020-11-24 07:12:59 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-24 07:13:00 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 4.125776
+INFO 2020-11-24 07:14:17 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 4.137276
+INFO 2020-11-24 07:15:35 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 4.081156
+INFO 2020-11-24 07:16:52 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 4.110230
+INFO 2020-11-24 07:18:09 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 4.123747
+INFO 2020-11-24 07:19:27 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 4.112608
+INFO 2020-11-24 07:20:44 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 4.106259
+INFO 2020-11-24 07:22:01 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 4.144594
+INFO 2020-11-24 07:23:19 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 4.135456
+INFO 2020-11-24 07:24:36 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 4.105270
+INFO 2020-11-24 07:25:54 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 4.108409
+INFO 2020-11-24 07:27:11 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 4.117318
+INFO 2020-11-24 07:28:29 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 4.089157
+INFO 2020-11-24 07:29:46 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 4.105667
+INFO 2020-11-24 07:31:03 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 4.105449
+INFO 2020-11-24 07:32:21 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-24 07:32:21 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 4.091707
+INFO 2020-11-24 07:33:39 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 4.051780
+INFO 2020-11-24 07:34:57 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 4.068963
+INFO 2020-11-24 07:35:03 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-24 07:35:05 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 3.997203
+INFO 2020-11-24 07:36:23 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 3.771892
+INFO 2020-11-24 07:37:40 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 3.740402
+INFO 2020-11-24 07:38:58 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 3.820428
+INFO 2020-11-24 07:40:17 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 3.868483
+INFO 2020-11-24 07:41:35 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 3.896072
+INFO 2020-11-24 07:42:53 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 3.925021
+INFO 2020-11-24 07:44:11 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 3.932392
+INFO 2020-11-24 07:45:30 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 3.972897
+INFO 2020-11-24 07:46:48 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 3.956087
+INFO 2020-11-24 07:48:06 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 3.988663
+INFO 2020-11-24 07:49:24 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 3.995037
+INFO 2020-11-24 07:50:42 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 3.995521
+INFO 2020-11-24 07:52:00 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 3.979017
+INFO 2020-11-24 07:53:18 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 3.993128
+INFO 2020-11-24 07:54:35 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-24 07:54:36 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 3.991411
+INFO 2020-11-24 07:55:54 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 4.018914
+INFO 2020-11-24 07:57:12 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 3.984513
+INFO 2020-11-24 07:58:30 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 3.998572
+INFO 2020-11-24 07:59:48 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 3.991734
+INFO 2020-11-24 08:01:06 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 3.996338
+INFO 2020-11-24 08:02:24 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 3.974797
+INFO 2020-11-24 08:03:43 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 3.952605
+INFO 2020-11-24 08:05:01 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 4.014193
+INFO 2020-11-24 08:06:19 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 3.977969
+INFO 2020-11-24 08:07:37 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 3.981304
+INFO 2020-11-24 08:08:55 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 3.964265
+INFO 2020-11-24 08:10:13 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 3.983255
+INFO 2020-11-24 08:11:31 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 3.969215
+INFO 2020-11-24 08:12:49 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 3.956603
+INFO 2020-11-24 08:14:07 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-24 08:14:07 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 3.966941
+INFO 2020-11-24 08:15:25 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 3.957199
+INFO 2020-11-24 08:16:43 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 3.967428
+INFO 2020-11-24 08:16:50 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-24 08:16:51 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 3.927234
+INFO 2020-11-24 08:18:09 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 3.668670
+INFO 2020-11-24 08:19:27 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 3.638747
+INFO 2020-11-24 08:20:44 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 3.713964
+INFO 2020-11-24 08:22:02 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 3.750934
+INFO 2020-11-24 08:23:19 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 3.792767
+INFO 2020-11-24 08:24:37 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 3.818018
+INFO 2020-11-24 08:25:54 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 3.831887
+INFO 2020-11-24 08:27:12 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 3.858248
+INFO 2020-11-24 08:28:29 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 3.857486
+INFO 2020-11-24 08:29:47 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 3.878654
+INFO 2020-11-24 08:31:04 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 3.900151
+INFO 2020-11-24 08:32:21 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 3.900703
+INFO 2020-11-24 08:33:39 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 3.899652
+INFO 2020-11-24 08:34:56 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 3.917318
+INFO 2020-11-24 08:36:13 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-24 08:36:13 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 3.864257
+INFO 2020-11-24 08:37:31 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 3.894497
+INFO 2020-11-24 08:38:50 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 3.915060
+INFO 2020-11-24 08:40:08 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 3.895507
+INFO 2020-11-24 08:41:26 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 3.909785
+INFO 2020-11-24 08:42:44 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 3.895703
+INFO 2020-11-24 08:44:02 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 3.882205
+INFO 2020-11-24 08:45:20 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 3.885509
+INFO 2020-11-24 08:46:38 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 3.858412
+INFO 2020-11-24 08:47:56 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 3.900832
+INFO 2020-11-24 08:49:14 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 3.874724
+INFO 2020-11-24 08:50:32 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 3.892915
+INFO 2020-11-24 08:51:50 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 3.873445
+INFO 2020-11-24 08:53:08 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 3.856133
+INFO 2020-11-24 08:54:26 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 3.888871
+INFO 2020-11-24 08:55:44 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-24 08:55:45 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 3.870069
+INFO 2020-11-24 08:57:03 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 3.890066
+INFO 2020-11-24 08:58:21 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 3.874906
+INFO 2020-11-24 08:58:27 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-24 08:58:28 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 3.769626
+INFO 2020-11-24 08:59:47 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 3.531501
+INFO 2020-11-24 09:01:05 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 3.536682
+INFO 2020-11-24 09:02:23 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 3.620693
+INFO 2020-11-24 09:03:42 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 3.669908
+INFO 2020-11-24 09:05:00 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 3.698737
+INFO 2020-11-24 09:06:18 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 3.743023
+INFO 2020-11-24 09:07:36 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 3.753198
+INFO 2020-11-24 09:08:54 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 3.772860
+INFO 2020-11-24 09:10:12 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 3.785721
+INFO 2020-11-24 09:11:31 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 3.795793
+INFO 2020-11-24 09:12:49 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 3.795237
+INFO 2020-11-24 09:14:07 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 3.820151
+INFO 2020-11-24 09:15:25 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 3.814036
+INFO 2020-11-24 09:16:43 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 3.832890
+INFO 2020-11-24 09:18:00 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-24 09:18:01 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 3.814459
+INFO 2020-11-24 09:19:18 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 3.805081
+INFO 2020-11-24 09:20:36 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 3.814949
+INFO 2020-11-24 09:21:53 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 3.798262
+INFO 2020-11-24 09:23:11 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 3.838674
+INFO 2020-11-24 09:24:28 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 3.815417
+INFO 2020-11-24 09:25:46 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 3.835098
+INFO 2020-11-24 09:27:03 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 3.826335
+INFO 2020-11-24 09:28:21 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 3.822815
+INFO 2020-11-24 09:29:38 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 3.854760
+INFO 2020-11-24 09:30:56 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 3.818180
+INFO 2020-11-24 09:32:13 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 3.793270
+INFO 2020-11-24 09:33:31 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 3.807816
+INFO 2020-11-24 09:34:48 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 3.812581
+INFO 2020-11-24 09:36:06 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 3.826077
+INFO 2020-11-24 09:37:23 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-24 09:37:23 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 3.800772
+INFO 2020-11-24 09:38:41 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 3.802567
+INFO 2020-11-24 09:39:59 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 3.809041
+INFO 2020-11-24 09:40:05 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-24 09:40:07 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 3.827116
+INFO 2020-11-24 09:41:25 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 3.510100
+INFO 2020-11-24 09:42:44 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 3.499867
+INFO 2020-11-24 09:44:02 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 3.532867
+INFO 2020-11-24 09:45:21 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 3.577693
+INFO 2020-11-24 09:46:39 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 3.632344
+INFO 2020-11-24 09:47:57 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 3.670031
+INFO 2020-11-24 09:49:15 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 3.696803
+INFO 2020-11-24 09:50:33 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 3.731199
+INFO 2020-11-24 09:51:52 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 3.725462
+INFO 2020-11-24 09:53:10 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 3.726511
+INFO 2020-11-24 09:54:28 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 3.738408
+INFO 2020-11-24 09:55:46 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 3.751172
+INFO 2020-11-24 09:57:04 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 3.729649
+INFO 2020-11-24 09:58:22 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 3.750047
+INFO 2020-11-24 09:59:40 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-24 09:59:40 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 3.786703
+INFO 2020-11-24 10:00:58 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 3.760409
+INFO 2020-11-24 10:02:16 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 3.747574
+INFO 2020-11-24 10:03:35 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 3.793872
+INFO 2020-11-24 10:04:53 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 3.744194
+INFO 2020-11-24 10:06:11 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 3.769263
+INFO 2020-11-24 10:07:29 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 3.735591
+INFO 2020-11-24 10:08:47 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 3.736867
+INFO 2020-11-24 10:10:05 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 3.743708
+INFO 2020-11-24 10:11:23 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 3.740603
+INFO 2020-11-24 10:12:41 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 3.759759
+INFO 2020-11-24 10:13:59 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 3.751569
+INFO 2020-11-24 10:15:17 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 3.753405
+INFO 2020-11-24 10:16:35 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 3.752440
+INFO 2020-11-24 10:17:53 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 3.742277
+INFO 2020-11-24 10:19:11 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-24 10:19:12 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 3.749094
+INFO 2020-11-24 10:20:29 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 3.727976
+INFO 2020-11-24 10:21:46 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 3.718605
+INFO 2020-11-24 10:21:52 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-24 10:21:54 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 3.746880
+INFO 2020-11-24 10:23:12 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 2.974117
+INFO 2020-11-24 10:24:30 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 2.808296
+INFO 2020-11-24 10:25:49 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 2.751615
+INFO 2020-11-24 10:27:07 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 2.711641
+INFO 2020-11-24 10:28:25 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 2.671832
+INFO 2020-11-24 10:29:44 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 2.651242
+INFO 2020-11-24 10:31:02 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 2.609904
+INFO 2020-11-24 10:32:20 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 2.597323
+INFO 2020-11-24 10:33:38 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 2.587481
+INFO 2020-11-24 10:34:56 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 2.557563
+INFO 2020-11-24 10:36:14 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 2.537237
+INFO 2020-11-24 10:37:32 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 2.522175
+INFO 2020-11-24 10:38:50 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 2.501540
+INFO 2020-11-24 10:40:08 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 2.501676
+INFO 2020-11-24 10:41:26 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-24 10:41:27 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 2.492330
+INFO 2020-11-24 10:42:45 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 2.461658
+INFO 2020-11-24 10:44:03 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 2.460148
+INFO 2020-11-24 10:45:21 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 2.448058
+INFO 2020-11-24 10:46:39 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 2.440698
+INFO 2020-11-24 10:47:57 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 2.428191
+INFO 2020-11-24 10:49:16 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 2.418324
+INFO 2020-11-24 10:50:34 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 2.409370
+INFO 2020-11-24 10:51:52 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 2.391027
+INFO 2020-11-24 10:53:10 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 2.402113
+INFO 2020-11-24 10:54:28 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 2.373241
+INFO 2020-11-24 10:55:46 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 2.381784
+INFO 2020-11-24 10:57:04 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 2.365681
+INFO 2020-11-24 10:58:23 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 2.354604
+INFO 2020-11-24 10:59:41 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 2.356477
+INFO 2020-11-24 11:00:59 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-24 11:00:59 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 2.352549
+INFO 2020-11-24 11:02:17 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 2.321264
+INFO 2020-11-24 11:03:35 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 2.335863
+INFO 2020-11-24 11:03:41 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-24 11:03:43 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 2.340185
+INFO 2020-11-24 11:05:01 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 2.118885
+INFO 2020-11-24 11:06:18 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 2.106990
+INFO 2020-11-24 11:07:36 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 2.124713
+INFO 2020-11-24 11:08:53 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 2.138400
+INFO 2020-11-24 11:10:11 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 2.121224
+INFO 2020-11-24 11:11:29 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 2.125041
+INFO 2020-11-24 11:12:46 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 2.116722
+INFO 2020-11-24 11:14:04 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 2.140145
+INFO 2020-11-24 11:15:21 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 2.143412
+INFO 2020-11-24 11:16:38 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 2.139413
+INFO 2020-11-24 11:17:56 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 2.152643
+INFO 2020-11-24 11:19:13 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 2.139739
+INFO 2020-11-24 11:20:31 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 2.143647
+INFO 2020-11-24 11:21:48 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 2.147597
+INFO 2020-11-24 11:23:05 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-24 11:23:06 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 2.153717
+INFO 2020-11-24 11:24:24 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 2.158236
+INFO 2020-11-24 11:25:42 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 2.149249
+INFO 2020-11-24 11:27:00 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 2.139474
+INFO 2020-11-24 11:28:18 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 2.146334
+INFO 2020-11-24 11:29:37 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 2.157116
+INFO 2020-11-24 11:30:55 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 2.152893
+INFO 2020-11-24 11:32:13 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 2.145313
+INFO 2020-11-24 11:33:31 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 2.165957
+INFO 2020-11-24 11:34:50 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 2.167014
+INFO 2020-11-24 11:36:08 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 2.172860
+INFO 2020-11-24 11:37:26 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 2.154261
+INFO 2020-11-24 11:38:44 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 2.163148
+INFO 2020-11-24 11:40:02 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 2.170224
+INFO 2020-11-24 11:41:21 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 2.178672
+INFO 2020-11-24 11:42:39 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-24 11:42:39 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 2.157969
+INFO 2020-11-24 11:43:57 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 2.171329
+INFO 2020-11-24 11:45:15 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 2.172228
+INFO 2020-11-24 11:45:21 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-24 11:45:23 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 2.098908
+INFO 2020-11-24 11:46:41 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 1.976808
+INFO 2020-11-24 11:48:00 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 1.957961
+INFO 2020-11-24 11:49:18 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 1.971902
+INFO 2020-11-24 11:50:36 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 1.985339
+INFO 2020-11-24 11:51:55 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 1.991969
+INFO 2020-11-24 11:53:13 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 1.992443
+INFO 2020-11-24 11:54:31 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 2.005499
+INFO 2020-11-24 11:55:49 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 2.019665
+INFO 2020-11-24 11:57:08 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 2.033139
+INFO 2020-11-24 11:58:26 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 2.028996
+INFO 2020-11-24 11:59:44 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 2.019801
+INFO 2020-11-24 12:01:02 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 2.045612
+INFO 2020-11-24 12:02:20 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 2.037465
+INFO 2020-11-24 12:03:38 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 2.035865
+INFO 2020-11-24 12:04:56 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-24 12:04:56 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 2.058705
+INFO 2020-11-24 12:06:14 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 2.057656
+INFO 2020-11-24 12:07:31 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 2.056521
+INFO 2020-11-24 12:08:49 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 2.059211
+INFO 2020-11-24 12:10:06 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 2.073577
+INFO 2020-11-24 12:11:23 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 2.068285
+INFO 2020-11-24 12:12:41 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 2.076814
+INFO 2020-11-24 12:13:58 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 2.081390
+INFO 2020-11-24 12:15:15 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 2.090879
+INFO 2020-11-24 12:16:33 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 2.093876
+INFO 2020-11-24 12:17:50 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 2.100984
+INFO 2020-11-24 12:19:08 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 2.109811
+INFO 2020-11-24 12:20:25 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 2.099833
+INFO 2020-11-24 12:21:42 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 2.101767
+INFO 2020-11-24 12:23:00 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 2.112344
+INFO 2020-11-24 12:24:17 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-24 12:24:17 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 2.126899
+INFO 2020-11-24 12:25:35 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 2.109422
+INFO 2020-11-24 12:26:54 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 2.116510
+INFO 2020-11-24 12:27:00 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-24 12:27:01 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 2.109774
+INFO 2020-11-24 12:28:19 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 1.850623
+INFO 2020-11-24 12:29:38 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 1.845698
+INFO 2020-11-24 12:30:56 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 1.832399
+INFO 2020-11-24 12:32:14 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 1.851899
+INFO 2020-11-24 12:33:33 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 1.831740
+INFO 2020-11-24 12:34:51 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 1.832839
+INFO 2020-11-24 12:36:09 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 1.825755
+INFO 2020-11-24 12:37:27 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 1.838505
+INFO 2020-11-24 12:38:46 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 1.826254
+INFO 2020-11-24 12:40:04 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 1.832688
+INFO 2020-11-24 12:41:22 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 1.824897
+INFO 2020-11-24 12:42:40 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 1.825613
+INFO 2020-11-24 12:43:58 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 1.833551
+INFO 2020-11-24 12:45:16 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 1.837937
+INFO 2020-11-24 12:46:34 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-24 12:46:34 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 1.833608
+INFO 2020-11-24 12:47:52 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 1.825760
+INFO 2020-11-24 12:49:11 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 1.822354
+INFO 2020-11-24 12:50:29 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 1.822294
+INFO 2020-11-24 12:51:47 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 1.831900
+INFO 2020-11-24 12:53:05 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 1.837293
+INFO 2020-11-24 12:54:23 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 1.833485
+INFO 2020-11-24 12:55:41 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 1.830787
+INFO 2020-11-24 12:56:59 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 1.836366
+INFO 2020-11-24 12:58:17 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 1.833278
+INFO 2020-11-24 12:59:35 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 1.831308
+INFO 2020-11-24 13:00:53 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 1.845200
+INFO 2020-11-24 13:02:11 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 1.833023
+INFO 2020-11-24 13:03:30 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 1.829591
+INFO 2020-11-24 13:04:48 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 1.846178
+INFO 2020-11-24 13:06:06 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-24 13:06:06 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 1.835252
+INFO 2020-11-24 13:07:24 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 1.840193
+INFO 2020-11-24 13:08:42 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 1.833468
+INFO 2020-11-24 13:08:48 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-24 13:08:50 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 1.819492
+INFO 2020-11-24 13:10:08 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 1.810554
+INFO 2020-11-24 13:11:26 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 1.820436
+INFO 2020-11-24 13:12:45 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 1.805539
+INFO 2020-11-24 13:14:03 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 1.788846
+INFO 2020-11-24 13:15:21 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 1.813469
+INFO 2020-11-24 13:16:39 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 1.806865
+INFO 2020-11-24 13:17:58 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 1.814105
+INFO 2020-11-24 13:19:16 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 1.817921
+INFO 2020-11-24 13:20:34 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 1.805090
+INFO 2020-11-24 13:21:52 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 1.810515
+INFO 2020-11-24 13:23:10 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 1.805530
+INFO 2020-11-24 13:24:28 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 1.823238
+INFO 2020-11-24 13:25:46 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 1.802504
+INFO 2020-11-24 13:27:04 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 1.813050
+INFO 2020-11-24 13:28:22 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-24 13:28:22 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 1.812065
+INFO 2020-11-24 13:29:40 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 1.798023
+INFO 2020-11-24 13:30:57 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 1.808276
+INFO 2020-11-24 13:32:15 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 1.830430
+INFO 2020-11-24 13:33:32 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 1.829250
+INFO 2020-11-24 13:34:49 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 1.808092
+INFO 2020-11-24 13:36:07 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 1.807351
+INFO 2020-11-24 13:37:24 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 1.811478
+INFO 2020-11-24 13:38:42 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 1.824016
+INFO 2020-11-24 13:39:59 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 1.828517
+INFO 2020-11-24 13:41:16 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 1.835607
+INFO 2020-11-24 13:42:34 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 1.824501
+INFO 2020-11-24 13:43:51 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 1.810710
+INFO 2020-11-24 13:45:09 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 1.826132
+INFO 2020-11-24 13:46:26 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 1.813428
+INFO 2020-11-24 13:47:43 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-24 13:47:43 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 1.814075
+INFO 2020-11-24 13:49:01 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 1.818595
+INFO 2020-11-24 13:50:19 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 1.833896
+INFO 2020-11-24 13:50:26 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-24 13:50:27 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 1.816874
+INFO 2020-11-24 13:51:45 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 1.786866
+INFO 2020-11-24 13:53:04 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 1.774457
+INFO 2020-11-24 13:54:22 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 1.799863
+INFO 2020-11-24 13:55:40 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 1.782401
+INFO 2020-11-24 13:56:58 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 1.797739
+INFO 2020-11-24 13:58:17 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 1.803113
+INFO 2020-11-24 13:59:35 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 1.802417
+INFO 2020-11-24 14:00:53 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 1.790440
+INFO 2020-11-24 14:02:11 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 1.794112
+INFO 2020-11-24 14:03:29 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 1.792146
+INFO 2020-11-24 14:04:47 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 1.791506
+INFO 2020-11-24 14:06:05 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 1.797008
+INFO 2020-11-24 14:07:23 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 1.797804
+INFO 2020-11-24 14:08:41 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 1.792445
+INFO 2020-11-24 14:09:59 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-24 14:10:00 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 1.808717
+INFO 2020-11-24 14:11:18 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 1.800370
+INFO 2020-11-24 14:12:36 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 1.810517
+INFO 2020-11-24 14:13:54 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 1.811586
+INFO 2020-11-24 14:15:12 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 1.796170
+INFO 2020-11-24 14:16:30 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 1.792518
+INFO 2020-11-24 14:17:48 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 1.802332
+INFO 2020-11-24 14:19:06 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 1.815958
+INFO 2020-11-24 14:20:25 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 1.807881
+INFO 2020-11-24 14:21:43 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 1.821013
+INFO 2020-11-24 14:23:01 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 1.809572
+INFO 2020-11-24 14:24:19 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 1.821744
+INFO 2020-11-24 14:25:37 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 1.808917
+INFO 2020-11-24 14:26:55 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 1.822763
+INFO 2020-11-24 14:28:13 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 1.823356
+INFO 2020-11-24 14:29:31 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-24 14:29:31 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 1.821922
+INFO 2020-11-24 14:30:50 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 1.834525
+INFO 2020-11-24 14:32:08 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 1.833749
+INFO 2020-11-24 14:32:14 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-24 14:32:15 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 1.798043
+INFO 2020-11-24 14:33:34 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 1.788470
+INFO 2020-11-24 14:34:52 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 1.770692
+INFO 2020-11-24 14:36:10 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 1.782898
+INFO 2020-11-24 14:37:29 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 1.781051
+INFO 2020-11-24 14:38:47 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 1.782889
+INFO 2020-11-24 14:40:05 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 1.765409
+INFO 2020-11-24 14:41:23 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 1.780779
+INFO 2020-11-24 14:42:41 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 1.783804
+INFO 2020-11-24 14:43:59 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 1.779461
+INFO 2020-11-24 14:45:17 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 1.766039
+INFO 2020-11-24 14:46:35 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 1.769282
+INFO 2020-11-24 14:47:54 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 1.762334
+INFO 2020-11-24 14:49:12 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 1.789894
+INFO 2020-11-24 14:50:30 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 1.781496
+INFO 2020-11-24 14:51:47 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-24 14:51:48 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 1.775365
+INFO 2020-11-24 14:53:06 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 1.782624
+INFO 2020-11-24 14:54:24 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 1.776455
+INFO 2020-11-24 14:55:42 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 1.788013
+INFO 2020-11-24 14:57:00 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 1.791255
+INFO 2020-11-24 14:58:18 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 1.774332
+INFO 2020-11-24 14:59:36 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 1.788021
+INFO 2020-11-24 15:00:54 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 1.778018
+INFO 2020-11-24 15:02:13 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 1.783425
+INFO 2020-11-24 15:03:31 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 1.774238
+INFO 2020-11-24 15:04:49 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 1.774257
+INFO 2020-11-24 15:06:07 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 1.780392
+INFO 2020-11-24 15:07:25 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 1.778083
+INFO 2020-11-24 15:08:43 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 1.778772
+INFO 2020-11-24 15:10:01 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 1.787950
+INFO 2020-11-24 15:11:19 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-24 15:11:19 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 1.763879
+INFO 2020-11-24 15:12:38 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 1.766053
+INFO 2020-11-24 15:13:56 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 1.774999
+INFO 2020-11-24 15:14:02 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-24 15:14:03 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 1.769013
+INFO 2020-11-24 15:15:22 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 1.777793
+INFO 2020-11-24 15:16:40 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 1.773711
+INFO 2020-11-24 15:17:58 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 1.765407
+INFO 2020-11-24 15:19:17 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 1.785018
+INFO 2020-11-24 15:20:35 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 1.774917
+INFO 2020-11-24 15:21:53 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 1.761761
+INFO 2020-11-24 15:23:11 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 1.768229
+INFO 2020-11-24 15:24:30 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 1.780249
+INFO 2020-11-24 15:25:48 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 1.783446
+INFO 2020-11-24 15:27:06 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 1.779726
+INFO 2020-11-24 15:28:24 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 1.778573
+INFO 2020-11-24 15:29:42 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 1.773806
+INFO 2020-11-24 15:31:00 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 1.780847
+INFO 2020-11-24 15:32:18 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 1.780268
+INFO 2020-11-24 15:33:36 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-24 15:33:36 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 1.779825
+INFO 2020-11-24 15:34:55 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 1.770760
+INFO 2020-11-24 15:36:13 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 1.769815
+INFO 2020-11-24 15:37:31 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 1.767904
+INFO 2020-11-24 15:38:49 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 1.779343
+INFO 2020-11-24 15:40:07 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 1.770879
+INFO 2020-11-24 15:41:25 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 1.774319
+INFO 2020-11-24 15:42:44 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 1.780236
+INFO 2020-11-24 15:44:02 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 1.775440
+INFO 2020-11-24 15:45:20 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 1.772610
+INFO 2020-11-24 15:46:38 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 1.778330
+INFO 2020-11-24 15:47:56 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 1.783649
+INFO 2020-11-24 15:49:15 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 1.777045
+INFO 2020-11-24 15:50:33 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 1.777206
+INFO 2020-11-24 15:51:51 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 1.780458
+INFO 2020-11-24 15:53:09 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-24 15:53:09 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 1.777346
+INFO 2020-11-24 15:54:27 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 1.776126
+INFO 2020-11-24 15:55:44 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 1.774096
+INFO 2020-11-24 15:55:50 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-24 15:55:51 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/.gitkeep b/bob/bio/facexzoo/models/heads/CircleLoss/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ef08cf330a231f4b2b9ab49099768448f7f29315
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_12_batch_5999.pt | 0.9573333333333334 | 0.002168802366215905 |
+| Epoch_13.pt | 0.9551666666666666 | 0.002114762923408253 |
+| Epoch_13_batch_2999.pt | 0.9540000000000001 | 0.0017427096823731277 |
+| Epoch_14.pt | 0.9535000000000002 | 0.002100117574603977 |
+| Epoch_12_batch_2999.pt | 0.9533333333333334 | 0.0020336672464136792 |
+| Epoch_11_batch_2999.pt | 0.9528333333333334 | 0.0019571016615342854 |
+| Epoch_13_batch_5999.pt | 0.9528333333333334 | 0.001909204475271064 |
+| Epoch_11_batch_5999.pt | 0.9528333333333332 | 0.0017924739783224065 |
+| Epoch_16_batch_5999.pt | 0.9526666666666668 | 0.0018121673811444562 |
+| Epoch_17.pt | 0.9525 | 0.0021118419787498 |
+| Epoch_17_batch_2999.pt | 0.9523333333333334 | 0.0017069212773041418 |
+| Epoch_17_batch_5999.pt | 0.9523333333333334 | 0.0017950549357114984 |
+| Epoch_14_batch_2999.pt | 0.9516666666666668 | 0.002108185106778921 |
+| Epoch_15.pt | 0.9516666666666668 | 0.0018425693279752202 |
+| Epoch_15_batch_5999.pt | 0.9516666666666665 | 0.002557969874049184 |
+| Epoch_11.pt | 0.9511666666666667 | 0.002472690342696435 |
+| Epoch_14_batch_5999.pt | 0.9511666666666667 | 0.0018765939726727201 |
+| Epoch_16.pt | 0.951 | 0.0017950549357115021 |
+| Epoch_16_batch_2999.pt | 0.9506666666666668 | 0.0019116278371205861 |
+| Epoch_15_batch_2999.pt | 0.9506666666666665 | 0.002372684056006955 |
+| Epoch_12.pt | 0.9505000000000001 | 0.0025098571106836683 |
+| Epoch_10_batch_2999.pt | 0.9496666666666667 | 0.002730712383876552 |
+| Epoch_10.pt | 0.9495000000000001 | 0.0024475989746567456 |
+| Epoch_10_batch_5999.pt | 0.9446666666666665 | 0.0021052550357218178 |
+| Epoch_9_batch_2999.pt | 0.9386666666666666 | 0.0031700761372638027 |
+| Epoch_8_batch_5999.pt | 0.9365 | 0.004405734198423739 |
+| Epoch_9_batch_5999.pt | 0.9361666666666666 | 0.0036855573979159965 |
+| Epoch_8_batch_2999.pt | 0.931 | 0.004225145187073672 |
+| Epoch_8.pt | 0.9306666666666666 | 0.0044527699798919745 |
+| Epoch_9.pt | 0.9303333333333335 | 0.003708515392950813 |
+| Epoch_6_batch_5999.pt | 0.9281666666666666 | 0.0044614258917586285 |
+| Epoch_5_batch_5999.pt | 0.9275 | 0.004702573500846545 |
+| Epoch_7_batch_2999.pt | 0.9266666666666665 | 0.003975231959999628 |
+| Epoch_5.pt | 0.9265000000000001 | 0.004543683717909407 |
+| Epoch_7_batch_5999.pt | 0.9261666666666667 | 0.004615125427986695 |
+| Epoch_6_batch_2999.pt | 0.9258333333333333 | 0.005478634145365555 |
+| Epoch_5_batch_2999.pt | 0.9254999999999999 | 0.00489425210871934 |
+| Epoch_7.pt | 0.9236666666666666 | 0.0033407325285273112 |
+| Epoch_4_batch_5999.pt | 0.9228333333333334 | 0.00418735604138028 |
+| Epoch_6.pt | 0.9206666666666667 | 0.004562325592810633 |
+| Epoch_3_batch_5999.pt | 0.9158333333333333 | 0.0056778212653259596 |
+| Epoch_3.pt | 0.9146666666666666 | 0.006986759965449996 |
+| Epoch_4_batch_2999.pt | 0.9126666666666667 | 0.005506449641495055 |
+| Epoch_4.pt | 0.9095000000000001 | 0.005931970297037765 |
+| Epoch_3_batch_2999.pt | 0.9081666666666667 | 0.005876559420041957 |
+| Epoch_2_batch_5999.pt | 0.8971666666666668 | 0.006611111111111107 |
+| Epoch_2.pt | 0.8943333333333333 | 0.0073753007890115995 |
+| Epoch_2_batch_2999.pt | 0.8844999999999998 | 0.006265967258054606 |
+| Epoch_1.pt | 0.8486666666666667 | 0.007030796453136162 |
+| Epoch_1_batch_5999.pt | 0.8463333333333333 | 0.007381993440832486 |
+| Epoch_1_batch_2999.pt | 0.8018333333333333 | 0.008679655508894721 |
+| Epoch_0_batch_5999.pt | 0.6713333333333333 | 0.007709221670534074 |
+| Epoch_0.pt | 0.6535 | 0.008056321802637614 |
+| Epoch_0_batch_2999.pt | 0.512 | 0.004518822090591723 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..ec9807419e16bdc08364f7f0afc26d5ee703853f
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.9400000000000001 | 0.0037843080813169698 |
+| Epoch_13_batch_2999.pt | 0.9398333333333333 | 0.003722222222222219 |
+| Epoch_14.pt | 0.9396666666666667 | 0.003903148460055621 |
+| Epoch_15_batch_2999.pt | 0.9396666666666667 | 0.004096068575814829 |
+| Epoch_13.pt | 0.9395 | 0.0038972133127323496 |
+| Epoch_16_batch_2999.pt | 0.9395 | 0.0036519063176766197 |
+| Epoch_15.pt | 0.9388333333333334 | 0.0035750334540435953 |
+| Epoch_17_batch_2999.pt | 0.9388333333333334 | 0.0035836563161933594 |
+| Epoch_16.pt | 0.9388333333333332 | 0.003634963956212353 |
+| Epoch_11_batch_5999.pt | 0.9381666666666668 | 0.0033559418467864334 |
+| Epoch_14_batch_5999.pt | 0.9381666666666666 | 0.003428727583157662 |
+| Epoch_15_batch_5999.pt | 0.9381666666666666 | 0.0037222222222222196 |
+| Epoch_11.pt | 0.9378333333333334 | 0.0039051248379533246 |
+| Epoch_12.pt | 0.9378333333333332 | 0.003944444444444442 |
+| Epoch_13_batch_5999.pt | 0.9376666666666665 | 0.0038425814368423534 |
+| Epoch_11_batch_2999.pt | 0.9375 | 0.003524604872347091 |
+| Epoch_14_batch_2999.pt | 0.9375 | 0.0037618126704955968 |
+| Epoch_17.pt | 0.9375 | 0.0033448873829978565 |
+| Epoch_16_batch_5999.pt | 0.9365 | 0.0035263557939018498 |
+| Epoch_10.pt | 0.9363333333333334 | 0.003749897117930263 |
+| Epoch_12_batch_5999.pt | 0.9355 | 0.003905124837953327 |
+| Epoch_10_batch_5999.pt | 0.9354999999999999 | 0.003743719019740318 |
+| Epoch_10_batch_2999.pt | 0.9344999999999999 | 0.003800990774021628 |
+| Epoch_12_batch_2999.pt | 0.9331666666666667 | 0.003482318654269959 |
+| Epoch_9_batch_5999.pt | 0.9279999999999999 | 0.004961580791503583 |
+| Epoch_9_batch_2999.pt | 0.9248333333333335 | 0.003884521356980743 |
+| Epoch_7_batch_2999.pt | 0.9238333333333333 | 0.0039833759489429745 |
+| Epoch_7_batch_5999.pt | 0.922 | 0.003749897117930269 |
+| Epoch_8_batch_5999.pt | 0.9211666666666666 | 0.004052510272185977 |
+| Epoch_5_batch_2999.pt | 0.9201666666666668 | 0.003037319318408148 |
+| Epoch_8_batch_2999.pt | 0.9201666666666666 | 0.004631147899192763 |
+| Epoch_6_batch_2999.pt | 0.9199999999999999 | 0.003634539385288034 |
+| Epoch_5_batch_5999.pt | 0.9179999999999999 | 0.004477653706579976 |
+| Epoch_6_batch_5999.pt | 0.9178333333333333 | 0.004216736202032426 |
+| Epoch_8.pt | 0.9168333333333333 | 0.004896773944315622 |
+| Epoch_9.pt | 0.9166666666666667 | 0.004288946459026401 |
+| Epoch_6.pt | 0.9165000000000001 | 0.003561193446786085 |
+| Epoch_7.pt | 0.9158333333333335 | 0.004144384867012949 |
+| Epoch_4_batch_5999.pt | 0.914 | 0.004046031434674025 |
+| Epoch_4_batch_2999.pt | 0.9093333333333333 | 0.0035935470286213847 |
+| Epoch_5.pt | 0.9091666666666667 | 0.0038349433012049563 |
+| Epoch_3_batch_5999.pt | 0.9086666666666667 | 0.004666666666666672 |
+| Epoch_4.pt | 0.9076666666666666 | 0.0031348302177035335 |
+| Epoch_3.pt | 0.9038333333333333 | 0.00520001187083166 |
+| Epoch_3_batch_2999.pt | 0.9014999999999999 | 0.0051606894936028404 |
+| Epoch_2_batch_5999.pt | 0.893 | 0.00502954235455834 |
+| Epoch_2.pt | 0.8891666666666668 | 0.004054033199812892 |
+| Epoch_2_batch_2999.pt | 0.8891666666666665 | 0.004857538369621563 |
+| Epoch_1_batch_5999.pt | 0.8545 | 0.0035316033500695106 |
+| Epoch_1.pt | 0.8438333333333332 | 0.00345205252953466 |
+| Epoch_1_batch_2999.pt | 0.7995 | 0.005858675112521096 |
+| Epoch_0_batch_5999.pt | 0.6233333333333334 | 0.007328281087929401 |
+| Epoch_0.pt | 0.6055 | 0.007088723439378912 |
+| Epoch_0_batch_2999.pt | 0.5035 | 0.006218500871304895 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..61139145fccc114d9b4797c97a2d5e2088569b76
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8341666666666667 | 0.006359837330010694 |
+| Epoch_16_batch_2999.pt | 0.834 | 0.006470254827351405 |
+| Epoch_15_batch_5999.pt | 0.8328333333333333 | 0.007412243791504118 |
+| Epoch_16_batch_5999.pt | 0.8326666666666667 | 0.006374137559672649 |
+| Epoch_14_batch_2999.pt | 0.8323333333333334 | 0.006904105059069322 |
+| Epoch_14.pt | 0.8323333333333333 | 0.0060868204110232586 |
+| Epoch_12_batch_2999.pt | 0.8320000000000001 | 0.005862624870343517 |
+| Epoch_13_batch_2999.pt | 0.8320000000000001 | 0.0060593769797493225 |
+| Epoch_12.pt | 0.8311666666666666 | 0.006181663446270991 |
+| Epoch_14_batch_5999.pt | 0.8308333333333333 | 0.006102266326618963 |
+| Epoch_15.pt | 0.8308333333333333 | 0.006704752322323599 |
+| Epoch_13_batch_5999.pt | 0.8306666666666667 | 0.006564964998543313 |
+| Epoch_11.pt | 0.8305 | 0.008247895354439079 |
+| Epoch_17_batch_5999.pt | 0.8305 | 0.006512333976936419 |
+| Epoch_13.pt | 0.8303333333333333 | 0.007039570693980961 |
+| Epoch_11_batch_5999.pt | 0.8296666666666667 | 0.0067713995489744805 |
+| Epoch_17_batch_2999.pt | 0.8293333333333333 | 0.006904105059069328 |
+| Epoch_12_batch_5999.pt | 0.8288333333333334 | 0.006568959930453315 |
+| Epoch_10_batch_5999.pt | 0.8286666666666666 | 0.00593587128408582 |
+| Epoch_10.pt | 0.8281666666666666 | 0.0059755158876415694 |
+| Epoch_15_batch_2999.pt | 0.8280000000000001 | 0.006906786785742247 |
+| Epoch_16.pt | 0.8263333333333334 | 0.006977919319158921 |
+| Epoch_10_batch_2999.pt | 0.8251666666666667 | 0.00608301622714859 |
+| Epoch_11_batch_2999.pt | 0.8245000000000001 | 0.0074703116103383245 |
+| Epoch_9_batch_2999.pt | 0.8011666666666667 | 0.007097426067957553 |
+| Epoch_7_batch_2999.pt | 0.8003333333333333 | 0.007148340047318588 |
+| Epoch_9_batch_5999.pt | 0.8003333333333332 | 0.006135305641012673 |
+| Epoch_6_batch_2999.pt | 0.7976666666666667 | 0.00562292485728 |
+| Epoch_8_batch_2999.pt | 0.7976666666666666 | 0.006086820411023254 |
+| Epoch_5_batch_2999.pt | 0.7935000000000001 | 0.006882837844241534 |
+| Epoch_8_batch_5999.pt | 0.7935000000000001 | 0.006552023530610908 |
+| Epoch_7_batch_5999.pt | 0.7935 | 0.005711424548251546 |
+| Epoch_9.pt | 0.793 | 0.00789123595758431 |
+| Epoch_5_batch_5999.pt | 0.7924999999999999 | 0.005032303058647612 |
+| Epoch_6_batch_5999.pt | 0.791 | 0.007141428428542845 |
+| Epoch_4_batch_5999.pt | 0.7849999999999999 | 0.008916623398995056 |
+| Epoch_6.pt | 0.7848333333333334 | 0.008286721492026852 |
+| Epoch_7.pt | 0.7838333333333333 | 0.006903434464627547 |
+| Epoch_5.pt | 0.7836666666666666 | 0.007757115411425356 |
+| Epoch_8.pt | 0.7828333333333333 | 0.007391394729642215 |
+| Epoch_4.pt | 0.7788333333333333 | 0.006101254677685179 |
+| Epoch_3_batch_5999.pt | 0.775 | 0.0065734219812217986 |
+| Epoch_4_batch_2999.pt | 0.7703333333333333 | 0.006425258797689314 |
+| Epoch_3.pt | 0.7696666666666666 | 0.005745637099487954 |
+| Epoch_3_batch_2999.pt | 0.7628333333333334 | 0.007710222491225363 |
+| Epoch_2_batch_5999.pt | 0.7498333333333334 | 0.008002507323132554 |
+| Epoch_2.pt | 0.7491666666666668 | 0.006621373920312353 |
+| Epoch_2_batch_2999.pt | 0.736 | 0.008520853142239035 |
+| Epoch_1_batch_5999.pt | 0.7126666666666667 | 0.008600172263563265 |
+| Epoch_1.pt | 0.7 | 0.008172522465970349 |
+| Epoch_1_batch_2999.pt | 0.6558333333333334 | 0.007943667405666719 |
+| Epoch_0.pt | 0.5785 | 0.008528999188679396 |
+| Epoch_0_batch_5999.pt | 0.5745 | 0.009490248406380875 |
+| Epoch_0_batch_2999.pt | 0.5053333333333333 | 0.002367475083629173 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..276eb6c3d90c56d1d4248d870244ab7c7ae529b3
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_11.pt | 0.9956666666666665 | 0.001059932446018833 |
+| Epoch_10_batch_2999.pt | 0.9950000000000001 | 0.0012422599874998821 |
+| Epoch_10_batch_5999.pt | 0.9948333333333335 | 0.0009444444444444486 |
+| Epoch_11_batch_5999.pt | 0.9948333333333332 | 0.0013252067157640635 |
+| Epoch_15_batch_2999.pt | 0.9946666666666666 | 0.0013333333333333344 |
+| Epoch_14.pt | 0.9945 | 0.0011399046960379555 |
+| Epoch_12_batch_2999.pt | 0.9944999999999998 | 0.0013158576980363357 |
+| Epoch_14_batch_5999.pt | 0.9943333333333333 | 0.0011166528467912097 |
+| Epoch_15.pt | 0.9943333333333333 | 0.0012957670877434004 |
+| Epoch_12_batch_5999.pt | 0.9941666666666666 | 0.0014540280364780504 |
+| Epoch_16_batch_5999.pt | 0.9941666666666666 | 0.001248455836346904 |
+| Epoch_16_batch_2999.pt | 0.9941666666666664 | 0.001248455836346904 |
+| Epoch_10.pt | 0.9940000000000001 | 0.001222222222222218 |
+| Epoch_13_batch_5999.pt | 0.994 | 0.0012472191289246452 |
+| Epoch_13.pt | 0.994 | 0.001196703290474334 |
+| Epoch_14_batch_2999.pt | 0.9938333333333335 | 0.001243501626977747 |
+| Epoch_17_batch_2999.pt | 0.9938333333333335 | 0.0013844373104863516 |
+| Epoch_11_batch_2999.pt | 0.9936666666666667 | 0.0014010578014353886 |
+| Epoch_13_batch_2999.pt | 0.9936666666666667 | 0.001309980680283513 |
+| Epoch_15_batch_5999.pt | 0.9936666666666667 | 0.001333333333333335 |
+| Epoch_16.pt | 0.9936666666666667 | 0.0012862041003100261 |
+| Epoch_17_batch_5999.pt | 0.9935 | 0.0015406027359846706 |
+| Epoch_17.pt | 0.9935 | 0.0013933262448871664 |
+| Epoch_12.pt | 0.993 | 0.0012619796324000619 |
+| Epoch_5_batch_5999.pt | 0.9928333333333332 | 0.0013844373104863487 |
+| Epoch_8_batch_2999.pt | 0.9923333333333334 | 0.0015947444549341471 |
+| Epoch_6_batch_5999.pt | 0.992 | 0.0015071844406945056 |
+| Epoch_5_batch_2999.pt | 0.9918333333333333 | 0.0017293758240303754 |
+| Epoch_7_batch_5999.pt | 0.9918333333333333 | 0.0014153043558729993 |
+| Epoch_9_batch_2999.pt | 0.9918333333333333 | 0.0017293758240303756 |
+| Epoch_6_batch_2999.pt | 0.9914999999999999 | 0.0016187558093703864 |
+| Epoch_8_batch_5999.pt | 0.9914999999999999 | 0.0016187558093703849 |
+| Epoch_9_batch_5999.pt | 0.9914999999999999 | 0.0015204369092671182 |
+| Epoch_4_batch_2999.pt | 0.9911666666666668 | 0.0010844011831079544 |
+| Epoch_9.pt | 0.9911666666666665 | 0.0019412672455173512 |
+| Epoch_8.pt | 0.9908333333333333 | 0.0012729376930432875 |
+| Epoch_4_batch_5999.pt | 0.9901666666666668 | 0.001899480110808745 |
+| Epoch_5.pt | 0.9901666666666668 | 0.001816420302696868 |
+| Epoch_7.pt | 0.9894999999999999 | 0.0017924739783224018 |
+| Epoch_6.pt | 0.9893333333333334 | 0.002140151142695358 |
+| Epoch_3_batch_5999.pt | 0.9893333333333333 | 0.001632993161855456 |
+| Epoch_7_batch_2999.pt | 0.9889999999999999 | 0.0015752718754175412 |
+| Epoch_4.pt | 0.9886666666666667 | 0.0017356110390903657 |
+| Epoch_3_batch_2999.pt | 0.9879999999999999 | 0.0014865653511399624 |
+| Epoch_3.pt | 0.9871666666666667 | 0.0022505143445032305 |
+| Epoch_2_batch_5999.pt | 0.9851666666666665 | 0.0019790570145063195 |
+| Epoch_2.pt | 0.9818333333333333 | 0.0016377114414426232 |
+| Epoch_2_batch_2999.pt | 0.9813333333333334 | 0.002045772515502434 |
+| Epoch_1_batch_5999.pt | 0.9736666666666665 | 0.0024570382652773243 |
+| Epoch_1.pt | 0.9733333333333333 | 0.0027329719724997394 |
+| Epoch_1_batch_2999.pt | 0.9536666666666667 | 0.003807075933113735 |
+| Epoch_0.pt | 0.898 | 0.004456926915584793 |
+| Epoch_0_batch_5999.pt | 0.89 | 0.004296136650929156 |
+| Epoch_0_batch_2999.pt | 0.6573333333333332 | 0.007054461857488875 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..923171cbea854fad6038afc61d6920c8c798c5f2
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8924999999999998 | 0.007045925186119637 |
+| Epoch_17_batch_5999.pt | 0.8908333333333335 | 0.00677345034894548 |
+| Epoch_17_batch_2999.pt | 0.89 | 0.006615544731824081 |
+| Epoch_16_batch_5999.pt | 0.8895000000000002 | 0.006450428256430573 |
+| Epoch_16_batch_2999.pt | 0.8895 | 0.006383089154985531 |
+| Epoch_15.pt | 0.8891666666666665 | 0.006579288490181673 |
+| Epoch_14_batch_2999.pt | 0.889 | 0.006550845765770526 |
+| Epoch_13.pt | 0.8886666666666667 | 0.006396372428721891 |
+| Epoch_15_batch_2999.pt | 0.8885 | 0.006523698488994466 |
+| Epoch_16.pt | 0.8868333333333334 | 0.007151145984308311 |
+| Epoch_13_batch_2999.pt | 0.8861666666666667 | 0.006662267066782581 |
+| Epoch_15_batch_5999.pt | 0.8859999999999999 | 0.005747785402834469 |
+| Epoch_14_batch_5999.pt | 0.8858333333333335 | 0.007050304257509516 |
+| Epoch_14.pt | 0.8855000000000001 | 0.006354011078848027 |
+| Epoch_13_batch_5999.pt | 0.8845000000000001 | 0.00744548085654548 |
+| Epoch_12_batch_2999.pt | 0.8838333333333332 | 0.0060809863559893275 |
+| Epoch_12_batch_5999.pt | 0.8838333333333332 | 0.0061113636311463946 |
+| Epoch_11_batch_5999.pt | 0.8836666666666666 | 0.006929094145630602 |
+| Epoch_12.pt | 0.8835000000000001 | 0.005860781981280208 |
+| Epoch_11_batch_2999.pt | 0.8831666666666667 | 0.006783467557139597 |
+| Epoch_10_batch_5999.pt | 0.882 | 0.0071483400473185904 |
+| Epoch_10.pt | 0.8808333333333334 | 0.006854979266251699 |
+| Epoch_11.pt | 0.8808333333333334 | 0.006904328576076009 |
+| Epoch_10_batch_2999.pt | 0.8781666666666667 | 0.006900751435196778 |
+| Epoch_9_batch_2999.pt | 0.8506666666666668 | 0.00579058409813128 |
+| Epoch_9_batch_5999.pt | 0.8483333333333334 | 0.006255861449004109 |
+| Epoch_8_batch_5999.pt | 0.8461666666666666 | 0.007158048220079806 |
+| Epoch_8_batch_2999.pt | 0.8431666666666666 | 0.005607809809916653 |
+| Epoch_9.pt | 0.8378333333333334 | 0.007629741998453347 |
+| Epoch_8.pt | 0.8346666666666666 | 0.005425135829214976 |
+| Epoch_5_batch_5999.pt | 0.834 | 0.0073333333333333315 |
+| Epoch_6_batch_5999.pt | 0.8328333333333333 | 0.006024896905014177 |
+| Epoch_7_batch_5999.pt | 0.8321666666666667 | 0.0070930760883457155 |
+| Epoch_5_batch_2999.pt | 0.8301666666666666 | 0.00660830939699555 |
+| Epoch_6_batch_2999.pt | 0.8300000000000001 | 0.006285393610547087 |
+| Epoch_4_batch_5999.pt | 0.8278333333333332 | 0.006717629284939481 |
+| Epoch_7_batch_2999.pt | 0.8278333333333332 | 0.007793832636161352 |
+| Epoch_6.pt | 0.8276666666666668 | 0.006246974576386935 |
+| Epoch_5.pt | 0.8225000000000001 | 0.006773450348945479 |
+| Epoch_7.pt | 0.8210000000000001 | 0.004166296279833929 |
+| Epoch_4.pt | 0.8191666666666666 | 0.007727016896801561 |
+| Epoch_4_batch_2999.pt | 0.8103333333333333 | 0.007293663747678362 |
+| Epoch_3_batch_5999.pt | 0.8045 | 0.007114799390732697 |
+| Epoch_3.pt | 0.7978333333333334 | 0.006817959218214394 |
+| Epoch_3_batch_2999.pt | 0.7951666666666667 | 0.007463698153292158 |
+| Epoch_2_batch_5999.pt | 0.7775000000000001 | 0.005165471785367927 |
+| Epoch_2.pt | 0.7708333333333333 | 0.008488372584647927 |
+| Epoch_2_batch_2999.pt | 0.7576666666666666 | 0.0066675925283011025 |
+| Epoch_1.pt | 0.7275 | 0.005634166057234542 |
+| Epoch_1_batch_5999.pt | 0.7231666666666667 | 0.0058765594200419664 |
+| Epoch_1_batch_2999.pt | 0.6628333333333333 | 0.0038733817807896395 |
+| Epoch_0.pt | 0.5853333333333333 | 0.0074195271221318 |
+| Epoch_0_batch_5999.pt | 0.5730000000000001 | 0.004777777777777773 |
+| Epoch_0_batch_2999.pt | 0.4916666666666667 | 0.004142522640512729 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..e632047ba5aedf8926bf7fc1315efc74972be9e4
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
+ +------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8826666666666668 | 0.00352416700603415 |
+| Epoch_11.pt | 0.8821666666666668 | 0.003952261424851648 |
+| Epoch_15_batch_2999.pt | 0.8808333333333334 | 0.003969404595088056 |
+| Epoch_16_batch_5999.pt | 0.8805 | 0.0038413764251250078 |
+| Epoch_12.pt | 0.8800000000000001 | 0.003415650255319866 |
+| Epoch_11_batch_2999.pt | 0.8796666666666667 | 0.004133572275052946 |
+| Epoch_13_batch_2999.pt | 0.8796666666666667 | 0.003965904066127723 |
+| Epoch_16_batch_2999.pt | 0.8796666666666667 | 0.004273085449562087 |
+| Epoch_17_batch_5999.pt | 0.8786666666666667 | 0.00453246178986025 |
+| Epoch_13.pt | 0.8785000000000001 | 0.0049156464729385885 |
+| Epoch_17_batch_2999.pt | 0.8779999999999999 | 0.004035337732741081 |
+| Epoch_11_batch_5999.pt | 0.8771666666666667 | 0.003469888104361606 |
+| Epoch_15.pt | 0.8770000000000001 | 0.004265856404544991 |
+| Epoch_12_batch_5999.pt | 0.8766666666666666 | 0.003967460238079363 |
+| Epoch_14_batch_5999.pt | 0.8765000000000001 | 0.0036468318738391817 |
+| Epoch_14_batch_2999.pt | 0.8763333333333334 | 0.004429140317332164 |
+| Epoch_14.pt | 0.8763333333333334 | 0.004422166387140531 |
+| Epoch_12_batch_2999.pt | 0.8758333333333332 | 0.004091922184071176 |
+| Epoch_15_batch_5999.pt | 0.8758333333333332 | 0.004521894610027697 |
+| Epoch_10_batch_2999.pt | 0.8755 | 0.004465574769801914 |
+| Epoch_13_batch_5999.pt | 0.875 | 0.0039047296424012394 |
+| Epoch_10.pt | 0.8746666666666668 | 0.004192880503136267 |
+| Epoch_10_batch_5999.pt | 0.8746666666666666 | 0.004699619107234803 |
+| Epoch_16.pt | 0.8724999999999999 | 0.004121982622566562 |
+| Epoch_9_batch_2999.pt | 0.8579999999999999 | 0.004498284995554925 |
+| Epoch_9_batch_5999.pt | 0.8548333333333332 | 0.004138422792547481 |
+| Epoch_8_batch_2999.pt | 0.8508333333333333 | 0.0051114130345606525 |
+| Epoch_7_batch_2999.pt | 0.8478333333333332 | 0.005079916884709396 |
+| Epoch_7_batch_5999.pt | 0.8408333333333333 | 0.004643128468533763 |
+| Epoch_6_batch_2999.pt | 0.8403333333333334 | 0.005408041566061317 |
+| Epoch_9.pt | 0.8371666666666666 | 0.004289306254879473 |
+| Epoch_4_batch_5999.pt | 0.836 | 0.003353641838397019 |
+| Epoch_5_batch_5999.pt | 0.8341666666666667 | 0.004933195693775633 |
+| Epoch_6_batch_5999.pt | 0.8341666666666667 | 0.005382011648325933 |
+| Epoch_6.pt | 0.834 | 0.0042975732457363825 |
+| Epoch_5_batch_2999.pt | 0.8331666666666667 | 0.005462837412242155 |
+| Epoch_7.pt | 0.8283333333333334 | 0.005910860479231283 |
+| Epoch_5.pt | 0.826 | 0.006451145940886575 |
+| Epoch_8_batch_5999.pt | 0.8253333333333334 | 0.004073400617738527 |
+| Epoch_8.pt | 0.8218333333333334 | 0.0027380492289670213 |
+| Epoch_4.pt | 0.8198333333333334 | 0.00617866700234977 |
+| Epoch_4_batch_2999.pt | 0.8150000000000001 | 0.005067446333773947 |
+| Epoch_3_batch_5999.pt | 0.8131666666666666 | 0.005558054993308537 |
+| Epoch_3_batch_2999.pt | 0.8066666666666666 | 0.005409182864283595 |
+| Epoch_3.pt | 0.8065 | 0.0034823186542699567 |
+| Epoch_2_batch_5999.pt | 0.7846666666666667 | 0.004229525846816509 |
+| Epoch_2.pt | 0.7771666666666667 | 0.005134307266916454 |
+| Epoch_2_batch_2999.pt | 0.776 | 0.004869912666846913 |
+| Epoch_1_batch_5999.pt | 0.7430000000000001 | 0.005011098792790969 |
+| Epoch_1.pt | 0.74 | 0.003617515688022156 |
+| Epoch_1_batch_2999.pt | 0.704 | 0.0070369395704751175 |
+| Epoch_0_batch_5999.pt | 0.6381666666666665 | 0.006355953755979359 |
+| Epoch_0.pt | 0.6379999999999999 | 0.007096338823038123 |
+| Epoch_0_batch_2999.pt | 0.5381666666666667 | 0.00688283784424153 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..42b7efb431ec1928a1476ef2bcac352df9250674
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.9531666666666666 | 0.0029128281619818486 |
+| Epoch_13_batch_2999.pt | 0.9528333333333334 | 0.002855036701673242 |
+| Epoch_16.pt | 0.952 | 0.002730712383876559 |
+| Epoch_14_batch_5999.pt | 0.9518333333333334 | 0.0026925824035672462 |
+| Epoch_15.pt | 0.9516666666666665 | 0.00276664435510861 |
+| Epoch_15_batch_5999.pt | 0.9515 | 0.0024400212506664213 |
+| Epoch_13_batch_5999.pt | 0.9513333333333334 | 0.0024062675364119727 |
+| Epoch_14_batch_2999.pt | 0.9513333333333334 | 0.002662033011269101 |
+| Epoch_17_batch_2999.pt | 0.9511666666666667 | 0.002304718723632395 |
+| Epoch_14.pt | 0.9508333333333333 | 0.002169513798862954 |
+| Epoch_16_batch_5999.pt | 0.9506666666666665 | 0.0025361582690029594 |
+| Epoch_17_batch_5999.pt | 0.9506666666666665 | 0.0023465235646603242 |
+| Epoch_13.pt | 0.9504999999999999 | 0.002485141027371677 |
+| Epoch_15_batch_2999.pt | 0.9503333333333334 | 0.0026270200927859793 |
+| Epoch_16_batch_2999.pt | 0.9501666666666667 | 0.0025391988626725435 |
+| Epoch_11_batch_5999.pt | 0.9494999999999999 | 0.0021379868227659866 |
+| Epoch_10_batch_5999.pt | 0.9486666666666667 | 0.0016442942874387479 |
+| Epoch_10.pt | 0.9478333333333333 | 0.0021523745142011668 |
+| Epoch_11_batch_2999.pt | 0.9478333333333333 | 0.003033251932159711 |
+| Epoch_12.pt | 0.9476666666666667 | 0.0029418227321941644 |
+| Epoch_12_batch_2999.pt | 0.9475 | 0.002253255532297028 |
+| Epoch_11.pt | 0.9468333333333334 | 0.0022559933893607715 |
+| Epoch_12_batch_5999.pt | 0.9441666666666666 | 0.003260784575739814 |
+| Epoch_10_batch_2999.pt | 0.9418333333333333 | 0.003224615969095876 |
+| Epoch_9_batch_5999.pt | 0.9283333333333333 | 0.0030832082056692404 |
+| Epoch_7_batch_5999.pt | 0.9215 | 0.003672133971035725 |
+| Epoch_9_batch_2999.pt | 0.9201666666666668 | 0.0019317042945237533 |
+| Epoch_7_batch_2999.pt | 0.9171666666666667 | 0.004216736202032427 |
+| Epoch_8_batch_5999.pt | 0.9168333333333333 | 0.0048460878965650815 |
+| Epoch_8_batch_2999.pt | 0.916 | 0.002081665999466134 |
+| Epoch_6_batch_5999.pt | 0.9146666666666666 | 0.0028738927014172414 |
+| Epoch_5_batch_5999.pt | 0.9126666666666667 | 0.0027352296944647084 |
+| Epoch_5_batch_2999.pt | 0.9111666666666665 | 0.004395915612710745 |
+| Epoch_9.pt | 0.9079999999999998 | 0.0019531550923607664 |
+| Epoch_6_batch_2999.pt | 0.9078333333333333 | 0.004513696577099912 |
+| Epoch_6.pt | 0.9075000000000001 | 0.004793578523580103 |
+| Epoch_7.pt | 0.9071666666666667 | 0.003239894069294079 |
+| Epoch_8.pt | 0.9068333333333334 | 0.0019790570145063243 |
+| Epoch_5.pt | 0.906 | 0.0035935470286213894 |
+| Epoch_4_batch_5999.pt | 0.9046666666666667 | 0.0031797973380564837 |
+| Epoch_4_batch_2999.pt | 0.9013333333333332 | 0.0033957126199858035 |
+| Epoch_3_batch_5999.pt | 0.8950000000000001 | 0.0034871899614389297 |
+| Epoch_3_batch_2999.pt | 0.8886666666666665 | 0.003996912388578875 |
+| Epoch_4.pt | 0.8879999999999999 | 0.005239922721134046 |
+| Epoch_3.pt | 0.8864999999999998 | 0.004370566536714871 |
+| Epoch_2_batch_5999.pt | 0.8833333333333334 | 0.003912625969257561 |
+| Epoch_2.pt | 0.8723333333333333 | 0.004876246279442601 |
+| Epoch_2_batch_2999.pt | 0.8676666666666668 | 0.003953432638812712 |
+| Epoch_1_batch_5999.pt | 0.845 | 0.004648111258522635 |
+| Epoch_1.pt | 0.828 | 0.0065111490328816 |
+| Epoch_1_batch_2999.pt | 0.7963333333333333 | 0.006439653391485378 |
+| Epoch_0.pt | 0.7131666666666667 | 0.005015100653815556 |
+| Epoch_0_batch_5999.pt | 0.6975 | 0.005347492778148068 |
+| Epoch_0_batch_2999.pt | 0.5478333333333334 | 0.007706218428589717 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..6f68ca80c079ae2141a6c052416343a74107fe22
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_2999.pt | 0.9148333333333334 | 0.002965334698907229 |
+| Epoch_16_batch_5999.pt | 0.9136666666666666 | 0.0027977062915587104 |
+| Epoch_14_batch_5999.pt | 0.9135 | 0.0038765677832043348 |
+| Epoch_17_batch_5999.pt | 0.9131666666666666 | 0.0027938424357067003 |
+| Epoch_14.pt | 0.9129999999999999 | 0.003359159212851326 |
+| Epoch_13_batch_2999.pt | 0.9128333333333332 | 0.003211188003692988 |
+| Epoch_15_batch_2999.pt | 0.9126666666666667 | 0.003937787810370969 |
+| Epoch_15_batch_5999.pt | 0.9126666666666667 | 0.003095197394929805 |
+| Epoch_14_batch_2999.pt | 0.9118333333333334 | 0.0038204291659898228 |
+| Epoch_13_batch_5999.pt | 0.9114999999999999 | 0.003195772670726588 |
+| Epoch_17.pt | 0.9113333333333336 | 0.003091206165165232 |
+| Epoch_15.pt | 0.9109999999999999 | 0.0035763282087624697 |
+| Epoch_16_batch_2999.pt | 0.9101666666666668 | 0.0029339435392246446 |
+| Epoch_13.pt | 0.9098333333333333 | 0.0032815420945827433 |
+| Epoch_11_batch_5999.pt | 0.909 | 0.0027910792637956617 |
+| Epoch_11_batch_2999.pt | 0.9088333333333333 | 0.0026764865489879836 |
+| Epoch_10_batch_5999.pt | 0.9083333333333334 | 0.003784308081316978 |
+| Epoch_16.pt | 0.9083333333333334 | 0.0031229931827900428 |
+| Epoch_12_batch_2999.pt | 0.9075 | 0.0025123153454655587 |
+| Epoch_11.pt | 0.9068333333333334 | 0.002681095224891926 |
+| Epoch_12.pt | 0.9065 | 0.0032909340488044233 |
+| Epoch_10_batch_2999.pt | 0.9051666666666666 | 0.003419714088113867 |
+| Epoch_12_batch_5999.pt | 0.9038333333333334 | 0.0032015621187164276 |
+| Epoch_10.pt | 0.9026666666666667 | 0.002802115602870774 |
+| Epoch_9_batch_5999.pt | 0.8816666666666666 | 0.0026057865332352347 |
+| Epoch_7_batch_2999.pt | 0.8798333333333334 | 0.005130699179846161 |
+| Epoch_8_batch_5999.pt | 0.8795 | 0.00428930625487947 |
+| Epoch_7_batch_5999.pt | 0.8786666666666667 | 0.0034228715112776366 |
+| Epoch_8_batch_2999.pt | 0.8786666666666665 | 0.004660048216780175 |
+| Epoch_9_batch_2999.pt | 0.8785000000000001 | 0.003812341880955059 |
+| Epoch_6_batch_5999.pt | 0.8744999999999999 | 0.003643444984904356 |
+| Epoch_9.pt | 0.8729999999999999 | 0.003887301263230198 |
+| Epoch_6_batch_2999.pt | 0.8705000000000002 | 0.004588296975545775 |
+| Epoch_5_batch_2999.pt | 0.8698333333333335 | 0.0038765677832043365 |
+| Epoch_8.pt | 0.8693333333333332 | 0.004582575694955842 |
+| Epoch_6.pt | 0.8661666666666668 | 0.004430882084797874 |
+| Epoch_5_batch_5999.pt | 0.8648333333333333 | 0.005190506528720619 |
+| Epoch_5.pt | 0.8634999999999999 | 0.0035956935833965347 |
+| Epoch_7.pt | 0.8628333333333333 | 0.004179978734661483 |
+| Epoch_4_batch_5999.pt | 0.8615 | 0.0032246159690958814 |
+| Epoch_4.pt | 0.8566666666666667 | 0.004634811914358712 |
+| Epoch_3_batch_5999.pt | 0.8528333333333332 | 0.004260426571119519 |
+| Epoch_4_batch_2999.pt | 0.8520000000000001 | 0.005425135829214979 |
+| Epoch_3_batch_2999.pt | 0.844 | 0.006531972647421811 |
+| Epoch_3.pt | 0.8378333333333332 | 0.005340562230540115 |
+| Epoch_2_batch_5999.pt | 0.8358333333333332 | 0.004933195693775631 |
+| Epoch_2.pt | 0.8240000000000001 | 0.004793256580027327 |
+| Epoch_2_batch_2999.pt | 0.8196666666666668 | 0.0061854069598741256 |
+| Epoch_1_batch_5999.pt | 0.7958333333333333 | 0.005948596680658032 |
+| Epoch_1.pt | 0.7865 | 0.004577521567803227 |
+| Epoch_1_batch_2999.pt | 0.7541666666666667 | 0.005330149512646014 |
+| Epoch_0.pt | 0.6646666666666667 | 0.008351092188494912 |
+| Epoch_0_batch_5999.pt | 0.6638333333333334 | 0.005505608812389579 |
+| Epoch_0_batch_2999.pt | 0.5306666666666666 | 0.0015555555555555546 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CircleLoss/log.log b/bob/bio/facexzoo/models/heads/CircleLoss/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..659d4ea68972a43ea48eff840b92c763088e9d7b
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CircleLoss/log.log
@@ -0,0 +1,655 @@
+INFO 2020-11-26 19:42:48 train.py: 172] Start optimization.
+INFO 2020-11-26 19:42:48 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='CircleLoss', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='arc-mobile', train_file='/export2/wangjun492/face_database/facex-zoo/private_file/train_data/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f864f978978>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'margin': 0.25, 'gamma': 256}
+INFO 2020-11-26 19:43:10 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 236.445282
+INFO 2020-11-26 19:44:36 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 229.148098
+INFO 2020-11-26 19:46:02 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 196.273600
+INFO 2020-11-26 19:47:29 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 179.734247
+INFO 2020-11-26 19:48:56 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 172.475471
+INFO 2020-11-26 19:50:23 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 167.090250
+INFO 2020-11-26 19:51:50 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 164.476667
+INFO 2020-11-26 19:53:17 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 163.304515
+INFO 2020-11-26 19:54:44 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 160.637086
+INFO 2020-11-26 19:56:11 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 158.921125
+INFO 2020-11-26 19:57:38 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 157.519713
+INFO 2020-11-26 19:59:05 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 156.050157
+INFO 2020-11-26 20:00:32 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 154.113322
+INFO 2020-11-26 20:01:59 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 152.476458
+INFO 2020-11-26 20:03:26 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 149.707236
+INFO 2020-11-26 20:04:53 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-26 20:04:53 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 149.043144
+INFO 2020-11-26 20:06:20 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 147.666695
+INFO 2020-11-26 20:07:47 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 146.025759
+INFO 2020-11-26 20:09:14 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 145.190045
+INFO 2020-11-26 20:10:41 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 144.774764
+INFO 2020-11-26 20:12:09 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 142.716846
+INFO 2020-11-26 20:13:36 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 141.269556
+INFO 2020-11-26 20:15:03 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 139.462846
+INFO 2020-11-26 20:16:30 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 137.888162
+INFO 2020-11-26 20:17:57 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 136.114996
+INFO 2020-11-26 20:19:24 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 135.194248
+INFO 2020-11-26 20:20:51 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 133.275941
+INFO 2020-11-26 20:22:18 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 132.376372
+INFO 2020-11-26 20:23:45 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 130.856675
+INFO 2020-11-26 20:25:12 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 130.048374
+INFO 2020-11-26 20:26:39 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-26 20:26:39 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 128.728865
+INFO 2020-11-26 20:28:05 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 127.498837
+INFO 2020-11-26 20:29:32 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 126.883197
+INFO 2020-11-26 20:29:39 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-26 20:29:40 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 128.330975
+INFO 2020-11-26 20:31:08 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 125.994783
+INFO 2020-11-26 20:32:35 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 124.406262
+INFO 2020-11-26 20:34:02 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 123.442757
+INFO 2020-11-26 20:35:29 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 121.986084
+INFO 2020-11-26 20:36:56 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 120.476022
+INFO 2020-11-26 20:38:23 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 119.497889
+INFO 2020-11-26 20:39:50 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 118.126316
+INFO 2020-11-26 20:41:17 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 116.712273
+INFO 2020-11-26 20:42:44 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 114.764139
+INFO 2020-11-26 20:44:11 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 113.326806
+INFO 2020-11-26 20:45:38 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 111.476230
+INFO 2020-11-26 20:47:05 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 109.822077
+INFO 2020-11-26 20:48:31 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 107.891609
+INFO 2020-11-26 20:49:58 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 105.960059
+INFO 2020-11-26 20:51:25 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-26 20:51:25 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 104.192442
+INFO 2020-11-26 20:52:52 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 102.446022
+INFO 2020-11-26 20:54:19 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 100.549216
+INFO 2020-11-26 20:55:46 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 98.592420
+INFO 2020-11-26 20:57:13 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 97.053783
+INFO 2020-11-26 20:58:40 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 95.074935
+INFO 2020-11-26 21:00:07 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 93.329992
+INFO 2020-11-26 21:01:34 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 91.466220
+INFO 2020-11-26 21:03:01 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 89.703424
+INFO 2020-11-26 21:04:27 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 88.214759
+INFO 2020-11-26 21:05:54 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 86.352812
+INFO 2020-11-26 21:07:21 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 84.603364
+INFO 2020-11-26 21:08:48 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 83.059662
+INFO 2020-11-26 21:10:15 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 81.656491
+INFO 2020-11-26 21:11:42 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 80.418262
+INFO 2020-11-26 21:13:09 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-26 21:13:09 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 78.667709
+INFO 2020-11-26 21:14:36 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 77.267353
+INFO 2020-11-26 21:16:03 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 76.273060
+INFO 2020-11-26 21:16:10 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-26 21:16:11 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 75.919550
+INFO 2020-11-26 21:17:38 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 72.338689
+INFO 2020-11-26 21:19:05 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 70.964829
+INFO 2020-11-26 21:20:31 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 70.029201
+INFO 2020-11-26 21:21:58 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 69.370619
+INFO 2020-11-26 21:23:24 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 68.796505
+INFO 2020-11-26 21:24:50 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 68.126790
+INFO 2020-11-26 21:26:16 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 67.509849
+INFO 2020-11-26 21:27:43 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 66.928115
+INFO 2020-11-26 21:29:09 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 65.845642
+INFO 2020-11-26 21:30:35 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 65.456259
+INFO 2020-11-26 21:32:01 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 64.792432
+INFO 2020-11-26 21:33:28 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 64.101271
+INFO 2020-11-26 21:34:54 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 63.566493
+INFO 2020-11-26 21:36:20 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 63.075800
+INFO 2020-11-26 21:37:46 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-26 21:37:47 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 62.403521
+INFO 2020-11-26 21:39:13 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 61.664559
+INFO 2020-11-26 21:40:40 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 61.256797
+INFO 2020-11-26 21:42:07 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 60.707795
+INFO 2020-11-26 21:43:34 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 60.303444
+INFO 2020-11-26 21:45:01 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 59.796985
+INFO 2020-11-26 21:46:28 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 59.280504
+INFO 2020-11-26 21:47:55 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 58.815001
+INFO 2020-11-26 21:49:22 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 58.356305
+INFO 2020-11-26 21:50:48 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 58.109190
+INFO 2020-11-26 21:52:15 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 57.581880
+INFO 2020-11-26 21:53:42 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 57.295455
+INFO 2020-11-26 21:55:09 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 56.750773
+INFO 2020-11-26 21:56:36 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 56.557938
+INFO 2020-11-26 21:58:03 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 56.018997
+INFO 2020-11-26 21:59:30 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-26 21:59:30 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 55.888020
+INFO 2020-11-26 22:00:57 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 55.547624
+INFO 2020-11-26 22:02:24 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 55.245033
+INFO 2020-11-26 22:02:31 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-26 22:02:33 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 54.789284
+INFO 2020-11-26 22:04:00 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 51.757208
+INFO 2020-11-26 22:05:27 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 51.327180
+INFO 2020-11-26 22:06:54 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 51.590362
+INFO 2020-11-26 22:08:21 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 51.538856
+INFO 2020-11-26 22:09:48 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 51.737381
+INFO 2020-11-26 22:11:15 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 51.802610
+INFO 2020-11-26 22:12:42 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 51.873690
+INFO 2020-11-26 22:14:09 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 51.511442
+INFO 2020-11-26 22:15:35 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 51.582274
+INFO 2020-11-26 22:17:02 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 51.628203
+INFO 2020-11-26 22:18:29 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 51.495475
+INFO 2020-11-26 22:19:56 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 51.277792
+INFO 2020-11-26 22:21:23 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 51.341286
+INFO 2020-11-26 22:22:50 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 51.088497
+INFO 2020-11-26 22:24:17 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-26 22:24:17 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 51.035089
+INFO 2020-11-26 22:25:43 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 50.865600
+INFO 2020-11-26 22:27:09 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 50.456885
+INFO 2020-11-26 22:28:36 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 50.663453
+INFO 2020-11-26 22:30:02 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 50.435186
+INFO 2020-11-26 22:31:28 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 50.138252
+INFO 2020-11-26 22:32:54 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 50.250664
+INFO 2020-11-26 22:34:20 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 50.062684
+INFO 2020-11-26 22:35:46 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 49.864215
+INFO 2020-11-26 22:37:13 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 49.808040
+INFO 2020-11-26 22:38:39 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 49.554442
+INFO 2020-11-26 22:40:05 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 49.521071
+INFO 2020-11-26 22:41:31 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 49.389865
+INFO 2020-11-26 22:42:57 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 49.155654
+INFO 2020-11-26 22:44:23 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 48.845312
+INFO 2020-11-26 22:45:49 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-26 22:45:50 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 48.927880
+INFO 2020-11-26 22:47:16 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 48.746228
+INFO 2020-11-26 22:48:43 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 48.643211
+INFO 2020-11-26 22:48:50 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-26 22:48:52 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 47.791201
+INFO 2020-11-26 22:50:19 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 45.423980
+INFO 2020-11-26 22:51:46 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 45.155928
+INFO 2020-11-26 22:53:14 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 45.462338
+INFO 2020-11-26 22:54:41 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 45.571803
+INFO 2020-11-26 22:56:08 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 45.996947
+INFO 2020-11-26 22:57:35 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 46.160358
+INFO 2020-11-26 22:59:02 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 46.215687
+INFO 2020-11-26 23:00:28 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 46.362514
+INFO 2020-11-26 23:01:55 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 46.569096
+INFO 2020-11-26 23:03:22 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 46.515524
+INFO 2020-11-26 23:04:49 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 46.557223
+INFO 2020-11-26 23:06:16 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 46.546508
+INFO 2020-11-26 23:07:43 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 46.586694
+INFO 2020-11-26 23:09:09 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 46.464597
+INFO 2020-11-26 23:10:36 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-26 23:10:36 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 46.376029
+INFO 2020-11-26 23:12:03 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 46.498382
+INFO 2020-11-26 23:13:30 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 46.463790
+INFO 2020-11-26 23:14:57 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 46.335010
+INFO 2020-11-26 23:16:23 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 46.402931
+INFO 2020-11-26 23:17:50 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 46.507156
+INFO 2020-11-26 23:19:17 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 46.240632
+INFO 2020-11-26 23:20:44 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 46.134378
+INFO 2020-11-26 23:22:11 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 46.091490
+INFO 2020-11-26 23:23:37 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 46.219833
+INFO 2020-11-26 23:25:04 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 46.022882
+INFO 2020-11-26 23:26:31 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 46.021027
+INFO 2020-11-26 23:27:58 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 45.855599
+INFO 2020-11-26 23:29:24 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 45.816725
+INFO 2020-11-26 23:30:51 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 45.683156
+INFO 2020-11-26 23:32:18 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-26 23:32:18 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 45.513741
+INFO 2020-11-26 23:33:44 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 45.623313
+INFO 2020-11-26 23:35:10 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 45.616495
+INFO 2020-11-26 23:35:17 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-26 23:35:19 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 44.961016
+INFO 2020-11-26 23:36:46 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 42.600657
+INFO 2020-11-26 23:38:13 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 42.189556
+INFO 2020-11-26 23:39:41 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 42.435889
+INFO 2020-11-26 23:41:08 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 42.928466
+INFO 2020-11-26 23:42:34 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 43.318748
+INFO 2020-11-26 23:44:01 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 43.399937
+INFO 2020-11-26 23:45:28 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 43.631693
+INFO 2020-11-26 23:46:55 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 43.738082
+INFO 2020-11-26 23:48:22 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 43.842576
+INFO 2020-11-26 23:49:48 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 43.930227
+INFO 2020-11-26 23:51:15 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 43.996763
+INFO 2020-11-26 23:52:42 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 44.191345
+INFO 2020-11-26 23:54:09 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 44.123106
+INFO 2020-11-26 23:55:36 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 44.040569
+INFO 2020-11-26 23:57:02 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-26 23:57:03 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 44.135542
+INFO 2020-11-26 23:58:29 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 44.269194
+INFO 2020-11-26 23:59:56 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 44.211707
+INFO 2020-11-27 00:01:23 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 44.172886
+INFO 2020-11-27 00:02:50 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 44.072519
+INFO 2020-11-27 00:04:17 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 44.014954
+INFO 2020-11-27 00:05:43 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 44.002913
+INFO 2020-11-27 00:07:10 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 44.034438
+INFO 2020-11-27 00:08:37 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 44.148964
+INFO 2020-11-27 00:10:04 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 44.019415
+INFO 2020-11-27 00:11:30 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 44.121731
+INFO 2020-11-27 00:12:57 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 43.993766
+INFO 2020-11-27 00:14:24 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 43.997096
+INFO 2020-11-27 00:15:51 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 43.758772
+INFO 2020-11-27 00:17:17 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 43.860515
+INFO 2020-11-27 00:18:44 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-27 00:18:44 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 43.794755
+INFO 2020-11-27 00:20:11 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 43.678361
+INFO 2020-11-27 00:21:38 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 43.598753
+INFO 2020-11-27 00:21:45 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-27 00:21:46 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 43.798932
+INFO 2020-11-27 00:23:13 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 40.667647
+INFO 2020-11-27 00:24:40 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 40.452621
+INFO 2020-11-27 00:26:06 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 40.909449
+INFO 2020-11-27 00:27:33 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 41.293824
+INFO 2020-11-27 00:28:59 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 41.443691
+INFO 2020-11-27 00:30:25 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 41.651462
+INFO 2020-11-27 00:31:51 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 41.854244
+INFO 2020-11-27 00:33:17 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 42.120480
+INFO 2020-11-27 00:34:44 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 42.384678
+INFO 2020-11-27 00:36:10 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 42.396282
+INFO 2020-11-27 00:37:36 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 42.510921
+INFO 2020-11-27 00:39:02 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 42.396719
+INFO 2020-11-27 00:40:28 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 42.582482
+INFO 2020-11-27 00:41:54 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 42.587796
+INFO 2020-11-27 00:43:20 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-27 00:43:20 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 42.671364
+INFO 2020-11-27 00:44:47 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 42.734600
+INFO 2020-11-27 00:46:14 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 42.531862
+INFO 2020-11-27 00:47:41 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 42.732665
+INFO 2020-11-27 00:49:07 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 42.690600
+INFO 2020-11-27 00:50:34 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 42.626764
+INFO 2020-11-27 00:52:01 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 42.647161
+INFO 2020-11-27 00:53:28 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 42.585500
+INFO 2020-11-27 00:54:55 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 42.752556
+INFO 2020-11-27 00:56:22 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 42.561006
+INFO 2020-11-27 00:57:49 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 42.452589
+INFO 2020-11-27 00:59:16 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 42.486131
+INFO 2020-11-27 01:00:42 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 42.673480
+INFO 2020-11-27 01:02:09 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 42.561921
+INFO 2020-11-27 01:03:36 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 42.619505
+INFO 2020-11-27 01:05:03 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-27 01:05:03 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 42.473971
+INFO 2020-11-27 01:06:30 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 42.544096
+INFO 2020-11-27 01:07:57 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 42.374096
+INFO 2020-11-27 01:08:04 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-27 01:08:05 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 42.654737
+INFO 2020-11-27 01:09:33 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 39.308887
+INFO 2020-11-27 01:11:00 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 39.105486
+INFO 2020-11-27 01:12:27 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 39.516438
+INFO 2020-11-27 01:13:54 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 39.954308
+INFO 2020-11-27 01:15:21 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 40.498375
+INFO 2020-11-27 01:16:48 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 40.779797
+INFO 2020-11-27 01:18:15 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 40.683996
+INFO 2020-11-27 01:19:42 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 40.913990
+INFO 2020-11-27 01:21:08 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 41.182009
+INFO 2020-11-27 01:22:35 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 41.414142
+INFO 2020-11-27 01:24:02 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 41.211606
+INFO 2020-11-27 01:25:28 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 41.477078
+INFO 2020-11-27 01:26:55 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 41.695661
+INFO 2020-11-27 01:28:22 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 41.531772
+INFO 2020-11-27 01:29:49 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-27 01:29:49 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 41.631660
+INFO 2020-11-27 01:31:16 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 41.471745
+INFO 2020-11-27 01:32:43 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 41.650587
+INFO 2020-11-27 01:34:09 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 41.530799
+INFO 2020-11-27 01:35:36 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 41.629342
+INFO 2020-11-27 01:37:03 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 41.652661
+INFO 2020-11-27 01:38:30 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 41.391060
+INFO 2020-11-27 01:39:57 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 41.587450
+INFO 2020-11-27 01:41:24 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 41.627298
+INFO 2020-11-27 01:42:50 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 41.771738
+INFO 2020-11-27 01:44:17 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 41.672905
+INFO 2020-11-27 01:45:44 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 41.543227
+INFO 2020-11-27 01:47:11 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 41.596790
+INFO 2020-11-27 01:48:38 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 41.685064
+INFO 2020-11-27 01:50:04 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 41.690941
+INFO 2020-11-27 01:51:31 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-27 01:51:31 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 41.400815
+INFO 2020-11-27 01:52:57 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 41.492015
+INFO 2020-11-27 01:54:24 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 41.510432
+INFO 2020-11-27 01:54:30 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-27 01:54:32 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 40.945515
+INFO 2020-11-27 01:55:59 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 38.564307
+INFO 2020-11-27 01:57:25 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 38.299250
+INFO 2020-11-27 01:58:52 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 38.680297
+INFO 2020-11-27 02:00:18 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 39.106062
+INFO 2020-11-27 02:01:45 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 39.401907
+INFO 2020-11-27 02:03:11 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 39.719832
+INFO 2020-11-27 02:04:37 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 39.913641
+INFO 2020-11-27 02:06:03 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 40.097178
+INFO 2020-11-27 02:07:29 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 40.297047
+INFO 2020-11-27 02:08:55 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 40.366074
+INFO 2020-11-27 02:10:21 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 40.576639
+INFO 2020-11-27 02:11:47 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 40.529459
+INFO 2020-11-27 02:13:14 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 40.643255
+INFO 2020-11-27 02:14:40 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 40.785482
+INFO 2020-11-27 02:16:06 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-27 02:16:06 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 40.737226
+INFO 2020-11-27 02:17:33 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 40.668111
+INFO 2020-11-27 02:18:59 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 40.934959
+INFO 2020-11-27 02:20:26 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 40.891842
+INFO 2020-11-27 02:21:53 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 40.715453
+INFO 2020-11-27 02:23:20 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 40.850994
+INFO 2020-11-27 02:24:46 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 40.872211
+INFO 2020-11-27 02:26:13 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 40.943621
+INFO 2020-11-27 02:27:40 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 40.939018
+INFO 2020-11-27 02:29:07 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 40.946464
+INFO 2020-11-27 02:30:34 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 40.871346
+INFO 2020-11-27 02:32:00 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 40.856844
+INFO 2020-11-27 02:33:27 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 40.813487
+INFO 2020-11-27 02:34:54 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 40.890397
+INFO 2020-11-27 02:36:21 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 40.743738
+INFO 2020-11-27 02:37:47 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-27 02:37:48 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 40.853788
+INFO 2020-11-27 02:39:14 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 40.517155
+INFO 2020-11-27 02:40:41 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 40.798125
+INFO 2020-11-27 02:40:48 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-27 02:40:50 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 40.668876
+INFO 2020-11-27 02:42:17 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 38.005270
+INFO 2020-11-27 02:43:44 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 37.604668
+INFO 2020-11-27 02:45:11 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 38.040666
+INFO 2020-11-27 02:46:38 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 38.510781
+INFO 2020-11-27 02:48:05 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 38.807474
+INFO 2020-11-27 02:49:32 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 38.966897
+INFO 2020-11-27 02:50:59 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 39.349659
+INFO 2020-11-27 02:52:26 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 39.414797
+INFO 2020-11-27 02:53:53 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 39.714779
+INFO 2020-11-27 02:55:20 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 39.778744
+INFO 2020-11-27 02:56:46 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 39.752089
+INFO 2020-11-27 02:58:13 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 39.914244
+INFO 2020-11-27 02:59:40 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 40.051592
+INFO 2020-11-27 03:01:07 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 39.892659
+INFO 2020-11-27 03:02:33 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-27 03:02:34 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 40.020097
+INFO 2020-11-27 03:04:00 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 40.222463
+INFO 2020-11-27 03:05:27 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 40.170754
+INFO 2020-11-27 03:06:54 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 40.248831
+INFO 2020-11-27 03:08:21 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 40.194239
+INFO 2020-11-27 03:09:47 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 40.247643
+INFO 2020-11-27 03:11:14 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 40.209278
+INFO 2020-11-27 03:12:41 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 40.371125
+INFO 2020-11-27 03:14:08 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 40.152585
+INFO 2020-11-27 03:15:35 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 40.227974
+INFO 2020-11-27 03:17:01 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 40.343418
+INFO 2020-11-27 03:18:28 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 40.202231
+INFO 2020-11-27 03:19:55 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 40.311432
+INFO 2020-11-27 03:21:22 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 40.331195
+INFO 2020-11-27 03:22:48 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 40.300664
+INFO 2020-11-27 03:24:15 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-27 03:24:15 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 40.281881
+INFO 2020-11-27 03:25:42 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 40.408143
+INFO 2020-11-27 03:27:09 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 40.120347
+INFO 2020-11-27 03:27:16 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-27 03:27:17 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 40.176020
+INFO 2020-11-27 03:28:45 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 33.909359
+INFO 2020-11-27 03:30:12 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 32.724347
+INFO 2020-11-27 03:31:39 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 32.177922
+INFO 2020-11-27 03:33:07 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 31.778316
+INFO 2020-11-27 03:34:34 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 31.529024
+INFO 2020-11-27 03:36:02 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 31.294768
+INFO 2020-11-27 03:37:30 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 31.149167
+INFO 2020-11-27 03:38:58 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 30.828602
+INFO 2020-11-27 03:40:26 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 30.668897
+INFO 2020-11-27 03:41:54 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 30.478664
+INFO 2020-11-27 03:43:21 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 30.465017
+INFO 2020-11-27 03:44:49 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 30.182491
+INFO 2020-11-27 03:46:17 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 30.111420
+INFO 2020-11-27 03:47:45 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 30.158345
+INFO 2020-11-27 03:49:13 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-27 03:49:13 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 29.883393
+INFO 2020-11-27 03:50:41 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 29.714406
+INFO 2020-11-27 03:52:09 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 29.773476
+INFO 2020-11-27 03:53:36 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 29.572188
+INFO 2020-11-27 03:55:04 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 29.425742
+INFO 2020-11-27 03:56:32 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 29.459554
+INFO 2020-11-27 03:58:00 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 29.259184
+INFO 2020-11-27 03:59:28 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 29.272473
+INFO 2020-11-27 04:00:56 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 29.062133
+INFO 2020-11-27 04:02:23 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 29.022064
+INFO 2020-11-27 04:03:51 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 28.983713
+INFO 2020-11-27 04:05:19 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 28.993835
+INFO 2020-11-27 04:06:47 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 28.904906
+INFO 2020-11-27 04:08:15 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 28.900576
+INFO 2020-11-27 04:09:42 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 28.579563
+INFO 2020-11-27 04:11:10 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-27 04:11:10 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 28.596133
+INFO 2020-11-27 04:12:37 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 28.573361
+INFO 2020-11-27 04:14:04 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 28.406788
+INFO 2020-11-27 04:14:11 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-27 04:14:13 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 28.448507
+INFO 2020-11-27 04:15:40 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 26.728660
+INFO 2020-11-27 04:17:08 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 26.788517
+INFO 2020-11-27 04:18:35 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 26.876938
+INFO 2020-11-27 04:20:02 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 26.723572
+INFO 2020-11-27 04:21:29 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 26.774317
+INFO 2020-11-27 04:22:56 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 26.746535
+INFO 2020-11-27 04:24:23 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 26.669073
+INFO 2020-11-27 04:25:50 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 26.790542
+INFO 2020-11-27 04:27:18 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 26.752551
+INFO 2020-11-27 04:28:45 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 26.758596
+INFO 2020-11-27 04:30:12 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 26.789020
+INFO 2020-11-27 04:31:39 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 26.755182
+INFO 2020-11-27 04:33:06 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 26.758084
+INFO 2020-11-27 04:34:33 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 26.700007
+INFO 2020-11-27 04:36:00 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-27 04:36:01 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 26.830555
+INFO 2020-11-27 04:37:29 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 26.815686
+INFO 2020-11-27 04:38:56 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 26.692639
+INFO 2020-11-27 04:40:24 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 26.795115
+INFO 2020-11-27 04:41:52 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 26.814807
+INFO 2020-11-27 04:43:20 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 26.743156
+INFO 2020-11-27 04:44:48 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 26.771831
+INFO 2020-11-27 04:46:15 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 26.753537
+INFO 2020-11-27 04:47:43 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 26.735943
+INFO 2020-11-27 04:49:11 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 26.711094
+INFO 2020-11-27 04:50:39 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 26.820783
+INFO 2020-11-27 04:52:06 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 26.785523
+INFO 2020-11-27 04:53:34 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 26.931479
+INFO 2020-11-27 04:55:02 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 26.742855
+INFO 2020-11-27 04:56:30 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 26.768739
+INFO 2020-11-27 04:57:57 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-27 04:57:58 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 26.890654
+INFO 2020-11-27 04:59:25 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 26.737629
+INFO 2020-11-27 05:00:53 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 26.890736
+INFO 2020-11-27 05:01:00 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-27 05:01:02 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 26.681655
+INFO 2020-11-27 05:02:30 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 25.090474
+INFO 2020-11-27 05:03:57 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 25.146984
+INFO 2020-11-27 05:05:25 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 25.154200
+INFO 2020-11-27 05:06:53 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 25.162126
+INFO 2020-11-27 05:08:21 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 25.237920
+INFO 2020-11-27 05:09:49 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 25.311664
+INFO 2020-11-27 05:11:16 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 25.354699
+INFO 2020-11-27 05:12:44 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 25.346918
+INFO 2020-11-27 05:14:12 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 25.440004
+INFO 2020-11-27 05:15:40 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 25.492017
+INFO 2020-11-27 05:17:08 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 25.549193
+INFO 2020-11-27 05:18:35 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 25.577543
+INFO 2020-11-27 05:20:03 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 25.643090
+INFO 2020-11-27 05:21:31 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 25.616621
+INFO 2020-11-27 05:22:58 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-27 05:22:59 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 25.710649
+INFO 2020-11-27 05:24:26 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 25.708967
+INFO 2020-11-27 05:25:52 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 25.684282
+INFO 2020-11-27 05:27:19 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 25.818975
+INFO 2020-11-27 05:28:46 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 25.798067
+INFO 2020-11-27 05:30:13 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 25.892547
+INFO 2020-11-27 05:31:40 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 25.849297
+INFO 2020-11-27 05:33:07 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 25.858340
+INFO 2020-11-27 05:34:33 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 25.829359
+INFO 2020-11-27 05:36:00 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 26.035046
+INFO 2020-11-27 05:37:27 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 25.930099
+INFO 2020-11-27 05:38:54 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 26.019671
+INFO 2020-11-27 05:40:21 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 25.960905
+INFO 2020-11-27 05:41:48 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 26.055500
+INFO 2020-11-27 05:43:15 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 26.076273
+INFO 2020-11-27 05:44:41 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-27 05:44:42 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 26.062323
+INFO 2020-11-27 05:46:09 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 26.219797
+INFO 2020-11-27 05:47:37 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 26.113504
+INFO 2020-11-27 05:47:43 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-27 05:47:45 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 25.922944
+INFO 2020-11-27 05:49:13 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 24.090297
+INFO 2020-11-27 05:50:40 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 23.873975
+INFO 2020-11-27 05:52:07 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 23.854976
+INFO 2020-11-27 05:53:34 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 23.919604
+INFO 2020-11-27 05:55:02 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 23.802342
+INFO 2020-11-27 05:56:30 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 23.931007
+INFO 2020-11-27 05:57:58 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 23.796777
+INFO 2020-11-27 05:59:26 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 23.837282
+INFO 2020-11-27 06:00:53 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 23.796395
+INFO 2020-11-27 06:02:21 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 23.758745
+INFO 2020-11-27 06:03:48 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 23.718186
+INFO 2020-11-27 06:05:16 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 23.672207
+INFO 2020-11-27 06:06:43 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 23.735637
+INFO 2020-11-27 06:08:10 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 23.733649
+INFO 2020-11-27 06:09:38 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-27 06:09:38 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 23.843300
+INFO 2020-11-27 06:11:05 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 23.780096
+INFO 2020-11-27 06:12:33 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 23.857551
+INFO 2020-11-27 06:14:00 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 23.777709
+INFO 2020-11-27 06:15:28 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 23.802056
+INFO 2020-11-27 06:16:55 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 23.683597
+INFO 2020-11-27 06:18:22 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 23.797441
+INFO 2020-11-27 06:19:50 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 23.793596
+INFO 2020-11-27 06:21:17 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 23.798867
+INFO 2020-11-27 06:22:44 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 23.901595
+INFO 2020-11-27 06:24:12 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 23.831785
+INFO 2020-11-27 06:25:39 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 23.794680
+INFO 2020-11-27 06:27:07 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 23.804153
+INFO 2020-11-27 06:28:34 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 23.756588
+INFO 2020-11-27 06:30:01 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 23.774126
+INFO 2020-11-27 06:31:28 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-27 06:31:29 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 23.585715
+INFO 2020-11-27 06:32:56 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 23.795388
+INFO 2020-11-27 06:34:23 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 23.846588
+INFO 2020-11-27 06:34:30 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-27 06:34:32 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 23.929906
+INFO 2020-11-27 06:36:00 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 23.518417
+INFO 2020-11-27 06:37:27 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 23.378059
+INFO 2020-11-27 06:38:54 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 23.525655
+INFO 2020-11-27 06:40:21 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 23.557490
+INFO 2020-11-27 06:41:48 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 23.516982
+INFO 2020-11-27 06:43:15 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 23.497585
+INFO 2020-11-27 06:44:42 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 23.417400
+INFO 2020-11-27 06:46:09 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 23.508198
+INFO 2020-11-27 06:47:35 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 23.550521
+INFO 2020-11-27 06:49:02 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 23.574154
+INFO 2020-11-27 06:50:29 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 23.588138
+INFO 2020-11-27 06:51:55 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 23.537958
+INFO 2020-11-27 06:53:22 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 23.513172
+INFO 2020-11-27 06:54:49 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 23.665900
+INFO 2020-11-27 06:56:15 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-27 06:56:15 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 23.617019
+INFO 2020-11-27 06:57:43 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 23.521042
+INFO 2020-11-27 06:59:10 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 23.607922
+INFO 2020-11-27 07:00:38 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 23.691289
+INFO 2020-11-27 07:02:05 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 23.544915
+INFO 2020-11-27 07:03:32 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 23.671088
+INFO 2020-11-27 07:05:00 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 23.677069
+INFO 2020-11-27 07:06:27 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 23.574303
+INFO 2020-11-27 07:07:54 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 23.577784
+INFO 2020-11-27 07:09:22 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 23.679696
+INFO 2020-11-27 07:10:49 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 23.669140
+INFO 2020-11-27 07:12:16 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 23.732464
+INFO 2020-11-27 07:13:44 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 23.584007
+INFO 2020-11-27 07:15:11 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 23.668555
+INFO 2020-11-27 07:16:38 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 23.592719
+INFO 2020-11-27 07:18:05 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-27 07:18:06 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 23.619880
+INFO 2020-11-27 07:19:33 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 23.767178
+INFO 2020-11-27 07:21:00 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 23.705099
+INFO 2020-11-27 07:21:07 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-27 07:21:09 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 23.711329
+INFO 2020-11-27 07:22:37 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 23.391213
+INFO 2020-11-27 07:24:05 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 23.319100
+INFO 2020-11-27 07:25:33 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 23.378035
+INFO 2020-11-27 07:27:00 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 23.392076
+INFO 2020-11-27 07:28:28 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 23.384959
+INFO 2020-11-27 07:29:56 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 23.477245
+INFO 2020-11-27 07:31:23 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 23.459945
+INFO 2020-11-27 07:32:50 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 23.452667
+INFO 2020-11-27 07:34:18 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 23.401626
+INFO 2020-11-27 07:35:45 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 23.446358
+INFO 2020-11-27 07:37:13 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 23.414836
+INFO 2020-11-27 07:38:40 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 23.444439
+INFO 2020-11-27 07:40:07 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 23.545497
+INFO 2020-11-27 07:41:34 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 23.381389
+INFO 2020-11-27 07:43:01 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-27 07:43:02 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 23.463254
+INFO 2020-11-27 07:44:29 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 23.428928
+INFO 2020-11-27 07:45:55 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 23.441934
+INFO 2020-11-27 07:47:22 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 23.352090
+INFO 2020-11-27 07:48:48 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 23.486887
+INFO 2020-11-27 07:50:15 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 23.442909
+INFO 2020-11-27 07:51:41 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 23.413164
+INFO 2020-11-27 07:53:08 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 23.555102
+INFO 2020-11-27 07:54:35 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 23.582662
+INFO 2020-11-27 07:56:01 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 23.477735
+INFO 2020-11-27 07:57:28 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 23.563900
+INFO 2020-11-27 07:58:54 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 23.485230
+INFO 2020-11-27 08:00:21 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 23.556081
+INFO 2020-11-27 08:01:47 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 23.543590
+INFO 2020-11-27 08:03:13 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 23.591393
+INFO 2020-11-27 08:04:40 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-27 08:04:40 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 23.549643
+INFO 2020-11-27 08:06:07 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 23.541022
+INFO 2020-11-27 08:07:35 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 23.537161
+INFO 2020-11-27 08:07:41 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-27 08:07:43 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 23.367138
+INFO 2020-11-27 08:09:11 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 23.243022
+INFO 2020-11-27 08:10:39 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 23.149845
+INFO 2020-11-27 08:12:07 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 23.253945
+INFO 2020-11-27 08:13:35 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 23.260012
+INFO 2020-11-27 08:15:02 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 23.219962
+INFO 2020-11-27 08:16:30 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 23.192115
+INFO 2020-11-27 08:17:57 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 23.202055
+INFO 2020-11-27 08:19:25 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 23.266561
+INFO 2020-11-27 08:20:52 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 23.282453
+INFO 2020-11-27 08:22:19 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 23.232793
+INFO 2020-11-27 08:23:47 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 23.129191
+INFO 2020-11-27 08:25:14 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 23.175048
+INFO 2020-11-27 08:26:41 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 23.217734
+INFO 2020-11-27 08:28:08 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 23.283132
+INFO 2020-11-27 08:29:35 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-27 08:29:35 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 23.301932
+INFO 2020-11-27 08:31:02 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 23.153000
+INFO 2020-11-27 08:32:29 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 23.151115
+INFO 2020-11-27 08:33:56 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 23.337621
+INFO 2020-11-27 08:35:23 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 23.141885
+INFO 2020-11-27 08:36:50 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 23.208935
+INFO 2020-11-27 08:38:17 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 23.323564
+INFO 2020-11-27 08:39:44 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 23.273823
+INFO 2020-11-27 08:41:11 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 23.291243
+INFO 2020-11-27 08:42:39 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 23.321186
+INFO 2020-11-27 08:44:06 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 23.277944
+INFO 2020-11-27 08:45:33 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 23.162413
+INFO 2020-11-27 08:47:00 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 23.176937
+INFO 2020-11-27 08:48:27 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 23.271643
+INFO 2020-11-27 08:49:54 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 23.267516
+INFO 2020-11-27 08:51:21 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-27 08:51:21 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 23.259937
+INFO 2020-11-27 08:52:47 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 23.181619
+INFO 2020-11-27 08:54:14 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 23.153317
+INFO 2020-11-27 08:54:21 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-27 08:54:23 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 23.140126
+INFO 2020-11-27 08:55:50 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 23.185751
+INFO 2020-11-27 08:57:18 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 23.235589
+INFO 2020-11-27 08:58:46 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 23.228196
+INFO 2020-11-27 09:00:14 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 23.190763
+INFO 2020-11-27 09:01:42 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 23.274936
+INFO 2020-11-27 09:03:09 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 23.264399
+INFO 2020-11-27 09:04:36 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 23.104942
+INFO 2020-11-27 09:06:04 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 23.289762
+INFO 2020-11-27 09:07:31 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 23.167957
+INFO 2020-11-27 09:08:58 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 23.241819
+INFO 2020-11-27 09:10:25 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 23.204760
+INFO 2020-11-27 09:11:52 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 23.216417
+INFO 2020-11-27 09:13:19 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 23.145435
+INFO 2020-11-27 09:14:46 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 23.227932
+INFO 2020-11-27 09:16:13 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-27 09:16:13 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 23.171197
+INFO 2020-11-27 09:17:40 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 23.165177
+INFO 2020-11-27 09:19:07 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 23.201884
+INFO 2020-11-27 09:20:33 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 23.300417
+INFO 2020-11-27 09:22:00 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 23.246319
+INFO 2020-11-27 09:23:27 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 23.299357
+INFO 2020-11-27 09:24:54 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 23.202319
+INFO 2020-11-27 09:26:21 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 23.171816
+INFO 2020-11-27 09:27:48 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 23.154298
+INFO 2020-11-27 09:29:15 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 23.194577
+INFO 2020-11-27 09:30:41 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 23.163668
+INFO 2020-11-27 09:32:08 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 23.219131
+INFO 2020-11-27 09:33:35 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 23.296676
+INFO 2020-11-27 09:35:02 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 23.271430
+INFO 2020-11-27 09:36:29 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 23.116875
+INFO 2020-11-27 09:37:56 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-27 09:37:56 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 23.200508
+INFO 2020-11-27 09:39:23 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 23.150287
+INFO 2020-11-27 09:40:50 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 23.116059
+INFO 2020-11-27 09:40:56 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-27 09:40:57 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/.gitkeep b/bob/bio/facexzoo/models/heads/CurricularFace/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..466e5913061c0771c6200120009fdd27ed53b3c5
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.9581666666666667 | 0.00434932945023329 |
+| Epoch_16_batch_2999.pt | 0.9578333333333333 | 0.004289306254879474 |
+| Epoch_16_batch_5999.pt | 0.9573333333333334 | 0.004403982485023646 |
+| Epoch_14.pt | 0.9571666666666667 | 0.004595018773722715 |
+| Epoch_14_batch_2999.pt | 0.9570000000000001 | 0.004365973934308554 |
+| Epoch_14_batch_5999.pt | 0.9569999999999999 | 0.004309048762147851 |
+| Epoch_17.pt | 0.9568333333333332 | 0.003884521356980745 |
+| Epoch_15_batch_2999.pt | 0.9566666666666667 | 0.00426730319326034 |
+| Epoch_13.pt | 0.9566666666666664 | 0.004353230672035494 |
+| Epoch_12_batch_5999.pt | 0.9564999999999999 | 0.004377622670687087 |
+| Epoch_15_batch_5999.pt | 0.9563333333333335 | 0.004358898943540675 |
+| Epoch_17_batch_5999.pt | 0.9563333333333333 | 0.003981438414926423 |
+| Epoch_11.pt | 0.9561666666666667 | 0.004120484809064149 |
+| Epoch_17_batch_2999.pt | 0.9560000000000001 | 0.004210510071443657 |
+| Epoch_13_batch_2999.pt | 0.9558333333333333 | 0.004319422114983371 |
+| Epoch_13_batch_5999.pt | 0.9551666666666667 | 0.0045909868758064595 |
+| Epoch_12_batch_2999.pt | 0.9549999999999998 | 0.00468778291327311 |
+| Epoch_15.pt | 0.9548333333333334 | 0.004130958098893621 |
+| Epoch_11_batch_2999.pt | 0.9546666666666667 | 0.004443055338473825 |
+| Epoch_10.pt | 0.9540000000000001 | 0.00386660280917896 |
+| Epoch_12.pt | 0.9538333333333332 | 0.004187356041380283 |
+| Epoch_11_batch_5999.pt | 0.9536666666666667 | 0.004505141096158507 |
+| Epoch_10_batch_2999.pt | 0.9518333333333333 | 0.003844588950399104 |
+| Epoch_10_batch_5999.pt | 0.9505000000000001 | 0.004906848325944789 |
+| Epoch_7_batch_2999.pt | 0.9441666666666666 | 0.00475478965795103 |
+| Epoch_7_batch_5999.pt | 0.942 | 0.004809969071536304 |
+| Epoch_8_batch_5999.pt | 0.9410000000000001 | 0.0042615130912811395 |
+| Epoch_8.pt | 0.9406666666666667 | 0.004715354483093987 |
+| Epoch_8_batch_2999.pt | 0.9403333333333332 | 0.004660048216780171 |
+| Epoch_9_batch_5999.pt | 0.9398333333333333 | 0.005325515102890151 |
+| Epoch_9_batch_2999.pt | 0.9386666666666666 | 0.006318696536939077 |
+| Epoch_6_batch_5999.pt | 0.9385 | 0.005130699179846159 |
+| Epoch_6.pt | 0.9378333333333334 | 0.005589065603568066 |
+| Epoch_5.pt | 0.9356666666666665 | 0.00530664619389314 |
+| Epoch_6_batch_2999.pt | 0.9343333333333333 | 0.005375986845956861 |
+| Epoch_5_batch_2999.pt | 0.9341666666666668 | 0.004466956872827818 |
+| Epoch_4_batch_2999.pt | 0.9338333333333333 | 0.0052355032066803715 |
+| Epoch_9.pt | 0.9329999999999998 | 0.004429140317332168 |
+| Epoch_4.pt | 0.9316666666666666 | 0.0049128202154298094 |
+| Epoch_7.pt | 0.9316666666666666 | 0.005363341504954298 |
+| Epoch_4_batch_5999.pt | 0.9315 | 0.005706018048982033 |
+| Epoch_3_batch_2999.pt | 0.9311666666666667 | 0.005205943896341165 |
+| Epoch_5_batch_5999.pt | 0.9306666666666666 | 0.00432620496309595 |
+| Epoch_3_batch_5999.pt | 0.9248333333333335 | 0.005807880980313002 |
+| Epoch_2_batch_5999.pt | 0.9246666666666666 | 0.005292669050584242 |
+| Epoch_3.pt | 0.9221666666666668 | 0.004733972309505041 |
+| Epoch_2.pt | 0.9215 | 0.00614862224856558 |
+| Epoch_2_batch_2999.pt | 0.9174999999999999 | 0.007433864787512108 |
+| Epoch_1_batch_5999.pt | 0.9135 | 0.006677999626560765 |
+| Epoch_1.pt | 0.9071666666666667 | 0.006934659773401975 |
+| Epoch_1_batch_2999.pt | 0.9033333333333333 | 0.005152010275275391 |
+| Epoch_0_batch_5999.pt | 0.8793333333333335 | 0.006574360974438664 |
+| Epoch_0.pt | 0.8790000000000001 | 0.007106769507976614 |
+| Epoch_0_batch_2999.pt | 0.8128333333333334 | 0.007962295406988378 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..fd394cc70dd69f7c5adb9fd9bf9c9399e45c96a4
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9374999999999998 | 0.0035158371849548353 |
+| Epoch_17_batch_5999.pt | 0.9371666666666666 | 0.003531603350069513 |
+| Epoch_15_batch_5999.pt | 0.937 | 0.003449816599168888 |
+| Epoch_14_batch_2999.pt | 0.9365 | 0.003646831873839173 |
+| Epoch_17.pt | 0.9360000000000002 | 0.0035416394334464927 |
+| Epoch_13_batch_5999.pt | 0.9358333333333334 | 0.0036025539637496605 |
+| Epoch_13.pt | 0.9358333333333334 | 0.003737117790853808 |
+| Epoch_16_batch_5999.pt | 0.9358333333333333 | 0.003390709893259361 |
+| Epoch_13_batch_2999.pt | 0.9356666666666668 | 0.0036106837353937532 |
+| Epoch_16.pt | 0.9355000000000002 | 0.0038654052859553277 |
+| Epoch_14_batch_5999.pt | 0.9355 | 0.0035490391674854295 |
+| Epoch_14.pt | 0.9353333333333333 | 0.003555555555555558 |
+| Epoch_15_batch_2999.pt | 0.9351666666666667 | 0.003755243248026936 |
+| Epoch_16_batch_2999.pt | 0.9346666666666668 | 0.0034228715112776345 |
+| Epoch_17_batch_2999.pt | 0.9345000000000001 | 0.003514081022415216 |
+| Epoch_10.pt | 0.9339999999999999 | 0.0038904758666692446 |
+| Epoch_11_batch_5999.pt | 0.9328333333333333 | 0.004332264825529316 |
+| Epoch_10_batch_5999.pt | 0.9326666666666668 | 0.004173697771331767 |
+| Epoch_11.pt | 0.9325000000000001 | 0.004000385783865485 |
+| Epoch_12_batch_2999.pt | 0.9323333333333332 | 0.0035763282087624693 |
+| Epoch_12_batch_5999.pt | 0.9313333333333332 | 0.003467663674294944 |
+| Epoch_11_batch_2999.pt | 0.9303333333333332 | 0.004280302285390712 |
+| Epoch_12.pt | 0.93 | 0.004437494566159076 |
+| Epoch_10_batch_2999.pt | 0.9273333333333333 | 0.0038022085849486 |
+| Epoch_9_batch_5999.pt | 0.9208333333333332 | 0.004989185836250435 |
+| Epoch_7_batch_5999.pt | 0.9204999999999999 | 0.00443784231856478 |
+| Epoch_9_batch_2999.pt | 0.9188333333333334 | 0.0048435396708615755 |
+| Epoch_8_batch_2999.pt | 0.9186666666666665 | 0.0051866400075483605 |
+| Epoch_8_batch_5999.pt | 0.9178333333333333 | 0.004416928713400209 |
+| Epoch_7_batch_2999.pt | 0.9176666666666667 | 0.004541985207993269 |
+| Epoch_8.pt | 0.9175000000000001 | 0.004754789657951019 |
+| Epoch_6_batch_5999.pt | 0.9173333333333333 | 0.004007708621526983 |
+| Epoch_6_batch_2999.pt | 0.9171666666666667 | 0.0035750334540436 |
+| Epoch_5_batch_2999.pt | 0.9168333333333333 | 0.005130699179846169 |
+| Epoch_9.pt | 0.9166666666666666 | 0.004906533814626577 |
+| Epoch_5_batch_5999.pt | 0.9148333333333334 | 0.004248819734444197 |
+| Epoch_6.pt | 0.914 | 0.0038506052113696522 |
+| Epoch_4.pt | 0.9128333333333334 | 0.0052236995546295585 |
+| Epoch_4_batch_2999.pt | 0.9116666666666667 | 0.004274529791482515 |
+| Epoch_5.pt | 0.9094999999999999 | 0.0054265579204683285 |
+| Epoch_4_batch_5999.pt | 0.9093333333333333 | 0.0043475550103364135 |
+| Epoch_3_batch_5999.pt | 0.908 | 0.005072316538005924 |
+| Epoch_7.pt | 0.906 | 0.00591190470863246 |
+| Epoch_3_batch_2999.pt | 0.9046666666666667 | 0.004358898943540678 |
+| Epoch_2_batch_5999.pt | 0.9046666666666665 | 0.0035468643776694134 |
+| Epoch_3.pt | 0.9024999999999999 | 0.004715681745091331 |
+| Epoch_2_batch_2999.pt | 0.8996666666666664 | 0.003823255674241167 |
+| Epoch_2.pt | 0.8995 | 0.004830778379628522 |
+| Epoch_1.pt | 0.8923333333333334 | 0.004901498888913943 |
+| Epoch_1_batch_5999.pt | 0.8903333333333334 | 0.0038873012632301943 |
+| Epoch_1_batch_2999.pt | 0.8828333333333334 | 0.005703854014373431 |
+| Epoch_0_batch_5999.pt | 0.8644999999999999 | 0.0047793924920007 |
+| Epoch_0.pt | 0.8620000000000001 | 0.005598721194375164 |
+| Epoch_0_batch_2999.pt | 0.8156666666666668 | 0.005410323921751043 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f0e93b0a5bcaa643b2512b9c85c0cf78ed6d2523
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.8303333333333333 | 0.006120195268409075 |
+| Epoch_17.pt | 0.8303333333333333 | 0.005830951894845296 |
+| Epoch_13.pt | 0.8296666666666667 | 0.006195378604200283 |
+| Epoch_14_batch_2999.pt | 0.829 | 0.0067914252399716175 |
+| Epoch_13_batch_5999.pt | 0.8288333333333334 | 0.006983446034624916 |
+| Epoch_15.pt | 0.8283333333333334 | 0.006749485577105528 |
+| Epoch_14.pt | 0.8278333333333332 | 0.006699226047061674 |
+| Epoch_15_batch_2999.pt | 0.8276666666666668 | 0.007154382236446779 |
+| Epoch_14_batch_5999.pt | 0.827 | 0.006743995957997119 |
+| Epoch_16_batch_5999.pt | 0.8261666666666667 | 0.006290547484465956 |
+| Epoch_17_batch_5999.pt | 0.826 | 0.006217259926016791 |
+| Epoch_15_batch_5999.pt | 0.8256666666666665 | 0.006422376007385967 |
+| Epoch_11_batch_5999.pt | 0.8254999999999999 | 0.006171669642231309 |
+| Epoch_16_batch_2999.pt | 0.8254999999999999 | 0.005968280352862992 |
+| Epoch_16.pt | 0.8244999999999999 | 0.006573656742100471 |
+| Epoch_17_batch_2999.pt | 0.8240000000000001 | 0.006541416013126674 |
+| Epoch_12_batch_5999.pt | 0.8238333333333333 | 0.007311626121678394 |
+| Epoch_12.pt | 0.8235000000000001 | 0.006499999999999999 |
+| Epoch_11_batch_2999.pt | 0.8223333333333332 | 0.006369293621318446 |
+| Epoch_11.pt | 0.8215 | 0.006760679708687345 |
+| Epoch_10.pt | 0.82 | 0.005510931896727668 |
+| Epoch_12_batch_2999.pt | 0.8198333333333334 | 0.006476214791426493 |
+| Epoch_10_batch_2999.pt | 0.8181666666666667 | 0.006128510583764106 |
+| Epoch_10_batch_5999.pt | 0.8176666666666665 | 0.006222222222222226 |
+| Epoch_8_batch_5999.pt | 0.8039999999999999 | 0.0068682486491224845 |
+| Epoch_9_batch_2999.pt | 0.7946666666666666 | 0.008243216440440623 |
+| Epoch_9_batch_5999.pt | 0.7939999999999999 | 0.008070675465482168 |
+| Epoch_6_batch_5999.pt | 0.7916666666666667 | 0.007453559924999304 |
+| Epoch_9.pt | 0.791 | 0.007350149070462764 |
+| Epoch_5.pt | 0.7901666666666667 | 0.006954215348341691 |
+| Epoch_5_batch_5999.pt | 0.7896666666666666 | 0.008380606654725816 |
+| Epoch_8_batch_2999.pt | 0.7891666666666667 | 0.007214739713655822 |
+| Epoch_6.pt | 0.7886666666666666 | 0.006008224815375902 |
+| Epoch_8.pt | 0.7883333333333333 | 0.007544108976341317 |
+| Epoch_6_batch_2999.pt | 0.7868333333333334 | 0.007897686799951205 |
+| Epoch_7_batch_5999.pt | 0.7846666666666666 | 0.007078920193865092 |
+| Epoch_5_batch_2999.pt | 0.7836666666666667 | 0.005740262823595327 |
+| Epoch_4_batch_5999.pt | 0.7821666666666667 | 0.00808156719919235 |
+| Epoch_4_batch_2999.pt | 0.7816666666666666 | 0.007062332703142532 |
+| Epoch_3_batch_5999.pt | 0.7806666666666666 | 0.007746763561579271 |
+| Epoch_7_batch_2999.pt | 0.7803333333333333 | 0.00705708644808907 |
+| Epoch_2_batch_5999.pt | 0.7801666666666665 | 0.00687835215695313 |
+| Epoch_7.pt | 0.7798333333333333 | 0.00934407450714418 |
+| Epoch_4.pt | 0.7793333333333333 | 0.007111111111111111 |
+| Epoch_3_batch_2999.pt | 0.7775000000000001 | 0.007806494605435357 |
+| Epoch_3.pt | 0.7676666666666667 | 0.00841515387944955 |
+| Epoch_2_batch_2999.pt | 0.7668333333333333 | 0.007586741196014837 |
+| Epoch_2.pt | 0.7630000000000001 | 0.0068709443705512515 |
+| Epoch_1_batch_5999.pt | 0.7583333333333334 | 0.007200822998230957 |
+| Epoch_1.pt | 0.7546666666666667 | 0.007152656417524928 |
+| Epoch_1_batch_2999.pt | 0.7390000000000001 | 0.007205964607595981 |
+| Epoch_0.pt | 0.7201666666666667 | 0.007151145984308311 |
+| Epoch_0_batch_5999.pt | 0.7156666666666667 | 0.006583743540900241 |
+| Epoch_0_batch_2999.pt | 0.6681666666666667 | 0.006673376253238234 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..7504ddcd87d2b58484b163cb4875d4568149d929
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_2999.pt | 0.9959999999999999 | 0.0012717247935843988 |
+| Epoch_13_batch_5999.pt | 0.9959999999999999 | 0.0011439589045541155 |
+| Epoch_14.pt | 0.9959999999999999 | 0.0011439589045541155 |
+| Epoch_15.pt | 0.9958333333333333 | 0.001320540480444965 |
+| Epoch_15_batch_2999.pt | 0.9958333333333332 | 0.0012729376930432875 |
+| Epoch_16_batch_5999.pt | 0.9956666666666665 | 0.0012472191289246467 |
+| Epoch_17.pt | 0.9956666666666665 | 0.001247219128924647 |
+| Epoch_12_batch_5999.pt | 0.9955 | 0.0013844373104863442 |
+| Epoch_13.pt | 0.9955 | 0.0012921892610681107 |
+| Epoch_14_batch_5999.pt | 0.9955 | 0.0012921892610681107 |
+| Epoch_16_batch_2999.pt | 0.9954999999999998 | 0.00129218926106811 |
+| Epoch_16.pt | 0.9954999999999998 | 0.00129218926106811 |
+| Epoch_10_batch_2999.pt | 0.9953333333333333 | 0.0012862041003100233 |
+| Epoch_11.pt | 0.9953333333333333 | 0.001333333333333335 |
+| Epoch_10.pt | 0.9951666666666666 | 0.0013252067157640591 |
+| Epoch_14_batch_2999.pt | 0.9951666666666666 | 0.0013709958532503363 |
+| Epoch_12_batch_2999.pt | 0.9949999999999999 | 0.0012909944487358084 |
+| Epoch_17_batch_2999.pt | 0.9949999999999999 | 0.0012909944487358035 |
+| Epoch_17_batch_5999.pt | 0.9948333333333335 | 0.0013482956777235082 |
+| Epoch_12.pt | 0.9946666666666667 | 0.0012372809695177798 |
+| Epoch_6_batch_2999.pt | 0.9945 | 0.0011666666666666618 |
+| Epoch_9_batch_5999.pt | 0.9944999999999998 | 0.0011928283640879936 |
+| Epoch_11_batch_5999.pt | 0.9944999999999998 | 0.0012680791345014764 |
+| Epoch_11_batch_2999.pt | 0.9943333333333333 | 0.001409841948938833 |
+| Epoch_15_batch_5999.pt | 0.9943333333333333 | 0.001409841948938833 |
+| Epoch_4_batch_2999.pt | 0.994 | 0.0013193713430042172 |
+| Epoch_8.pt | 0.994 | 0.0013877773329774184 |
+| Epoch_10_batch_5999.pt | 0.994 | 0.0012717247935843984 |
+| Epoch_7_batch_2999.pt | 0.9936666666666667 | 0.0014010578014353886 |
+| Epoch_8_batch_2999.pt | 0.9936666666666667 | 0.0010482201257840645 |
+| Epoch_9_batch_2999.pt | 0.9936666666666666 | 0.001378852627332317 |
+| Epoch_4_batch_5999.pt | 0.9933333333333334 | 0.0011653431646335057 |
+| Epoch_7.pt | 0.9931666666666665 | 0.0014369463507086146 |
+| Epoch_6.pt | 0.993 | 0.001378852627332321 |
+| Epoch_5.pt | 0.9928333333333332 | 0.0011928283640879928 |
+| Epoch_9.pt | 0.9925 | 0.0017258027296676735 |
+| Epoch_5_batch_5999.pt | 0.9924999999999999 | 0.0012969575033254142 |
+| Epoch_6_batch_5999.pt | 0.9921666666666666 | 0.0014497764834111042 |
+| Epoch_8_batch_5999.pt | 0.992 | 0.0011863420280034797 |
+| Epoch_3_batch_5999.pt | 0.9919999999999998 | 0.0010772621905369634 |
+| Epoch_4.pt | 0.9916666666666668 | 0.001290994448735813 |
+| Epoch_7_batch_5999.pt | 0.9915 | 0.0018164203026968613 |
+| Epoch_2.pt | 0.9914999999999999 | 0.0008031573497111617 |
+| Epoch_3_batch_2999.pt | 0.9913333333333334 | 0.0012862041003100302 |
+| Epoch_5_batch_2999.pt | 0.9911666666666668 | 0.0013844373104863472 |
+| Epoch_3.pt | 0.991 | 0.001000000000000003 |
+| Epoch_2_batch_5999.pt | 0.9905000000000002 | 0.0014497764834110977 |
+| Epoch_2_batch_2999.pt | 0.9891666666666667 | 0.0012729376930432908 |
+| Epoch_1_batch_5999.pt | 0.9884999999999999 | 0.0014153043558730012 |
+| Epoch_1.pt | 0.9878333333333333 | 0.0018434066692263383 |
+| Epoch_1_batch_2999.pt | 0.9866666666666667 | 0.002330686329267001 |
+| Epoch_0_batch_5999.pt | 0.9821666666666667 | 0.0019253026056848261 |
+| Epoch_0.pt | 0.9811666666666665 | 0.0022367580154663736 |
+| Epoch_0_batch_2999.pt | 0.9680000000000002 | 0.0031210159789306956 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..2b1d98fe3762173e59eebc0edfe9bf5065c4a6c0
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_African.txt
@@ -0,0 +1,57 @@
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.882 | 0.0037826765623442636 |
+| Epoch_17.pt | 0.8798333333333332 | 0.003491170520747769 |
+| Epoch_14_batch_2999.pt | 0.8793333333333333 | 0.004389943614475201 |
+| Epoch_15_batch_5999.pt | 0.8793333333333333 | 0.00459602619554365 |
+| Epoch_16_batch_5999.pt | 0.8791666666666667 | 0.0043265616605416765 |
+| Epoch_17_batch_5999.pt | 0.8789999999999999 | 0.004403982485023646 |
+| Epoch_13.pt | 0.8786666666666667 | 0.004546060565661954 |
+| Epoch_14_batch_5999.pt | 0.8785000000000001 | 0.003931904950937888 |
+| Epoch_12_batch_2999.pt | 0.8775000000000001 | 0.0037863464914197212 |
+| Epoch_13_batch_2999.pt | 0.8775000000000001 | 0.004023465124111088 |
+| Epoch_17_batch_2999.pt | 0.8765000000000001 | 0.004775516260631467 |
+| Epoch_15_batch_2999.pt | 0.8760000000000001 | 0.004347555010336411 |
+| Epoch_15.pt | 0.876 | 0.004283185614976973 |
+| Epoch_11_batch_5999.pt | 0.8756666666666666 | 0.004741464065189311 |
+| Epoch_13_batch_5999.pt | 0.8755000000000001 | 0.0040448870331705345 |
+| Epoch_16_batch_2999.pt | 0.8755000000000001 | 0.004165184921717006 |
+| Epoch_16.pt | 0.8748333333333334 | 0.003763453234319985 |
+| Epoch_11.pt | 0.8736666666666666 | 0.00421490594186459 |
+| Epoch_12_batch_5999.pt | 0.8723333333333333 | 0.003969015799887053 |
+| Epoch_10_batch_2999.pt | 0.8721666666666665 | 0.004513696577099911 |
+| Epoch_12.pt | 0.8708333333333333 | 0.0034359213546813826 |
+| Epoch_11_batch_2999.pt | 0.8698333333333332 | 0.004530077807088045 |
+| Epoch_10.pt | 0.8666666666666666 | 0.0036260375271290465 |
+| Epoch_10_batch_5999.pt | 0.8661666666666668 | 0.003944444444444448 |
+| Epoch_9_batch_5999.pt | 0.8468333333333333 | 0.004489012649559059 |
+| Epoch_8_batch_5999.pt | 0.8436666666666668 | 0.006038968107422752 |
+| Epoch_7_batch_2999.pt | 0.8425 | 0.00474178952519348 |
+| Epoch_6_batch_2999.pt | 0.8365 | 0.005902761437885877 |
+| Epoch_5_batch_2999.pt | 0.8353333333333334 | 0.004155912046503343 |
+| Epoch_5_batch_5999.pt | 0.8333333333333334 | 0.005746711351549222 |
+| Epoch_9_batch_2999.pt | 0.8333333333333334 | 0.0047466687473986325 |
+| Epoch_7_batch_5999.pt | 0.8293333333333333 | 0.005364492313143695 |
+| Epoch_4_batch_5999.pt | 0.8291666666666668 | 0.0057157460655998145 |
+| Epoch_6_batch_5999.pt | 0.8291666666666668 | 0.003636661744234031 |
+| Epoch_9.pt | 0.825 | 0.0049190985824841505 |
+| Epoch_7.pt | 0.8228333333333333 | 0.005578010176654772 |
+| Epoch_8_batch_2999.pt | 0.8226666666666667 | 0.005259911279353162 |
+| Epoch_8.pt | 0.8226666666666667 | 0.00452155332208351 |
+| Epoch_4_batch_2999.pt | 0.8215 | 0.003195772670726584 |
+| Epoch_4.pt | 0.8203333333333334 | 0.0045188220905917256 |
+| Epoch_3.pt | 0.8193333333333334 | 0.005271633850789535 |
+| Epoch_3_batch_2999.pt | 0.8186666666666668 | 0.003911047979288449 |
+| Epoch_3_batch_5999.pt | 0.8184999999999999 | 0.004971834249295637 |
+| Epoch_5.pt | 0.8181666666666667 | 0.004657729537494042 |
+| Epoch_6.pt | 0.8133333333333332 | 0.005252865179365693 |
+| Epoch_2_batch_5999.pt | 0.8123333333333334 | 0.004158881616047307 |
+| Epoch_2.pt | 0.8048333333333334 | 0.005406043714397374 |
+| Epoch_2_batch_2999.pt | 0.7948333333333334 | 0.00525550857390323 |
+| Epoch_1_batch_5999.pt | 0.7875 | 0.002846375212766558 |
+| Epoch_1.pt | 0.7735000000000001 | 0.005732999559621233 |
+| Epoch_1_batch_2999.pt | 0.7653333333333334 | 0.003359159212851328 |
+| Epoch_0_batch_5999.pt | 0.7391666666666666 | 0.0062817096831364215 |
+| Epoch_0.pt | 0.7383333333333333 | 0.003042903097250919 |
+| Epoch_0_batch_2999.pt | 0.6805 | 0.006921294747925009 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0623899e3d54a6c141c3984d79ed6c2bf8bbf018
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8733333333333334 | 0.0031229931827900484 |
+| Epoch_13_batch_5999.pt | 0.8733333333333333 | 0.0035136418446315362 |
+| Epoch_12_batch_5999.pt | 0.873 | 0.004103596736137647 |
+| Epoch_16_batch_2999.pt | 0.8729999999999999 | 0.0031505437508350725 |
+| Epoch_14_batch_2999.pt | 0.8726666666666668 | 0.0038022085849485922 |
+| Epoch_16_batch_5999.pt | 0.8724999999999999 | 0.0039849252978731734 |
+| Epoch_15.pt | 0.8716666666666667 | 0.0043673875571185634 |
+| Epoch_15_batch_2999.pt | 0.8714999999999999 | 0.0034106767729268524 |
+| Epoch_15_batch_5999.pt | 0.8711666666666668 | 0.0037519542233117797 |
+| Epoch_17_batch_2999.pt | 0.8711666666666668 | 0.0042604265711195174 |
+| Epoch_16.pt | 0.8710000000000001 | 0.003678432301610412 |
+| Epoch_11_batch_5999.pt | 0.8703333333333333 | 0.003485419364746252 |
+| Epoch_14_batch_5999.pt | 0.8699999999999999 | 0.003379312516832348 |
+| Epoch_13.pt | 0.8691666666666666 | 0.004061639272540946 |
+| Epoch_17_batch_5999.pt | 0.8691666666666666 | 0.004069231128273328 |
+| Epoch_14.pt | 0.8686666666666667 | 0.003708515392950805 |
+| Epoch_12.pt | 0.8673333333333334 | 0.003984538017120245 |
+| Epoch_11.pt | 0.8666666666666666 | 0.004587960627125013 |
+| Epoch_12_batch_2999.pt | 0.8664999999999999 | 0.003526355793901858 |
+| Epoch_13_batch_2999.pt | 0.8661666666666668 | 0.003460981806038339 |
+| Epoch_10.pt | 0.8656666666666666 | 0.0032317865716108853 |
+| Epoch_11_batch_2999.pt | 0.8631666666666666 | 0.003994208770670818 |
+| Epoch_10_batch_5999.pt | 0.8611666666666666 | 0.0036434449849043547 |
+| Epoch_10_batch_2999.pt | 0.8558333333333333 | 0.003471666622215104 |
+| Epoch_8_batch_2999.pt | 0.8373333333333333 | 0.004254264382017037 |
+| Epoch_7_batch_5999.pt | 0.8344999999999999 | 0.005340562230540111 |
+| Epoch_8.pt | 0.8311666666666667 | 0.004681523968396056 |
+| Epoch_9.pt | 0.8311666666666667 | 0.0048879418274609164 |
+| Epoch_7_batch_2999.pt | 0.8305 | 0.005146315954275888 |
+| Epoch_8_batch_5999.pt | 0.8303333333333335 | 0.0062301536750560774 |
+| Epoch_9_batch_5999.pt | 0.8288333333333332 | 0.005259031033128386 |
+| Epoch_9_batch_2999.pt | 0.8286666666666667 | 0.004573136806057042 |
+| Epoch_6_batch_2999.pt | 0.8255000000000001 | 0.005720064318036039 |
+| Epoch_4_batch_5999.pt | 0.8240000000000001 | 0.004151453709393204 |
+| Epoch_7.pt | 0.8236666666666667 | 0.007784917358045929 |
+| Epoch_6_batch_5999.pt | 0.8233333333333333 | 0.004654746681256319 |
+| Epoch_3_batch_2999.pt | 0.8216666666666667 | 0.005947299418254503 |
+| Epoch_5_batch_5999.pt | 0.8211666666666666 | 0.005784451274166792 |
+| Epoch_6.pt | 0.8206666666666667 | 0.004876246279442601 |
+| Epoch_5.pt | 0.8201666666666666 | 0.006998456619979034 |
+| Epoch_5_batch_2999.pt | 0.8191666666666666 | 0.0045626638309838145 |
+| Epoch_2.pt | 0.8188333333333334 | 0.004014249311000021 |
+| Epoch_3.pt | 0.8186666666666665 | 0.0047127355689885585 |
+| Epoch_4.pt | 0.8166666666666667 | 0.005784184482142391 |
+| Epoch_4_batch_2999.pt | 0.8126666666666666 | 0.0047284272957673805 |
+| Epoch_3_batch_5999.pt | 0.8116666666666668 | 0.0054997194092287005 |
+| Epoch_2_batch_2999.pt | 0.8076666666666666 | 0.00516517302185309 |
+| Epoch_2_batch_5999.pt | 0.8058333333333334 | 0.005694105680980364 |
+| Epoch_1.pt | 0.8004999999999999 | 0.00544359403789286 |
+| Epoch_1_batch_5999.pt | 0.7969999999999999 | 0.006008224815375895 |
+| Epoch_1_batch_2999.pt | 0.7856666666666667 | 0.0035066075195687827 |
+| Epoch_0.pt | 0.7631666666666665 | 0.0056789083458002685 |
+| Epoch_0_batch_5999.pt | 0.7575 | 0.004276334532872399 |
+| Epoch_0_batch_2999.pt | 0.7103333333333333 | 0.006449231938191758 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b7ecf814a43229545ee8dcfffeaaccf08d1891d6
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9526666666666666 | 0.003399346342395195 |
+| Epoch_16_batch_5999.pt | 0.9521666666666666 | 0.0036855573979159917 |
+| Epoch_13_batch_5999.pt | 0.9513333333333334 | 0.003503085060096543 |
+| Epoch_17_batch_2999.pt | 0.9513333333333334 | 0.003725123247608935 |
+| Epoch_17_batch_5999.pt | 0.951 | 0.0032317865716108896 |
+| Epoch_14_batch_5999.pt | 0.9506666666666665 | 0.003802208584948601 |
+| Epoch_17.pt | 0.9504999999999999 | 0.0038413764251250095 |
+| Epoch_14.pt | 0.9503333333333334 | 0.003942487777602261 |
+| Epoch_15_batch_5999.pt | 0.9498333333333333 | 0.0036552853665768833 |
+| Epoch_16.pt | 0.9498333333333333 | 0.003908284963769235 |
+| Epoch_14_batch_2999.pt | 0.9495000000000001 | 0.0036008400940219583 |
+| Epoch_15.pt | 0.9494999999999999 | 0.0038171963080670876 |
+| Epoch_15_batch_2999.pt | 0.9490000000000001 | 0.0034977947197104894 |
+| Epoch_12_batch_5999.pt | 0.9471666666666667 | 0.003944444444444445 |
+| Epoch_13_batch_2999.pt | 0.9468333333333335 | 0.003763453234319988 |
+| Epoch_12.pt | 0.9466666666666667 | 0.003600411499115475 |
+| Epoch_13.pt | 0.9465000000000001 | 0.0037552432480269246 |
+| Epoch_11.pt | 0.9461666666666666 | 0.0038813418832685733 |
+| Epoch_12_batch_2999.pt | 0.946 | 0.0034174569998288236 |
+| Epoch_11_batch_2999.pt | 0.9453333333333335 | 0.003349958540373624 |
+| Epoch_11_batch_5999.pt | 0.9448333333333332 | 0.0035088072610306937 |
+| Epoch_10.pt | 0.9438333333333333 | 0.004423910900359651 |
+| Epoch_10_batch_5999.pt | 0.9421666666666667 | 0.004245913067618999 |
+| Epoch_10_batch_2999.pt | 0.9418333333333333 | 0.003908284963769236 |
+| Epoch_9_batch_5999.pt | 0.9268333333333334 | 0.00461111111111111 |
+| Epoch_9_batch_2999.pt | 0.9251666666666667 | 0.00410848264321874 |
+| Epoch_8_batch_5999.pt | 0.9228333333333334 | 0.004568072235704825 |
+| Epoch_7_batch_5999.pt | 0.9223333333333334 | 0.003906310184721549 |
+| Epoch_8_batch_2999.pt | 0.9204999999999999 | 0.005346338310781807 |
+| Epoch_7_batch_2999.pt | 0.9179999999999999 | 0.004974006507910976 |
+| Epoch_9.pt | 0.9174999999999999 | 0.003938179688543838 |
+| Epoch_5_batch_5999.pt | 0.9173333333333333 | 0.003826483412827242 |
+| Epoch_8.pt | 0.9166666666666666 | 0.002821872244266731 |
+| Epoch_3_batch_2999.pt | 0.9156666666666669 | 0.004702245326555297 |
+| Epoch_6_batch_5999.pt | 0.9155 | 0.004120484809064152 |
+| Epoch_6_batch_2999.pt | 0.9143333333333332 | 0.003785938897200179 |
+| Epoch_5_batch_2999.pt | 0.9133333333333334 | 0.005103859590158154 |
+| Epoch_7.pt | 0.9119999999999999 | 0.0033683347536053575 |
+| Epoch_4_batch_2999.pt | 0.9113333333333333 | 0.004012326685615065 |
+| Epoch_4_batch_5999.pt | 0.9106666666666667 | 0.005105068892293398 |
+| Epoch_3_batch_5999.pt | 0.9096666666666667 | 0.003911047979288462 |
+| Epoch_6.pt | 0.9095000000000001 | 0.0040904133631555475 |
+| Epoch_5.pt | 0.9091666666666667 | 0.004562663830983817 |
+| Epoch_2_batch_5999.pt | 0.9085000000000001 | 0.0038123418809550558 |
+| Epoch_3.pt | 0.9076666666666666 | 0.004015402444353925 |
+| Epoch_4.pt | 0.9058333333333332 | 0.0032322640461753017 |
+| Epoch_2.pt | 0.9026666666666667 | 0.005044248650140517 |
+| Epoch_2_batch_2999.pt | 0.8975000000000002 | 0.0035070475782414752 |
+| Epoch_1.pt | 0.8931666666666667 | 0.005519046706018155 |
+| Epoch_1_batch_5999.pt | 0.8911666666666669 | 0.00348763247015989 |
+| Epoch_1_batch_2999.pt | 0.8801666666666665 | 0.0052672409438427785 |
+| Epoch_0_batch_5999.pt | 0.8571666666666667 | 0.005736228812351746 |
+| Epoch_0.pt | 0.857 | 0.004456926915584794 |
+| Epoch_0_batch_2999.pt | 0.8089999999999999 | 0.00579591171366443 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..2533154dfbc2ea058f9a9d098720a3d28f35d188
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.9056666666666666 | 0.003222222222222226 |
+| Epoch_13_batch_2999.pt | 0.9055 | 0.003370624736026116 |
+| Epoch_17_batch_5999.pt | 0.9053333333333333 | 0.002808716591058786 |
+| Epoch_15.pt | 0.9051666666666668 | 0.002388888888888885 |
+| Epoch_13.pt | 0.9048333333333332 | 0.003604267018513017 |
+| Epoch_15_batch_2999.pt | 0.9046666666666667 | 0.004058218303944376 |
+| Epoch_16_batch_5999.pt | 0.9043333333333333 | 0.0025724082006200453 |
+| Epoch_16.pt | 0.9043333333333333 | 0.0030852096393144054 |
+| Epoch_14_batch_2999.pt | 0.9035 | 0.0031763983019828432 |
+| Epoch_16_batch_2999.pt | 0.9030000000000001 | 0.0026504134315281213 |
+| Epoch_17.pt | 0.9029999999999999 | 0.00274199170650068 |
+| Epoch_14_batch_5999.pt | 0.9019999999999999 | 0.003485419364746249 |
+| Epoch_14.pt | 0.9019999999999999 | 0.0033222036417169007 |
+| Epoch_15_batch_5999.pt | 0.9018333333333335 | 0.003337497399083457 |
+| Epoch_17_batch_2999.pt | 0.9018333333333333 | 0.003156905032261189 |
+| Epoch_12.pt | 0.901 | 0.0030751894451329005 |
+| Epoch_11_batch_5999.pt | 0.9 | 0.0029292058503253486 |
+| Epoch_12_batch_5999.pt | 0.8993333333333334 | 0.004173697771331766 |
+| Epoch_11.pt | 0.8981666666666666 | 0.002923404914871745 |
+| Epoch_11_batch_2999.pt | 0.8966666666666667 | 0.0028760398012321726 |
+| Epoch_12_batch_2999.pt | 0.8963333333333333 | 0.0033407325285273086 |
+| Epoch_10_batch_5999.pt | 0.8951666666666667 | 0.0034911705207477735 |
+| Epoch_10_batch_2999.pt | 0.8938333333333333 | 0.004981756842176049 |
+| Epoch_10.pt | 0.8916666666666666 | 0.004186987484759283 |
+| Epoch_9_batch_2999.pt | 0.8705 | 0.004465574769801907 |
+| Epoch_7_batch_5999.pt | 0.8678333333333332 | 0.004388888888888892 |
+| Epoch_8_batch_5999.pt | 0.8678333333333332 | 0.004894252108719336 |
+| Epoch_7_batch_2999.pt | 0.8676666666666666 | 0.004121608220220309 |
+| Epoch_9_batch_5999.pt | 0.8658333333333333 | 0.0038349433012049615 |
+| Epoch_9.pt | 0.8643333333333334 | 0.004715354483093987 |
+| Epoch_6_batch_5999.pt | 0.8641666666666667 | 0.0036111111111111157 |
+| Epoch_5_batch_5999.pt | 0.8633333333333333 | 0.004281744192888371 |
+| Epoch_6_batch_2999.pt | 0.8618333333333335 | 0.004093430448841924 |
+| Epoch_8.pt | 0.8613333333333333 | 0.003871389197818748 |
+| Epoch_8_batch_2999.pt | 0.8608333333333335 | 0.004174067501391522 |
+| Epoch_6.pt | 0.8596666666666666 | 0.0037498971179302674 |
+| Epoch_4.pt | 0.8571666666666667 | 0.002344549760667684 |
+| Epoch_7.pt | 0.8571666666666665 | 0.004458657828274064 |
+| Epoch_3_batch_2999.pt | 0.8558333333333333 | 0.004225510415512541 |
+| Epoch_5.pt | 0.8556666666666667 | 0.004068851871911232 |
+| Epoch_4_batch_5999.pt | 0.8551666666666667 | 0.004363498992413095 |
+| Epoch_5_batch_2999.pt | 0.8551666666666666 | 0.005524636181530111 |
+| Epoch_4_batch_2999.pt | 0.85 | 0.004759655553467528 |
+| Epoch_3.pt | 0.8498333333333334 | 0.005208314814781898 |
+| Epoch_3_batch_5999.pt | 0.8484999999999999 | 0.004677566635509284 |
+| Epoch_2_batch_5999.pt | 0.8431666666666666 | 0.003963179294891506 |
+| Epoch_2.pt | 0.8418333333333333 | 0.004205008771148062 |
+| Epoch_2_batch_2999.pt | 0.8413333333333334 | 0.004686465935325141 |
+| Epoch_1_batch_5999.pt | 0.8391666666666667 | 0.004312270748961122 |
+| Epoch_1.pt | 0.8336666666666666 | 0.00349426337633698 |
+| Epoch_1_batch_2999.pt | 0.8241666666666667 | 0.0038429830234491787 |
+| Epoch_0_batch_5999.pt | 0.8068333333333333 | 0.00712953342888441 |
+| Epoch_0.pt | 0.8003333333333333 | 0.004141032256352058 |
+| Epoch_0_batch_2999.pt | 0.7623333333333333 | 0.0033993463423951913 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/CurricularFace/log.log b/bob/bio/facexzoo/models/heads/CurricularFace/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..47d8c914734134b76245a7b24f678d588058129d
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/CurricularFace/log.log
@@ -0,0 +1,655 @@
+INFO 2020-11-24 23:06:11 train.py: 172] Start optimization.
+INFO 2020-11-24 23:06:11 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='curricular_face', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='arc-mobile', train_file='/export/home/wangjun492/wj_data/faceX-Zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f78f02d6c88>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'margin': 0.5, 'scale': 64}
+INFO 2020-11-24 23:06:31 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 42.042370
+INFO 2020-11-24 23:07:54 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 40.287617
+INFO 2020-11-24 23:09:17 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 34.400746
+INFO 2020-11-24 23:10:40 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 29.962811
+INFO 2020-11-24 23:12:03 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 26.778901
+INFO 2020-11-24 23:13:25 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 24.399180
+INFO 2020-11-24 23:14:47 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 22.380460
+INFO 2020-11-24 23:16:09 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 20.666360
+INFO 2020-11-24 23:17:30 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 19.215193
+INFO 2020-11-24 23:18:50 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 17.884455
+INFO 2020-11-24 23:20:10 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 16.757137
+INFO 2020-11-24 23:21:30 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 15.721476
+INFO 2020-11-24 23:22:49 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 14.769793
+INFO 2020-11-24 23:24:08 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 13.934270
+INFO 2020-11-24 23:25:26 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 13.236284
+INFO 2020-11-24 23:26:44 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-24 23:26:45 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.597119
+INFO 2020-11-24 23:28:03 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.038828
+INFO 2020-11-24 23:29:21 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 11.497521
+INFO 2020-11-24 23:30:39 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 11.076711
+INFO 2020-11-24 23:31:56 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 10.634422
+INFO 2020-11-24 23:33:14 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 10.340265
+INFO 2020-11-24 23:34:31 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 9.994956
+INFO 2020-11-24 23:35:49 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 9.760416
+INFO 2020-11-24 23:37:06 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 9.492944
+INFO 2020-11-24 23:38:23 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 9.274274
+INFO 2020-11-24 23:39:41 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 9.085262
+INFO 2020-11-24 23:40:58 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 8.806890
+INFO 2020-11-24 23:42:15 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 8.673344
+INFO 2020-11-24 23:43:33 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 8.488318
+INFO 2020-11-24 23:44:50 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 8.377889
+INFO 2020-11-24 23:46:07 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-24 23:46:07 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 8.211444
+INFO 2020-11-24 23:47:24 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 8.094617
+INFO 2020-11-24 23:48:42 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 7.990195
+INFO 2020-11-24 23:48:48 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-24 23:48:49 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 7.960975
+INFO 2020-11-24 23:50:07 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 7.365597
+INFO 2020-11-24 23:51:24 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 7.369541
+INFO 2020-11-24 23:52:41 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 7.333354
+INFO 2020-11-24 23:53:58 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 7.335843
+INFO 2020-11-24 23:55:15 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 7.298988
+INFO 2020-11-24 23:56:33 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 7.230117
+INFO 2020-11-24 23:57:50 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 7.222714
+INFO 2020-11-24 23:59:07 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 7.173081
+INFO 2020-11-25 00:00:24 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 7.144652
+INFO 2020-11-25 00:01:41 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 7.090705
+INFO 2020-11-25 00:02:58 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 7.049920
+INFO 2020-11-25 00:04:16 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 7.025612
+INFO 2020-11-25 00:05:33 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 6.986696
+INFO 2020-11-25 00:06:50 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 6.927468
+INFO 2020-11-25 00:08:07 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-25 00:08:07 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 6.901348
+INFO 2020-11-25 00:09:24 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 6.854008
+INFO 2020-11-25 00:10:41 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 6.800296
+INFO 2020-11-25 00:11:58 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 6.793969
+INFO 2020-11-25 00:13:14 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 6.726662
+INFO 2020-11-25 00:14:31 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 6.696375
+INFO 2020-11-25 00:15:48 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 6.683155
+INFO 2020-11-25 00:17:05 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 6.650281
+INFO 2020-11-25 00:18:22 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 6.586107
+INFO 2020-11-25 00:19:38 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 6.564484
+INFO 2020-11-25 00:20:55 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 6.527291
+INFO 2020-11-25 00:22:12 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 6.522084
+INFO 2020-11-25 00:23:29 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 6.495036
+INFO 2020-11-25 00:24:46 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 6.437946
+INFO 2020-11-25 00:26:03 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 6.407512
+INFO 2020-11-25 00:27:19 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-25 00:27:20 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 6.379827
+INFO 2020-11-25 00:28:37 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 6.347942
+INFO 2020-11-25 00:29:54 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 6.310871
+INFO 2020-11-25 00:30:00 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-25 00:30:02 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 6.375245
+INFO 2020-11-25 00:31:19 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 5.938911
+INFO 2020-11-25 00:32:37 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 5.936828
+INFO 2020-11-25 00:33:54 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 6.027498
+INFO 2020-11-25 00:35:11 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 6.042055
+INFO 2020-11-25 00:36:28 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 6.077025
+INFO 2020-11-25 00:37:45 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 6.129691
+INFO 2020-11-25 00:39:02 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 6.099179
+INFO 2020-11-25 00:40:19 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 6.113452
+INFO 2020-11-25 00:41:36 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 6.085214
+INFO 2020-11-25 00:42:53 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 6.095888
+INFO 2020-11-25 00:44:10 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 6.118369
+INFO 2020-11-25 00:45:27 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 6.077740
+INFO 2020-11-25 00:46:44 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 6.086304
+INFO 2020-11-25 00:48:01 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 6.081509
+INFO 2020-11-25 00:49:18 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-25 00:49:18 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 6.036793
+INFO 2020-11-25 00:50:36 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 6.016242
+INFO 2020-11-25 00:51:53 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 6.013468
+INFO 2020-11-25 00:53:11 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 6.016957
+INFO 2020-11-25 00:54:28 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 5.980253
+INFO 2020-11-25 00:55:45 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 5.977148
+INFO 2020-11-25 00:57:03 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 5.951558
+INFO 2020-11-25 00:58:20 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 5.930721
+INFO 2020-11-25 00:59:38 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 5.909010
+INFO 2020-11-25 01:00:55 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 5.953338
+INFO 2020-11-25 01:02:12 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 5.922677
+INFO 2020-11-25 01:03:30 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 5.892111
+INFO 2020-11-25 01:04:47 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 5.886261
+INFO 2020-11-25 01:06:05 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 5.875677
+INFO 2020-11-25 01:07:22 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 5.861309
+INFO 2020-11-25 01:08:39 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-25 01:08:40 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 5.863802
+INFO 2020-11-25 01:09:57 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 5.846492
+INFO 2020-11-25 01:11:15 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 5.835188
+INFO 2020-11-25 01:11:21 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-25 01:11:22 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 5.881510
+INFO 2020-11-25 01:12:40 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 5.455216
+INFO 2020-11-25 01:13:57 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 5.478229
+INFO 2020-11-25 01:15:14 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 5.552012
+INFO 2020-11-25 01:16:31 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 5.592359
+INFO 2020-11-25 01:17:48 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 5.662363
+INFO 2020-11-25 01:19:05 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 5.636334
+INFO 2020-11-25 01:20:22 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 5.682642
+INFO 2020-11-25 01:21:39 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 5.672076
+INFO 2020-11-25 01:22:56 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 5.677696
+INFO 2020-11-25 01:24:13 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 5.700102
+INFO 2020-11-25 01:25:30 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 5.666122
+INFO 2020-11-25 01:26:47 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 5.684393
+INFO 2020-11-25 01:28:04 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 5.657063
+INFO 2020-11-25 01:29:21 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 5.675353
+INFO 2020-11-25 01:30:38 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-25 01:30:38 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 5.666831
+INFO 2020-11-25 01:31:55 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 5.646569
+INFO 2020-11-25 01:33:11 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 5.661226
+INFO 2020-11-25 01:34:28 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 5.621876
+INFO 2020-11-25 01:35:45 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 5.649912
+INFO 2020-11-25 01:37:02 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 5.689381
+INFO 2020-11-25 01:38:19 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 5.635298
+INFO 2020-11-25 01:39:37 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 5.601900
+INFO 2020-11-25 01:40:54 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 5.631656
+INFO 2020-11-25 01:42:11 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 5.606029
+INFO 2020-11-25 01:43:28 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 5.596862
+INFO 2020-11-25 01:44:45 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 5.573140
+INFO 2020-11-25 01:46:02 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 5.603300
+INFO 2020-11-25 01:47:20 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 5.591047
+INFO 2020-11-25 01:48:37 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 5.550213
+INFO 2020-11-25 01:49:54 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-25 01:49:55 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 5.509061
+INFO 2020-11-25 01:51:12 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 5.543779
+INFO 2020-11-25 01:52:29 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 5.539869
+INFO 2020-11-25 01:52:35 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-25 01:52:37 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 5.531301
+INFO 2020-11-25 01:53:54 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 5.193623
+INFO 2020-11-25 01:55:12 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 5.208746
+INFO 2020-11-25 01:56:29 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 5.282292
+INFO 2020-11-25 01:57:46 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 5.348818
+INFO 2020-11-25 01:59:03 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 5.417137
+INFO 2020-11-25 02:00:20 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 5.371798
+INFO 2020-11-25 02:01:37 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 5.393708
+INFO 2020-11-25 02:02:54 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 5.416946
+INFO 2020-11-25 02:04:11 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 5.412619
+INFO 2020-11-25 02:05:28 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 5.436570
+INFO 2020-11-25 02:06:45 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 5.431568
+INFO 2020-11-25 02:08:02 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 5.438205
+INFO 2020-11-25 02:09:19 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 5.407525
+INFO 2020-11-25 02:10:36 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 5.459442
+INFO 2020-11-25 02:11:53 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-25 02:11:53 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 5.430193
+INFO 2020-11-25 02:13:10 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 5.418758
+INFO 2020-11-25 02:14:26 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 5.438002
+INFO 2020-11-25 02:15:42 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 5.416004
+INFO 2020-11-25 02:16:59 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 5.435309
+INFO 2020-11-25 02:18:15 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 5.413711
+INFO 2020-11-25 02:19:32 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 5.424407
+INFO 2020-11-25 02:20:48 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 5.398459
+INFO 2020-11-25 02:22:05 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 5.415027
+INFO 2020-11-25 02:23:21 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 5.403962
+INFO 2020-11-25 02:24:38 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 5.398726
+INFO 2020-11-25 02:25:54 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 5.383455
+INFO 2020-11-25 02:27:11 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 5.373630
+INFO 2020-11-25 02:28:27 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 5.390030
+INFO 2020-11-25 02:29:44 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 5.341767
+INFO 2020-11-25 02:31:00 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-25 02:31:00 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 5.329285
+INFO 2020-11-25 02:32:18 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 5.353299
+INFO 2020-11-25 02:33:35 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 5.375048
+INFO 2020-11-25 02:33:41 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-25 02:33:43 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 5.360528
+INFO 2020-11-25 02:35:00 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 5.070175
+INFO 2020-11-25 02:36:18 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 5.057510
+INFO 2020-11-25 02:37:35 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 5.098853
+INFO 2020-11-25 02:38:52 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 5.130238
+INFO 2020-11-25 02:40:09 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 5.197936
+INFO 2020-11-25 02:41:26 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 5.206038
+INFO 2020-11-25 02:42:43 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 5.238030
+INFO 2020-11-25 02:44:00 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 5.282627
+INFO 2020-11-25 02:45:17 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 5.285867
+INFO 2020-11-25 02:46:34 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 5.290083
+INFO 2020-11-25 02:47:51 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 5.251016
+INFO 2020-11-25 02:49:08 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 5.272353
+INFO 2020-11-25 02:50:25 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 5.274801
+INFO 2020-11-25 02:51:42 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 5.279719
+INFO 2020-11-25 02:52:59 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-25 02:52:59 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 5.278217
+INFO 2020-11-25 02:54:16 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 5.249032
+INFO 2020-11-25 02:55:33 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 5.262226
+INFO 2020-11-25 02:56:50 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 5.251608
+INFO 2020-11-25 02:58:07 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 5.257571
+INFO 2020-11-25 02:59:24 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 5.252832
+INFO 2020-11-25 03:00:41 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 5.249336
+INFO 2020-11-25 03:01:58 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 5.252232
+INFO 2020-11-25 03:03:15 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 5.250526
+INFO 2020-11-25 03:04:33 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 5.229238
+INFO 2020-11-25 03:05:50 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 5.242230
+INFO 2020-11-25 03:07:07 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 5.231566
+INFO 2020-11-25 03:08:24 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 5.203700
+INFO 2020-11-25 03:09:41 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 5.220206
+INFO 2020-11-25 03:10:58 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 5.257583
+INFO 2020-11-25 03:12:16 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-25 03:12:16 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 5.212092
+INFO 2020-11-25 03:13:34 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 5.238987
+INFO 2020-11-25 03:14:52 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 5.231315
+INFO 2020-11-25 03:14:58 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-25 03:15:00 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 5.254541
+INFO 2020-11-25 03:16:17 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 4.913384
+INFO 2020-11-25 03:17:35 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 4.914545
+INFO 2020-11-25 03:18:53 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 4.975958
+INFO 2020-11-25 03:20:10 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 5.017715
+INFO 2020-11-25 03:21:27 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 5.044479
+INFO 2020-11-25 03:22:44 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 5.073014
+INFO 2020-11-25 03:24:02 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 5.095653
+INFO 2020-11-25 03:25:19 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 5.114384
+INFO 2020-11-25 03:26:36 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 5.129287
+INFO 2020-11-25 03:27:54 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 5.127077
+INFO 2020-11-25 03:29:11 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 5.127640
+INFO 2020-11-25 03:30:28 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 5.158664
+INFO 2020-11-25 03:31:45 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 5.164996
+INFO 2020-11-25 03:33:02 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 5.148390
+INFO 2020-11-25 03:34:19 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-25 03:34:20 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 5.161559
+INFO 2020-11-25 03:35:36 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 5.169790
+INFO 2020-11-25 03:36:53 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 5.154975
+INFO 2020-11-25 03:38:09 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 5.158896
+INFO 2020-11-25 03:39:26 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 5.133687
+INFO 2020-11-25 03:40:43 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 5.161926
+INFO 2020-11-25 03:41:59 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 5.138321
+INFO 2020-11-25 03:43:16 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 5.156313
+INFO 2020-11-25 03:44:32 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 5.144607
+INFO 2020-11-25 03:45:49 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 5.124832
+INFO 2020-11-25 03:47:05 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 5.130217
+INFO 2020-11-25 03:48:22 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 5.111167
+INFO 2020-11-25 03:49:38 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 5.131899
+INFO 2020-11-25 03:50:55 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 5.119307
+INFO 2020-11-25 03:52:11 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 5.093508
+INFO 2020-11-25 03:53:28 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-25 03:53:28 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 5.117291
+INFO 2020-11-25 03:54:45 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 5.107929
+INFO 2020-11-25 03:56:03 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 5.117152
+INFO 2020-11-25 03:56:09 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-25 03:56:10 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 5.170720
+INFO 2020-11-25 03:57:28 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 4.815692
+INFO 2020-11-25 03:58:45 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 4.819999
+INFO 2020-11-25 04:00:03 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 4.875074
+INFO 2020-11-25 04:01:20 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 4.936592
+INFO 2020-11-25 04:02:37 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 4.950650
+INFO 2020-11-25 04:03:54 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 4.957513
+INFO 2020-11-25 04:05:12 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 5.014659
+INFO 2020-11-25 04:06:29 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 5.015863
+INFO 2020-11-25 04:07:46 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 5.056468
+INFO 2020-11-25 04:09:03 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 5.050543
+INFO 2020-11-25 04:10:21 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 5.049507
+INFO 2020-11-25 04:11:38 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 5.046941
+INFO 2020-11-25 04:12:55 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 5.066842
+INFO 2020-11-25 04:14:12 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 5.070015
+INFO 2020-11-25 04:15:29 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-25 04:15:29 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 5.068537
+INFO 2020-11-25 04:16:46 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 5.051572
+INFO 2020-11-25 04:18:03 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 5.040140
+INFO 2020-11-25 04:19:20 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 5.045315
+INFO 2020-11-25 04:20:37 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 5.068163
+INFO 2020-11-25 04:21:54 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 5.052730
+INFO 2020-11-25 04:23:11 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 5.024690
+INFO 2020-11-25 04:24:28 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 5.023289
+INFO 2020-11-25 04:25:45 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 5.045918
+INFO 2020-11-25 04:27:02 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 5.086260
+INFO 2020-11-25 04:28:20 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 5.054334
+INFO 2020-11-25 04:29:37 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 5.049156
+INFO 2020-11-25 04:30:54 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 5.047981
+INFO 2020-11-25 04:32:11 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 5.036965
+INFO 2020-11-25 04:33:28 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 5.046502
+INFO 2020-11-25 04:34:45 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-25 04:34:45 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 5.034760
+INFO 2020-11-25 04:36:03 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 5.016964
+INFO 2020-11-25 04:37:20 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 5.038388
+INFO 2020-11-25 04:37:26 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-25 04:37:27 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 4.960764
+INFO 2020-11-25 04:38:45 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 4.736723
+INFO 2020-11-25 04:40:03 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 4.730425
+INFO 2020-11-25 04:41:20 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 4.791884
+INFO 2020-11-25 04:42:38 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 4.855293
+INFO 2020-11-25 04:43:55 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 4.870950
+INFO 2020-11-25 04:45:12 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 4.935642
+INFO 2020-11-25 04:46:29 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 4.921344
+INFO 2020-11-25 04:47:46 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 4.952218
+INFO 2020-11-25 04:49:04 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 4.960001
+INFO 2020-11-25 04:50:21 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 4.965463
+INFO 2020-11-25 04:51:38 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 4.980658
+INFO 2020-11-25 04:52:55 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 4.945666
+INFO 2020-11-25 04:54:12 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 4.968713
+INFO 2020-11-25 04:55:29 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 4.972122
+INFO 2020-11-25 04:56:46 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-25 04:56:47 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 4.992543
+INFO 2020-11-25 04:58:03 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 4.976044
+INFO 2020-11-25 04:59:20 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 4.985466
+INFO 2020-11-25 05:00:37 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 4.967078
+INFO 2020-11-25 05:01:53 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 4.956671
+INFO 2020-11-25 05:03:10 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 4.983818
+INFO 2020-11-25 05:04:26 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 4.970311
+INFO 2020-11-25 05:05:43 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 4.986590
+INFO 2020-11-25 05:07:00 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 4.996357
+INFO 2020-11-25 05:08:16 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 4.984256
+INFO 2020-11-25 05:09:33 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 4.980496
+INFO 2020-11-25 05:10:50 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 4.938101
+INFO 2020-11-25 05:12:06 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 4.995547
+INFO 2020-11-25 05:13:23 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 4.980469
+INFO 2020-11-25 05:14:40 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 4.965259
+INFO 2020-11-25 05:15:56 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-25 05:15:57 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 4.968681
+INFO 2020-11-25 05:17:14 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 4.978511
+INFO 2020-11-25 05:18:31 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 4.957441
+INFO 2020-11-25 05:18:37 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-25 05:18:39 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 5.009809
+INFO 2020-11-25 05:19:56 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 4.657794
+INFO 2020-11-25 05:21:14 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 4.648524
+INFO 2020-11-25 05:22:31 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 4.750580
+INFO 2020-11-25 05:23:49 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 4.795163
+INFO 2020-11-25 05:25:06 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 4.807642
+INFO 2020-11-25 05:26:23 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 4.832407
+INFO 2020-11-25 05:27:40 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 4.840196
+INFO 2020-11-25 05:28:58 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 4.867023
+INFO 2020-11-25 05:30:15 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 4.931373
+INFO 2020-11-25 05:31:32 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 4.915685
+INFO 2020-11-25 05:32:49 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 4.916253
+INFO 2020-11-25 05:34:06 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 4.919844
+INFO 2020-11-25 05:35:24 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 4.919813
+INFO 2020-11-25 05:36:41 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 4.940619
+INFO 2020-11-25 05:37:58 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-25 05:37:58 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 4.911692
+INFO 2020-11-25 05:39:15 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 4.916604
+INFO 2020-11-25 05:40:33 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 4.902658
+INFO 2020-11-25 05:41:50 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 4.920947
+INFO 2020-11-25 05:43:07 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 4.907323
+INFO 2020-11-25 05:44:24 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 4.917991
+INFO 2020-11-25 05:45:42 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 4.905762
+INFO 2020-11-25 05:46:59 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 4.930231
+INFO 2020-11-25 05:48:16 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 4.925420
+INFO 2020-11-25 05:49:33 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 4.918059
+INFO 2020-11-25 05:50:50 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 4.907311
+INFO 2020-11-25 05:52:08 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 4.906141
+INFO 2020-11-25 05:53:25 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 4.922259
+INFO 2020-11-25 05:54:42 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 4.921680
+INFO 2020-11-25 05:55:59 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 4.920889
+INFO 2020-11-25 05:57:16 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-25 05:57:17 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 4.901847
+INFO 2020-11-25 05:58:33 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 4.891255
+INFO 2020-11-25 05:59:50 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 4.897839
+INFO 2020-11-25 05:59:56 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-25 05:59:57 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 4.829921
+INFO 2020-11-25 06:01:15 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 4.103516
+INFO 2020-11-25 06:02:32 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 3.892278
+INFO 2020-11-25 06:03:50 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 3.838229
+INFO 2020-11-25 06:05:07 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 3.796969
+INFO 2020-11-25 06:06:24 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 3.745677
+INFO 2020-11-25 06:07:41 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 3.720655
+INFO 2020-11-25 06:08:58 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 3.659097
+INFO 2020-11-25 06:10:15 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 3.642434
+INFO 2020-11-25 06:11:33 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 3.635211
+INFO 2020-11-25 06:12:50 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 3.614383
+INFO 2020-11-25 06:14:07 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 3.576103
+INFO 2020-11-25 06:15:24 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 3.529414
+INFO 2020-11-25 06:16:41 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 3.531614
+INFO 2020-11-25 06:17:58 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 3.510305
+INFO 2020-11-25 06:19:15 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-25 06:19:15 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 3.481845
+INFO 2020-11-25 06:20:32 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 3.500755
+INFO 2020-11-25 06:21:49 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 3.460110
+INFO 2020-11-25 06:23:06 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 3.434048
+INFO 2020-11-25 06:24:24 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 3.438628
+INFO 2020-11-25 06:25:41 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 3.402299
+INFO 2020-11-25 06:26:58 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 3.421369
+INFO 2020-11-25 06:28:15 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 3.381788
+INFO 2020-11-25 06:29:33 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 3.393569
+INFO 2020-11-25 06:30:50 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 3.376034
+INFO 2020-11-25 06:32:07 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 3.363850
+INFO 2020-11-25 06:33:24 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 3.351688
+INFO 2020-11-25 06:34:41 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 3.354001
+INFO 2020-11-25 06:35:59 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 3.343582
+INFO 2020-11-25 06:37:16 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 3.340157
+INFO 2020-11-25 06:38:33 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-25 06:38:33 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 3.313840
+INFO 2020-11-25 06:39:50 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 3.323283
+INFO 2020-11-25 06:41:07 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 3.302351
+INFO 2020-11-25 06:41:14 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-25 06:41:15 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 3.303349
+INFO 2020-11-25 06:42:33 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 3.094353
+INFO 2020-11-25 06:43:50 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 3.083311
+INFO 2020-11-25 06:45:07 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 3.093790
+INFO 2020-11-25 06:46:25 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 3.077252
+INFO 2020-11-25 06:47:42 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 3.095194
+INFO 2020-11-25 06:48:59 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 3.125393
+INFO 2020-11-25 06:50:16 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 3.088273
+INFO 2020-11-25 06:51:33 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 3.099029
+INFO 2020-11-25 06:52:50 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 3.092995
+INFO 2020-11-25 06:54:07 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 3.113773
+INFO 2020-11-25 06:55:24 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 3.105556
+INFO 2020-11-25 06:56:41 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 3.106959
+INFO 2020-11-25 06:57:58 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 3.117923
+INFO 2020-11-25 06:59:15 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 3.118559
+INFO 2020-11-25 07:00:32 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-25 07:00:32 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 3.119714
+INFO 2020-11-25 07:01:49 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 3.110444
+INFO 2020-11-25 07:03:06 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 3.120379
+INFO 2020-11-25 07:04:24 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 3.126779
+INFO 2020-11-25 07:05:41 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 3.107868
+INFO 2020-11-25 07:06:58 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 3.128272
+INFO 2020-11-25 07:08:15 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 3.133930
+INFO 2020-11-25 07:09:32 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 3.129913
+INFO 2020-11-25 07:10:49 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 3.134143
+INFO 2020-11-25 07:12:06 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 3.119687
+INFO 2020-11-25 07:13:24 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 3.122087
+INFO 2020-11-25 07:14:41 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 3.140185
+INFO 2020-11-25 07:15:58 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 3.136302
+INFO 2020-11-25 07:17:15 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 3.122025
+INFO 2020-11-25 07:18:32 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 3.156345
+INFO 2020-11-25 07:19:49 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-25 07:19:49 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 3.135009
+INFO 2020-11-25 07:21:07 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 3.156921
+INFO 2020-11-25 07:22:24 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 3.143265
+INFO 2020-11-25 07:22:30 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-25 07:22:31 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 3.142571
+INFO 2020-11-25 07:23:49 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 2.919096
+INFO 2020-11-25 07:25:06 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 2.903563
+INFO 2020-11-25 07:26:23 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 2.926760
+INFO 2020-11-25 07:27:41 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 2.964709
+INFO 2020-11-25 07:28:58 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 2.953673
+INFO 2020-11-25 07:30:15 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 2.955584
+INFO 2020-11-25 07:31:32 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 2.974594
+INFO 2020-11-25 07:32:49 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 3.008352
+INFO 2020-11-25 07:34:06 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 2.998404
+INFO 2020-11-25 07:35:23 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 2.991091
+INFO 2020-11-25 07:36:40 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 3.007094
+INFO 2020-11-25 07:37:57 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 3.024265
+INFO 2020-11-25 07:39:14 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 3.015459
+INFO 2020-11-25 07:40:31 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 3.032809
+INFO 2020-11-25 07:41:48 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-25 07:41:48 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 3.039170
+INFO 2020-11-25 07:43:05 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 3.037112
+INFO 2020-11-25 07:44:22 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 3.030745
+INFO 2020-11-25 07:45:39 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 3.048100
+INFO 2020-11-25 07:46:56 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 3.046298
+INFO 2020-11-25 07:48:13 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 3.069485
+INFO 2020-11-25 07:49:30 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 3.055643
+INFO 2020-11-25 07:50:47 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 3.078445
+INFO 2020-11-25 07:52:04 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 3.065766
+INFO 2020-11-25 07:53:21 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 3.058420
+INFO 2020-11-25 07:54:38 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 3.072879
+INFO 2020-11-25 07:55:56 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 3.087009
+INFO 2020-11-25 07:57:12 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 3.073453
+INFO 2020-11-25 07:58:30 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 3.081941
+INFO 2020-11-25 07:59:47 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 3.113620
+INFO 2020-11-25 08:01:04 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-25 08:01:04 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 3.097797
+INFO 2020-11-25 08:02:20 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 3.095820
+INFO 2020-11-25 08:03:37 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 3.104302
+INFO 2020-11-25 08:03:43 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-25 08:03:44 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 3.024425
+INFO 2020-11-25 08:05:01 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 2.817303
+INFO 2020-11-25 08:06:18 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 2.789431
+INFO 2020-11-25 08:07:34 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 2.790243
+INFO 2020-11-25 08:08:51 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 2.781607
+INFO 2020-11-25 08:10:08 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 2.774322
+INFO 2020-11-25 08:11:24 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 2.774867
+INFO 2020-11-25 08:12:40 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 2.775057
+INFO 2020-11-25 08:13:57 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 2.762426
+INFO 2020-11-25 08:15:13 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 2.775696
+INFO 2020-11-25 08:16:30 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 2.780679
+INFO 2020-11-25 08:17:46 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 2.776697
+INFO 2020-11-25 08:19:02 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 2.779701
+INFO 2020-11-25 08:20:19 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 2.780407
+INFO 2020-11-25 08:21:35 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 2.783996
+INFO 2020-11-25 08:22:51 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-25 08:22:51 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 2.776334
+INFO 2020-11-25 08:24:08 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 2.769242
+INFO 2020-11-25 08:25:25 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 2.766997
+INFO 2020-11-25 08:26:42 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 2.772093
+INFO 2020-11-25 08:27:59 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 2.779261
+INFO 2020-11-25 08:29:16 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 2.791191
+INFO 2020-11-25 08:30:34 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 2.774417
+INFO 2020-11-25 08:31:51 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 2.772727
+INFO 2020-11-25 08:33:08 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 2.764403
+INFO 2020-11-25 08:34:25 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 2.769649
+INFO 2020-11-25 08:35:42 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 2.779951
+INFO 2020-11-25 08:36:59 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 2.786055
+INFO 2020-11-25 08:38:16 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 2.797189
+INFO 2020-11-25 08:39:33 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 2.801276
+INFO 2020-11-25 08:40:50 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 2.765005
+INFO 2020-11-25 08:42:06 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-25 08:42:07 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 2.791018
+INFO 2020-11-25 08:43:24 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 2.768172
+INFO 2020-11-25 08:44:41 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 2.791562
+INFO 2020-11-25 08:44:47 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-25 08:44:48 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 2.852801
+INFO 2020-11-25 08:46:06 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 2.730884
+INFO 2020-11-25 08:47:23 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 2.747798
+INFO 2020-11-25 08:48:40 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 2.755886
+INFO 2020-11-25 08:49:57 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 2.742973
+INFO 2020-11-25 08:51:15 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 2.748884
+INFO 2020-11-25 08:52:32 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 2.760553
+INFO 2020-11-25 08:53:49 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 2.726681
+INFO 2020-11-25 08:55:06 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 2.758610
+INFO 2020-11-25 08:56:23 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 2.741889
+INFO 2020-11-25 08:57:40 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 2.751718
+INFO 2020-11-25 08:58:57 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 2.763134
+INFO 2020-11-25 09:00:14 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 2.754917
+INFO 2020-11-25 09:01:32 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 2.741878
+INFO 2020-11-25 09:02:49 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 2.747592
+INFO 2020-11-25 09:04:05 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-25 09:04:06 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 2.739146
+INFO 2020-11-25 09:05:23 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 2.765652
+INFO 2020-11-25 09:06:40 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 2.762120
+INFO 2020-11-25 09:07:57 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 2.763000
+INFO 2020-11-25 09:09:14 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 2.752892
+INFO 2020-11-25 09:10:31 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 2.750705
+INFO 2020-11-25 09:11:49 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 2.753437
+INFO 2020-11-25 09:13:06 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 2.771864
+INFO 2020-11-25 09:14:23 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 2.767045
+INFO 2020-11-25 09:15:40 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 2.758935
+INFO 2020-11-25 09:16:57 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 2.752333
+INFO 2020-11-25 09:18:14 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 2.773884
+INFO 2020-11-25 09:19:31 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 2.760020
+INFO 2020-11-25 09:20:48 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 2.773259
+INFO 2020-11-25 09:22:05 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 2.769311
+INFO 2020-11-25 09:23:22 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-25 09:23:23 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 2.767211
+INFO 2020-11-25 09:24:39 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 2.763966
+INFO 2020-11-25 09:25:56 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 2.773697
+INFO 2020-11-25 09:26:02 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-25 09:26:03 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 2.742708
+INFO 2020-11-25 09:27:20 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 2.723138
+INFO 2020-11-25 09:28:37 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 2.729672
+INFO 2020-11-25 09:29:53 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 2.725044
+INFO 2020-11-25 09:31:11 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 2.731488
+INFO 2020-11-25 09:32:28 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 2.738319
+INFO 2020-11-25 09:33:45 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 2.731070
+INFO 2020-11-25 09:35:02 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 2.745769
+INFO 2020-11-25 09:36:19 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 2.755175
+INFO 2020-11-25 09:37:36 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 2.717457
+INFO 2020-11-25 09:38:53 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 2.733845
+INFO 2020-11-25 09:40:10 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 2.753019
+INFO 2020-11-25 09:41:28 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 2.735876
+INFO 2020-11-25 09:42:45 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 2.736461
+INFO 2020-11-25 09:44:02 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 2.736138
+INFO 2020-11-25 09:45:18 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-25 09:45:19 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 2.736460
+INFO 2020-11-25 09:46:36 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 2.751560
+INFO 2020-11-25 09:47:53 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 2.731188
+INFO 2020-11-25 09:49:10 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 2.755524
+INFO 2020-11-25 09:50:27 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 2.737660
+INFO 2020-11-25 09:51:44 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 2.748002
+INFO 2020-11-25 09:53:01 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 2.753457
+INFO 2020-11-25 09:54:19 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 2.747317
+INFO 2020-11-25 09:55:36 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 2.755721
+INFO 2020-11-25 09:56:53 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 2.752439
+INFO 2020-11-25 09:58:10 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 2.739484
+INFO 2020-11-25 09:59:27 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 2.752466
+INFO 2020-11-25 10:00:44 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 2.777040
+INFO 2020-11-25 10:02:01 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 2.733310
+INFO 2020-11-25 10:03:18 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 2.780046
+INFO 2020-11-25 10:04:35 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-25 10:04:36 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 2.759378
+INFO 2020-11-25 10:05:52 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 2.763482
+INFO 2020-11-25 10:07:09 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 2.755439
+INFO 2020-11-25 10:07:14 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-25 10:07:16 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 2.756847
+INFO 2020-11-25 10:08:33 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 2.705241
+INFO 2020-11-25 10:09:49 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 2.710645
+INFO 2020-11-25 10:11:06 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 2.709752
+INFO 2020-11-25 10:12:23 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 2.717110
+INFO 2020-11-25 10:13:39 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 2.699355
+INFO 2020-11-25 10:14:56 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 2.703827
+INFO 2020-11-25 10:16:12 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 2.714331
+INFO 2020-11-25 10:17:29 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 2.690827
+INFO 2020-11-25 10:18:45 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 2.696545
+INFO 2020-11-25 10:20:02 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 2.721342
+INFO 2020-11-25 10:21:18 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 2.730122
+INFO 2020-11-25 10:22:35 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 2.714163
+INFO 2020-11-25 10:23:51 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 2.721998
+INFO 2020-11-25 10:25:07 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 2.724338
+INFO 2020-11-25 10:26:24 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-25 10:26:24 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 2.715077
+INFO 2020-11-25 10:27:41 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 2.716743
+INFO 2020-11-25 10:28:58 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 2.713927
+INFO 2020-11-25 10:30:15 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 2.714817
+INFO 2020-11-25 10:31:32 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 2.704207
+INFO 2020-11-25 10:32:49 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 2.721077
+INFO 2020-11-25 10:34:07 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 2.710396
+INFO 2020-11-25 10:35:24 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 2.710642
+INFO 2020-11-25 10:36:41 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 2.715925
+INFO 2020-11-25 10:37:58 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 2.716740
+INFO 2020-11-25 10:39:15 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 2.714157
+INFO 2020-11-25 10:40:32 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 2.697949
+INFO 2020-11-25 10:41:49 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 2.714547
+INFO 2020-11-25 10:43:06 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 2.723439
+INFO 2020-11-25 10:44:23 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 2.710402
+INFO 2020-11-25 10:45:40 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-25 10:45:40 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 2.707711
+INFO 2020-11-25 10:46:57 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 2.722624
+INFO 2020-11-25 10:48:14 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 2.707112
+INFO 2020-11-25 10:48:21 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-25 10:48:22 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 2.652284
+INFO 2020-11-25 10:49:40 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 2.684191
+INFO 2020-11-25 10:50:57 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 2.724746
+INFO 2020-11-25 10:52:14 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 2.706552
+INFO 2020-11-25 10:53:31 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 2.712692
+INFO 2020-11-25 10:54:49 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 2.718098
+INFO 2020-11-25 10:56:06 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 2.712688
+INFO 2020-11-25 10:57:23 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 2.705111
+INFO 2020-11-25 10:58:40 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 2.724551
+INFO 2020-11-25 10:59:57 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 2.707902
+INFO 2020-11-25 11:01:15 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 2.719900
+INFO 2020-11-25 11:02:32 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 2.698277
+INFO 2020-11-25 11:03:49 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 2.723926
+INFO 2020-11-25 11:05:06 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 2.712355
+INFO 2020-11-25 11:06:23 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 2.712863
+INFO 2020-11-25 11:07:39 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-25 11:07:40 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 2.705217
+INFO 2020-11-25 11:08:56 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 2.699086
+INFO 2020-11-25 11:10:13 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 2.712169
+INFO 2020-11-25 11:11:29 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 2.702075
+INFO 2020-11-25 11:12:45 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 2.723592
+INFO 2020-11-25 11:14:02 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 2.718578
+INFO 2020-11-25 11:15:18 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 2.712310
+INFO 2020-11-25 11:16:34 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 2.718530
+INFO 2020-11-25 11:17:51 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 2.714387
+INFO 2020-11-25 11:19:07 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 2.717614
+INFO 2020-11-25 11:20:23 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 2.699235
+INFO 2020-11-25 11:21:40 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 2.697020
+INFO 2020-11-25 11:22:56 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 2.734353
+INFO 2020-11-25 11:24:12 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 2.683748
+INFO 2020-11-25 11:25:29 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 2.703836
+INFO 2020-11-25 11:26:45 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-25 11:26:45 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 2.706376
+INFO 2020-11-25 11:28:03 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 2.708875
+INFO 2020-11-25 11:29:20 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 2.700593
+INFO 2020-11-25 11:29:26 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-25 11:29:26 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/.gitkeep b/bob/bio/facexzoo/models/heads/MV-Softmax/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..0db4976e32282361fe102cc7e26e28dd760be91c
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9596666666666666 | 0.002808716591058783 |
+| Epoch_16_batch_5999.pt | 0.9595 | 0.0024726903426964316 |
+| Epoch_13.pt | 0.9588333333333333 | 0.002236758015466374 |
+| Epoch_14.pt | 0.9586666666666666 | 0.002830608711745998 |
+| Epoch_17.pt | 0.9584999999999999 | 0.0026810952248919255 |
+| Epoch_15.pt | 0.9583333333333334 | 0.002277100170213248 |
+| Epoch_14_batch_5999.pt | 0.9580000000000002 | 0.0025067809272618815 |
+| Epoch_16.pt | 0.958 | 0.0029376231258671785 |
+| Epoch_17_batch_2999.pt | 0.9578333333333333 | 0.0027448042948968053 |
+| Epoch_14_batch_2999.pt | 0.9576666666666667 | 0.002678215731820883 |
+| Epoch_17_batch_5999.pt | 0.9576666666666667 | 0.0027011657291429437 |
+| Epoch_13_batch_5999.pt | 0.9570000000000001 | 0.0027307123838765626 |
+| Epoch_13_batch_2999.pt | 0.9570000000000001 | 0.0026620330112690966 |
+| Epoch_15_batch_5999.pt | 0.9565000000000001 | 0.0026229048075806453 |
+| Epoch_15_batch_2999.pt | 0.9565000000000001 | 0.0027154109799666527 |
+| Epoch_12_batch_5999.pt | 0.9563333333333333 | 0.002662033011269096 |
+| Epoch_11_batch_5999.pt | 0.9560000000000001 | 0.003353641838397019 |
+| Epoch_10_batch_2999.pt | 0.9558333333333333 | 0.0029000851413640456 |
+| Epoch_12.pt | 0.9556666666666667 | 0.003288588574877501 |
+| Epoch_12_batch_2999.pt | 0.9553333333333331 | 0.003237511618740774 |
+| Epoch_11.pt | 0.9550000000000001 | 0.0028544961285922577 |
+| Epoch_10_batch_5999.pt | 0.9536666666666666 | 0.0031407320055783453 |
+| Epoch_11_batch_2999.pt | 0.9531666666666666 | 0.0029234049148717527 |
+| Epoch_10.pt | 0.9531666666666666 | 0.0030169275516412093 |
+| Epoch_8_batch_2999.pt | 0.9455 | 0.003505287012077192 |
+| Epoch_8_batch_5999.pt | 0.9445 | 0.0031822229981377076 |
+| Epoch_7_batch_2999.pt | 0.9418333333333331 | 0.00419766248885858 |
+| Epoch_9.pt | 0.9413333333333332 | 0.003749897117930266 |
+| Epoch_9_batch_5999.pt | 0.9411666666666665 | 0.004245913067618996 |
+| Epoch_6_batch_5999.pt | 0.9405000000000001 | 0.00441692871340021 |
+| Epoch_7_batch_5999.pt | 0.9403333333333332 | 0.003839367231815777 |
+| Epoch_5_batch_5999.pt | 0.9401666666666667 | 0.004461425891758629 |
+| Epoch_6_batch_2999.pt | 0.9396666666666667 | 0.003798960221617497 |
+| Epoch_9_batch_2999.pt | 0.9383333333333332 | 0.003583225665910461 |
+| Epoch_5_batch_2999.pt | 0.9356666666666668 | 0.0043899436144752 |
+| Epoch_8.pt | 0.9335000000000001 | 0.0038845213569807416 |
+| Epoch_6.pt | 0.9328333333333333 | 0.005500000000000003 |
+| Epoch_4_batch_2999.pt | 0.9305 | 0.004486261607382757 |
+| Epoch_4_batch_5999.pt | 0.9303333333333332 | 0.004373037478700983 |
+| Epoch_4.pt | 0.9301666666666666 | 0.0031666666666666623 |
+| Epoch_3_batch_2999.pt | 0.9298333333333334 | 0.004047938051543746 |
+| Epoch_7.pt | 0.9286666666666668 | 0.0035030850600965497 |
+| Epoch_3_batch_5999.pt | 0.9275 | 0.004603071974772747 |
+| Epoch_2_batch_5999.pt | 0.9241666666666667 | 0.006087073938749425 |
+| Epoch_3.pt | 0.9235 | 0.00513671125031725 |
+| Epoch_5.pt | 0.9218333333333332 | 0.004419722911821962 |
+| Epoch_2_batch_2999.pt | 0.921 | 0.0051830683509736 |
+| Epoch_2.pt | 0.9126666666666667 | 0.00476095228569523 |
+| Epoch_1_batch_5999.pt | 0.9028333333333333 | 0.007777976187945486 |
+| Epoch_1.pt | 0.9019999999999999 | 0.005664487598458173 |
+| Epoch_1_batch_2999.pt | 0.893 | 0.006185406959874126 |
+| Epoch_0_batch_5999.pt | 0.8415000000000001 | 0.007224145043182491 |
+| Epoch_0.pt | 0.8406666666666667 | 0.006602468674109144 |
+| Epoch_0_batch_2999.pt | 0.7585 | 0.005871304984602846 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..77bce39df7a606a31ef76f7d74719a0a4d8ef50f
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9381666666666666 | 0.003713921091857389 |
+| Epoch_14_batch_2999.pt | 0.9378333333333334 | 0.0034609818060383395 |
+| Epoch_14.pt | 0.9376666666666666 | 0.0033166247903553946 |
+| Epoch_15.pt | 0.9375 | 0.0037944891814509227 |
+| Epoch_13.pt | 0.9375 | 0.003593976442141298 |
+| Epoch_16_batch_2999.pt | 0.9375 | 0.004031128874149273 |
+| Epoch_12_batch_2999.pt | 0.9373333333333334 | 0.004015402444353916 |
+| Epoch_10_batch_5999.pt | 0.9373333333333334 | 0.0036868133384526766 |
+| Epoch_11_batch_2999.pt | 0.9371666666666668 | 0.0035140810224152056 |
+| Epoch_17_batch_5999.pt | 0.9369999999999999 | 0.0040046269535422415 |
+| Epoch_16.pt | 0.9368333333333334 | 0.0036552853665768807 |
+| Epoch_15_batch_5999.pt | 0.9366666666666668 | 0.0039518709430619355 |
+| Epoch_12.pt | 0.9366666666666665 | 0.003944053188733076 |
+| Epoch_17_batch_2999.pt | 0.9365 | 0.0035870996571906477 |
+| Epoch_12_batch_5999.pt | 0.9361666666666666 | 0.0035316033500695124 |
+| Epoch_17.pt | 0.9359999999999999 | 0.0037035185138886554 |
+| Epoch_15_batch_2999.pt | 0.9359999999999999 | 0.0036700321255363913 |
+| Epoch_13_batch_2999.pt | 0.9359999999999999 | 0.0037284359412361094 |
+| Epoch_14_batch_5999.pt | 0.9356666666666668 | 0.0034623192113072995 |
+| Epoch_11_batch_5999.pt | 0.9356666666666668 | 0.004099081499043714 |
+| Epoch_10_batch_2999.pt | 0.9354999999999999 | 0.0037271940261598894 |
+| Epoch_13_batch_5999.pt | 0.9353333333333333 | 0.003503085060096543 |
+| Epoch_11.pt | 0.9351666666666667 | 0.003156905032261193 |
+| Epoch_10.pt | 0.9334999999999999 | 0.00389245867902296 |
+| Epoch_9_batch_5999.pt | 0.9271666666666667 | 0.004082860894816248 |
+| Epoch_7_batch_5999.pt | 0.9256666666666667 | 0.003644715437079277 |
+| Epoch_8_batch_2999.pt | 0.9255000000000001 | 0.00458156531555982 |
+| Epoch_9_batch_2999.pt | 0.9243333333333332 | 0.0038666028091789584 |
+| Epoch_7_batch_2999.pt | 0.9238333333333333 | 0.003324525400083818 |
+| Epoch_6_batch_5999.pt | 0.9236666666666666 | 0.0038393672318157825 |
+| Epoch_5_batch_5999.pt | 0.9225 | 0.0030251007533471062 |
+| Epoch_9.pt | 0.9221666666666666 | 0.0033152286106301523 |
+| Epoch_8.pt | 0.9214999999999998 | 0.003482318654269966 |
+| Epoch_6_batch_2999.pt | 0.9213333333333334 | 0.0034318767136623358 |
+| Epoch_8_batch_5999.pt | 0.9211666666666666 | 0.0034161020310318276 |
+| Epoch_4_batch_5999.pt | 0.9185000000000001 | 0.004040306185683719 |
+| Epoch_6.pt | 0.9185000000000001 | 0.0036972629182166314 |
+| Epoch_4.pt | 0.9178333333333333 | 0.004289306254879471 |
+| Epoch_5_batch_2999.pt | 0.9173333333333332 | 0.0035241670060341605 |
+| Epoch_3_batch_5999.pt | 0.915 | 0.004201704533597563 |
+| Epoch_3_batch_2999.pt | 0.9148333333333334 | 0.003796115623523669 |
+| Epoch_4_batch_2999.pt | 0.9145 | 0.003677173498527313 |
+| Epoch_3.pt | 0.9115 | 0.004769048873779113 |
+| Epoch_7.pt | 0.9108333333333334 | 0.0037122586382862485 |
+| Epoch_2_batch_2999.pt | 0.9106666666666667 | 0.005364492313143695 |
+| Epoch_2_batch_5999.pt | 0.908 | 0.005310134731266008 |
+| Epoch_2.pt | 0.9076666666666666 | 0.004268749491621903 |
+| Epoch_5.pt | 0.9049999999999999 | 0.002897423291201178 |
+| Epoch_1.pt | 0.9023333333333333 | 0.004876246279442598 |
+| Epoch_1_batch_5999.pt | 0.8965 | 0.004915646472938589 |
+| Epoch_1_batch_2999.pt | 0.8821666666666668 | 0.005832275036275763 |
+| Epoch_0_batch_5999.pt | 0.8460000000000001 | 0.005410323921751042 |
+| Epoch_0.pt | 0.8458333333333332 | 0.006513281776967183 |
+| Epoch_0_batch_2999.pt | 0.7321666666666666 | 0.008913334441397197 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..795f9f8bb3803af502a7f3832dca0e3791ee72c1
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.8333333333333333 | 0.007260581890090914 |
+| Epoch_16_batch_5999.pt | 0.8308333333333333 | 0.006541651922606642 |
+| Epoch_10_batch_5999.pt | 0.8305 | 0.006898962169221229 |
+| Epoch_15.pt | 0.8303333333333333 | 0.006544246365612105 |
+| Epoch_14.pt | 0.8301666666666667 | 0.006394683362656983 |
+| Epoch_12_batch_2999.pt | 0.8298333333333334 | 0.0068693719949452685 |
+| Epoch_12_batch_5999.pt | 0.8298333333333334 | 0.006806179109755634 |
+| Epoch_17_batch_2999.pt | 0.8296666666666667 | 0.006343072433863149 |
+| Epoch_17_batch_5999.pt | 0.8295 | 0.006912370374736211 |
+| Epoch_15_batch_2999.pt | 0.8285 | 0.006385022984963992 |
+| Epoch_16_batch_2999.pt | 0.8281666666666666 | 0.0066223061143082865 |
+| Epoch_16.pt | 0.828 | 0.006220237778790407 |
+| Epoch_13_batch_2999.pt | 0.8278333333333334 | 0.006781647347123024 |
+| Epoch_13.pt | 0.8275 | 0.006723140433157013 |
+| Epoch_14_batch_2999.pt | 0.8271666666666666 | 0.006804364975223588 |
+| Epoch_12.pt | 0.8266666666666665 | 0.007657804862272348 |
+| Epoch_15_batch_5999.pt | 0.8265 | 0.006622306114308287 |
+| Epoch_17.pt | 0.8265 | 0.005765210691812374 |
+| Epoch_11_batch_5999.pt | 0.8263333333333334 | 0.007473409653688988 |
+| Epoch_14_batch_5999.pt | 0.8256666666666665 | 0.007058835633008262 |
+| Epoch_11_batch_2999.pt | 0.825 | 0.007395360574352775 |
+| Epoch_10_batch_2999.pt | 0.8243333333333333 | 0.007400367024598542 |
+| Epoch_11.pt | 0.8234999999999999 | 0.006336500182900753 |
+| Epoch_10.pt | 0.8203333333333334 | 0.006059376979749321 |
+| Epoch_9_batch_5999.pt | 0.8091666666666667 | 0.005785518319236597 |
+| Epoch_8.pt | 0.8038333333333332 | 0.008326107978766194 |
+| Epoch_7_batch_5999.pt | 0.8023333333333333 | 0.006291283407507587 |
+| Epoch_9_batch_2999.pt | 0.8021666666666667 | 0.007833333333333335 |
+| Epoch_8_batch_2999.pt | 0.8015000000000001 | 0.006570839057983835 |
+| Epoch_8_batch_5999.pt | 0.8013333333333333 | 0.00795744856435284 |
+| Epoch_6_batch_5999.pt | 0.8001666666666667 | 0.007897686799951202 |
+| Epoch_9.pt | 0.7971666666666667 | 0.008927174493894722 |
+| Epoch_5_batch_5999.pt | 0.7953333333333333 | 0.008042326301865584 |
+| Epoch_5_batch_2999.pt | 0.7938333333333334 | 0.007856938429279016 |
+| Epoch_6_batch_2999.pt | 0.7936666666666666 | 0.0081187178942807 |
+| Epoch_7_batch_2999.pt | 0.7928333333333334 | 0.008516686598512587 |
+| Epoch_6.pt | 0.7921666666666667 | 0.0050738375181408595 |
+| Epoch_4_batch_2999.pt | 0.7898333333333334 | 0.005496631965391444 |
+| Epoch_4_batch_5999.pt | 0.7896666666666666 | 0.006269660435407063 |
+| Epoch_3.pt | 0.7871666666666667 | 0.00937902177109724 |
+| Epoch_3_batch_5999.pt | 0.7851666666666667 | 0.0068423607939488125 |
+| Epoch_2.pt | 0.7841666666666666 | 0.009048054152723514 |
+| Epoch_4.pt | 0.7835 | 0.006806179109755635 |
+| Epoch_7.pt | 0.7829999999999999 | 0.006235105709599193 |
+| Epoch_3_batch_2999.pt | 0.7828333333333333 | 0.00753612698118491 |
+| Epoch_5.pt | 0.7808333333333334 | 0.007646713164636313 |
+| Epoch_2_batch_5999.pt | 0.7789999999999999 | 0.008632410933135572 |
+| Epoch_2_batch_2999.pt | 0.7759999999999999 | 0.008274607894182611 |
+| Epoch_1_batch_5999.pt | 0.7626666666666667 | 0.007106769507976613 |
+| Epoch_1_batch_2999.pt | 0.7535000000000001 | 0.008799726426724021 |
+| Epoch_1.pt | 0.7526666666666667 | 0.006877230277975012 |
+| Epoch_0.pt | 0.7073333333333334 | 0.009102299261826758 |
+| Epoch_0_batch_5999.pt | 0.7026666666666668 | 0.007790465841314707 |
+| Epoch_0_batch_2999.pt | 0.6346666666666667 | 0.00881146839684748 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..265655b2c338ca798d30df93c9eef55630fb5125
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9956666666666665 | 0.0011706281947614146 |
+| Epoch_12_batch_2999.pt | 0.9956666666666665 | 0.0010886621079036374 |
+| Epoch_12.pt | 0.9956666666666665 | 0.0010599324460188284 |
+| Epoch_16_batch_2999.pt | 0.9956666666666665 | 0.0011706281947614146 |
+| Epoch_15_batch_5999.pt | 0.9955 | 0.0011399046960379555 |
+| Epoch_15_batch_2999.pt | 0.9955 | 0.0011399046960379555 |
+| Epoch_11_batch_5999.pt | 0.9955 | 0.0011666666666666642 |
+| Epoch_16.pt | 0.9953333333333333 | 0.001105541596785135 |
+| Epoch_13.pt | 0.9953333333333332 | 0.001133115447465063 |
+| Epoch_14_batch_2999.pt | 0.9951666666666666 | 0.0010957268290731133 |
+| Epoch_10.pt | 0.9951666666666666 | 0.0011772011166898393 |
+| Epoch_17_batch_2999.pt | 0.9951666666666666 | 0.0010378634273483006 |
+| Epoch_13_batch_2999.pt | 0.9949999999999999 | 0.0011111111111111128 |
+| Epoch_17_batch_5999.pt | 0.9949999999999999 | 0.0011111111111111128 |
+| Epoch_13_batch_5999.pt | 0.9948333333333332 | 0.001228519132638667 |
+| Epoch_17.pt | 0.9948333333333332 | 0.0011235415786753737 |
+| Epoch_14_batch_5999.pt | 0.9948333333333332 | 0.0010671873729054782 |
+| Epoch_12_batch_5999.pt | 0.9946666666666667 | 0.001018350154434636 |
+| Epoch_14.pt | 0.9945 | 0.0011399046960379542 |
+| Epoch_16_batch_5999.pt | 0.9944999999999998 | 0.0011666666666666696 |
+| Epoch_11_batch_2999.pt | 0.9943333333333333 | 0.0012222222222222198 |
+| Epoch_10_batch_5999.pt | 0.9941666666666666 | 0.0010613873985857113 |
+| Epoch_11.pt | 0.9940000000000001 | 0.0013193713430042124 |
+| Epoch_10_batch_2999.pt | 0.994 | 0.0012957670877434026 |
+| Epoch_7_batch_5999.pt | 0.9935 | 0.0013017082793177735 |
+| Epoch_8.pt | 0.9931666666666666 | 0.001301708279317778 |
+| Epoch_9_batch_2999.pt | 0.9931666666666666 | 0.0012533904636309486 |
+| Epoch_9_batch_5999.pt | 0.993 | 0.0011600340565456162 |
+| Epoch_5_batch_5999.pt | 0.993 | 0.0017177360926378114 |
+| Epoch_8_batch_5999.pt | 0.9926666666666666 | 0.0013653561919382778 |
+| Epoch_6_batch_5999.pt | 0.9924999999999999 | 0.0017078251276599358 |
+| Epoch_9.pt | 0.9923333333333332 | 0.0007934920476158749 |
+| Epoch_4_batch_5999.pt | 0.9921666666666666 | 0.0016859989894992859 |
+| Epoch_3.pt | 0.992 | 0.0017533037597843894 |
+| Epoch_3_batch_2999.pt | 0.9918333333333335 | 0.0012031337682059848 |
+| Epoch_8_batch_2999.pt | 0.9918333333333333 | 0.0016187558093703858 |
+| Epoch_6_batch_2999.pt | 0.9916666666666668 | 0.0015908690070307063 |
+| Epoch_5_batch_2999.pt | 0.9915 | 0.001479280772854929 |
+| Epoch_2_batch_2999.pt | 0.9915 | 0.0018333333333333322 |
+| Epoch_2_batch_5999.pt | 0.9913333333333334 | 0.0017177360926378127 |
+| Epoch_6.pt | 0.9911666666666665 | 0.0012435016269777464 |
+| Epoch_4_batch_2999.pt | 0.991 | 0.001742709682373126 |
+| Epoch_3_batch_5999.pt | 0.9909999999999999 | 0.0020964402515681346 |
+| Epoch_5.pt | 0.9908333333333333 | 0.001614937983749851 |
+| Epoch_4.pt | 0.9908333333333333 | 0.0015762512176790149 |
+| Epoch_7.pt | 0.9906666666666668 | 0.0017603310575283158 |
+| Epoch_7_batch_2999.pt | 0.9905000000000002 | 0.0017751717009633857 |
+| Epoch_2.pt | 0.9884999999999999 | 0.0016749792701868215 |
+| Epoch_1_batch_5999.pt | 0.9864999999999998 | 0.0015204369092671074 |
+| Epoch_1.pt | 0.9841666666666666 | 0.002620550314460163 |
+| Epoch_1_batch_2999.pt | 0.9823333333333334 | 0.0018954135676924433 |
+| Epoch_0.pt | 0.9703333333333333 | 0.0022879178091082257 |
+| Epoch_0_batch_5999.pt | 0.9655000000000001 | 0.0030434102055116375 |
+| Epoch_0_batch_2999.pt | 0.9373333333333334 | 0.003390254733864971 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..fdd509ff22c09c8d778b6039b9ea42abb155b89c
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17.pt | 0.8873333333333333 | 0.005117146195408648 |
+| Epoch_17_batch_2999.pt | 0.8868333333333334 | 0.004292183535937484 |
+| Epoch_16_batch_2999.pt | 0.8868333333333333 | 0.004657729537494039 |
+| Epoch_16_batch_5999.pt | 0.8866666666666667 | 0.00450651106459406 |
+| Epoch_14_batch_2999.pt | 0.8859999999999999 | 0.0049140765305546565 |
+| Epoch_15_batch_5999.pt | 0.8859999999999999 | 0.004622809791714378 |
+| Epoch_15_batch_2999.pt | 0.8844999999999998 | 0.0044655747698019105 |
+| Epoch_14_batch_5999.pt | 0.8836666666666666 | 0.0055154105093292724 |
+| Epoch_13.pt | 0.8835 | 0.004965622559901967 |
+| Epoch_13_batch_2999.pt | 0.8835 | 0.004703885969055426 |
+| Epoch_17_batch_5999.pt | 0.8831666666666667 | 0.004377622670687092 |
+| Epoch_12_batch_2999.pt | 0.8825 | 0.005617707893376856 |
+| Epoch_13_batch_5999.pt | 0.8818333333333334 | 0.005722222222222222 |
+| Epoch_16.pt | 0.8818333333333334 | 0.004677566635509283 |
+| Epoch_14.pt | 0.8815 | 0.0056023033182018836 |
+| Epoch_15.pt | 0.8813333333333334 | 0.004955356249106172 |
+| Epoch_12_batch_5999.pt | 0.8805 | 0.005443594037892865 |
+| Epoch_11_batch_5999.pt | 0.8796666666666667 | 0.005315943872846971 |
+| Epoch_12.pt | 0.8796666666666665 | 0.0045119867787215395 |
+| Epoch_11.pt | 0.8786666666666667 | 0.004936635531449571 |
+| Epoch_10.pt | 0.8785000000000001 | 0.00465109837021135 |
+| Epoch_11_batch_2999.pt | 0.8781666666666668 | 0.004577521567803227 |
+| Epoch_10_batch_2999.pt | 0.876 | 0.004375859703892269 |
+| Epoch_10_batch_5999.pt | 0.8758333333333332 | 0.005573581865803604 |
+| Epoch_9_batch_5999.pt | 0.8578333333333333 | 0.0053115876122834475 |
+| Epoch_9_batch_2999.pt | 0.8526666666666667 | 0.006522515609528517 |
+| Epoch_8_batch_2999.pt | 0.8496666666666666 | 0.004841946348777979 |
+| Epoch_7_batch_5999.pt | 0.8451666666666666 | 0.004530077807088045 |
+| Epoch_6_batch_5999.pt | 0.8445 | 0.007248030809625737 |
+| Epoch_4_batch_5999.pt | 0.8413333333333334 | 0.007070194786636901 |
+| Epoch_7_batch_2999.pt | 0.8393333333333333 | 0.0061824123303304696 |
+| Epoch_8_batch_5999.pt | 0.8391666666666667 | 0.0062274283775512725 |
+| Epoch_9.pt | 0.8381666666666666 | 0.003986474044646718 |
+| Epoch_5_batch_2999.pt | 0.8370000000000001 | 0.006807766080906841 |
+| Epoch_6_batch_2999.pt | 0.8348333333333333 | 0.0043493294502332976 |
+| Epoch_4_batch_2999.pt | 0.8328333333333333 | 0.006206577554358889 |
+| Epoch_5.pt | 0.8301666666666666 | 0.006302311941063307 |
+| Epoch_4.pt | 0.8298333333333334 | 0.006336500182900745 |
+| Epoch_6.pt | 0.8283333333333334 | 0.005746711351549225 |
+| Epoch_5_batch_5999.pt | 0.8283333333333334 | 0.00569275042553311 |
+| Epoch_3.pt | 0.8258333333333334 | 0.0055235187389253235 |
+| Epoch_8.pt | 0.8256666666666665 | 0.0067868791351446455 |
+| Epoch_7.pt | 0.8245000000000001 | 0.006554849302822761 |
+| Epoch_3_batch_2999.pt | 0.8243333333333334 | 0.00799768485019027 |
+| Epoch_3_batch_5999.pt | 0.8193333333333334 | 0.005726266100766724 |
+| Epoch_2_batch_5999.pt | 0.8168333333333335 | 0.005738380637922078 |
+| Epoch_2_batch_2999.pt | 0.8146666666666667 | 0.0063089198073681685 |
+| Epoch_2.pt | 0.7996666666666666 | 0.005788451678947156 |
+| Epoch_1_batch_5999.pt | 0.7956666666666667 | 0.0048317366455908695 |
+| Epoch_1.pt | 0.7776666666666666 | 0.00639347661369591 |
+| Epoch_1_batch_2999.pt | 0.7581666666666667 | 0.006471447259279004 |
+| Epoch_0.pt | 0.7028333333333332 | 0.005905897868201724 |
+| Epoch_0_batch_5999.pt | 0.7021666666666667 | 0.00636856671279407 |
+| Epoch_0_batch_2999.pt | 0.6210000000000001 | 0.008047697316547477 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..661f912477becb9158263e521e0795d01aca6912
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,57 @@
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.8801666666666665 | 0.004248819734444195 |
+| Epoch_16_batch_2999.pt | 0.8779999999999999 | 0.004653420353638202 |
+| Epoch_14_batch_5999.pt | 0.8775000000000001 | 0.0038908725099762597 |
+| Epoch_17_batch_5999.pt | 0.8771666666666667 | 0.004635144863025806 |
+| Epoch_15.pt | 0.8768333333333335 | 0.004953175811263309 |
+| Epoch_17_batch_2999.pt | 0.8765000000000001 | 0.004160736520745062 |
+| Epoch_15_batch_2999.pt | 0.8761666666666666 | 0.0035836563161933607 |
+| Epoch_12_batch_2999.pt | 0.8756666666666666 | 0.005056471223547888 |
+| Epoch_11_batch_5999.pt | 0.8755 | 0.0038574122970755527 |
+| Epoch_13_batch_2999.pt | 0.8753333333333334 | 0.00437303747870098 |
+| Epoch_16_batch_5999.pt | 0.8753333333333334 | 0.004552844721899529 |
+| Epoch_15_batch_5999.pt | 0.8751666666666666 | 0.004447568346583431 |
+| Epoch_13_batch_5999.pt | 0.8745 | 0.0035228530805528203 |
+| Epoch_13.pt | 0.8741666666666668 | 0.004596361953749566 |
+| Epoch_10_batch_5999.pt | 0.8741666666666668 | 0.0031549490810001577 |
+| Epoch_12_batch_5999.pt | 0.8736666666666668 | 0.003440858348267117 |
+| Epoch_17.pt | 0.8728333333333333 | 0.0047012606662311185 |
+| Epoch_11.pt | 0.8718333333333333 | 0.003955383891406126 |
+| Epoch_14.pt | 0.8718333333333333 | 0.004996603784843862 |
+| Epoch_14_batch_2999.pt | 0.8709999999999999 | 0.004417976744904675 |
+| Epoch_10.pt | 0.8706666666666667 | 0.004361730316975672 |
+| Epoch_11_batch_2999.pt | 0.8698333333333335 | 0.0038204291659898184 |
+| Epoch_10_batch_2999.pt | 0.8676666666666666 | 0.004121608220220315 |
+| Epoch_12.pt | 0.8655000000000002 | 0.003991116679035955 |
+| Epoch_9_batch_5999.pt | 0.849 | 0.004863570806275395 |
+| Epoch_8_batch_2999.pt | 0.8474999999999999 | 0.004844813951249549 |
+| Epoch_9_batch_2999.pt | 0.8471666666666667 | 0.0032303537244681383 |
+| Epoch_6_batch_5999.pt | 0.8446666666666667 | 0.00385541146064388 |
+| Epoch_4_batch_5999.pt | 0.8445 | 0.004759979769642659 |
+| Epoch_8.pt | 0.8425 | 0.005742681870148159 |
+| Epoch_5_batch_5999.pt | 0.8401666666666667 | 0.004509591971906815 |
+| Epoch_8_batch_5999.pt | 0.8398333333333332 | 0.004801298692624985 |
+| Epoch_4_batch_2999.pt | 0.8386666666666667 | 0.003218388523991166 |
+| Epoch_7_batch_2999.pt | 0.8383333333333333 | 0.005773502691896265 |
+| Epoch_6_batch_2999.pt | 0.8380000000000001 | 0.0038554114606438802 |
+| Epoch_6.pt | 0.8380000000000001 | 0.006175419214040115 |
+| Epoch_5_batch_2999.pt | 0.8346666666666668 | 0.0043871304494221405 |
+| Epoch_7_batch_5999.pt | 0.8341666666666667 | 0.004181455237263045 |
+| Epoch_3_batch_2999.pt | 0.8300000000000001 | 0.0031720227608044898 |
+| Epoch_3.pt | 0.8288333333333332 | 0.0050126383482031215 |
+| Epoch_4.pt | 0.8286666666666667 | 0.004930379496074179 |
+| Epoch_3_batch_5999.pt | 0.825 | 0.005085685551709239 |
+| Epoch_2_batch_2999.pt | 0.8246666666666667 | 0.0039031484600556194 |
+| Epoch_2_batch_5999.pt | 0.8236666666666667 | 0.0039346513799168375 |
+| Epoch_9.pt | 0.8233333333333335 | 0.004303314829119351 |
+| Epoch_7.pt | 0.8216666666666667 | 0.004164814403109187 |
+| Epoch_2.pt | 0.8099999999999999 | 0.005097808776424204 |
+| Epoch_5.pt | 0.8035 | 0.00479486608162738 |
+| Epoch_1.pt | 0.7941666666666667 | 0.006257094738606828 |
+| Epoch_1_batch_5999.pt | 0.7921666666666666 | 0.0047729303574985185 |
+| Epoch_1_batch_2999.pt | 0.7901666666666666 | 0.005342873411966292 |
+| Epoch_0.pt | 0.7403333333333333 | 0.005362190449783839 |
+| Epoch_0_batch_5999.pt | 0.7344999999999999 | 0.0041725883845834194 |
+| Epoch_0_batch_2999.pt | 0.6815 | 0.0070072713556599085 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..49ffb997aec3b5089eed91009926c5e6aa036ea4
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_5999.pt | 0.9570000000000001 | 0.00356422554052122 |
+| Epoch_16_batch_2999.pt | 0.9558333333333333 | 0.0036111111111111153 |
+| Epoch_14_batch_5999.pt | 0.9558333333333332 | 0.003326381639982827 |
+| Epoch_17_batch_2999.pt | 0.9551666666666667 | 0.0035000000000000005 |
+| Epoch_17.pt | 0.9551666666666666 | 0.00387656778320434 |
+| Epoch_14.pt | 0.9548333333333334 | 0.0031861002130977017 |
+| Epoch_17_batch_5999.pt | 0.9546666666666667 | 0.003546864377669419 |
+| Epoch_15_batch_5999.pt | 0.9545000000000001 | 0.0041577682760150425 |
+| Epoch_13_batch_2999.pt | 0.9545 | 0.003833333333333336 |
+| Epoch_15_batch_2999.pt | 0.9543333333333333 | 0.0035676876351116303 |
+| Epoch_13.pt | 0.9538333333333332 | 0.003735465660892057 |
+| Epoch_13_batch_5999.pt | 0.9533333333333334 | 0.0036767538017276236 |
+| Epoch_14_batch_2999.pt | 0.9531666666666666 | 0.0034645470728153086 |
+| Epoch_15.pt | 0.9523333333333334 | 0.0037531879453454498 |
+| Epoch_12_batch_2999.pt | 0.9523333333333331 | 0.0033993463423951887 |
+| Epoch_16.pt | 0.9521666666666666 | 0.003702268240343592 |
+| Epoch_11_batch_5999.pt | 0.9521666666666666 | 0.004127968441835539 |
+| Epoch_12_batch_5999.pt | 0.9516666666666665 | 0.003591828861165471 |
+| Epoch_12.pt | 0.9515 | 0.0034911705207477744 |
+| Epoch_11_batch_2999.pt | 0.9513333333333331 | 0.00288461221905493 |
+| Epoch_11.pt | 0.95 | 0.0034066021592790915 |
+| Epoch_10_batch_5999.pt | 0.9496666666666667 | 0.003879353388910243 |
+| Epoch_10.pt | 0.9488333333333335 | 0.0034964708839021266 |
+| Epoch_10_batch_2999.pt | 0.9481666666666667 | 0.003621352997273837 |
+| Epoch_8_batch_5999.pt | 0.9299999999999999 | 0.004201704533597566 |
+| Epoch_9_batch_5999.pt | 0.9295 | 0.004423910900359653 |
+| Epoch_9.pt | 0.9289999999999999 | 0.003154459903684086 |
+| Epoch_7_batch_5999.pt | 0.9263333333333333 | 0.0028087165910587845 |
+| Epoch_8_batch_2999.pt | 0.9256666666666666 | 0.005672110674711205 |
+| Epoch_9_batch_2999.pt | 0.9248333333333333 | 0.0038526085439079573 |
+| Epoch_6_batch_5999.pt | 0.9244999999999999 | 0.004245913067618998 |
+| Epoch_8.pt | 0.9225 | 0.002952818281315182 |
+| Epoch_5_batch_5999.pt | 0.9206666666666667 | 0.0039999999999999975 |
+| Epoch_6_batch_2999.pt | 0.9203333333333333 | 0.003208784239598597 |
+| Epoch_7_batch_2999.pt | 0.9196666666666667 | 0.004096068575814838 |
+| Epoch_5_batch_2999.pt | 0.9141666666666668 | 0.0031155721984022153 |
+| Epoch_4.pt | 0.9126666666666667 | 0.0015947444549341489 |
+| Epoch_6.pt | 0.9120000000000001 | 0.0028414915227876533 |
+| Epoch_4_batch_2999.pt | 0.9114999999999999 | 0.003787976429717954 |
+| Epoch_4_batch_5999.pt | 0.9106666666666667 | 0.004812535072823938 |
+| Epoch_7.pt | 0.9095000000000001 | 0.0029085866917129815 |
+| Epoch_3_batch_2999.pt | 0.9081666666666666 | 0.0029022128636908826 |
+| Epoch_3_batch_5999.pt | 0.9065000000000001 | 0.003057575090181559 |
+| Epoch_2_batch_2999.pt | 0.9025000000000001 | 0.0031353224576716543 |
+| Epoch_2_batch_5999.pt | 0.9010000000000001 | 0.003014368881389011 |
+| Epoch_3.pt | 0.8998333333333333 | 0.0016933056282364609 |
+| Epoch_5.pt | 0.8995000000000001 | 0.003668770440244237 |
+| Epoch_2.pt | 0.8978333333333332 | 0.003960062974932603 |
+| Epoch_1_batch_5999.pt | 0.8835000000000001 | 0.003621352997273838 |
+| Epoch_1.pt | 0.8811666666666668 | 0.0033059056770692987 |
+| Epoch_1_batch_2999.pt | 0.8671666666666666 | 0.004339383241150778 |
+| Epoch_0.pt | 0.8361666666666666 | 0.004289306254879478 |
+| Epoch_0_batch_5999.pt | 0.8266666666666668 | 0.006211299937499414 |
+| Epoch_0_batch_2999.pt | 0.7578333333333334 | 0.005931970297037765 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..b2ac0c099ba21fef0489cd0b0d150457751a9e66
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_5999.pt | 0.9085000000000001 | 0.003852608543907959 |
+| Epoch_16.pt | 0.9075 | 0.003930334701294229 |
+| Epoch_17.pt | 0.9066666666666666 | 0.003397529966982369 |
+| Epoch_17_batch_2999.pt | 0.9066666666666666 | 0.003659926464889028 |
+| Epoch_13_batch_5999.pt | 0.9065 | 0.003446683859832462 |
+| Epoch_12_batch_5999.pt | 0.9061666666666668 | 0.00407529443002074 |
+| Epoch_13.pt | 0.906 | 0.004046031434674032 |
+| Epoch_16_batch_5999.pt | 0.9059999999999999 | 0.003761402417735721 |
+| Epoch_12_batch_2999.pt | 0.9058333333333334 | 0.0037781862524265083 |
+| Epoch_17_batch_5999.pt | 0.9056666666666666 | 0.0036784323016104117 |
+| Epoch_15_batch_2999.pt | 0.9053333333333334 | 0.0034676636742949378 |
+| Epoch_16_batch_2999.pt | 0.9053333333333333 | 0.003839367231815781 |
+| Epoch_10.pt | 0.9045 | 0.003549039167485427 |
+| Epoch_14_batch_5999.pt | 0.9043333333333334 | 0.00375318794534545 |
+| Epoch_13_batch_2999.pt | 0.9041666666666666 | 0.0033724556023947074 |
+| Epoch_15.pt | 0.9033333333333333 | 0.003583225665910463 |
+| Epoch_14_batch_2999.pt | 0.9030000000000001 | 0.00391104797928845 |
+| Epoch_10_batch_2999.pt | 0.9024999999999999 | 0.003193840522623292 |
+| Epoch_14.pt | 0.9024999999999999 | 0.003859012219291616 |
+| Epoch_12.pt | 0.9021666666666667 | 0.004668319813010156 |
+| Epoch_10_batch_5999.pt | 0.9019999999999999 | 0.004192880503136267 |
+| Epoch_11_batch_5999.pt | 0.8996666666666668 | 0.004004626953542241 |
+| Epoch_11_batch_2999.pt | 0.899 | 0.0038984010650775177 |
+| Epoch_11.pt | 0.8986666666666666 | 0.004344714399114921 |
+| Epoch_9.pt | 0.8821666666666668 | 0.0052531589555567075 |
+| Epoch_9_batch_5999.pt | 0.8811666666666668 | 0.004981756842176049 |
+| Epoch_8_batch_2999.pt | 0.8796666666666665 | 0.004103596736137634 |
+| Epoch_7_batch_5999.pt | 0.8793333333333333 | 0.0050442486501405155 |
+| Epoch_7_batch_2999.pt | 0.8791666666666668 | 0.005710343657903436 |
+| Epoch_9_batch_2999.pt | 0.8789999999999999 | 0.003610683735393763 |
+| Epoch_6_batch_2999.pt | 0.8786666666666667 | 0.004148478822798768 |
+| Epoch_5_batch_5999.pt | 0.8785000000000001 | 0.004890466917040397 |
+| Epoch_6_batch_5999.pt | 0.8768333333333335 | 0.005196449405098938 |
+| Epoch_8_batch_5999.pt | 0.8756666666666666 | 0.003636237371545237 |
+| Epoch_8.pt | 0.8733333333333333 | 0.004694362260950578 |
+| Epoch_6.pt | 0.8720000000000001 | 0.005162782291328417 |
+| Epoch_4_batch_2999.pt | 0.8716666666666667 | 0.00477260702109212 |
+| Epoch_5_batch_2999.pt | 0.8705 | 0.00606574067007016 |
+| Epoch_4_batch_5999.pt | 0.8680000000000001 | 0.004816381511487389 |
+| Epoch_4.pt | 0.867 | 0.0035294178165041355 |
+| Epoch_5.pt | 0.8666666666666666 | 0.005594309277855158 |
+| Epoch_3_batch_5999.pt | 0.8621666666666666 | 0.005079916884709384 |
+| Epoch_2_batch_5999.pt | 0.8618333333333335 | 0.0044890126495590555 |
+| Epoch_3_batch_2999.pt | 0.8616666666666667 | 0.004296136650929155 |
+| Epoch_3.pt | 0.8536666666666666 | 0.004294699575575041 |
+| Epoch_2_batch_2999.pt | 0.8508333333333333 | 0.0049705925324307915 |
+| Epoch_2.pt | 0.8498333333333333 | 0.006238322424070969 |
+| Epoch_7.pt | 0.8488333333333333 | 0.00534633831078181 |
+| Epoch_1_batch_5999.pt | 0.8373333333333333 | 0.004528374193847736 |
+| Epoch_1.pt | 0.8318333333333333 | 0.005377421934967231 |
+| Epoch_1_batch_2999.pt | 0.8140000000000001 | 0.005495228008390329 |
+| Epoch_0_batch_5999.pt | 0.7881666666666666 | 0.004370566536714872 |
+| Epoch_0.pt | 0.7849999999999999 | 0.00481125224324688 |
+| Epoch_0_batch_2999.pt | 0.7068333333333332 | 0.005278947238855226 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MV-Softmax/log.log b/bob/bio/facexzoo/models/heads/MV-Softmax/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..d38e0a8f66c6c904fb3541adc812d096a3bddfdb
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MV-Softmax/log.log
@@ -0,0 +1,657 @@
+INFO 2020-11-24 17:02:43 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2020-11-24 17:02:43 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2020-11-24 17:02:43 train.py: 172] Start optimization.
+INFO 2020-11-24 17:02:43 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/home/wangjun492/wj_data/facex-zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='mv-softmax', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mv-mobile', train_file='/home/wangjun492/wj_data/facex-zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7fa027e56b00>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'is_am': 1, 'margin': 0.35, 'mv_weight': 1.12, 'scale': 32}
+INFO 2020-11-24 17:03:10 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 16.358618
+INFO 2020-11-24 17:04:32 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 15.827444
+INFO 2020-11-24 17:05:55 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 15.327547
+INFO 2020-11-24 17:07:17 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 15.027858
+INFO 2020-11-24 17:08:39 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 14.677421
+INFO 2020-11-24 17:10:01 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 14.355350
+INFO 2020-11-24 17:11:23 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 14.027037
+INFO 2020-11-24 17:12:45 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 13.690001
+INFO 2020-11-24 17:14:07 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 13.346130
+INFO 2020-11-24 17:15:29 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 13.024926
+INFO 2020-11-24 17:16:51 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 12.701884
+INFO 2020-11-24 17:18:13 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 12.395277
+INFO 2020-11-24 17:19:35 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 12.130202
+INFO 2020-11-24 17:20:57 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 11.987308
+INFO 2020-11-24 17:22:18 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 11.955337
+INFO 2020-11-24 17:23:39 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-24 17:23:39 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 12.067932
+INFO 2020-11-24 17:25:00 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 12.314615
+INFO 2020-11-24 17:26:21 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 12.647381
+INFO 2020-11-24 17:27:42 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 13.030897
+INFO 2020-11-24 17:29:02 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 13.441236
+INFO 2020-11-24 17:30:22 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 13.796649
+INFO 2020-11-24 17:31:41 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 14.154292
+INFO 2020-11-24 17:33:00 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 14.411892
+INFO 2020-11-24 17:34:19 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 14.619047
+INFO 2020-11-24 17:35:38 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 14.832919
+INFO 2020-11-24 17:36:56 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 14.915812
+INFO 2020-11-24 17:38:14 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 14.976967
+INFO 2020-11-24 17:39:32 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 14.973729
+INFO 2020-11-24 17:40:50 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 14.912079
+INFO 2020-11-24 17:42:08 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 14.853437
+INFO 2020-11-24 17:43:25 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-24 17:43:26 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 14.786114
+INFO 2020-11-24 17:44:43 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 14.640276
+INFO 2020-11-24 17:46:00 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 14.499856
+INFO 2020-11-24 17:46:07 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-24 17:46:08 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 14.468562
+INFO 2020-11-24 17:47:26 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 14.190097
+INFO 2020-11-24 17:48:43 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 14.034737
+INFO 2020-11-24 17:50:00 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 13.867991
+INFO 2020-11-24 17:51:17 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 13.711440
+INFO 2020-11-24 17:52:35 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 13.581579
+INFO 2020-11-24 17:53:51 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 13.396562
+INFO 2020-11-24 17:55:08 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 13.189663
+INFO 2020-11-24 17:56:25 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 13.039445
+INFO 2020-11-24 17:57:42 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 12.903281
+INFO 2020-11-24 17:58:59 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 12.729310
+INFO 2020-11-24 18:00:15 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 12.598813
+INFO 2020-11-24 18:01:32 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 12.446454
+INFO 2020-11-24 18:02:49 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 12.294402
+INFO 2020-11-24 18:04:06 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 12.151741
+INFO 2020-11-24 18:05:22 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-24 18:05:23 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 12.006855
+INFO 2020-11-24 18:06:39 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 11.907829
+INFO 2020-11-24 18:07:55 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 11.759257
+INFO 2020-11-24 18:09:11 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 11.656780
+INFO 2020-11-24 18:10:27 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 11.537830
+INFO 2020-11-24 18:11:43 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 11.460143
+INFO 2020-11-24 18:13:00 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 11.327951
+INFO 2020-11-24 18:14:16 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 11.268308
+INFO 2020-11-24 18:15:32 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 11.173050
+INFO 2020-11-24 18:16:48 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 11.074287
+INFO 2020-11-24 18:18:04 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 10.989137
+INFO 2020-11-24 18:19:21 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 10.935530
+INFO 2020-11-24 18:20:37 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 10.858660
+INFO 2020-11-24 18:21:53 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 10.765018
+INFO 2020-11-24 18:23:10 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 10.696050
+INFO 2020-11-24 18:24:26 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-24 18:24:26 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 10.650471
+INFO 2020-11-24 18:25:43 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 10.606789
+INFO 2020-11-24 18:27:00 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 10.507497
+INFO 2020-11-24 18:27:06 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-24 18:27:08 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 10.401783
+INFO 2020-11-24 18:28:25 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 9.958792
+INFO 2020-11-24 18:29:42 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 9.970688
+INFO 2020-11-24 18:30:58 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 10.018418
+INFO 2020-11-24 18:32:15 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 10.043185
+INFO 2020-11-24 18:33:32 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 10.026329
+INFO 2020-11-24 18:34:48 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 10.008143
+INFO 2020-11-24 18:36:05 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 10.019662
+INFO 2020-11-24 18:37:22 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 9.995654
+INFO 2020-11-24 18:38:38 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 9.985731
+INFO 2020-11-24 18:39:55 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 9.926832
+INFO 2020-11-24 18:41:11 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 9.920456
+INFO 2020-11-24 18:42:28 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 9.864497
+INFO 2020-11-24 18:43:45 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 9.836358
+INFO 2020-11-24 18:45:01 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 9.801993
+INFO 2020-11-24 18:46:18 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-24 18:46:18 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 9.784279
+INFO 2020-11-24 18:47:35 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 9.777114
+INFO 2020-11-24 18:48:51 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 9.716017
+INFO 2020-11-24 18:50:08 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 9.685481
+INFO 2020-11-24 18:51:25 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 9.638493
+INFO 2020-11-24 18:52:42 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 9.615591
+INFO 2020-11-24 18:53:58 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 9.573484
+INFO 2020-11-24 18:55:15 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 9.557758
+INFO 2020-11-24 18:56:32 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 9.530840
+INFO 2020-11-24 18:57:49 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 9.470006
+INFO 2020-11-24 18:59:06 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 9.479366
+INFO 2020-11-24 19:00:23 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 9.434593
+INFO 2020-11-24 19:01:40 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 9.401267
+INFO 2020-11-24 19:02:56 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 9.366447
+INFO 2020-11-24 19:04:13 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 9.358310
+INFO 2020-11-24 19:05:30 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-24 19:05:31 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 9.291458
+INFO 2020-11-24 19:06:47 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 9.317040
+INFO 2020-11-24 19:08:03 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 9.263851
+INFO 2020-11-24 19:08:09 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-24 19:08:11 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 9.158338
+INFO 2020-11-24 19:09:27 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 8.738770
+INFO 2020-11-24 19:10:43 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 8.747253
+INFO 2020-11-24 19:12:00 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 8.834588
+INFO 2020-11-24 19:13:16 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 8.904287
+INFO 2020-11-24 19:14:33 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 8.929899
+INFO 2020-11-24 19:15:50 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 8.972742
+INFO 2020-11-24 19:17:06 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 8.967419
+INFO 2020-11-24 19:18:23 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 8.950048
+INFO 2020-11-24 19:19:39 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 9.018943
+INFO 2020-11-24 19:20:56 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 8.952868
+INFO 2020-11-24 19:22:12 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 8.942784
+INFO 2020-11-24 19:23:29 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 8.956252
+INFO 2020-11-24 19:24:45 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 8.937395
+INFO 2020-11-24 19:26:02 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 8.942570
+INFO 2020-11-24 19:27:18 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-24 19:27:19 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 8.916036
+INFO 2020-11-24 19:28:35 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 8.912676
+INFO 2020-11-24 19:29:52 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 8.899460
+INFO 2020-11-24 19:31:09 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 8.846656
+INFO 2020-11-24 19:32:25 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 8.852470
+INFO 2020-11-24 19:33:42 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 8.813011
+INFO 2020-11-24 19:34:59 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 8.826520
+INFO 2020-11-24 19:36:15 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 8.782351
+INFO 2020-11-24 19:37:32 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 8.789090
+INFO 2020-11-24 19:38:49 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 8.784844
+INFO 2020-11-24 19:40:06 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 8.752145
+INFO 2020-11-24 19:41:22 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 8.755990
+INFO 2020-11-24 19:42:39 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 8.748737
+INFO 2020-11-24 19:43:56 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 8.742908
+INFO 2020-11-24 19:45:13 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 8.702594
+INFO 2020-11-24 19:46:30 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-24 19:46:30 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 8.683531
+INFO 2020-11-24 19:47:47 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 8.677554
+INFO 2020-11-24 19:49:04 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 8.687340
+INFO 2020-11-24 19:49:10 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-24 19:49:12 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 8.645050
+INFO 2020-11-24 19:50:29 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 8.138176
+INFO 2020-11-24 19:51:45 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 8.133244
+INFO 2020-11-24 19:53:02 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 8.266590
+INFO 2020-11-24 19:54:18 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 8.264567
+INFO 2020-11-24 19:55:35 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 8.324405
+INFO 2020-11-24 19:56:52 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 8.387397
+INFO 2020-11-24 19:58:08 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 8.381465
+INFO 2020-11-24 19:59:25 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 8.419101
+INFO 2020-11-24 20:00:41 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 8.418929
+INFO 2020-11-24 20:01:58 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 8.439405
+INFO 2020-11-24 20:03:14 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 8.440718
+INFO 2020-11-24 20:04:31 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 8.468213
+INFO 2020-11-24 20:05:47 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 8.412492
+INFO 2020-11-24 20:07:04 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 8.401329
+INFO 2020-11-24 20:08:20 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-24 20:08:20 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 8.462322
+INFO 2020-11-24 20:09:36 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 8.428143
+INFO 2020-11-24 20:10:52 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 8.421280
+INFO 2020-11-24 20:12:09 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 8.391416
+INFO 2020-11-24 20:13:25 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 8.353408
+INFO 2020-11-24 20:14:42 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 8.400384
+INFO 2020-11-24 20:15:59 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 8.367021
+INFO 2020-11-24 20:17:15 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 8.395899
+INFO 2020-11-24 20:18:32 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 8.344757
+INFO 2020-11-24 20:19:49 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 8.359471
+INFO 2020-11-24 20:21:06 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 8.358082
+INFO 2020-11-24 20:22:22 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 8.314536
+INFO 2020-11-24 20:23:39 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 8.286721
+INFO 2020-11-24 20:24:56 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 8.331915
+INFO 2020-11-24 20:26:13 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 8.285538
+INFO 2020-11-24 20:27:29 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-24 20:27:30 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 8.278215
+INFO 2020-11-24 20:28:47 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 8.267316
+INFO 2020-11-24 20:30:04 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 8.247428
+INFO 2020-11-24 20:30:10 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-24 20:30:12 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 8.371536
+INFO 2020-11-24 20:31:28 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 7.767822
+INFO 2020-11-24 20:32:45 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 7.743762
+INFO 2020-11-24 20:34:02 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 7.862978
+INFO 2020-11-24 20:35:18 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 7.930466
+INFO 2020-11-24 20:36:35 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 7.993809
+INFO 2020-11-24 20:37:51 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 8.014129
+INFO 2020-11-24 20:39:08 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 8.067885
+INFO 2020-11-24 20:40:24 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 8.064815
+INFO 2020-11-24 20:41:41 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 8.070696
+INFO 2020-11-24 20:42:57 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 8.081757
+INFO 2020-11-24 20:44:14 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 8.069307
+INFO 2020-11-24 20:45:30 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 8.108222
+INFO 2020-11-24 20:46:47 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 8.123253
+INFO 2020-11-24 20:48:03 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 8.108594
+INFO 2020-11-24 20:49:20 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-24 20:49:20 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 8.108960
+INFO 2020-11-24 20:50:36 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 8.088804
+INFO 2020-11-24 20:51:52 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 8.077876
+INFO 2020-11-24 20:53:08 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 8.086027
+INFO 2020-11-24 20:54:24 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 8.072146
+INFO 2020-11-24 20:55:40 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 8.084796
+INFO 2020-11-24 20:56:56 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 8.068487
+INFO 2020-11-24 20:58:12 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 8.052808
+INFO 2020-11-24 20:59:28 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 8.054618
+INFO 2020-11-24 21:00:44 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 8.083125
+INFO 2020-11-24 21:02:00 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 8.037382
+INFO 2020-11-24 21:03:16 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 8.025006
+INFO 2020-11-24 21:04:32 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 8.020393
+INFO 2020-11-24 21:05:48 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 7.998549
+INFO 2020-11-24 21:07:04 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 7.991002
+INFO 2020-11-24 21:08:20 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-24 21:08:21 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 8.000411
+INFO 2020-11-24 21:09:38 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 7.964763
+INFO 2020-11-24 21:10:55 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 7.980003
+INFO 2020-11-24 21:11:01 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-24 21:11:03 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 7.949017
+INFO 2020-11-24 21:12:19 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 7.518950
+INFO 2020-11-24 21:13:36 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 7.480380
+INFO 2020-11-24 21:14:53 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 7.594087
+INFO 2020-11-24 21:16:09 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 7.652602
+INFO 2020-11-24 21:17:26 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 7.707925
+INFO 2020-11-24 21:18:42 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 7.781615
+INFO 2020-11-24 21:19:59 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 7.807427
+INFO 2020-11-24 21:21:15 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 7.810504
+INFO 2020-11-24 21:22:32 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 7.841157
+INFO 2020-11-24 21:23:48 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 7.837509
+INFO 2020-11-24 21:25:05 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 7.858138
+INFO 2020-11-24 21:26:21 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 7.860692
+INFO 2020-11-24 21:27:38 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 7.850410
+INFO 2020-11-24 21:28:54 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 7.864597
+INFO 2020-11-24 21:30:10 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-24 21:30:11 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 7.829093
+INFO 2020-11-24 21:31:28 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 7.873271
+INFO 2020-11-24 21:32:44 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 7.859303
+INFO 2020-11-24 21:34:01 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 7.833756
+INFO 2020-11-24 21:35:17 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 7.868141
+INFO 2020-11-24 21:36:34 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 7.832145
+INFO 2020-11-24 21:37:51 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 7.827492
+INFO 2020-11-24 21:39:07 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 7.822211
+INFO 2020-11-24 21:40:24 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 7.803100
+INFO 2020-11-24 21:41:41 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 7.839389
+INFO 2020-11-24 21:42:57 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 7.828123
+INFO 2020-11-24 21:44:14 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 7.807350
+INFO 2020-11-24 21:45:31 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 7.840542
+INFO 2020-11-24 21:46:48 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 7.790327
+INFO 2020-11-24 21:48:05 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 7.800599
+INFO 2020-11-24 21:49:21 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-24 21:49:22 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 7.794503
+INFO 2020-11-24 21:50:38 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 7.811972
+INFO 2020-11-24 21:51:55 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 7.816250
+INFO 2020-11-24 21:52:01 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-24 21:52:03 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 7.721017
+INFO 2020-11-24 21:53:20 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 7.263689
+INFO 2020-11-24 21:54:37 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 7.302127
+INFO 2020-11-24 21:55:53 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 7.388004
+INFO 2020-11-24 21:57:10 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 7.476136
+INFO 2020-11-24 21:58:26 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 7.525854
+INFO 2020-11-24 21:59:43 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 7.569437
+INFO 2020-11-24 22:00:59 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 7.606572
+INFO 2020-11-24 22:02:16 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 7.646732
+INFO 2020-11-24 22:03:32 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 7.663091
+INFO 2020-11-24 22:04:49 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 7.641256
+INFO 2020-11-24 22:06:05 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 7.667531
+INFO 2020-11-24 22:07:21 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 7.688095
+INFO 2020-11-24 22:08:38 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 7.702703
+INFO 2020-11-24 22:09:54 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 7.673153
+INFO 2020-11-24 22:11:11 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-24 22:11:11 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 7.691864
+INFO 2020-11-24 22:12:28 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 7.671905
+INFO 2020-11-24 22:13:44 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 7.709454
+INFO 2020-11-24 22:15:01 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 7.666515
+INFO 2020-11-24 22:16:18 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 7.645696
+INFO 2020-11-24 22:17:34 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 7.697766
+INFO 2020-11-24 22:18:51 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 7.661309
+INFO 2020-11-24 22:20:08 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 7.704096
+INFO 2020-11-24 22:21:24 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 7.703111
+INFO 2020-11-24 22:22:41 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 7.686914
+INFO 2020-11-24 22:23:58 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 7.652473
+INFO 2020-11-24 22:25:15 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 7.657636
+INFO 2020-11-24 22:26:31 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 7.659919
+INFO 2020-11-24 22:27:48 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 7.633489
+INFO 2020-11-24 22:29:05 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 7.652294
+INFO 2020-11-24 22:30:22 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-24 22:30:22 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 7.629084
+INFO 2020-11-24 22:31:39 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 7.587239
+INFO 2020-11-24 22:32:56 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 7.620722
+INFO 2020-11-24 22:33:02 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-24 22:33:04 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 7.599243
+INFO 2020-11-24 22:34:21 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 7.169748
+INFO 2020-11-24 22:35:37 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 7.161929
+INFO 2020-11-24 22:36:54 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 7.237029
+INFO 2020-11-24 22:38:10 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 7.324340
+INFO 2020-11-24 22:39:27 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 7.391021
+INFO 2020-11-24 22:40:43 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 7.419794
+INFO 2020-11-24 22:42:00 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 7.462434
+INFO 2020-11-24 22:43:16 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 7.477895
+INFO 2020-11-24 22:44:33 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 7.508406
+INFO 2020-11-24 22:45:49 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 7.522996
+INFO 2020-11-24 22:47:06 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 7.525577
+INFO 2020-11-24 22:48:22 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 7.525456
+INFO 2020-11-24 22:49:38 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 7.531138
+INFO 2020-11-24 22:50:55 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 7.542735
+INFO 2020-11-24 22:52:11 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-24 22:52:12 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 7.531534
+INFO 2020-11-24 22:53:28 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 7.529461
+INFO 2020-11-24 22:54:45 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 7.560661
+INFO 2020-11-24 22:56:01 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 7.550087
+INFO 2020-11-24 22:57:18 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 7.508024
+INFO 2020-11-24 22:58:35 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 7.549892
+INFO 2020-11-24 22:59:51 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 7.559926
+INFO 2020-11-24 23:01:08 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 7.509007
+INFO 2020-11-24 23:02:25 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 7.568892
+INFO 2020-11-24 23:03:41 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 7.538015
+INFO 2020-11-24 23:04:58 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 7.519274
+INFO 2020-11-24 23:06:15 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 7.505941
+INFO 2020-11-24 23:07:32 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 7.555411
+INFO 2020-11-24 23:08:49 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 7.507525
+INFO 2020-11-24 23:10:05 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 7.528209
+INFO 2020-11-24 23:11:22 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-24 23:11:22 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 7.515858
+INFO 2020-11-24 23:12:39 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 7.534646
+INFO 2020-11-24 23:13:56 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 7.508126
+INFO 2020-11-24 23:14:02 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-24 23:14:04 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 7.443227
+INFO 2020-11-24 23:15:20 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 7.009048
+INFO 2020-11-24 23:16:36 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 7.035446
+INFO 2020-11-24 23:17:52 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 7.069120
+INFO 2020-11-24 23:19:08 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 7.197210
+INFO 2020-11-24 23:20:24 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 7.304525
+INFO 2020-11-24 23:21:40 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 7.299744
+INFO 2020-11-24 23:22:56 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 7.324403
+INFO 2020-11-24 23:24:11 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 7.391143
+INFO 2020-11-24 23:25:27 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 7.417595
+INFO 2020-11-24 23:26:43 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 7.439957
+INFO 2020-11-24 23:27:59 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 7.406044
+INFO 2020-11-24 23:29:15 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 7.424606
+INFO 2020-11-24 23:30:31 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 7.398502
+INFO 2020-11-24 23:31:47 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 7.423678
+INFO 2020-11-24 23:33:02 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-24 23:33:03 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 7.436243
+INFO 2020-11-24 23:34:19 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 7.473204
+INFO 2020-11-24 23:35:36 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 7.443741
+INFO 2020-11-24 23:36:53 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 7.441603
+INFO 2020-11-24 23:38:09 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 7.417943
+INFO 2020-11-24 23:39:26 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 7.423790
+INFO 2020-11-24 23:40:42 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 7.446940
+INFO 2020-11-24 23:41:59 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 7.406042
+INFO 2020-11-24 23:43:16 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 7.427912
+INFO 2020-11-24 23:44:33 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 7.437229
+INFO 2020-11-24 23:45:49 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 7.434042
+INFO 2020-11-24 23:47:06 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 7.401678
+INFO 2020-11-24 23:48:23 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 7.439736
+INFO 2020-11-24 23:49:40 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 7.417320
+INFO 2020-11-24 23:50:56 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 7.431844
+INFO 2020-11-24 23:52:13 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-24 23:52:14 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 7.436903
+INFO 2020-11-24 23:53:30 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 7.398268
+INFO 2020-11-24 23:54:47 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 7.397418
+INFO 2020-11-24 23:54:53 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-24 23:54:55 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 7.353138
+INFO 2020-11-24 23:56:12 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 6.257766
+INFO 2020-11-24 23:57:29 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 6.017529
+INFO 2020-11-24 23:58:45 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 5.915038
+INFO 2020-11-25 00:00:02 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 5.836611
+INFO 2020-11-25 00:01:18 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 5.783414
+INFO 2020-11-25 00:02:34 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 5.734762
+INFO 2020-11-25 00:03:51 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 5.738091
+INFO 2020-11-25 00:05:07 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 5.682976
+INFO 2020-11-25 00:06:23 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 5.647589
+INFO 2020-11-25 00:07:40 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 5.622531
+INFO 2020-11-25 00:08:56 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 5.601900
+INFO 2020-11-25 00:10:12 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 5.571118
+INFO 2020-11-25 00:11:29 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 5.534947
+INFO 2020-11-25 00:12:45 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 5.509576
+INFO 2020-11-25 00:14:01 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-25 00:14:02 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 5.511195
+INFO 2020-11-25 00:15:17 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 5.482364
+INFO 2020-11-25 00:16:33 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 5.489383
+INFO 2020-11-25 00:17:49 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 5.448496
+INFO 2020-11-25 00:19:05 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 5.419645
+INFO 2020-11-25 00:20:21 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 5.423076
+INFO 2020-11-25 00:21:37 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 5.412503
+INFO 2020-11-25 00:22:53 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 5.375420
+INFO 2020-11-25 00:24:08 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 5.357363
+INFO 2020-11-25 00:25:24 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 5.346682
+INFO 2020-11-25 00:26:40 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 5.338702
+INFO 2020-11-25 00:27:56 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 5.338754
+INFO 2020-11-25 00:29:12 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 5.340596
+INFO 2020-11-25 00:30:28 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 5.277478
+INFO 2020-11-25 00:31:44 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 5.315326
+INFO 2020-11-25 00:33:00 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-25 00:33:01 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 5.293450
+INFO 2020-11-25 00:34:17 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 5.272774
+INFO 2020-11-25 00:35:34 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 5.283323
+INFO 2020-11-25 00:35:40 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-25 00:35:42 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 5.223241
+INFO 2020-11-25 00:36:58 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 4.948623
+INFO 2020-11-25 00:38:15 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 4.906858
+INFO 2020-11-25 00:39:31 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 4.907461
+INFO 2020-11-25 00:40:48 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 4.904690
+INFO 2020-11-25 00:42:04 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 4.928326
+INFO 2020-11-25 00:43:21 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 4.924131
+INFO 2020-11-25 00:44:37 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 4.936636
+INFO 2020-11-25 00:45:53 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 4.918629
+INFO 2020-11-25 00:47:09 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 4.919805
+INFO 2020-11-25 00:48:26 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 4.960497
+INFO 2020-11-25 00:49:42 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 4.957431
+INFO 2020-11-25 00:50:58 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 4.956089
+INFO 2020-11-25 00:52:15 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 4.945728
+INFO 2020-11-25 00:53:31 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 4.957325
+INFO 2020-11-25 00:54:47 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-25 00:54:47 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 4.980482
+INFO 2020-11-25 00:56:04 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 4.983182
+INFO 2020-11-25 00:57:20 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 4.966438
+INFO 2020-11-25 00:58:37 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 4.987589
+INFO 2020-11-25 00:59:53 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 4.968704
+INFO 2020-11-25 01:01:10 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 4.980256
+INFO 2020-11-25 01:02:26 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 4.970413
+INFO 2020-11-25 01:03:43 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 5.002899
+INFO 2020-11-25 01:04:59 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 5.007192
+INFO 2020-11-25 01:06:16 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 4.998616
+INFO 2020-11-25 01:07:33 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 5.009392
+INFO 2020-11-25 01:08:49 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 4.986240
+INFO 2020-11-25 01:10:06 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 4.972138
+INFO 2020-11-25 01:11:23 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 4.976395
+INFO 2020-11-25 01:12:39 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 4.971344
+INFO 2020-11-25 01:13:56 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-25 01:13:56 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 4.987184
+INFO 2020-11-25 01:15:12 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 4.976141
+INFO 2020-11-25 01:16:28 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 5.002322
+INFO 2020-11-25 01:16:35 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-25 01:16:36 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 4.983368
+INFO 2020-11-25 01:17:53 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 4.611268
+INFO 2020-11-25 01:19:09 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 4.640412
+INFO 2020-11-25 01:20:26 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 4.673774
+INFO 2020-11-25 01:21:42 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 4.692424
+INFO 2020-11-25 01:22:59 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 4.691139
+INFO 2020-11-25 01:24:15 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 4.690612
+INFO 2020-11-25 01:25:31 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 4.695927
+INFO 2020-11-25 01:26:47 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 4.736963
+INFO 2020-11-25 01:28:04 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 4.741829
+INFO 2020-11-25 01:29:20 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 4.736865
+INFO 2020-11-25 01:30:36 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 4.762031
+INFO 2020-11-25 01:31:53 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 4.767115
+INFO 2020-11-25 01:33:09 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 4.782443
+INFO 2020-11-25 01:34:25 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 4.760673
+INFO 2020-11-25 01:35:41 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-25 01:35:42 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 4.802496
+INFO 2020-11-25 01:36:58 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 4.796419
+INFO 2020-11-25 01:38:15 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 4.792343
+INFO 2020-11-25 01:39:31 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 4.832982
+INFO 2020-11-25 01:40:48 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 4.821779
+INFO 2020-11-25 01:42:04 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 4.881370
+INFO 2020-11-25 01:43:20 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 4.884917
+INFO 2020-11-25 01:44:37 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 4.857210
+INFO 2020-11-25 01:45:53 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 4.858853
+INFO 2020-11-25 01:47:10 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 4.859920
+INFO 2020-11-25 01:48:27 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 4.852360
+INFO 2020-11-25 01:49:43 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 4.872837
+INFO 2020-11-25 01:51:00 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 4.871405
+INFO 2020-11-25 01:52:17 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 4.908140
+INFO 2020-11-25 01:53:33 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 4.893963
+INFO 2020-11-25 01:54:50 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-25 01:54:50 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 4.906605
+INFO 2020-11-25 01:56:07 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 4.916764
+INFO 2020-11-25 01:57:23 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 4.909206
+INFO 2020-11-25 01:57:30 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-25 01:57:31 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 4.876686
+INFO 2020-11-25 01:58:47 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 4.442451
+INFO 2020-11-25 02:00:03 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 4.408205
+INFO 2020-11-25 02:01:19 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 4.419447
+INFO 2020-11-25 02:02:35 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 4.387983
+INFO 2020-11-25 02:03:51 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 4.412583
+INFO 2020-11-25 02:05:06 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 4.399296
+INFO 2020-11-25 02:06:22 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 4.394198
+INFO 2020-11-25 02:07:38 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 4.380451
+INFO 2020-11-25 02:08:53 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 4.419471
+INFO 2020-11-25 02:10:09 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 4.381618
+INFO 2020-11-25 02:11:24 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 4.409688
+INFO 2020-11-25 02:12:40 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 4.408339
+INFO 2020-11-25 02:13:56 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 4.411750
+INFO 2020-11-25 02:15:11 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 4.413968
+INFO 2020-11-25 02:16:27 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-25 02:16:27 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 4.390786
+INFO 2020-11-25 02:17:44 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 4.429628
+INFO 2020-11-25 02:19:00 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 4.420643
+INFO 2020-11-25 02:20:16 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 4.388773
+INFO 2020-11-25 02:21:33 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 4.397255
+INFO 2020-11-25 02:22:49 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 4.405686
+INFO 2020-11-25 02:24:06 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 4.395911
+INFO 2020-11-25 02:25:22 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 4.432924
+INFO 2020-11-25 02:26:39 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 4.400759
+INFO 2020-11-25 02:27:55 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 4.384009
+INFO 2020-11-25 02:29:12 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 4.438662
+INFO 2020-11-25 02:30:29 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 4.423308
+INFO 2020-11-25 02:31:45 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 4.411874
+INFO 2020-11-25 02:33:02 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 4.426057
+INFO 2020-11-25 02:34:19 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 4.420076
+INFO 2020-11-25 02:35:35 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-25 02:35:35 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 4.426545
+INFO 2020-11-25 02:36:52 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 4.424432
+INFO 2020-11-25 02:38:09 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 4.414508
+INFO 2020-11-25 02:38:15 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-25 02:38:17 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 4.443413
+INFO 2020-11-25 02:39:33 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 4.337744
+INFO 2020-11-25 02:40:50 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 4.345555
+INFO 2020-11-25 02:42:06 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 4.335841
+INFO 2020-11-25 02:43:23 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 4.349915
+INFO 2020-11-25 02:44:39 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 4.377623
+INFO 2020-11-25 02:45:55 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 4.354851
+INFO 2020-11-25 02:47:12 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 4.369637
+INFO 2020-11-25 02:48:28 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 4.347162
+INFO 2020-11-25 02:49:44 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 4.375702
+INFO 2020-11-25 02:51:00 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 4.355511
+INFO 2020-11-25 02:52:17 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 4.351502
+INFO 2020-11-25 02:53:33 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 4.346628
+INFO 2020-11-25 02:54:49 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 4.385360
+INFO 2020-11-25 02:56:06 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 4.382407
+INFO 2020-11-25 02:57:22 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-25 02:57:22 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 4.368762
+INFO 2020-11-25 02:58:38 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 4.392034
+INFO 2020-11-25 02:59:54 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 4.361917
+INFO 2020-11-25 03:01:09 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 4.373660
+INFO 2020-11-25 03:02:25 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 4.387465
+INFO 2020-11-25 03:03:41 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 4.354591
+INFO 2020-11-25 03:04:57 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 4.383426
+INFO 2020-11-25 03:06:13 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 4.364166
+INFO 2020-11-25 03:07:29 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 4.376958
+INFO 2020-11-25 03:08:44 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 4.372547
+INFO 2020-11-25 03:10:00 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 4.359061
+INFO 2020-11-25 03:11:16 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 4.379103
+INFO 2020-11-25 03:12:32 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 4.410229
+INFO 2020-11-25 03:13:48 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 4.386184
+INFO 2020-11-25 03:15:04 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 4.415015
+INFO 2020-11-25 03:16:20 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-25 03:16:20 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 4.391684
+INFO 2020-11-25 03:17:37 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 4.421325
+INFO 2020-11-25 03:18:53 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 4.391710
+INFO 2020-11-25 03:19:00 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-25 03:19:01 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 4.447200
+INFO 2020-11-25 03:20:18 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 4.308515
+INFO 2020-11-25 03:21:34 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 4.339306
+INFO 2020-11-25 03:22:51 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 4.358757
+INFO 2020-11-25 03:24:07 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 4.343400
+INFO 2020-11-25 03:25:23 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 4.327294
+INFO 2020-11-25 03:26:40 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 4.339487
+INFO 2020-11-25 03:27:56 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 4.330323
+INFO 2020-11-25 03:29:12 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 4.349277
+INFO 2020-11-25 03:30:28 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 4.335158
+INFO 2020-11-25 03:31:45 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 4.341329
+INFO 2020-11-25 03:33:01 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 4.328359
+INFO 2020-11-25 03:34:17 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 4.342991
+INFO 2020-11-25 03:35:34 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 4.365566
+INFO 2020-11-25 03:36:50 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 4.348621
+INFO 2020-11-25 03:38:06 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-25 03:38:06 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 4.341549
+INFO 2020-11-25 03:39:23 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 4.353372
+INFO 2020-11-25 03:40:39 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 4.341762
+INFO 2020-11-25 03:41:56 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 4.357957
+INFO 2020-11-25 03:43:12 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 4.356401
+INFO 2020-11-25 03:44:28 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 4.337076
+INFO 2020-11-25 03:45:45 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 4.354352
+INFO 2020-11-25 03:47:01 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 4.380453
+INFO 2020-11-25 03:48:18 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 4.356988
+INFO 2020-11-25 03:49:34 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 4.371785
+INFO 2020-11-25 03:50:51 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 4.359285
+INFO 2020-11-25 03:52:07 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 4.362331
+INFO 2020-11-25 03:53:24 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 4.364020
+INFO 2020-11-25 03:54:41 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 4.396912
+INFO 2020-11-25 03:55:57 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 4.379699
+INFO 2020-11-25 03:57:14 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-25 03:57:14 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 4.372220
+INFO 2020-11-25 03:58:31 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 4.375687
+INFO 2020-11-25 03:59:47 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 4.387501
+INFO 2020-11-25 03:59:54 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-25 03:59:55 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 4.437425
+INFO 2020-11-25 04:01:12 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 4.312835
+INFO 2020-11-25 04:02:28 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 4.326099
+INFO 2020-11-25 04:03:45 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 4.280351
+INFO 2020-11-25 04:05:01 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 4.285148
+INFO 2020-11-25 04:06:18 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 4.303608
+INFO 2020-11-25 04:07:34 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 4.304034
+INFO 2020-11-25 04:08:50 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 4.293425
+INFO 2020-11-25 04:10:07 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 4.328923
+INFO 2020-11-25 04:11:23 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 4.299053
+INFO 2020-11-25 04:12:39 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 4.304958
+INFO 2020-11-25 04:13:55 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 4.297158
+INFO 2020-11-25 04:15:12 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 4.286107
+INFO 2020-11-25 04:16:28 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 4.298167
+INFO 2020-11-25 04:17:45 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 4.299586
+INFO 2020-11-25 04:19:01 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-25 04:19:01 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 4.286830
+INFO 2020-11-25 04:20:17 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 4.326733
+INFO 2020-11-25 04:21:33 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 4.311498
+INFO 2020-11-25 04:22:48 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 4.313252
+INFO 2020-11-25 04:24:04 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 4.304375
+INFO 2020-11-25 04:25:20 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 4.305652
+INFO 2020-11-25 04:26:36 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 4.332660
+INFO 2020-11-25 04:27:52 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 4.300166
+INFO 2020-11-25 04:29:08 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 4.277409
+INFO 2020-11-25 04:30:23 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 4.320323
+INFO 2020-11-25 04:31:39 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 4.266749
+INFO 2020-11-25 04:32:55 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 4.300358
+INFO 2020-11-25 04:34:11 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 4.313041
+INFO 2020-11-25 04:35:27 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 4.324255
+INFO 2020-11-25 04:36:43 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 4.296991
+INFO 2020-11-25 04:37:59 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-25 04:37:59 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 4.308512
+INFO 2020-11-25 04:39:16 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 4.318126
+INFO 2020-11-25 04:40:33 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 4.305930
+INFO 2020-11-25 04:40:39 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-25 04:40:41 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 4.295313
+INFO 2020-11-25 04:41:57 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 4.302628
+INFO 2020-11-25 04:43:14 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 4.303336
+INFO 2020-11-25 04:44:30 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 4.289142
+INFO 2020-11-25 04:45:46 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 4.317333
+INFO 2020-11-25 04:47:03 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 4.305242
+INFO 2020-11-25 04:48:19 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 4.287574
+INFO 2020-11-25 04:49:36 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 4.306370
+INFO 2020-11-25 04:50:52 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 4.315201
+INFO 2020-11-25 04:52:08 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 4.301431
+INFO 2020-11-25 04:53:24 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 4.299300
+INFO 2020-11-25 04:54:41 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 4.293645
+INFO 2020-11-25 04:55:57 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 4.291119
+INFO 2020-11-25 04:57:13 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 4.308664
+INFO 2020-11-25 04:58:29 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 4.319460
+INFO 2020-11-25 04:59:46 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-25 04:59:46 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 4.309168
+INFO 2020-11-25 05:01:02 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 4.292737
+INFO 2020-11-25 05:02:19 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 4.293627
+INFO 2020-11-25 05:03:35 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 4.306646
+INFO 2020-11-25 05:04:52 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 4.282552
+INFO 2020-11-25 05:06:08 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 4.314200
+INFO 2020-11-25 05:07:25 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 4.296638
+INFO 2020-11-25 05:08:41 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 4.324167
+INFO 2020-11-25 05:09:58 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 4.305802
+INFO 2020-11-25 05:11:14 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 4.278409
+INFO 2020-11-25 05:12:31 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 4.306388
+INFO 2020-11-25 05:13:47 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 4.309003
+INFO 2020-11-25 05:15:04 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 4.266980
+INFO 2020-11-25 05:16:21 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 4.294368
+INFO 2020-11-25 05:17:37 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 4.304988
+INFO 2020-11-25 05:18:54 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-25 05:18:54 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 4.306873
+INFO 2020-11-25 05:20:11 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 4.316005
+INFO 2020-11-25 05:21:28 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 4.300295
+INFO 2020-11-25 05:21:34 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-25 05:21:34 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/MagFace/.gitkeep b/bob/bio/facexzoo/models/heads/MagFace/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..3cde9bf024262280e418714b1394efe96105ab46
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9581666666666667 | 0.003491170520747777 |
+| Epoch_14_batch_5999.pt | 0.9581666666666665 | 0.0030877096070200598 |
+| Epoch_14.pt | 0.958 | 0.0034498165991689016 |
+| Epoch_17_batch_5999.pt | 0.958 | 0.0029585615457098603 |
+| Epoch_16_batch_5999.pt | 0.9578333333333333 | 0.0032777777777777783 |
+| Epoch_15.pt | 0.9574999999999999 | 0.0030454377941479425 |
+| Epoch_12.pt | 0.9573333333333334 | 0.0037284359412361107 |
+| Epoch_11.pt | 0.9573333333333333 | 0.0031739681904634953 |
+| Epoch_13_batch_5999.pt | 0.9571666666666665 | 0.00336145544609019 |
+| Epoch_14_batch_2999.pt | 0.9571666666666665 | 0.003287180487219336 |
+| Epoch_17_batch_2999.pt | 0.9569999999999999 | 0.0030812054719693465 |
+| Epoch_13.pt | 0.9568333333333332 | 0.003243702350443333 |
+| Epoch_15_batch_2999.pt | 0.9566666666666668 | 0.0031426968052735457 |
+| Epoch_17.pt | 0.9555000000000001 | 0.0032683480177961616 |
+| Epoch_12_batch_2999.pt | 0.9555000000000001 | 0.002558573097704321 |
+| Epoch_16.pt | 0.9553333333333333 | 0.0033129003368612417 |
+| Epoch_15_batch_5999.pt | 0.9553333333333333 | 0.0032659863237109 |
+| Epoch_10_batch_2999.pt | 0.9551666666666667 | 0.0032341732395173126 |
+| Epoch_12_batch_5999.pt | 0.9551666666666666 | 0.0032721231828060572 |
+| Epoch_11_batch_2999.pt | 0.9550000000000001 | 0.0034426518632954877 |
+| Epoch_10.pt | 0.9550000000000001 | 0.003693504475243256 |
+| Epoch_11_batch_5999.pt | 0.9550000000000001 | 0.003152502435358025 |
+| Epoch_13_batch_2999.pt | 0.9546666666666667 | 0.0029793528172871006 |
+| Epoch_10_batch_5999.pt | 0.95 | 0.0036599264648890227 |
+| Epoch_9_batch_2999.pt | 0.9433333333333334 | 0.002496911672693802 |
+| Epoch_9_batch_5999.pt | 0.942 | 0.004784233364802438 |
+| Epoch_9.pt | 0.9385 | 0.005434514754537328 |
+| Epoch_8_batch_2999.pt | 0.9373333333333334 | 0.003999999999999995 |
+| Epoch_5_batch_5999.pt | 0.9366666666666665 | 0.00427452979148252 |
+| Epoch_6_batch_2999.pt | 0.9363333333333334 | 0.005228129047119373 |
+| Epoch_7_batch_2999.pt | 0.9361666666666668 | 0.005720064318036043 |
+| Epoch_7_batch_5999.pt | 0.9360000000000002 | 0.004438885412319597 |
+| Epoch_8_batch_5999.pt | 0.9349999999999998 | 0.0048939367878989455 |
+| Epoch_6.pt | 0.9348333333333333 | 0.004263323342877706 |
+| Epoch_4.pt | 0.9339999999999999 | 0.00586473032017629 |
+| Epoch_5_batch_2999.pt | 0.9339999999999999 | 0.003984538017120243 |
+| Epoch_6_batch_5999.pt | 0.9336666666666666 | 0.004050605807457376 |
+| Epoch_7.pt | 0.9313333333333335 | 0.0037908271353848805 |
+| Epoch_5.pt | 0.9308333333333334 | 0.0045963619537495665 |
+| Epoch_4_batch_5999.pt | 0.9303333333333332 | 0.00530431922768 |
+| Epoch_8.pt | 0.9291666666666666 | 0.0043832593994609734 |
+| Epoch_4_batch_2999.pt | 0.9271666666666667 | 0.0048243850756365505 |
+| Epoch_3_batch_5999.pt | 0.9266666666666667 | 0.006231144396804422 |
+| Epoch_3.pt | 0.9260000000000002 | 0.00551764845241562 |
+| Epoch_3_batch_2999.pt | 0.9251666666666667 | 0.0051306991798461595 |
+| Epoch_2_batch_5999.pt | 0.9241666666666666 | 0.0048511803323813855 |
+| Epoch_2.pt | 0.9229999999999998 | 0.005470459388929404 |
+| Epoch_2_batch_2999.pt | 0.9205000000000002 | 0.005968280352862987 |
+| Epoch_1_batch_5999.pt | 0.9096666666666667 | 0.005593205754956984 |
+| Epoch_1.pt | 0.907 | 0.006501661705923461 |
+| Epoch_1_batch_2999.pt | 0.8968333333333334 | 0.006668749674580854 |
+| Epoch_0.pt | 0.8556666666666667 | 0.006635110500814945 |
+| Epoch_0_batch_5999.pt | 0.8525 | 0.007662841235629844 |
+| Epoch_0_batch_2999.pt | 0.7725 | 0.0097444191055036 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..cce5244d71474c3c68d3730e42b7b71e854a5422
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.9403333333333332 | 0.0035986966090448126 |
+| Epoch_15_batch_2999.pt | 0.9398333333333333 | 0.0036298658275376898 |
+| Epoch_12.pt | 0.9395 | 0.0037765520887460737 |
+| Epoch_16.pt | 0.9386666666666666 | 0.003981438414926421 |
+| Epoch_13_batch_5999.pt | 0.9385000000000001 | 0.003852608543907954 |
+| Epoch_15_batch_5999.pt | 0.9385 | 0.0042050087711480565 |
+| Epoch_11_batch_2999.pt | 0.9381666666666666 | 0.0034734442294090256 |
+| Epoch_10.pt | 0.9380000000000001 | 0.00365823947403429 |
+| Epoch_11_batch_5999.pt | 0.9380000000000001 | 0.003950308629918039 |
+| Epoch_11.pt | 0.9380000000000001 | 0.003624334762288903 |
+| Epoch_12_batch_2999.pt | 0.9380000000000001 | 0.0037251232476089337 |
+| Epoch_10_batch_5999.pt | 0.9375000000000002 | 0.0037700083505163434 |
+| Epoch_16_batch_5999.pt | 0.9375 | 0.0039067052203887 |
+| Epoch_13_batch_2999.pt | 0.9373333333333335 | 0.0040536525215454835 |
+| Epoch_14_batch_5999.pt | 0.9373333333333334 | 0.003644715437079272 |
+| Epoch_16_batch_2999.pt | 0.9369999999999999 | 0.003798960221617501 |
+| Epoch_15.pt | 0.9369999999999999 | 0.0039891829046702805 |
+| Epoch_17_batch_2999.pt | 0.9368333333333334 | 0.0036383587400073154 |
+| Epoch_17_batch_5999.pt | 0.9368333333333334 | 0.003947573094109001 |
+| Epoch_13.pt | 0.9366666666666668 | 0.0037761434373258717 |
+| Epoch_17.pt | 0.9365 | 0.0035611934467860715 |
+| Epoch_14_batch_2999.pt | 0.9363333333333334 | 0.003823255674241166 |
+| Epoch_12_batch_5999.pt | 0.9355 | 0.003776552088746069 |
+| Epoch_10_batch_2999.pt | 0.9343333333333332 | 0.0035849479566448152 |
+| Epoch_9_batch_2999.pt | 0.9271666666666667 | 0.004289306254879467 |
+| Epoch_8_batch_5999.pt | 0.925 | 0.004667989230577778 |
+| Epoch_9_batch_5999.pt | 0.9248333333333335 | 0.004657729537494042 |
+| Epoch_9.pt | 0.9248333333333333 | 0.004248819734444205 |
+| Epoch_6_batch_2999.pt | 0.924 | 0.003929942040850537 |
+| Epoch_7_batch_5999.pt | 0.9235 | 0.004248819734444197 |
+| Epoch_6_batch_5999.pt | 0.9231666666666667 | 0.0037470152730927193 |
+| Epoch_8_batch_2999.pt | 0.923 | 0.0041111111111111105 |
+| Epoch_7_batch_2999.pt | 0.9218333333333334 | 0.004263323342877713 |
+| Epoch_8.pt | 0.9216666666666666 | 0.004766135686561598 |
+| Epoch_5_batch_5999.pt | 0.9213333333333334 | 0.0030912061651652335 |
+| Epoch_4_batch_5999.pt | 0.9211666666666666 | 0.003983375948942965 |
+| Epoch_5_batch_2999.pt | 0.9188333333333333 | 0.0037105954398881247 |
+| Epoch_6.pt | 0.9181666666666667 | 0.004040306185683718 |
+| Epoch_7.pt | 0.9168333333333333 | 0.0037798197095942365 |
+| Epoch_3_batch_2999.pt | 0.9149999999999998 | 0.0042236839574995985 |
+| Epoch_5.pt | 0.9143333333333332 | 0.004541985207993271 |
+| Epoch_3.pt | 0.9125 | 0.0037288498210110753 |
+| Epoch_2_batch_5999.pt | 0.9121666666666666 | 0.0038654052859553247 |
+| Epoch_4_batch_2999.pt | 0.9118333333333334 | 0.0052319649071953375 |
+| Epoch_3_batch_5999.pt | 0.9111666666666667 | 0.003897213312732349 |
+| Epoch_4.pt | 0.9106666666666667 | 0.004203173404306168 |
+| Epoch_2.pt | 0.9076666666666666 | 0.004926622061643414 |
+| Epoch_2_batch_2999.pt | 0.9063333333333334 | 0.004880042501332836 |
+| Epoch_1_batch_5999.pt | 0.8953333333333333 | 0.00586262487034352 |
+| Epoch_1.pt | 0.8919999999999998 | 0.004633479880442902 |
+| Epoch_1_batch_2999.pt | 0.8816666666666666 | 0.006014386045655275 |
+| Epoch_0_batch_5999.pt | 0.8471666666666667 | 0.004913134324327892 |
+| Epoch_0.pt | 0.844 | 0.004747969026493668 |
+| Epoch_0_batch_2999.pt | 0.7553333333333333 | 0.005966987369821134 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..8aa0e0541ab45b2dfb6d08dce8d6cdec93843ea3
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_17_batch_5999.pt | 0.8431666666666666 | 0.0072113165524578865 |
+| Epoch_16_batch_5999.pt | 0.8416666666666666 | 0.007179359990733382 |
+| Epoch_16_batch_2999.pt | 0.8415000000000001 | 0.007253989962108468 |
+| Epoch_17_batch_2999.pt | 0.8413333333333334 | 0.007276717480014645 |
+| Epoch_14_batch_2999.pt | 0.8403333333333333 | 0.008103497187428811 |
+| Epoch_13_batch_5999.pt | 0.8398333333333333 | 0.007480220832549242 |
+| Epoch_15.pt | 0.8396666666666667 | 0.007673105186942018 |
+| Epoch_16.pt | 0.8395000000000001 | 0.007576971283247717 |
+| Epoch_17.pt | 0.8390000000000001 | 0.007483314773547881 |
+| Epoch_14_batch_5999.pt | 0.8388333333333333 | 0.00778590844862501 |
+| Epoch_15_batch_2999.pt | 0.8385 | 0.007611111111111109 |
+| Epoch_13.pt | 0.8383333333333333 | 0.007503085784948498 |
+| Epoch_15_batch_5999.pt | 0.8378333333333334 | 0.00791174310833875 |
+| Epoch_14.pt | 0.8360000000000001 | 0.007201680188043912 |
+| Epoch_12_batch_2999.pt | 0.8358333333333332 | 0.006649282890881017 |
+| Epoch_13_batch_2999.pt | 0.8356666666666666 | 0.007670691375152532 |
+| Epoch_11_batch_2999.pt | 0.8348333333333333 | 0.0064043291685207315 |
+| Epoch_12.pt | 0.8336666666666668 | 0.007518700964763194 |
+| Epoch_11.pt | 0.8336666666666666 | 0.006830396839690526 |
+| Epoch_12_batch_5999.pt | 0.8335000000000001 | 0.0073259643110255325 |
+| Epoch_11_batch_5999.pt | 0.8308333333333333 | 0.0076386363594622576 |
+| Epoch_10.pt | 0.8301666666666667 | 0.006778916115700483 |
+| Epoch_10_batch_5999.pt | 0.827 | 0.007013215039137553 |
+| Epoch_10_batch_2999.pt | 0.825 | 0.0071836577223309455 |
+| Epoch_9_batch_2999.pt | 0.8086666666666668 | 0.00790296084537035 |
+| Epoch_8_batch_2999.pt | 0.8073333333333335 | 0.005704665623552639 |
+| Epoch_8_batch_5999.pt | 0.806 | 0.007549016778428008 |
+| Epoch_9_batch_5999.pt | 0.8055 | 0.007345318488922295 |
+| Epoch_8.pt | 0.8025 | 0.007845931555609477 |
+| Epoch_7.pt | 0.8009999999999999 | 0.0060461190490723495 |
+| Epoch_7_batch_5999.pt | 0.8008333333333333 | 0.006489545248387156 |
+| Epoch_6_batch_2999.pt | 0.7975 | 0.005081131885177546 |
+| Epoch_7_batch_2999.pt | 0.7973333333333334 | 0.006460707448936282 |
+| Epoch_5_batch_5999.pt | 0.7965000000000002 | 0.005255508573903235 |
+| Epoch_6_batch_5999.pt | 0.7961666666666667 | 0.005572474238235866 |
+| Epoch_5_batch_2999.pt | 0.7958333333333333 | 0.006441809806277032 |
+| Epoch_4_batch_5999.pt | 0.7943333333333332 | 0.008308109975080049 |
+| Epoch_9.pt | 0.7925000000000001 | 0.006588664039633566 |
+| Epoch_5.pt | 0.7901666666666667 | 0.006985213660049581 |
+| Epoch_6.pt | 0.7851666666666668 | 0.007513157183514636 |
+| Epoch_4.pt | 0.7845000000000001 | 0.007018713961903018 |
+| Epoch_3_batch_5999.pt | 0.7833333333333333 | 0.0071922454811921156 |
+| Epoch_3.pt | 0.7813333333333333 | 0.008737601445038885 |
+| Epoch_2_batch_5999.pt | 0.7761666666666667 | 0.008923716493108844 |
+| Epoch_4_batch_2999.pt | 0.7753333333333334 | 0.007212814385385545 |
+| Epoch_3_batch_2999.pt | 0.7733333333333333 | 0.008577892918052788 |
+| Epoch_2_batch_2999.pt | 0.7688333333333334 | 0.00884729853446941 |
+| Epoch_2.pt | 0.7663333333333333 | 0.006623937138128397 |
+| Epoch_1.pt | 0.756 | 0.006818638218918336 |
+| Epoch_1_batch_5999.pt | 0.754 | 0.008248456645763317 |
+| Epoch_1_batch_2999.pt | 0.7431666666666666 | 0.008778656777367186 |
+| Epoch_0_batch_5999.pt | 0.7070000000000001 | 0.007965202096899021 |
+| Epoch_0.pt | 0.7025 | 0.007278625904779586 |
+| Epoch_0_batch_2999.pt | 0.6276666666666666 | 0.011386043004865298 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_ijbc.txt b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_ijbc.txt
new file mode 100644
index 0000000000000000000000000000000000000000..c1ed42c8f153795421d7f8823565bf1b0eccbfd4
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_ijbc.txt
@@ -0,0 +1,5 @@
++-------------+--------+--------+--------+--------+--------+--------+
+| model_name | 1e-6 | 1e-5 | 1e-4 | 1e-3 | 1e-2 | 1e-1 |
++-------------+--------+--------+--------+--------+--------+--------+
+| Epoch_17.pt | 0.7009 | 0.8046 | 0.8840 | 0.9363 | 0.9688 | 0.9881 |
++-------------+--------+--------+--------+--------+--------+--------+
diff --git a/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..d0049b1d0fb8353ae1fabdbd5b8917c661af4223
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_12.pt | 0.9953333333333335 | 0.0011055415967851294 |
+| Epoch_15.pt | 0.9953333333333333 | 0.0012619796324000608 |
+| Epoch_10_batch_5999.pt | 0.9953333333333333 | 0.0013099806802835115 |
+| Epoch_15_batch_5999.pt | 0.9951666666666666 | 0.0013017082793177787 |
+| Epoch_16_batch_5999.pt | 0.9950000000000001 | 0.0012422599874998854 |
+| Epoch_11_batch_2999.pt | 0.9950000000000001 | 0.0011915339216404044 |
+| Epoch_16_batch_2999.pt | 0.9950000000000001 | 0.001290994448735802 |
+| Epoch_16.pt | 0.9948333333333335 | 0.0012777777777777785 |
+| Epoch_10.pt | 0.9948333333333335 | 0.0010378634273483002 |
+| Epoch_14_batch_5999.pt | 0.9946666666666667 | 0.001333333333333335 |
+| Epoch_12_batch_5999.pt | 0.9946666666666667 | 0.0012862041003100229 |
+| Epoch_15_batch_2999.pt | 0.9946666666666667 | 0.001333333333333335 |
+| Epoch_11.pt | 0.9946666666666667 | 0.0013333333333333307 |
+| Epoch_10_batch_2999.pt | 0.9945 | 0.0013158576980363402 |
+| Epoch_17_batch_2999.pt | 0.9945 | 0.0013391078659104373 |
+| Epoch_17_batch_5999.pt | 0.9945 | 0.0013158576980363357 |
+| Epoch_13.pt | 0.9945 | 0.0012184284555256302 |
+| Epoch_14.pt | 0.9943333333333333 | 0.0014529663145135623 |
+| Epoch_14_batch_2999.pt | 0.9943333333333333 | 0.001319371343004211 |
+| Epoch_13_batch_5999.pt | 0.9941666666666666 | 0.0013888888888888892 |
+| Epoch_12_batch_2999.pt | 0.9940000000000001 | 0.0013653561919382854 |
+| Epoch_13_batch_2999.pt | 0.994 | 0.0014315665251916818 |
+| Epoch_17.pt | 0.994 | 0.0014529663145135606 |
+| Epoch_8_batch_2999.pt | 0.994 | 0.0010599324460188319 |
+| Epoch_9_batch_2999.pt | 0.9936666666666667 | 0.001717736092637809 |
+| Epoch_6_batch_2999.pt | 0.9933333333333334 | 0.0016480441082434864 |
+| Epoch_6_batch_5999.pt | 0.9933333333333334 | 0.0016292086998461281 |
+| Epoch_7.pt | 0.9933333333333334 | 0.001531560972454468 |
+| Epoch_9_batch_5999.pt | 0.9931666666666666 | 0.001711435755638817 |
+| Epoch_11_batch_5999.pt | 0.9931666666666666 | 0.00152043690926711 |
+| Epoch_4_batch_5999.pt | 0.9928333333333332 | 0.0012184284555256332 |
+| Epoch_3_batch_5999.pt | 0.9926666666666668 | 0.0014529663145135586 |
+| Epoch_5_batch_2999.pt | 0.9924999999999999 | 0.0013205404804449705 |
+| Epoch_5.pt | 0.9924999999999999 | 0.0013664859862498745 |
+| Epoch_8_batch_5999.pt | 0.992 | 0.001444444444444449 |
+| Epoch_6.pt | 0.992 | 0.0015275252316519416 |
+| Epoch_7_batch_2999.pt | 0.9918333333333335 | 0.0014369463507086146 |
+| Epoch_7_batch_5999.pt | 0.9918333333333333 | 0.0013017082793177733 |
+| Epoch_9.pt | 0.9918333333333333 | 0.0015605079894653485 |
+| Epoch_8.pt | 0.9916666666666666 | 0.001571348402636772 |
+| Epoch_4_batch_2999.pt | 0.9911666666666668 | 0.0015525765124980153 |
+| Epoch_3.pt | 0.991 | 0.0009686442096756996 |
+| Epoch_4.pt | 0.9908333333333333 | 0.0014751020052613051 |
+| Epoch_5_batch_5999.pt | 0.9904999999999999 | 0.0017042068500197727 |
+| Epoch_2_batch_5999.pt | 0.9898333333333333 | 0.0017114357556388157 |
+| Epoch_2.pt | 0.9898333333333333 | 0.0016749792701868144 |
+| Epoch_3_batch_2999.pt | 0.9896666666666667 | 0.0019372884193514107 |
+| Epoch_2_batch_2999.pt | 0.9886666666666667 | 0.0015071844406945043 |
+| Epoch_1_batch_5999.pt | 0.9881666666666666 | 0.0018500917561496313 |
+| Epoch_1.pt | 0.9875 | 0.0019444444444444429 |
+| Epoch_1_batch_2999.pt | 0.9858333333333335 | 0.0019285061064121896 |
+| Epoch_0.pt | 0.9734999999999999 | 0.002669558617051992 |
+| Epoch_0_batch_5999.pt | 0.9724999999999999 | 0.0029632523006876852 |
+| Epoch_0_batch_2999.pt | 0.9391666666666667 | 0.004061639272540938 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_megaface.txt b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_megaface.txt
new file mode 100644
index 0000000000000000000000000000000000000000..82d51d444bccba2291693936ad08e97eb6e5705e
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MagFace/accu_files/accu_megaface.txt
@@ -0,0 +1,14 @@
++------+--------------------+
+| rank | accuracy |
++------+--------------------+
+| 1 | 0.8984599763073278 |
+| 2 | 0.916756707499642 |
+| 3 | 0.9257456031867946 |
+| 4 | 0.9312391788276033 |
+| 5 | 0.9349688220055457 |
+| 6 | 0.9380215316922036 |
+| 7 | 0.9406055951156646 |
+| 8 | 0.9426298866136402 |
+| 9 | 0.944185531848419 |
+| 10 | 0.9456305244932762 |
++------+--------------------+
diff --git a/bob/bio/facexzoo/models/heads/MagFace/log.log b/bob/bio/facexzoo/models/heads/MagFace/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..008634839bfd69921509259751b2633605ee755f
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/MagFace/log.log
@@ -0,0 +1,657 @@
+INFO 2021-03-18 11:49:18 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2021-03-18 11:49:18 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2021-03-18 11:49:19 train.py: 180] Start optimization.
+INFO 2021-03-18 11:49:19 train.py: 181] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='MobileFaceNet', batch_size=512, data_root='/home/wangjun492/wj_data/facex-zoo/private_file/train_data/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='MagFace', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='mag-mobile', train_file='/home/wangjun492/wj_data/facex-zoo/private_file/train_data/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7f044c5ca7b8>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'margin_am': 0.0, 'scale': 64, 'l_a': 10, 'u_a': 110, 'l_margin': 0.45, 'u_margin': 0.8, 'lamda': 20}
+INFO 2021-03-18 11:49:42 train.py: 82] Epoch 0, iter 0/6416, lr 0.100000, loss 44.941360
+INFO 2021-03-18 11:50:59 train.py: 82] Epoch 0, iter 200/6416, lr 0.100000, loss 43.128643
+INFO 2021-03-18 11:52:16 train.py: 82] Epoch 0, iter 400/6416, lr 0.100000, loss 42.320675
+INFO 2021-03-18 11:53:33 train.py: 82] Epoch 0, iter 600/6416, lr 0.100000, loss 42.100863
+INFO 2021-03-18 11:54:50 train.py: 82] Epoch 0, iter 800/6416, lr 0.100000, loss 41.848822
+INFO 2021-03-18 11:56:07 train.py: 82] Epoch 0, iter 1000/6416, lr 0.100000, loss 41.520204
+INFO 2021-03-18 11:57:24 train.py: 82] Epoch 0, iter 1200/6416, lr 0.100000, loss 41.108999
+INFO 2021-03-18 11:58:41 train.py: 82] Epoch 0, iter 1400/6416, lr 0.100000, loss 40.648215
+INFO 2021-03-18 11:59:58 train.py: 82] Epoch 0, iter 1600/6416, lr 0.100000, loss 40.151882
+INFO 2021-03-18 12:01:15 train.py: 82] Epoch 0, iter 1800/6416, lr 0.100000, loss 39.656723
+INFO 2021-03-18 12:02:32 train.py: 82] Epoch 0, iter 2000/6416, lr 0.100000, loss 39.152646
+INFO 2021-03-18 12:03:49 train.py: 82] Epoch 0, iter 2200/6416, lr 0.100000, loss 38.640248
+INFO 2021-03-18 12:05:06 train.py: 82] Epoch 0, iter 2400/6416, lr 0.100000, loss 38.065436
+INFO 2021-03-18 12:06:23 train.py: 82] Epoch 0, iter 2600/6416, lr 0.100000, loss 37.453125
+INFO 2021-03-18 12:07:40 train.py: 82] Epoch 0, iter 2800/6416, lr 0.100000, loss 36.798310
+INFO 2021-03-18 12:08:57 train.py: 95] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2021-03-18 12:08:57 train.py: 82] Epoch 0, iter 3000/6416, lr 0.100000, loss 36.098991
+INFO 2021-03-18 12:10:15 train.py: 82] Epoch 0, iter 3200/6416, lr 0.100000, loss 35.371497
+INFO 2021-03-18 12:11:33 train.py: 82] Epoch 0, iter 3400/6416, lr 0.100000, loss 34.595981
+INFO 2021-03-18 12:12:50 train.py: 82] Epoch 0, iter 3600/6416, lr 0.100000, loss 33.769410
+INFO 2021-03-18 12:14:08 train.py: 82] Epoch 0, iter 3800/6416, lr 0.100000, loss 32.885128
+INFO 2021-03-18 12:15:25 train.py: 82] Epoch 0, iter 4000/6416, lr 0.100000, loss 32.060547
+INFO 2021-03-18 12:16:43 train.py: 82] Epoch 0, iter 4200/6416, lr 0.100000, loss 31.117111
+INFO 2021-03-18 12:18:00 train.py: 82] Epoch 0, iter 4400/6416, lr 0.100000, loss 30.136257
+INFO 2021-03-18 12:19:18 train.py: 82] Epoch 0, iter 4600/6416, lr 0.100000, loss 29.154507
+INFO 2021-03-18 12:20:36 train.py: 82] Epoch 0, iter 4800/6416, lr 0.100000, loss 28.264849
+INFO 2021-03-18 12:21:53 train.py: 82] Epoch 0, iter 5000/6416, lr 0.100000, loss 27.239139
+INFO 2021-03-18 12:23:11 train.py: 82] Epoch 0, iter 5200/6416, lr 0.100000, loss 26.256088
+INFO 2021-03-18 12:24:28 train.py: 82] Epoch 0, iter 5400/6416, lr 0.100000, loss 25.363886
+INFO 2021-03-18 12:25:46 train.py: 82] Epoch 0, iter 5600/6416, lr 0.100000, loss 24.460719
+INFO 2021-03-18 12:27:03 train.py: 82] Epoch 0, iter 5800/6416, lr 0.100000, loss 23.676551
+INFO 2021-03-18 12:28:21 train.py: 95] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2021-03-18 12:28:22 train.py: 82] Epoch 0, iter 6000/6416, lr 0.100000, loss 22.855685
+INFO 2021-03-18 12:29:39 train.py: 82] Epoch 0, iter 6200/6416, lr 0.100000, loss 22.193575
+INFO 2021-03-18 12:30:57 train.py: 82] Epoch 0, iter 6400/6416, lr 0.100000, loss 21.538951
+INFO 2021-03-18 12:31:03 train.py: 100] Save checkpoint Epoch_0.pt to disk...
+INFO 2021-03-18 12:31:05 train.py: 82] Epoch 1, iter 0/6416, lr 0.100000, loss 21.302389
+INFO 2021-03-18 12:32:23 train.py: 82] Epoch 1, iter 200/6416, lr 0.100000, loss 20.206784
+INFO 2021-03-18 12:33:41 train.py: 82] Epoch 1, iter 400/6416, lr 0.100000, loss 19.691075
+INFO 2021-03-18 12:34:59 train.py: 82] Epoch 1, iter 600/6416, lr 0.100000, loss 19.407650
+INFO 2021-03-18 12:36:16 train.py: 82] Epoch 1, iter 800/6416, lr 0.100000, loss 19.060496
+INFO 2021-03-18 12:37:34 train.py: 82] Epoch 1, iter 1000/6416, lr 0.100000, loss 18.812180
+INFO 2021-03-18 12:38:52 train.py: 82] Epoch 1, iter 1200/6416, lr 0.100000, loss 18.574332
+INFO 2021-03-18 12:40:09 train.py: 82] Epoch 1, iter 1400/6416, lr 0.100000, loss 18.334928
+INFO 2021-03-18 12:41:27 train.py: 82] Epoch 1, iter 1600/6416, lr 0.100000, loss 18.052895
+INFO 2021-03-18 12:42:45 train.py: 82] Epoch 1, iter 1800/6416, lr 0.100000, loss 17.859909
+INFO 2021-03-18 12:44:02 train.py: 82] Epoch 1, iter 2000/6416, lr 0.100000, loss 17.602214
+INFO 2021-03-18 12:45:20 train.py: 82] Epoch 1, iter 2200/6416, lr 0.100000, loss 17.440505
+INFO 2021-03-18 12:46:38 train.py: 82] Epoch 1, iter 2400/6416, lr 0.100000, loss 17.153511
+INFO 2021-03-18 12:47:55 train.py: 82] Epoch 1, iter 2600/6416, lr 0.100000, loss 17.022853
+INFO 2021-03-18 12:49:13 train.py: 82] Epoch 1, iter 2800/6416, lr 0.100000, loss 16.827922
+INFO 2021-03-18 12:50:31 train.py: 95] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2021-03-18 12:50:31 train.py: 82] Epoch 1, iter 3000/6416, lr 0.100000, loss 16.717812
+INFO 2021-03-18 12:51:48 train.py: 82] Epoch 1, iter 3200/6416, lr 0.100000, loss 16.578337
+INFO 2021-03-18 12:53:05 train.py: 82] Epoch 1, iter 3400/6416, lr 0.100000, loss 16.412496
+INFO 2021-03-18 12:54:22 train.py: 82] Epoch 1, iter 3600/6416, lr 0.100000, loss 16.303695
+INFO 2021-03-18 12:55:39 train.py: 82] Epoch 1, iter 3800/6416, lr 0.100000, loss 16.167215
+INFO 2021-03-18 12:56:56 train.py: 82] Epoch 1, iter 4000/6416, lr 0.100000, loss 16.043823
+INFO 2021-03-18 12:58:13 train.py: 82] Epoch 1, iter 4200/6416, lr 0.100000, loss 15.874951
+INFO 2021-03-18 12:59:30 train.py: 82] Epoch 1, iter 4400/6416, lr 0.100000, loss 15.825704
+INFO 2021-03-18 13:00:47 train.py: 82] Epoch 1, iter 4600/6416, lr 0.100000, loss 15.702083
+INFO 2021-03-18 13:02:04 train.py: 82] Epoch 1, iter 4800/6416, lr 0.100000, loss 15.577482
+INFO 2021-03-18 13:03:22 train.py: 82] Epoch 1, iter 5000/6416, lr 0.100000, loss 15.565388
+INFO 2021-03-18 13:04:39 train.py: 82] Epoch 1, iter 5200/6416, lr 0.100000, loss 15.411686
+INFO 2021-03-18 13:05:56 train.py: 82] Epoch 1, iter 5400/6416, lr 0.100000, loss 15.361888
+INFO 2021-03-18 13:07:13 train.py: 82] Epoch 1, iter 5600/6416, lr 0.100000, loss 15.302795
+INFO 2021-03-18 13:08:30 train.py: 82] Epoch 1, iter 5800/6416, lr 0.100000, loss 15.180263
+INFO 2021-03-18 13:09:47 train.py: 95] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2021-03-18 13:09:47 train.py: 82] Epoch 1, iter 6000/6416, lr 0.100000, loss 15.076002
+INFO 2021-03-18 13:11:05 train.py: 82] Epoch 1, iter 6200/6416, lr 0.100000, loss 15.072046
+INFO 2021-03-18 13:12:23 train.py: 82] Epoch 1, iter 6400/6416, lr 0.100000, loss 14.944845
+INFO 2021-03-18 13:12:29 train.py: 100] Save checkpoint Epoch_1.pt to disk...
+INFO 2021-03-18 13:12:31 train.py: 82] Epoch 2, iter 0/6416, lr 0.100000, loss 14.917471
+INFO 2021-03-18 13:13:49 train.py: 82] Epoch 2, iter 200/6416, lr 0.100000, loss 14.065084
+INFO 2021-03-18 13:15:07 train.py: 82] Epoch 2, iter 400/6416, lr 0.100000, loss 14.046457
+INFO 2021-03-18 13:16:24 train.py: 82] Epoch 2, iter 600/6416, lr 0.100000, loss 14.158617
+INFO 2021-03-18 13:17:42 train.py: 82] Epoch 2, iter 800/6416, lr 0.100000, loss 14.179159
+INFO 2021-03-18 13:19:00 train.py: 82] Epoch 2, iter 1000/6416, lr 0.100000, loss 14.286912
+INFO 2021-03-18 13:20:18 train.py: 82] Epoch 2, iter 1200/6416, lr 0.100000, loss 14.269518
+INFO 2021-03-18 13:21:35 train.py: 82] Epoch 2, iter 1400/6416, lr 0.100000, loss 14.277494
+INFO 2021-03-18 13:22:53 train.py: 82] Epoch 2, iter 1600/6416, lr 0.100000, loss 14.283750
+INFO 2021-03-18 13:24:11 train.py: 82] Epoch 2, iter 1800/6416, lr 0.100000, loss 14.300519
+INFO 2021-03-18 13:25:28 train.py: 82] Epoch 2, iter 2000/6416, lr 0.100000, loss 14.253421
+INFO 2021-03-18 13:26:46 train.py: 82] Epoch 2, iter 2200/6416, lr 0.100000, loss 14.243677
+INFO 2021-03-18 13:28:04 train.py: 82] Epoch 2, iter 2400/6416, lr 0.100000, loss 14.244384
+INFO 2021-03-18 13:29:22 train.py: 82] Epoch 2, iter 2600/6416, lr 0.100000, loss 14.191393
+INFO 2021-03-18 13:30:39 train.py: 82] Epoch 2, iter 2800/6416, lr 0.100000, loss 14.152997
+INFO 2021-03-18 13:31:57 train.py: 95] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2021-03-18 13:31:57 train.py: 82] Epoch 2, iter 3000/6416, lr 0.100000, loss 14.092379
+INFO 2021-03-18 13:33:15 train.py: 82] Epoch 2, iter 3200/6416, lr 0.100000, loss 14.081347
+INFO 2021-03-18 13:34:33 train.py: 82] Epoch 2, iter 3400/6416, lr 0.100000, loss 14.086740
+INFO 2021-03-18 13:35:50 train.py: 82] Epoch 2, iter 3600/6416, lr 0.100000, loss 13.955012
+INFO 2021-03-18 13:37:08 train.py: 82] Epoch 2, iter 3800/6416, lr 0.100000, loss 13.989396
+INFO 2021-03-18 13:38:26 train.py: 82] Epoch 2, iter 4000/6416, lr 0.100000, loss 13.956590
+INFO 2021-03-18 13:39:44 train.py: 82] Epoch 2, iter 4200/6416, lr 0.100000, loss 13.888980
+INFO 2021-03-18 13:41:01 train.py: 82] Epoch 2, iter 4400/6416, lr 0.100000, loss 13.818482
+INFO 2021-03-18 13:42:19 train.py: 82] Epoch 2, iter 4600/6416, lr 0.100000, loss 13.818384
+INFO 2021-03-18 13:43:37 train.py: 82] Epoch 2, iter 4800/6416, lr 0.100000, loss 13.805937
+INFO 2021-03-18 13:44:54 train.py: 82] Epoch 2, iter 5000/6416, lr 0.100000, loss 13.783898
+INFO 2021-03-18 13:46:12 train.py: 82] Epoch 2, iter 5200/6416, lr 0.100000, loss 13.734794
+INFO 2021-03-18 13:47:30 train.py: 82] Epoch 2, iter 5400/6416, lr 0.100000, loss 13.678580
+INFO 2021-03-18 13:48:47 train.py: 82] Epoch 2, iter 5600/6416, lr 0.100000, loss 13.648223
+INFO 2021-03-18 13:50:05 train.py: 82] Epoch 2, iter 5800/6416, lr 0.100000, loss 13.609823
+INFO 2021-03-18 13:51:23 train.py: 95] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2021-03-18 13:51:23 train.py: 82] Epoch 2, iter 6000/6416, lr 0.100000, loss 13.628657
+INFO 2021-03-18 13:52:40 train.py: 82] Epoch 2, iter 6200/6416, lr 0.100000, loss 13.533296
+INFO 2021-03-18 13:53:57 train.py: 82] Epoch 2, iter 6400/6416, lr 0.100000, loss 13.455896
+INFO 2021-03-18 13:54:04 train.py: 100] Save checkpoint Epoch_2.pt to disk...
+INFO 2021-03-18 13:54:06 train.py: 82] Epoch 3, iter 0/6416, lr 0.100000, loss 13.628212
+INFO 2021-03-18 13:55:24 train.py: 82] Epoch 3, iter 200/6416, lr 0.100000, loss 12.619045
+INFO 2021-03-18 13:56:42 train.py: 82] Epoch 3, iter 400/6416, lr 0.100000, loss 12.682672
+INFO 2021-03-18 13:57:59 train.py: 82] Epoch 3, iter 600/6416, lr 0.100000, loss 12.867427
+INFO 2021-03-18 13:59:17 train.py: 82] Epoch 3, iter 800/6416, lr 0.100000, loss 12.879549
+INFO 2021-03-18 14:00:35 train.py: 82] Epoch 3, iter 1000/6416, lr 0.100000, loss 13.029148
+INFO 2021-03-18 14:01:53 train.py: 82] Epoch 3, iter 1200/6416, lr 0.100000, loss 13.061333
+INFO 2021-03-18 14:03:11 train.py: 82] Epoch 3, iter 1400/6416, lr 0.100000, loss 13.039559
+INFO 2021-03-18 14:04:28 train.py: 82] Epoch 3, iter 1600/6416, lr 0.100000, loss 13.042433
+INFO 2021-03-18 14:05:46 train.py: 82] Epoch 3, iter 1800/6416, lr 0.100000, loss 13.127904
+INFO 2021-03-18 14:07:04 train.py: 82] Epoch 3, iter 2000/6416, lr 0.100000, loss 13.150322
+INFO 2021-03-18 14:08:22 train.py: 82] Epoch 3, iter 2200/6416, lr 0.100000, loss 13.133953
+INFO 2021-03-18 14:09:39 train.py: 82] Epoch 3, iter 2400/6416, lr 0.100000, loss 13.079338
+INFO 2021-03-18 14:10:57 train.py: 82] Epoch 3, iter 2600/6416, lr 0.100000, loss 13.189715
+INFO 2021-03-18 14:12:15 train.py: 82] Epoch 3, iter 2800/6416, lr 0.100000, loss 13.106944
+INFO 2021-03-18 14:13:33 train.py: 95] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2021-03-18 14:13:33 train.py: 82] Epoch 3, iter 3000/6416, lr 0.100000, loss 13.044181
+INFO 2021-03-18 14:14:51 train.py: 82] Epoch 3, iter 3200/6416, lr 0.100000, loss 13.038052
+INFO 2021-03-18 14:16:09 train.py: 82] Epoch 3, iter 3400/6416, lr 0.100000, loss 13.062270
+INFO 2021-03-18 14:17:26 train.py: 82] Epoch 3, iter 3600/6416, lr 0.100000, loss 13.053553
+INFO 2021-03-18 14:18:44 train.py: 82] Epoch 3, iter 3800/6416, lr 0.100000, loss 13.034928
+INFO 2021-03-18 14:20:02 train.py: 82] Epoch 3, iter 4000/6416, lr 0.100000, loss 12.995936
+INFO 2021-03-18 14:21:20 train.py: 82] Epoch 3, iter 4200/6416, lr 0.100000, loss 12.951277
+INFO 2021-03-18 14:22:38 train.py: 82] Epoch 3, iter 4400/6416, lr 0.100000, loss 12.940024
+INFO 2021-03-18 14:23:56 train.py: 82] Epoch 3, iter 4600/6416, lr 0.100000, loss 12.916850
+INFO 2021-03-18 14:25:13 train.py: 82] Epoch 3, iter 4800/6416, lr 0.100000, loss 12.877921
+INFO 2021-03-18 14:26:31 train.py: 82] Epoch 3, iter 5000/6416, lr 0.100000, loss 12.893714
+INFO 2021-03-18 14:27:49 train.py: 82] Epoch 3, iter 5200/6416, lr 0.100000, loss 12.833438
+INFO 2021-03-18 14:29:07 train.py: 82] Epoch 3, iter 5400/6416, lr 0.100000, loss 12.850828
+INFO 2021-03-18 14:30:24 train.py: 82] Epoch 3, iter 5600/6416, lr 0.100000, loss 12.808592
+INFO 2021-03-18 14:31:42 train.py: 82] Epoch 3, iter 5800/6416, lr 0.100000, loss 12.778366
+INFO 2021-03-18 14:33:00 train.py: 95] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2021-03-18 14:33:01 train.py: 82] Epoch 3, iter 6000/6416, lr 0.100000, loss 12.792225
+INFO 2021-03-18 14:34:18 train.py: 82] Epoch 3, iter 6200/6416, lr 0.100000, loss 12.786476
+INFO 2021-03-18 14:35:36 train.py: 82] Epoch 3, iter 6400/6416, lr 0.100000, loss 12.720869
+INFO 2021-03-18 14:35:43 train.py: 100] Save checkpoint Epoch_3.pt to disk...
+INFO 2021-03-18 14:35:45 train.py: 82] Epoch 4, iter 0/6416, lr 0.100000, loss 12.764198
+INFO 2021-03-18 14:37:02 train.py: 82] Epoch 4, iter 200/6416, lr 0.100000, loss 11.924315
+INFO 2021-03-18 14:38:19 train.py: 82] Epoch 4, iter 400/6416, lr 0.100000, loss 11.916345
+INFO 2021-03-18 14:39:37 train.py: 82] Epoch 4, iter 600/6416, lr 0.100000, loss 12.073752
+INFO 2021-03-18 14:40:54 train.py: 82] Epoch 4, iter 800/6416, lr 0.100000, loss 12.217385
+INFO 2021-03-18 14:42:11 train.py: 82] Epoch 4, iter 1000/6416, lr 0.100000, loss 12.335926
+INFO 2021-03-18 14:43:28 train.py: 82] Epoch 4, iter 1200/6416, lr 0.100000, loss 12.324938
+INFO 2021-03-18 14:44:45 train.py: 82] Epoch 4, iter 1400/6416, lr 0.100000, loss 12.368517
+INFO 2021-03-18 14:46:02 train.py: 82] Epoch 4, iter 1600/6416, lr 0.100000, loss 12.404423
+INFO 2021-03-18 14:47:20 train.py: 82] Epoch 4, iter 1800/6416, lr 0.100000, loss 12.442831
+INFO 2021-03-18 14:48:37 train.py: 82] Epoch 4, iter 2000/6416, lr 0.100000, loss 12.476674
+INFO 2021-03-18 14:49:54 train.py: 82] Epoch 4, iter 2200/6416, lr 0.100000, loss 12.471686
+INFO 2021-03-18 14:51:11 train.py: 82] Epoch 4, iter 2400/6416, lr 0.100000, loss 12.447021
+INFO 2021-03-18 14:52:28 train.py: 82] Epoch 4, iter 2600/6416, lr 0.100000, loss 12.460190
+INFO 2021-03-18 14:53:45 train.py: 82] Epoch 4, iter 2800/6416, lr 0.100000, loss 12.472314
+INFO 2021-03-18 14:55:02 train.py: 95] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2021-03-18 14:55:03 train.py: 82] Epoch 4, iter 3000/6416, lr 0.100000, loss 12.465331
+INFO 2021-03-18 14:56:20 train.py: 82] Epoch 4, iter 3200/6416, lr 0.100000, loss 12.465115
+INFO 2021-03-18 14:57:38 train.py: 82] Epoch 4, iter 3400/6416, lr 0.100000, loss 12.427757
+INFO 2021-03-18 14:58:56 train.py: 82] Epoch 4, iter 3600/6416, lr 0.100000, loss 12.461125
+INFO 2021-03-18 15:00:14 train.py: 82] Epoch 4, iter 3800/6416, lr 0.100000, loss 12.436430
+INFO 2021-03-18 15:01:31 train.py: 82] Epoch 4, iter 4000/6416, lr 0.100000, loss 12.392376
+INFO 2021-03-18 15:02:49 train.py: 82] Epoch 4, iter 4200/6416, lr 0.100000, loss 12.405286
+INFO 2021-03-18 15:04:07 train.py: 82] Epoch 4, iter 4400/6416, lr 0.100000, loss 12.329487
+INFO 2021-03-18 15:05:25 train.py: 82] Epoch 4, iter 4600/6416, lr 0.100000, loss 12.372520
+INFO 2021-03-18 15:06:42 train.py: 82] Epoch 4, iter 4800/6416, lr 0.100000, loss 12.366381
+INFO 2021-03-18 15:08:00 train.py: 82] Epoch 4, iter 5000/6416, lr 0.100000, loss 12.278925
+INFO 2021-03-18 15:09:18 train.py: 82] Epoch 4, iter 5200/6416, lr 0.100000, loss 12.345700
+INFO 2021-03-18 15:10:35 train.py: 82] Epoch 4, iter 5400/6416, lr 0.100000, loss 12.305708
+INFO 2021-03-18 15:11:53 train.py: 82] Epoch 4, iter 5600/6416, lr 0.100000, loss 12.244926
+INFO 2021-03-18 15:13:11 train.py: 82] Epoch 4, iter 5800/6416, lr 0.100000, loss 12.340714
+INFO 2021-03-18 15:14:28 train.py: 95] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2021-03-18 15:14:29 train.py: 82] Epoch 4, iter 6000/6416, lr 0.100000, loss 12.286311
+INFO 2021-03-18 15:15:47 train.py: 82] Epoch 4, iter 6200/6416, lr 0.100000, loss 12.286924
+INFO 2021-03-18 15:17:04 train.py: 82] Epoch 4, iter 6400/6416, lr 0.100000, loss 12.303652
+INFO 2021-03-18 15:17:11 train.py: 100] Save checkpoint Epoch_4.pt to disk...
+INFO 2021-03-18 15:17:13 train.py: 82] Epoch 5, iter 0/6416, lr 0.100000, loss 12.322933
+INFO 2021-03-18 15:18:30 train.py: 82] Epoch 5, iter 200/6416, lr 0.100000, loss 11.437115
+INFO 2021-03-18 15:19:47 train.py: 82] Epoch 5, iter 400/6416, lr 0.100000, loss 11.467058
+INFO 2021-03-18 15:21:05 train.py: 82] Epoch 5, iter 600/6416, lr 0.100000, loss 11.607058
+INFO 2021-03-18 15:22:22 train.py: 82] Epoch 5, iter 800/6416, lr 0.100000, loss 11.716540
+INFO 2021-03-18 15:23:39 train.py: 82] Epoch 5, iter 1000/6416, lr 0.100000, loss 11.828646
+INFO 2021-03-18 15:24:56 train.py: 82] Epoch 5, iter 1200/6416, lr 0.100000, loss 11.830711
+INFO 2021-03-18 15:26:13 train.py: 82] Epoch 5, iter 1400/6416, lr 0.100000, loss 11.876649
+INFO 2021-03-18 15:27:30 train.py: 82] Epoch 5, iter 1600/6416, lr 0.100000, loss 11.975076
+INFO 2021-03-18 15:28:47 train.py: 82] Epoch 5, iter 1800/6416, lr 0.100000, loss 11.972674
+INFO 2021-03-18 15:30:05 train.py: 82] Epoch 5, iter 2000/6416, lr 0.100000, loss 12.004308
+INFO 2021-03-18 15:31:22 train.py: 82] Epoch 5, iter 2200/6416, lr 0.100000, loss 12.046119
+INFO 2021-03-18 15:32:39 train.py: 82] Epoch 5, iter 2400/6416, lr 0.100000, loss 12.123801
+INFO 2021-03-18 15:33:56 train.py: 82] Epoch 5, iter 2600/6416, lr 0.100000, loss 12.028132
+INFO 2021-03-18 15:35:13 train.py: 82] Epoch 5, iter 2800/6416, lr 0.100000, loss 12.039813
+INFO 2021-03-18 15:36:30 train.py: 95] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2021-03-18 15:36:31 train.py: 82] Epoch 5, iter 3000/6416, lr 0.100000, loss 12.045182
+INFO 2021-03-18 15:37:49 train.py: 82] Epoch 5, iter 3200/6416, lr 0.100000, loss 12.073116
+INFO 2021-03-18 15:39:06 train.py: 82] Epoch 5, iter 3400/6416, lr 0.100000, loss 12.023968
+INFO 2021-03-18 15:40:24 train.py: 82] Epoch 5, iter 3600/6416, lr 0.100000, loss 12.038412
+INFO 2021-03-18 15:41:42 train.py: 82] Epoch 5, iter 3800/6416, lr 0.100000, loss 12.042168
+INFO 2021-03-18 15:42:59 train.py: 82] Epoch 5, iter 4000/6416, lr 0.100000, loss 11.977067
+INFO 2021-03-18 15:44:17 train.py: 82] Epoch 5, iter 4200/6416, lr 0.100000, loss 11.993285
+INFO 2021-03-18 15:45:35 train.py: 82] Epoch 5, iter 4400/6416, lr 0.100000, loss 12.035413
+INFO 2021-03-18 15:46:53 train.py: 82] Epoch 5, iter 4600/6416, lr 0.100000, loss 11.962388
+INFO 2021-03-18 15:48:10 train.py: 82] Epoch 5, iter 4800/6416, lr 0.100000, loss 11.961586
+INFO 2021-03-18 15:49:28 train.py: 82] Epoch 5, iter 5000/6416, lr 0.100000, loss 11.981703
+INFO 2021-03-18 15:50:46 train.py: 82] Epoch 5, iter 5200/6416, lr 0.100000, loss 11.963891
+INFO 2021-03-18 15:52:04 train.py: 82] Epoch 5, iter 5400/6416, lr 0.100000, loss 11.983260
+INFO 2021-03-18 15:53:21 train.py: 82] Epoch 5, iter 5600/6416, lr 0.100000, loss 11.948841
+INFO 2021-03-18 15:54:39 train.py: 82] Epoch 5, iter 5800/6416, lr 0.100000, loss 11.955416
+INFO 2021-03-18 15:55:57 train.py: 95] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2021-03-18 15:55:57 train.py: 82] Epoch 5, iter 6000/6416, lr 0.100000, loss 11.935622
+INFO 2021-03-18 15:57:15 train.py: 82] Epoch 5, iter 6200/6416, lr 0.100000, loss 11.926018
+INFO 2021-03-18 15:58:33 train.py: 82] Epoch 5, iter 6400/6416, lr 0.100000, loss 11.921697
+INFO 2021-03-18 15:58:39 train.py: 100] Save checkpoint Epoch_5.pt to disk...
+INFO 2021-03-18 15:58:41 train.py: 82] Epoch 6, iter 0/6416, lr 0.100000, loss 11.713128
+INFO 2021-03-18 15:59:59 train.py: 82] Epoch 6, iter 200/6416, lr 0.100000, loss 11.104023
+INFO 2021-03-18 16:01:17 train.py: 82] Epoch 6, iter 400/6416, lr 0.100000, loss 11.096907
+INFO 2021-03-18 16:02:35 train.py: 82] Epoch 6, iter 600/6416, lr 0.100000, loss 11.241163
+INFO 2021-03-18 16:03:53 train.py: 82] Epoch 6, iter 800/6416, lr 0.100000, loss 11.387452
+INFO 2021-03-18 16:05:10 train.py: 82] Epoch 6, iter 1000/6416, lr 0.100000, loss 11.475449
+INFO 2021-03-18 16:06:28 train.py: 82] Epoch 6, iter 1200/6416, lr 0.100000, loss 11.496409
+INFO 2021-03-18 16:07:46 train.py: 82] Epoch 6, iter 1400/6416, lr 0.100000, loss 11.596607
+INFO 2021-03-18 16:09:04 train.py: 82] Epoch 6, iter 1600/6416, lr 0.100000, loss 11.612384
+INFO 2021-03-18 16:10:22 train.py: 82] Epoch 6, iter 1800/6416, lr 0.100000, loss 11.674685
+INFO 2021-03-18 16:11:39 train.py: 82] Epoch 6, iter 2000/6416, lr 0.100000, loss 11.691938
+INFO 2021-03-18 16:12:57 train.py: 82] Epoch 6, iter 2200/6416, lr 0.100000, loss 11.703235
+INFO 2021-03-18 16:14:15 train.py: 82] Epoch 6, iter 2400/6416, lr 0.100000, loss 11.702025
+INFO 2021-03-18 16:15:32 train.py: 82] Epoch 6, iter 2600/6416, lr 0.100000, loss 11.820112
+INFO 2021-03-18 16:16:50 train.py: 82] Epoch 6, iter 2800/6416, lr 0.100000, loss 11.715783
+INFO 2021-03-18 16:18:08 train.py: 95] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2021-03-18 16:18:08 train.py: 82] Epoch 6, iter 3000/6416, lr 0.100000, loss 11.739233
+INFO 2021-03-18 16:19:26 train.py: 82] Epoch 6, iter 3200/6416, lr 0.100000, loss 11.709885
+INFO 2021-03-18 16:20:43 train.py: 82] Epoch 6, iter 3400/6416, lr 0.100000, loss 11.742900
+INFO 2021-03-18 16:22:00 train.py: 82] Epoch 6, iter 3600/6416, lr 0.100000, loss 11.794743
+INFO 2021-03-18 16:23:17 train.py: 82] Epoch 6, iter 3800/6416, lr 0.100000, loss 11.666821
+INFO 2021-03-18 16:24:34 train.py: 82] Epoch 6, iter 4000/6416, lr 0.100000, loss 11.737986
+INFO 2021-03-18 16:25:51 train.py: 82] Epoch 6, iter 4200/6416, lr 0.100000, loss 11.743747
+INFO 2021-03-18 16:27:08 train.py: 82] Epoch 6, iter 4400/6416, lr 0.100000, loss 11.706075
+INFO 2021-03-18 16:28:25 train.py: 82] Epoch 6, iter 4600/6416, lr 0.100000, loss 11.646327
+INFO 2021-03-18 16:29:43 train.py: 82] Epoch 6, iter 4800/6416, lr 0.100000, loss 11.780442
+INFO 2021-03-18 16:31:00 train.py: 82] Epoch 6, iter 5000/6416, lr 0.100000, loss 11.729160
+INFO 2021-03-18 16:32:17 train.py: 82] Epoch 6, iter 5200/6416, lr 0.100000, loss 11.708937
+INFO 2021-03-18 16:33:34 train.py: 82] Epoch 6, iter 5400/6416, lr 0.100000, loss 11.713377
+INFO 2021-03-18 16:34:51 train.py: 82] Epoch 6, iter 5600/6416, lr 0.100000, loss 11.694175
+INFO 2021-03-18 16:36:08 train.py: 82] Epoch 6, iter 5800/6416, lr 0.100000, loss 11.644919
+INFO 2021-03-18 16:37:25 train.py: 95] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2021-03-18 16:37:26 train.py: 82] Epoch 6, iter 6000/6416, lr 0.100000, loss 11.671565
+INFO 2021-03-18 16:38:44 train.py: 82] Epoch 6, iter 6200/6416, lr 0.100000, loss 11.660993
+INFO 2021-03-18 16:40:01 train.py: 82] Epoch 6, iter 6400/6416, lr 0.100000, loss 11.664161
+INFO 2021-03-18 16:40:08 train.py: 100] Save checkpoint Epoch_6.pt to disk...
+INFO 2021-03-18 16:40:10 train.py: 82] Epoch 7, iter 0/6416, lr 0.100000, loss 11.548596
+INFO 2021-03-18 16:41:28 train.py: 82] Epoch 7, iter 200/6416, lr 0.100000, loss 10.884601
+INFO 2021-03-18 16:42:46 train.py: 82] Epoch 7, iter 400/6416, lr 0.100000, loss 10.800916
+INFO 2021-03-18 16:44:04 train.py: 82] Epoch 7, iter 600/6416, lr 0.100000, loss 11.000092
+INFO 2021-03-18 16:45:21 train.py: 82] Epoch 7, iter 800/6416, lr 0.100000, loss 11.140886
+INFO 2021-03-18 16:46:39 train.py: 82] Epoch 7, iter 1000/6416, lr 0.100000, loss 11.198542
+INFO 2021-03-18 16:47:57 train.py: 82] Epoch 7, iter 1200/6416, lr 0.100000, loss 11.292385
+INFO 2021-03-18 16:49:15 train.py: 82] Epoch 7, iter 1400/6416, lr 0.100000, loss 11.354598
+INFO 2021-03-18 16:50:33 train.py: 82] Epoch 7, iter 1600/6416, lr 0.100000, loss 11.297116
+INFO 2021-03-18 16:51:51 train.py: 82] Epoch 7, iter 1800/6416, lr 0.100000, loss 11.410804
+INFO 2021-03-18 16:53:08 train.py: 82] Epoch 7, iter 2000/6416, lr 0.100000, loss 11.439772
+INFO 2021-03-18 16:54:26 train.py: 82] Epoch 7, iter 2200/6416, lr 0.100000, loss 11.412165
+INFO 2021-03-18 16:55:44 train.py: 82] Epoch 7, iter 2400/6416, lr 0.100000, loss 11.490373
+INFO 2021-03-18 16:57:02 train.py: 82] Epoch 7, iter 2600/6416, lr 0.100000, loss 11.492009
+INFO 2021-03-18 16:58:19 train.py: 82] Epoch 7, iter 2800/6416, lr 0.100000, loss 11.485011
+INFO 2021-03-18 16:59:37 train.py: 95] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2021-03-18 16:59:37 train.py: 82] Epoch 7, iter 3000/6416, lr 0.100000, loss 11.462375
+INFO 2021-03-18 17:00:55 train.py: 82] Epoch 7, iter 3200/6416, lr 0.100000, loss 11.499691
+INFO 2021-03-18 17:02:13 train.py: 82] Epoch 7, iter 3400/6416, lr 0.100000, loss 11.469779
+INFO 2021-03-18 17:03:31 train.py: 82] Epoch 7, iter 3600/6416, lr 0.100000, loss 11.465497
+INFO 2021-03-18 17:04:49 train.py: 82] Epoch 7, iter 3800/6416, lr 0.100000, loss 11.517237
+INFO 2021-03-18 17:06:06 train.py: 82] Epoch 7, iter 4000/6416, lr 0.100000, loss 11.446964
+INFO 2021-03-18 17:07:24 train.py: 82] Epoch 7, iter 4200/6416, lr 0.100000, loss 11.525941
+INFO 2021-03-18 17:08:42 train.py: 82] Epoch 7, iter 4400/6416, lr 0.100000, loss 11.519700
+INFO 2021-03-18 17:10:00 train.py: 82] Epoch 7, iter 4600/6416, lr 0.100000, loss 11.524622
+INFO 2021-03-18 17:11:18 train.py: 82] Epoch 7, iter 4800/6416, lr 0.100000, loss 11.491251
+INFO 2021-03-18 17:12:36 train.py: 82] Epoch 7, iter 5000/6416, lr 0.100000, loss 11.502971
+INFO 2021-03-18 17:13:53 train.py: 82] Epoch 7, iter 5200/6416, lr 0.100000, loss 11.419860
+INFO 2021-03-18 17:15:11 train.py: 82] Epoch 7, iter 5400/6416, lr 0.100000, loss 11.485724
+INFO 2021-03-18 17:16:29 train.py: 82] Epoch 7, iter 5600/6416, lr 0.100000, loss 11.468143
+INFO 2021-03-18 17:17:47 train.py: 82] Epoch 7, iter 5800/6416, lr 0.100000, loss 11.496728
+INFO 2021-03-18 17:19:05 train.py: 95] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2021-03-18 17:19:05 train.py: 82] Epoch 7, iter 6000/6416, lr 0.100000, loss 11.425808
+INFO 2021-03-18 17:20:23 train.py: 82] Epoch 7, iter 6200/6416, lr 0.100000, loss 11.472280
+INFO 2021-03-18 17:21:41 train.py: 82] Epoch 7, iter 6400/6416, lr 0.100000, loss 11.400418
+INFO 2021-03-18 17:21:47 train.py: 100] Save checkpoint Epoch_7.pt to disk...
+INFO 2021-03-18 17:21:49 train.py: 82] Epoch 8, iter 0/6416, lr 0.100000, loss 11.356438
+INFO 2021-03-18 17:23:07 train.py: 82] Epoch 8, iter 200/6416, lr 0.100000, loss 10.668524
+INFO 2021-03-18 17:24:25 train.py: 82] Epoch 8, iter 400/6416, lr 0.100000, loss 10.624209
+INFO 2021-03-18 17:25:43 train.py: 82] Epoch 8, iter 600/6416, lr 0.100000, loss 10.755585
+INFO 2021-03-18 17:27:01 train.py: 82] Epoch 8, iter 800/6416, lr 0.100000, loss 10.885146
+INFO 2021-03-18 17:28:19 train.py: 82] Epoch 8, iter 1000/6416, lr 0.100000, loss 11.020628
+INFO 2021-03-18 17:29:37 train.py: 82] Epoch 8, iter 1200/6416, lr 0.100000, loss 11.086525
+INFO 2021-03-18 17:30:54 train.py: 82] Epoch 8, iter 1400/6416, lr 0.100000, loss 11.131171
+INFO 2021-03-18 17:32:12 train.py: 82] Epoch 8, iter 1600/6416, lr 0.100000, loss 11.169323
+INFO 2021-03-18 17:33:30 train.py: 82] Epoch 8, iter 1800/6416, lr 0.100000, loss 11.188725
+INFO 2021-03-18 17:34:48 train.py: 82] Epoch 8, iter 2000/6416, lr 0.100000, loss 11.194794
+INFO 2021-03-18 17:36:05 train.py: 82] Epoch 8, iter 2200/6416, lr 0.100000, loss 11.230506
+INFO 2021-03-18 17:37:23 train.py: 82] Epoch 8, iter 2400/6416, lr 0.100000, loss 11.251766
+INFO 2021-03-18 17:38:41 train.py: 82] Epoch 8, iter 2600/6416, lr 0.100000, loss 11.289051
+INFO 2021-03-18 17:39:59 train.py: 82] Epoch 8, iter 2800/6416, lr 0.100000, loss 11.295398
+INFO 2021-03-18 17:41:16 train.py: 95] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2021-03-18 17:41:17 train.py: 82] Epoch 8, iter 3000/6416, lr 0.100000, loss 11.315131
+INFO 2021-03-18 17:42:34 train.py: 82] Epoch 8, iter 3200/6416, lr 0.100000, loss 11.303687
+INFO 2021-03-18 17:43:51 train.py: 82] Epoch 8, iter 3400/6416, lr 0.100000, loss 11.281648
+INFO 2021-03-18 17:45:08 train.py: 82] Epoch 8, iter 3600/6416, lr 0.100000, loss 11.300683
+INFO 2021-03-18 17:46:25 train.py: 82] Epoch 8, iter 3800/6416, lr 0.100000, loss 11.297300
+INFO 2021-03-18 17:47:43 train.py: 82] Epoch 8, iter 4000/6416, lr 0.100000, loss 11.327196
+INFO 2021-03-18 17:49:00 train.py: 82] Epoch 8, iter 4200/6416, lr 0.100000, loss 11.283500
+INFO 2021-03-18 17:50:17 train.py: 82] Epoch 8, iter 4400/6416, lr 0.100000, loss 11.298061
+INFO 2021-03-18 17:51:34 train.py: 82] Epoch 8, iter 4600/6416, lr 0.100000, loss 11.302732
+INFO 2021-03-18 17:52:51 train.py: 82] Epoch 8, iter 4800/6416, lr 0.100000, loss 11.311950
+INFO 2021-03-18 17:54:08 train.py: 82] Epoch 8, iter 5000/6416, lr 0.100000, loss 11.306220
+INFO 2021-03-18 17:55:25 train.py: 82] Epoch 8, iter 5200/6416, lr 0.100000, loss 11.303066
+INFO 2021-03-18 17:56:43 train.py: 82] Epoch 8, iter 5400/6416, lr 0.100000, loss 11.314745
+INFO 2021-03-18 17:58:00 train.py: 82] Epoch 8, iter 5600/6416, lr 0.100000, loss 11.342180
+INFO 2021-03-18 17:59:17 train.py: 82] Epoch 8, iter 5800/6416, lr 0.100000, loss 11.256254
+INFO 2021-03-18 18:00:34 train.py: 95] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2021-03-18 18:00:34 train.py: 82] Epoch 8, iter 6000/6416, lr 0.100000, loss 11.251138
+INFO 2021-03-18 18:01:52 train.py: 82] Epoch 8, iter 6200/6416, lr 0.100000, loss 11.217863
+INFO 2021-03-18 18:03:10 train.py: 82] Epoch 8, iter 6400/6416, lr 0.100000, loss 11.227782
+INFO 2021-03-18 18:03:17 train.py: 100] Save checkpoint Epoch_8.pt to disk...
+INFO 2021-03-18 18:03:18 train.py: 82] Epoch 9, iter 0/6416, lr 0.100000, loss 11.146408
+INFO 2021-03-18 18:04:36 train.py: 82] Epoch 9, iter 200/6416, lr 0.100000, loss 10.455902
+INFO 2021-03-18 18:05:55 train.py: 82] Epoch 9, iter 400/6416, lr 0.100000, loss 10.473977
+INFO 2021-03-18 18:07:12 train.py: 82] Epoch 9, iter 600/6416, lr 0.100000, loss 10.613014
+INFO 2021-03-18 18:08:30 train.py: 82] Epoch 9, iter 800/6416, lr 0.100000, loss 10.681427
+INFO 2021-03-18 18:09:48 train.py: 82] Epoch 9, iter 1000/6416, lr 0.100000, loss 10.811079
+INFO 2021-03-18 18:11:06 train.py: 82] Epoch 9, iter 1200/6416, lr 0.100000, loss 10.868071
+INFO 2021-03-18 18:12:24 train.py: 82] Epoch 9, iter 1400/6416, lr 0.100000, loss 10.977432
+INFO 2021-03-18 18:13:41 train.py: 82] Epoch 9, iter 1600/6416, lr 0.100000, loss 11.016141
+INFO 2021-03-18 18:14:59 train.py: 82] Epoch 9, iter 1800/6416, lr 0.100000, loss 11.076815
+INFO 2021-03-18 18:16:17 train.py: 82] Epoch 9, iter 2000/6416, lr 0.100000, loss 11.043664
+INFO 2021-03-18 18:17:35 train.py: 82] Epoch 9, iter 2200/6416, lr 0.100000, loss 11.096883
+INFO 2021-03-18 18:18:52 train.py: 82] Epoch 9, iter 2400/6416, lr 0.100000, loss 11.100446
+INFO 2021-03-18 18:20:10 train.py: 82] Epoch 9, iter 2600/6416, lr 0.100000, loss 11.087793
+INFO 2021-03-18 18:21:28 train.py: 82] Epoch 9, iter 2800/6416, lr 0.100000, loss 11.107951
+INFO 2021-03-18 18:22:46 train.py: 95] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2021-03-18 18:22:46 train.py: 82] Epoch 9, iter 3000/6416, lr 0.100000, loss 11.051496
+INFO 2021-03-18 18:24:03 train.py: 82] Epoch 9, iter 3200/6416, lr 0.100000, loss 11.108010
+INFO 2021-03-18 18:25:20 train.py: 82] Epoch 9, iter 3400/6416, lr 0.100000, loss 11.123284
+INFO 2021-03-18 18:26:38 train.py: 82] Epoch 9, iter 3600/6416, lr 0.100000, loss 11.198730
+INFO 2021-03-18 18:27:55 train.py: 82] Epoch 9, iter 3800/6416, lr 0.100000, loss 11.175144
+INFO 2021-03-18 18:29:12 train.py: 82] Epoch 9, iter 4000/6416, lr 0.100000, loss 11.097541
+INFO 2021-03-18 18:30:29 train.py: 82] Epoch 9, iter 4200/6416, lr 0.100000, loss 11.098678
+INFO 2021-03-18 18:31:46 train.py: 82] Epoch 9, iter 4400/6416, lr 0.100000, loss 11.141439
+INFO 2021-03-18 18:33:03 train.py: 82] Epoch 9, iter 4600/6416, lr 0.100000, loss 11.072770
+INFO 2021-03-18 18:34:21 train.py: 82] Epoch 9, iter 4800/6416, lr 0.100000, loss 11.126307
+INFO 2021-03-18 18:35:38 train.py: 82] Epoch 9, iter 5000/6416, lr 0.100000, loss 11.134760
+INFO 2021-03-18 18:36:55 train.py: 82] Epoch 9, iter 5200/6416, lr 0.100000, loss 11.117253
+INFO 2021-03-18 18:38:12 train.py: 82] Epoch 9, iter 5400/6416, lr 0.100000, loss 11.107631
+INFO 2021-03-18 18:39:29 train.py: 82] Epoch 9, iter 5600/6416, lr 0.100000, loss 11.139527
+INFO 2021-03-18 18:40:46 train.py: 82] Epoch 9, iter 5800/6416, lr 0.100000, loss 11.119905
+INFO 2021-03-18 18:42:03 train.py: 95] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2021-03-18 18:42:04 train.py: 82] Epoch 9, iter 6000/6416, lr 0.100000, loss 11.158419
+INFO 2021-03-18 18:43:22 train.py: 82] Epoch 9, iter 6200/6416, lr 0.100000, loss 11.108142
+INFO 2021-03-18 18:44:39 train.py: 82] Epoch 9, iter 6400/6416, lr 0.100000, loss 11.110868
+INFO 2021-03-18 18:44:46 train.py: 100] Save checkpoint Epoch_9.pt to disk...
+INFO 2021-03-18 18:44:48 train.py: 82] Epoch 10, iter 0/6416, lr 0.010000, loss 11.247077
+INFO 2021-03-18 18:46:06 train.py: 82] Epoch 10, iter 200/6416, lr 0.010000, loss 9.202994
+INFO 2021-03-18 18:47:24 train.py: 82] Epoch 10, iter 400/6416, lr 0.010000, loss 8.855631
+INFO 2021-03-18 18:48:42 train.py: 82] Epoch 10, iter 600/6416, lr 0.010000, loss 8.674940
+INFO 2021-03-18 18:50:00 train.py: 82] Epoch 10, iter 800/6416, lr 0.010000, loss 8.603190
+INFO 2021-03-18 18:51:18 train.py: 82] Epoch 10, iter 1000/6416, lr 0.010000, loss 8.471477
+INFO 2021-03-18 18:52:35 train.py: 82] Epoch 10, iter 1200/6416, lr 0.010000, loss 8.397812
+INFO 2021-03-18 18:53:53 train.py: 82] Epoch 10, iter 1400/6416, lr 0.010000, loss 8.373116
+INFO 2021-03-18 18:55:11 train.py: 82] Epoch 10, iter 1600/6416, lr 0.010000, loss 8.282594
+INFO 2021-03-18 18:56:29 train.py: 82] Epoch 10, iter 1800/6416, lr 0.010000, loss 8.258688
+INFO 2021-03-18 18:57:46 train.py: 82] Epoch 10, iter 2000/6416, lr 0.010000, loss 8.221287
+INFO 2021-03-18 18:59:04 train.py: 82] Epoch 10, iter 2200/6416, lr 0.010000, loss 8.147146
+INFO 2021-03-18 19:00:22 train.py: 82] Epoch 10, iter 2400/6416, lr 0.010000, loss 8.135890
+INFO 2021-03-18 19:01:40 train.py: 82] Epoch 10, iter 2600/6416, lr 0.010000, loss 8.067604
+INFO 2021-03-18 19:02:57 train.py: 82] Epoch 10, iter 2800/6416, lr 0.010000, loss 8.071922
+INFO 2021-03-18 19:04:15 train.py: 95] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2021-03-18 19:04:15 train.py: 82] Epoch 10, iter 3000/6416, lr 0.010000, loss 8.021073
+INFO 2021-03-18 19:05:33 train.py: 82] Epoch 10, iter 3200/6416, lr 0.010000, loss 7.991662
+INFO 2021-03-18 19:06:50 train.py: 82] Epoch 10, iter 3400/6416, lr 0.010000, loss 7.960448
+INFO 2021-03-18 19:08:07 train.py: 82] Epoch 10, iter 3600/6416, lr 0.010000, loss 7.870281
+INFO 2021-03-18 19:09:24 train.py: 82] Epoch 10, iter 3800/6416, lr 0.010000, loss 7.997557
+INFO 2021-03-18 19:10:41 train.py: 82] Epoch 10, iter 4000/6416, lr 0.010000, loss 7.901649
+INFO 2021-03-18 19:11:58 train.py: 82] Epoch 10, iter 4200/6416, lr 0.010000, loss 7.827317
+INFO 2021-03-18 19:13:16 train.py: 82] Epoch 10, iter 4400/6416, lr 0.010000, loss 7.820432
+INFO 2021-03-18 19:14:33 train.py: 82] Epoch 10, iter 4600/6416, lr 0.010000, loss 7.782658
+INFO 2021-03-18 19:15:50 train.py: 82] Epoch 10, iter 4800/6416, lr 0.010000, loss 7.811718
+INFO 2021-03-18 19:17:07 train.py: 82] Epoch 10, iter 5000/6416, lr 0.010000, loss 7.745914
+INFO 2021-03-18 19:18:24 train.py: 82] Epoch 10, iter 5200/6416, lr 0.010000, loss 7.739341
+INFO 2021-03-18 19:19:41 train.py: 82] Epoch 10, iter 5400/6416, lr 0.010000, loss 7.730650
+INFO 2021-03-18 19:20:58 train.py: 82] Epoch 10, iter 5600/6416, lr 0.010000, loss 7.752667
+INFO 2021-03-18 19:22:16 train.py: 82] Epoch 10, iter 5800/6416, lr 0.010000, loss 7.690258
+INFO 2021-03-18 19:23:33 train.py: 95] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2021-03-18 19:23:33 train.py: 82] Epoch 10, iter 6000/6416, lr 0.010000, loss 7.641223
+INFO 2021-03-18 19:24:51 train.py: 82] Epoch 10, iter 6200/6416, lr 0.010000, loss 7.657668
+INFO 2021-03-18 19:26:09 train.py: 82] Epoch 10, iter 6400/6416, lr 0.010000, loss 7.678712
+INFO 2021-03-18 19:26:15 train.py: 100] Save checkpoint Epoch_10.pt to disk...
+INFO 2021-03-18 19:26:17 train.py: 82] Epoch 11, iter 0/6416, lr 0.010000, loss 7.579871
+INFO 2021-03-18 19:27:35 train.py: 82] Epoch 11, iter 200/6416, lr 0.010000, loss 7.064480
+INFO 2021-03-18 19:28:53 train.py: 82] Epoch 11, iter 400/6416, lr 0.010000, loss 7.041594
+INFO 2021-03-18 19:30:11 train.py: 82] Epoch 11, iter 600/6416, lr 0.010000, loss 7.089375
+INFO 2021-03-18 19:31:29 train.py: 82] Epoch 11, iter 800/6416, lr 0.010000, loss 7.069124
+INFO 2021-03-18 19:32:47 train.py: 82] Epoch 11, iter 1000/6416, lr 0.010000, loss 7.128865
+INFO 2021-03-18 19:34:05 train.py: 82] Epoch 11, iter 1200/6416, lr 0.010000, loss 7.087839
+INFO 2021-03-18 19:35:23 train.py: 82] Epoch 11, iter 1400/6416, lr 0.010000, loss 7.089903
+INFO 2021-03-18 19:36:40 train.py: 82] Epoch 11, iter 1600/6416, lr 0.010000, loss 7.099175
+INFO 2021-03-18 19:37:58 train.py: 82] Epoch 11, iter 1800/6416, lr 0.010000, loss 7.120688
+INFO 2021-03-18 19:39:16 train.py: 82] Epoch 11, iter 2000/6416, lr 0.010000, loss 7.148376
+INFO 2021-03-18 19:40:34 train.py: 82] Epoch 11, iter 2200/6416, lr 0.010000, loss 7.141556
+INFO 2021-03-18 19:41:51 train.py: 82] Epoch 11, iter 2400/6416, lr 0.010000, loss 7.137837
+INFO 2021-03-18 19:43:09 train.py: 82] Epoch 11, iter 2600/6416, lr 0.010000, loss 7.167497
+INFO 2021-03-18 19:44:27 train.py: 82] Epoch 11, iter 2800/6416, lr 0.010000, loss 7.154820
+INFO 2021-03-18 19:45:45 train.py: 95] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2021-03-18 19:45:45 train.py: 82] Epoch 11, iter 3000/6416, lr 0.010000, loss 7.128011
+INFO 2021-03-18 19:47:02 train.py: 82] Epoch 11, iter 3200/6416, lr 0.010000, loss 7.153761
+INFO 2021-03-18 19:48:19 train.py: 82] Epoch 11, iter 3400/6416, lr 0.010000, loss 7.167526
+INFO 2021-03-18 19:49:37 train.py: 82] Epoch 11, iter 3600/6416, lr 0.010000, loss 7.177234
+INFO 2021-03-18 19:50:54 train.py: 82] Epoch 11, iter 3800/6416, lr 0.010000, loss 7.151658
+INFO 2021-03-18 19:52:11 train.py: 82] Epoch 11, iter 4000/6416, lr 0.010000, loss 7.101844
+INFO 2021-03-18 19:53:28 train.py: 82] Epoch 11, iter 4200/6416, lr 0.010000, loss 7.165015
+INFO 2021-03-18 19:54:45 train.py: 82] Epoch 11, iter 4400/6416, lr 0.010000, loss 7.141874
+INFO 2021-03-18 19:56:02 train.py: 82] Epoch 11, iter 4600/6416, lr 0.010000, loss 7.160193
+INFO 2021-03-18 19:57:19 train.py: 82] Epoch 11, iter 4800/6416, lr 0.010000, loss 7.184668
+INFO 2021-03-18 19:58:37 train.py: 82] Epoch 11, iter 5000/6416, lr 0.010000, loss 7.174349
+INFO 2021-03-18 19:59:54 train.py: 82] Epoch 11, iter 5200/6416, lr 0.010000, loss 7.169127
+INFO 2021-03-18 20:01:11 train.py: 82] Epoch 11, iter 5400/6416, lr 0.010000, loss 7.165051
+INFO 2021-03-18 20:02:28 train.py: 82] Epoch 11, iter 5600/6416, lr 0.010000, loss 7.170341
+INFO 2021-03-18 20:03:45 train.py: 82] Epoch 11, iter 5800/6416, lr 0.010000, loss 7.162791
+INFO 2021-03-18 20:05:02 train.py: 95] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2021-03-18 20:05:03 train.py: 82] Epoch 11, iter 6000/6416, lr 0.010000, loss 7.197751
+INFO 2021-03-18 20:06:21 train.py: 82] Epoch 11, iter 6200/6416, lr 0.010000, loss 7.196192
+INFO 2021-03-18 20:07:38 train.py: 82] Epoch 11, iter 6400/6416, lr 0.010000, loss 7.195508
+INFO 2021-03-18 20:07:45 train.py: 100] Save checkpoint Epoch_11.pt to disk...
+INFO 2021-03-18 20:07:47 train.py: 82] Epoch 12, iter 0/6416, lr 0.010000, loss 7.139786
+INFO 2021-03-18 20:09:05 train.py: 82] Epoch 12, iter 200/6416, lr 0.010000, loss 6.700989
+INFO 2021-03-18 20:10:23 train.py: 82] Epoch 12, iter 400/6416, lr 0.010000, loss 6.654858
+INFO 2021-03-18 20:11:41 train.py: 82] Epoch 12, iter 600/6416, lr 0.010000, loss 6.640231
+INFO 2021-03-18 20:12:59 train.py: 82] Epoch 12, iter 800/6416, lr 0.010000, loss 6.709312
+INFO 2021-03-18 20:14:17 train.py: 82] Epoch 12, iter 1000/6416, lr 0.010000, loss 6.711276
+INFO 2021-03-18 20:15:35 train.py: 82] Epoch 12, iter 1200/6416, lr 0.010000, loss 6.755749
+INFO 2021-03-18 20:16:53 train.py: 82] Epoch 12, iter 1400/6416, lr 0.010000, loss 6.746229
+INFO 2021-03-18 20:18:10 train.py: 82] Epoch 12, iter 1600/6416, lr 0.010000, loss 6.763264
+INFO 2021-03-18 20:19:28 train.py: 82] Epoch 12, iter 1800/6416, lr 0.010000, loss 6.760377
+INFO 2021-03-18 20:20:46 train.py: 82] Epoch 12, iter 2000/6416, lr 0.010000, loss 6.779121
+INFO 2021-03-18 20:22:04 train.py: 82] Epoch 12, iter 2200/6416, lr 0.010000, loss 6.800692
+INFO 2021-03-18 20:23:22 train.py: 82] Epoch 12, iter 2400/6416, lr 0.010000, loss 6.848898
+INFO 2021-03-18 20:24:39 train.py: 82] Epoch 12, iter 2600/6416, lr 0.010000, loss 6.799859
+INFO 2021-03-18 20:25:57 train.py: 82] Epoch 12, iter 2800/6416, lr 0.010000, loss 6.855125
+INFO 2021-03-18 20:27:15 train.py: 95] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2021-03-18 20:27:15 train.py: 82] Epoch 12, iter 3000/6416, lr 0.010000, loss 6.870399
+INFO 2021-03-18 20:28:33 train.py: 82] Epoch 12, iter 3200/6416, lr 0.010000, loss 6.816895
+INFO 2021-03-18 20:29:50 train.py: 82] Epoch 12, iter 3400/6416, lr 0.010000, loss 6.903950
+INFO 2021-03-18 20:31:07 train.py: 82] Epoch 12, iter 3600/6416, lr 0.010000, loss 6.873216
+INFO 2021-03-18 20:32:25 train.py: 82] Epoch 12, iter 3800/6416, lr 0.010000, loss 6.942008
+INFO 2021-03-18 20:33:42 train.py: 82] Epoch 12, iter 4000/6416, lr 0.010000, loss 6.938401
+INFO 2021-03-18 20:34:59 train.py: 82] Epoch 12, iter 4200/6416, lr 0.010000, loss 6.961632
+INFO 2021-03-18 20:36:17 train.py: 82] Epoch 12, iter 4400/6416, lr 0.010000, loss 6.955409
+INFO 2021-03-18 20:37:34 train.py: 82] Epoch 12, iter 4600/6416, lr 0.010000, loss 6.977312
+INFO 2021-03-18 20:38:51 train.py: 82] Epoch 12, iter 4800/6416, lr 0.010000, loss 6.997378
+INFO 2021-03-18 20:40:08 train.py: 82] Epoch 12, iter 5000/6416, lr 0.010000, loss 6.972436
+INFO 2021-03-18 20:41:26 train.py: 82] Epoch 12, iter 5200/6416, lr 0.010000, loss 6.983730
+INFO 2021-03-18 20:42:43 train.py: 82] Epoch 12, iter 5400/6416, lr 0.010000, loss 7.026159
+INFO 2021-03-18 20:44:00 train.py: 82] Epoch 12, iter 5600/6416, lr 0.010000, loss 7.007557
+INFO 2021-03-18 20:45:18 train.py: 82] Epoch 12, iter 5800/6416, lr 0.010000, loss 7.052989
+INFO 2021-03-18 20:46:35 train.py: 95] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2021-03-18 20:46:35 train.py: 82] Epoch 12, iter 6000/6416, lr 0.010000, loss 7.006707
+INFO 2021-03-18 20:47:53 train.py: 82] Epoch 12, iter 6200/6416, lr 0.010000, loss 7.045581
+INFO 2021-03-18 20:49:11 train.py: 82] Epoch 12, iter 6400/6416, lr 0.010000, loss 7.016183
+INFO 2021-03-18 20:49:18 train.py: 100] Save checkpoint Epoch_12.pt to disk...
+INFO 2021-03-18 20:49:20 train.py: 82] Epoch 13, iter 0/6416, lr 0.001000, loss 6.984809
+INFO 2021-03-18 20:50:38 train.py: 82] Epoch 13, iter 200/6416, lr 0.001000, loss 6.380268
+INFO 2021-03-18 20:51:56 train.py: 82] Epoch 13, iter 400/6416, lr 0.001000, loss 6.373197
+INFO 2021-03-18 20:53:14 train.py: 82] Epoch 13, iter 600/6416, lr 0.001000, loss 6.333324
+INFO 2021-03-18 20:54:32 train.py: 82] Epoch 13, iter 800/6416, lr 0.001000, loss 6.309630
+INFO 2021-03-18 20:55:50 train.py: 82] Epoch 13, iter 1000/6416, lr 0.001000, loss 6.282385
+INFO 2021-03-18 20:57:07 train.py: 82] Epoch 13, iter 1200/6416, lr 0.001000, loss 6.310199
+INFO 2021-03-18 20:58:25 train.py: 82] Epoch 13, iter 1400/6416, lr 0.001000, loss 6.291609
+INFO 2021-03-18 20:59:43 train.py: 82] Epoch 13, iter 1600/6416, lr 0.001000, loss 6.287858
+INFO 2021-03-18 21:01:01 train.py: 82] Epoch 13, iter 1800/6416, lr 0.001000, loss 6.277865
+INFO 2021-03-18 21:02:19 train.py: 82] Epoch 13, iter 2000/6416, lr 0.001000, loss 6.270192
+INFO 2021-03-18 21:03:37 train.py: 82] Epoch 13, iter 2200/6416, lr 0.001000, loss 6.317798
+INFO 2021-03-18 21:04:55 train.py: 82] Epoch 13, iter 2400/6416, lr 0.001000, loss 6.248097
+INFO 2021-03-18 21:06:12 train.py: 82] Epoch 13, iter 2600/6416, lr 0.001000, loss 6.273139
+INFO 2021-03-18 21:07:30 train.py: 82] Epoch 13, iter 2800/6416, lr 0.001000, loss 6.302427
+INFO 2021-03-18 21:08:48 train.py: 95] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2021-03-18 21:08:48 train.py: 82] Epoch 13, iter 3000/6416, lr 0.001000, loss 6.289605
+INFO 2021-03-18 21:10:06 train.py: 82] Epoch 13, iter 3200/6416, lr 0.001000, loss 6.260822
+INFO 2021-03-18 21:11:24 train.py: 82] Epoch 13, iter 3400/6416, lr 0.001000, loss 6.277920
+INFO 2021-03-18 21:12:42 train.py: 82] Epoch 13, iter 3600/6416, lr 0.001000, loss 6.307444
+INFO 2021-03-18 21:13:59 train.py: 82] Epoch 13, iter 3800/6416, lr 0.001000, loss 6.283028
+INFO 2021-03-18 21:15:17 train.py: 82] Epoch 13, iter 4000/6416, lr 0.001000, loss 6.303678
+INFO 2021-03-18 21:16:35 train.py: 82] Epoch 13, iter 4200/6416, lr 0.001000, loss 6.294402
+INFO 2021-03-18 21:17:53 train.py: 82] Epoch 13, iter 4400/6416, lr 0.001000, loss 6.315266
+INFO 2021-03-18 21:19:11 train.py: 82] Epoch 13, iter 4600/6416, lr 0.001000, loss 6.255787
+INFO 2021-03-18 21:20:29 train.py: 82] Epoch 13, iter 4800/6416, lr 0.001000, loss 6.297829
+INFO 2021-03-18 21:21:46 train.py: 82] Epoch 13, iter 5000/6416, lr 0.001000, loss 6.317875
+INFO 2021-03-18 21:23:04 train.py: 82] Epoch 13, iter 5200/6416, lr 0.001000, loss 6.318076
+INFO 2021-03-18 21:24:22 train.py: 82] Epoch 13, iter 5400/6416, lr 0.001000, loss 6.303325
+INFO 2021-03-18 21:25:40 train.py: 82] Epoch 13, iter 5600/6416, lr 0.001000, loss 6.261842
+INFO 2021-03-18 21:26:58 train.py: 82] Epoch 13, iter 5800/6416, lr 0.001000, loss 6.296454
+INFO 2021-03-18 21:28:16 train.py: 95] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2021-03-18 21:28:16 train.py: 82] Epoch 13, iter 6000/6416, lr 0.001000, loss 6.289716
+INFO 2021-03-18 21:29:33 train.py: 82] Epoch 13, iter 6200/6416, lr 0.001000, loss 6.272198
+INFO 2021-03-18 21:30:50 train.py: 82] Epoch 13, iter 6400/6416, lr 0.001000, loss 6.302922
+INFO 2021-03-18 21:30:57 train.py: 100] Save checkpoint Epoch_13.pt to disk...
+INFO 2021-03-18 21:30:59 train.py: 82] Epoch 14, iter 0/6416, lr 0.001000, loss 6.306069
+INFO 2021-03-18 21:32:17 train.py: 82] Epoch 14, iter 200/6416, lr 0.001000, loss 6.192837
+INFO 2021-03-18 21:33:35 train.py: 82] Epoch 14, iter 400/6416, lr 0.001000, loss 6.170409
+INFO 2021-03-18 21:34:53 train.py: 82] Epoch 14, iter 600/6416, lr 0.001000, loss 6.239060
+INFO 2021-03-18 21:36:11 train.py: 82] Epoch 14, iter 800/6416, lr 0.001000, loss 6.225698
+INFO 2021-03-18 21:37:29 train.py: 82] Epoch 14, iter 1000/6416, lr 0.001000, loss 6.184798
+INFO 2021-03-18 21:38:47 train.py: 82] Epoch 14, iter 1200/6416, lr 0.001000, loss 6.231239
+INFO 2021-03-18 21:40:04 train.py: 82] Epoch 14, iter 1400/6416, lr 0.001000, loss 6.231826
+INFO 2021-03-18 21:41:22 train.py: 82] Epoch 14, iter 1600/6416, lr 0.001000, loss 6.256388
+INFO 2021-03-18 21:42:40 train.py: 82] Epoch 14, iter 1800/6416, lr 0.001000, loss 6.221277
+INFO 2021-03-18 21:43:58 train.py: 82] Epoch 14, iter 2000/6416, lr 0.001000, loss 6.236577
+INFO 2021-03-18 21:45:16 train.py: 82] Epoch 14, iter 2200/6416, lr 0.001000, loss 6.236276
+INFO 2021-03-18 21:46:33 train.py: 82] Epoch 14, iter 2400/6416, lr 0.001000, loss 6.238536
+INFO 2021-03-18 21:47:51 train.py: 82] Epoch 14, iter 2600/6416, lr 0.001000, loss 6.205701
+INFO 2021-03-18 21:49:09 train.py: 82] Epoch 14, iter 2800/6416, lr 0.001000, loss 6.254074
+INFO 2021-03-18 21:50:27 train.py: 95] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2021-03-18 21:50:27 train.py: 82] Epoch 14, iter 3000/6416, lr 0.001000, loss 6.272589
+INFO 2021-03-18 21:51:45 train.py: 82] Epoch 14, iter 3200/6416, lr 0.001000, loss 6.184874
+INFO 2021-03-18 21:53:03 train.py: 82] Epoch 14, iter 3400/6416, lr 0.001000, loss 6.264461
+INFO 2021-03-18 21:54:20 train.py: 82] Epoch 14, iter 3600/6416, lr 0.001000, loss 6.254376
+INFO 2021-03-18 21:55:38 train.py: 82] Epoch 14, iter 3800/6416, lr 0.001000, loss 6.252648
+INFO 2021-03-18 21:56:56 train.py: 82] Epoch 14, iter 4000/6416, lr 0.001000, loss 6.235493
+INFO 2021-03-18 21:58:14 train.py: 82] Epoch 14, iter 4200/6416, lr 0.001000, loss 6.260520
+INFO 2021-03-18 21:59:32 train.py: 82] Epoch 14, iter 4400/6416, lr 0.001000, loss 6.257240
+INFO 2021-03-18 22:00:50 train.py: 82] Epoch 14, iter 4600/6416, lr 0.001000, loss 6.242652
+INFO 2021-03-18 22:02:08 train.py: 82] Epoch 14, iter 4800/6416, lr 0.001000, loss 6.276694
+INFO 2021-03-18 22:03:25 train.py: 82] Epoch 14, iter 5000/6416, lr 0.001000, loss 6.261944
+INFO 2021-03-18 22:04:43 train.py: 82] Epoch 14, iter 5200/6416, lr 0.001000, loss 6.289712
+INFO 2021-03-18 22:06:01 train.py: 82] Epoch 14, iter 5400/6416, lr 0.001000, loss 6.226915
+INFO 2021-03-18 22:07:19 train.py: 82] Epoch 14, iter 5600/6416, lr 0.001000, loss 6.299423
+INFO 2021-03-18 22:08:37 train.py: 82] Epoch 14, iter 5800/6416, lr 0.001000, loss 6.261087
+INFO 2021-03-18 22:09:55 train.py: 95] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2021-03-18 22:09:55 train.py: 82] Epoch 14, iter 6000/6416, lr 0.001000, loss 6.271331
+INFO 2021-03-18 22:11:13 train.py: 82] Epoch 14, iter 6200/6416, lr 0.001000, loss 6.283215
+INFO 2021-03-18 22:12:31 train.py: 82] Epoch 14, iter 6400/6416, lr 0.001000, loss 6.227340
+INFO 2021-03-18 22:12:38 train.py: 100] Save checkpoint Epoch_14.pt to disk...
+INFO 2021-03-18 22:12:39 train.py: 82] Epoch 15, iter 0/6416, lr 0.001000, loss 6.216057
+INFO 2021-03-18 22:13:57 train.py: 82] Epoch 15, iter 200/6416, lr 0.001000, loss 6.177841
+INFO 2021-03-18 22:15:14 train.py: 82] Epoch 15, iter 400/6416, lr 0.001000, loss 6.174004
+INFO 2021-03-18 22:16:32 train.py: 82] Epoch 15, iter 600/6416, lr 0.001000, loss 6.182670
+INFO 2021-03-18 22:17:49 train.py: 82] Epoch 15, iter 800/6416, lr 0.001000, loss 6.184497
+INFO 2021-03-18 22:19:06 train.py: 82] Epoch 15, iter 1000/6416, lr 0.001000, loss 6.167211
+INFO 2021-03-18 22:20:23 train.py: 82] Epoch 15, iter 1200/6416, lr 0.001000, loss 6.185973
+INFO 2021-03-18 22:21:40 train.py: 82] Epoch 15, iter 1400/6416, lr 0.001000, loss 6.221166
+INFO 2021-03-18 22:22:58 train.py: 82] Epoch 15, iter 1600/6416, lr 0.001000, loss 6.189931
+INFO 2021-03-18 22:24:15 train.py: 82] Epoch 15, iter 1800/6416, lr 0.001000, loss 6.232997
+INFO 2021-03-18 22:25:32 train.py: 82] Epoch 15, iter 2000/6416, lr 0.001000, loss 6.231943
+INFO 2021-03-18 22:26:49 train.py: 82] Epoch 15, iter 2200/6416, lr 0.001000, loss 6.246964
+INFO 2021-03-18 22:28:06 train.py: 82] Epoch 15, iter 2400/6416, lr 0.001000, loss 6.214535
+INFO 2021-03-18 22:29:24 train.py: 82] Epoch 15, iter 2600/6416, lr 0.001000, loss 6.214586
+INFO 2021-03-18 22:30:41 train.py: 82] Epoch 15, iter 2800/6416, lr 0.001000, loss 6.164661
+INFO 2021-03-18 22:31:58 train.py: 95] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2021-03-18 22:31:58 train.py: 82] Epoch 15, iter 3000/6416, lr 0.001000, loss 6.210753
+INFO 2021-03-18 22:33:16 train.py: 82] Epoch 15, iter 3200/6416, lr 0.001000, loss 6.234213
+INFO 2021-03-18 22:34:34 train.py: 82] Epoch 15, iter 3400/6416, lr 0.001000, loss 6.211773
+INFO 2021-03-18 22:35:52 train.py: 82] Epoch 15, iter 3600/6416, lr 0.001000, loss 6.205386
+INFO 2021-03-18 22:37:10 train.py: 82] Epoch 15, iter 3800/6416, lr 0.001000, loss 6.233970
+INFO 2021-03-18 22:38:27 train.py: 82] Epoch 15, iter 4000/6416, lr 0.001000, loss 6.207277
+INFO 2021-03-18 22:39:45 train.py: 82] Epoch 15, iter 4200/6416, lr 0.001000, loss 6.229234
+INFO 2021-03-18 22:41:03 train.py: 82] Epoch 15, iter 4400/6416, lr 0.001000, loss 6.201804
+INFO 2021-03-18 22:42:21 train.py: 82] Epoch 15, iter 4600/6416, lr 0.001000, loss 6.214722
+INFO 2021-03-18 22:43:39 train.py: 82] Epoch 15, iter 4800/6416, lr 0.001000, loss 6.241902
+INFO 2021-03-18 22:44:57 train.py: 82] Epoch 15, iter 5000/6416, lr 0.001000, loss 6.204927
+INFO 2021-03-18 22:46:14 train.py: 82] Epoch 15, iter 5200/6416, lr 0.001000, loss 6.223426
+INFO 2021-03-18 22:47:32 train.py: 82] Epoch 15, iter 5400/6416, lr 0.001000, loss 6.236838
+INFO 2021-03-18 22:48:50 train.py: 82] Epoch 15, iter 5600/6416, lr 0.001000, loss 6.220132
+INFO 2021-03-18 22:50:08 train.py: 82] Epoch 15, iter 5800/6416, lr 0.001000, loss 6.278170
+INFO 2021-03-18 22:51:26 train.py: 95] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2021-03-18 22:51:26 train.py: 82] Epoch 15, iter 6000/6416, lr 0.001000, loss 6.200431
+INFO 2021-03-18 22:52:44 train.py: 82] Epoch 15, iter 6200/6416, lr 0.001000, loss 6.244218
+INFO 2021-03-18 22:54:02 train.py: 82] Epoch 15, iter 6400/6416, lr 0.001000, loss 6.226380
+INFO 2021-03-18 22:54:08 train.py: 100] Save checkpoint Epoch_15.pt to disk...
+INFO 2021-03-18 22:54:10 train.py: 82] Epoch 16, iter 0/6416, lr 0.000100, loss 6.172353
+INFO 2021-03-18 22:55:28 train.py: 82] Epoch 16, iter 200/6416, lr 0.000100, loss 6.137570
+INFO 2021-03-18 22:56:46 train.py: 82] Epoch 16, iter 400/6416, lr 0.000100, loss 6.113624
+INFO 2021-03-18 22:58:04 train.py: 82] Epoch 16, iter 600/6416, lr 0.000100, loss 6.157844
+INFO 2021-03-18 22:59:22 train.py: 82] Epoch 16, iter 800/6416, lr 0.000100, loss 6.151621
+INFO 2021-03-18 23:00:40 train.py: 82] Epoch 16, iter 1000/6416, lr 0.000100, loss 6.119920
+INFO 2021-03-18 23:01:58 train.py: 82] Epoch 16, iter 1200/6416, lr 0.000100, loss 6.114026
+INFO 2021-03-18 23:03:16 train.py: 82] Epoch 16, iter 1400/6416, lr 0.000100, loss 6.155079
+INFO 2021-03-18 23:04:33 train.py: 82] Epoch 16, iter 1600/6416, lr 0.000100, loss 6.175717
+INFO 2021-03-18 23:05:51 train.py: 82] Epoch 16, iter 1800/6416, lr 0.000100, loss 6.168410
+INFO 2021-03-18 23:07:09 train.py: 82] Epoch 16, iter 2000/6416, lr 0.000100, loss 6.145844
+INFO 2021-03-18 23:08:27 train.py: 82] Epoch 16, iter 2200/6416, lr 0.000100, loss 6.104518
+INFO 2021-03-18 23:09:44 train.py: 82] Epoch 16, iter 2400/6416, lr 0.000100, loss 6.146550
+INFO 2021-03-18 23:11:02 train.py: 82] Epoch 16, iter 2600/6416, lr 0.000100, loss 6.132901
+INFO 2021-03-18 23:12:20 train.py: 82] Epoch 16, iter 2800/6416, lr 0.000100, loss 6.138385
+INFO 2021-03-18 23:13:38 train.py: 95] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2021-03-18 23:13:38 train.py: 82] Epoch 16, iter 3000/6416, lr 0.000100, loss 6.150673
+INFO 2021-03-18 23:14:56 train.py: 82] Epoch 16, iter 3200/6416, lr 0.000100, loss 6.145843
+INFO 2021-03-18 23:16:13 train.py: 82] Epoch 16, iter 3400/6416, lr 0.000100, loss 6.094719
+INFO 2021-03-18 23:17:30 train.py: 82] Epoch 16, iter 3600/6416, lr 0.000100, loss 6.128690
+INFO 2021-03-18 23:18:47 train.py: 82] Epoch 16, iter 3800/6416, lr 0.000100, loss 6.137684
+INFO 2021-03-18 23:20:04 train.py: 82] Epoch 16, iter 4000/6416, lr 0.000100, loss 6.189636
+INFO 2021-03-18 23:21:22 train.py: 82] Epoch 16, iter 4200/6416, lr 0.000100, loss 6.152111
+INFO 2021-03-18 23:22:39 train.py: 82] Epoch 16, iter 4400/6416, lr 0.000100, loss 6.124304
+INFO 2021-03-18 23:23:56 train.py: 82] Epoch 16, iter 4600/6416, lr 0.000100, loss 6.146148
+INFO 2021-03-18 23:25:13 train.py: 82] Epoch 16, iter 4800/6416, lr 0.000100, loss 6.136418
+INFO 2021-03-18 23:26:30 train.py: 82] Epoch 16, iter 5000/6416, lr 0.000100, loss 6.132527
+INFO 2021-03-18 23:27:47 train.py: 82] Epoch 16, iter 5200/6416, lr 0.000100, loss 6.136409
+INFO 2021-03-18 23:29:05 train.py: 82] Epoch 16, iter 5400/6416, lr 0.000100, loss 6.153092
+INFO 2021-03-18 23:30:22 train.py: 82] Epoch 16, iter 5600/6416, lr 0.000100, loss 6.168946
+INFO 2021-03-18 23:31:39 train.py: 82] Epoch 16, iter 5800/6416, lr 0.000100, loss 6.143332
+INFO 2021-03-18 23:32:56 train.py: 95] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2021-03-18 23:32:56 train.py: 82] Epoch 16, iter 6000/6416, lr 0.000100, loss 6.150599
+INFO 2021-03-18 23:34:14 train.py: 82] Epoch 16, iter 6200/6416, lr 0.000100, loss 6.143510
+INFO 2021-03-18 23:35:32 train.py: 82] Epoch 16, iter 6400/6416, lr 0.000100, loss 6.143740
+INFO 2021-03-18 23:35:39 train.py: 100] Save checkpoint Epoch_16.pt to disk...
+INFO 2021-03-18 23:35:41 train.py: 82] Epoch 17, iter 0/6416, lr 0.000100, loss 6.238836
+INFO 2021-03-18 23:36:59 train.py: 82] Epoch 17, iter 200/6416, lr 0.000100, loss 6.129740
+INFO 2021-03-18 23:38:17 train.py: 82] Epoch 17, iter 400/6416, lr 0.000100, loss 6.152919
+INFO 2021-03-18 23:39:35 train.py: 82] Epoch 17, iter 600/6416, lr 0.000100, loss 6.134320
+INFO 2021-03-18 23:40:53 train.py: 82] Epoch 17, iter 800/6416, lr 0.000100, loss 6.104474
+INFO 2021-03-18 23:42:11 train.py: 82] Epoch 17, iter 1000/6416, lr 0.000100, loss 6.118261
+INFO 2021-03-18 23:43:28 train.py: 82] Epoch 17, iter 1200/6416, lr 0.000100, loss 6.112352
+INFO 2021-03-18 23:44:46 train.py: 82] Epoch 17, iter 1400/6416, lr 0.000100, loss 6.168288
+INFO 2021-03-18 23:46:04 train.py: 82] Epoch 17, iter 1600/6416, lr 0.000100, loss 6.112389
+INFO 2021-03-18 23:47:22 train.py: 82] Epoch 17, iter 1800/6416, lr 0.000100, loss 6.138920
+INFO 2021-03-18 23:48:40 train.py: 82] Epoch 17, iter 2000/6416, lr 0.000100, loss 6.135342
+INFO 2021-03-18 23:49:58 train.py: 82] Epoch 17, iter 2200/6416, lr 0.000100, loss 6.115831
+INFO 2021-03-18 23:51:15 train.py: 82] Epoch 17, iter 2400/6416, lr 0.000100, loss 6.128751
+INFO 2021-03-18 23:52:33 train.py: 82] Epoch 17, iter 2600/6416, lr 0.000100, loss 6.131120
+INFO 2021-03-18 23:53:51 train.py: 82] Epoch 17, iter 2800/6416, lr 0.000100, loss 6.166421
+INFO 2021-03-18 23:55:09 train.py: 95] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2021-03-18 23:55:09 train.py: 82] Epoch 17, iter 3000/6416, lr 0.000100, loss 6.120512
+INFO 2021-03-18 23:56:26 train.py: 82] Epoch 17, iter 3200/6416, lr 0.000100, loss 6.148808
+INFO 2021-03-18 23:57:43 train.py: 82] Epoch 17, iter 3400/6416, lr 0.000100, loss 6.160245
+INFO 2021-03-18 23:59:01 train.py: 82] Epoch 17, iter 3600/6416, lr 0.000100, loss 6.135905
+INFO 2021-03-19 00:00:18 train.py: 82] Epoch 17, iter 3800/6416, lr 0.000100, loss 6.134643
+INFO 2021-03-19 00:01:35 train.py: 82] Epoch 17, iter 4000/6416, lr 0.000100, loss 6.095920
+INFO 2021-03-19 00:02:52 train.py: 82] Epoch 17, iter 4200/6416, lr 0.000100, loss 6.106539
+INFO 2021-03-19 00:04:09 train.py: 82] Epoch 17, iter 4400/6416, lr 0.000100, loss 6.174105
+INFO 2021-03-19 00:05:27 train.py: 82] Epoch 17, iter 4600/6416, lr 0.000100, loss 6.111957
+INFO 2021-03-19 00:06:44 train.py: 82] Epoch 17, iter 4800/6416, lr 0.000100, loss 6.142558
+INFO 2021-03-19 00:08:01 train.py: 82] Epoch 17, iter 5000/6416, lr 0.000100, loss 6.171786
+INFO 2021-03-19 00:09:18 train.py: 82] Epoch 17, iter 5200/6416, lr 0.000100, loss 6.160346
+INFO 2021-03-19 00:10:35 train.py: 82] Epoch 17, iter 5400/6416, lr 0.000100, loss 6.140907
+INFO 2021-03-19 00:11:52 train.py: 82] Epoch 17, iter 5600/6416, lr 0.000100, loss 6.120508
+INFO 2021-03-19 00:13:10 train.py: 82] Epoch 17, iter 5800/6416, lr 0.000100, loss 6.125139
+INFO 2021-03-19 00:14:27 train.py: 95] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2021-03-19 00:14:27 train.py: 82] Epoch 17, iter 6000/6416, lr 0.000100, loss 6.118235
+INFO 2021-03-19 00:15:45 train.py: 82] Epoch 17, iter 6200/6416, lr 0.000100, loss 6.180168
+INFO 2021-03-19 00:17:03 train.py: 82] Epoch 17, iter 6400/6416, lr 0.000100, loss 6.150227
+INFO 2021-03-19 00:17:09 train.py: 100] Save checkpoint Epoch_17.pt to disk...
+INFO 2021-03-19 00:17:10 train.py: 183] Optimization done!
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/.gitkeep b/bob/bio/facexzoo/models/heads/NPCFace/.gitkeep
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_agedb.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_agedb.txt
new file mode 100644
index 0000000000000000000000000000000000000000..f9d0713b9fe1b24e293b3edbac6a84412e934876
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_agedb.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16.pt | 0.9586666666666668 | 0.003855411460643884 |
+| Epoch_17_batch_5999.pt | 0.9585000000000001 | 0.0031274373211194564 |
+| Epoch_15_batch_5999.pt | 0.9583333333333334 | 0.003073181485764294 |
+| Epoch_13_batch_5999.pt | 0.9578333333333333 | 0.003549039167485437 |
+| Epoch_12_batch_2999.pt | 0.9578333333333333 | 0.0033152286106301588 |
+| Epoch_17_batch_2999.pt | 0.9578333333333333 | 0.003651906317676623 |
+| Epoch_15.pt | 0.9576666666666667 | 0.0036362373715452486 |
+| Epoch_15_batch_2999.pt | 0.9576666666666667 | 0.0030852096393144094 |
+| Epoch_14.pt | 0.9571666666666665 | 0.0034161020310318267 |
+| Epoch_14_batch_2999.pt | 0.9570000000000001 | 0.003431876713662336 |
+| Epoch_14_batch_5999.pt | 0.9570000000000001 | 0.00325652241994508 |
+| Epoch_13.pt | 0.9564999999999999 | 0.003771645349443043 |
+| Epoch_13_batch_2999.pt | 0.9563333333333333 | 0.0033222036417169015 |
+| Epoch_16_batch_5999.pt | 0.9560000000000001 | 0.003297960462145741 |
+| Epoch_16_batch_2999.pt | 0.9560000000000001 | 0.0035590260840104365 |
+| Epoch_17.pt | 0.9558333333333333 | 0.003712258638286252 |
+| Epoch_12.pt | 0.9556666666666667 | 0.003444444444444452 |
+| Epoch_12_batch_5999.pt | 0.9553333333333333 | 0.003494263376336977 |
+| Epoch_10_batch_2999.pt | 0.9546666666666667 | 0.003683463212055729 |
+| Epoch_10_batch_5999.pt | 0.9546666666666667 | 0.0030611060697679844 |
+| Epoch_11_batch_5999.pt | 0.9541666666666668 | 0.0040768088464349615 |
+| Epoch_11.pt | 0.9535 | 0.004100963451991243 |
+| Epoch_10.pt | 0.9531666666666666 | 0.0034197140881138703 |
+| Epoch_11_batch_2999.pt | 0.9525 | 0.0035070475782414713 |
+| Epoch_8_batch_2999.pt | 0.9438333333333334 | 0.003416102031031817 |
+| Epoch_9_batch_5999.pt | 0.9406666666666668 | 0.005105068892293398 |
+| Epoch_8.pt | 0.9401666666666666 | 0.0032437023504433365 |
+| Epoch_7_batch_5999.pt | 0.9398333333333333 | 0.0039553838914061215 |
+| Epoch_6.pt | 0.9386666666666666 | 0.004803548071381615 |
+| Epoch_9.pt | 0.938 | 0.004273085449562086 |
+| Epoch_6_batch_2999.pt | 0.9376666666666666 | 0.004129089816902671 |
+| Epoch_7_batch_2999.pt | 0.9369999999999999 | 0.0043800896320948905 |
+| Epoch_8_batch_5999.pt | 0.9369999999999999 | 0.004273085449562093 |
+| Epoch_5_batch_5999.pt | 0.9368333333333334 | 0.00432798815634718 |
+| Epoch_9_batch_2999.pt | 0.9368333333333332 | 0.005513451563961381 |
+| Epoch_4_batch_5999.pt | 0.9336666666666666 | 0.003879353388910243 |
+| Epoch_5_batch_2999.pt | 0.9331666666666667 | 0.004205008771148052 |
+| Epoch_4.pt | 0.932 | 0.00478423336480244 |
+| Epoch_6_batch_5999.pt | 0.9318333333333333 | 0.005689767727255955 |
+| Epoch_5.pt | 0.9306666666666666 | 0.004970282054743339 |
+| Epoch_7.pt | 0.9293333333333333 | 0.005129195061148352 |
+| Epoch_3.pt | 0.9288333333333334 | 0.0039051248379533255 |
+| Epoch_3_batch_5999.pt | 0.9285 | 0.0051845568402352 |
+| Epoch_3_batch_2999.pt | 0.9271666666666667 | 0.006310142726956971 |
+| Epoch_4_batch_2999.pt | 0.9259999999999998 | 0.0042759736455319705 |
+| Epoch_2_batch_5999.pt | 0.9248333333333333 | 0.005015100653815553 |
+| Epoch_2.pt | 0.923 | 0.006924638422614617 |
+| Epoch_2_batch_2999.pt | 0.9180000000000001 | 0.007531005867924685 |
+| Epoch_1.pt | 0.9094999999999999 | 0.004969350505291861 |
+| Epoch_1_batch_5999.pt | 0.8964999999999999 | 0.005923639605246036 |
+| Epoch_1_batch_2999.pt | 0.8881666666666665 | 0.005743756674924583 |
+| Epoch_0_batch_5999.pt | 0.851 | 0.006122212139535962 |
+| Epoch_0.pt | 0.8476666666666667 | 0.007888888888888888 |
+| Epoch_0_batch_2999.pt | 0.756 | 0.006010279260577128 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_calfw.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_calfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..003bcd4674704387621d2a45cc3661362cd9ecd4
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_calfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_16_batch_2999.pt | 0.9413333333333334 | 0.0035555555555555514 |
+| Epoch_17.pt | 0.9408333333333333 | 0.003326381639982814 |
+| Epoch_15_batch_2999.pt | 0.9404999999999999 | 0.0036687704402442265 |
+| Epoch_15_batch_5999.pt | 0.9403333333333332 | 0.0033499585403736266 |
+| Epoch_13_batch_2999.pt | 0.9403333333333332 | 0.003649792823479297 |
+| Epoch_14.pt | 0.9401666666666666 | 0.0037387691907536046 |
+| Epoch_17_batch_5999.pt | 0.9398333333333333 | 0.0036637193541772294 |
+| Epoch_12_batch_2999.pt | 0.9396666666666667 | 0.0037908271353848904 |
+| Epoch_14_batch_2999.pt | 0.9396666666666664 | 0.0036413265795942036 |
+| Epoch_13_batch_5999.pt | 0.9395 | 0.0036687704402442313 |
+| Epoch_16_batch_5999.pt | 0.9395 | 0.003967849185704244 |
+| Epoch_14_batch_5999.pt | 0.9393333333333332 | 0.0033536418383970134 |
+| Epoch_17_batch_2999.pt | 0.9391666666666667 | 0.0036111111111111057 |
+| Epoch_11_batch_5999.pt | 0.9391666666666666 | 0.004031128874149276 |
+| Epoch_13.pt | 0.9388333333333335 | 0.0037683706404692424 |
+| Epoch_12.pt | 0.9388333333333332 | 0.0037928620419329555 |
+| Epoch_10.pt | 0.9383333333333332 | 0.004216370213557835 |
+| Epoch_15.pt | 0.9381666666666668 | 0.003595693583396533 |
+| Epoch_16.pt | 0.9378333333333332 | 0.0033430414185178633 |
+| Epoch_11.pt | 0.9371666666666666 | 0.00388134188326857 |
+| Epoch_10_batch_5999.pt | 0.9365 | 0.003916174121906021 |
+| Epoch_11_batch_2999.pt | 0.9363333333333334 | 0.004163331998932266 |
+| Epoch_12_batch_5999.pt | 0.9355 | 0.0037601713908928204 |
+| Epoch_10_batch_2999.pt | 0.9345000000000001 | 0.003881341883268571 |
+| Epoch_8_batch_2999.pt | 0.9254999999999999 | 0.004187356041380288 |
+| Epoch_5_batch_5999.pt | 0.9251666666666667 | 0.0051845568402352 |
+| Epoch_9_batch_5999.pt | 0.925 | 0.003710179523792 |
+| Epoch_9_batch_2999.pt | 0.9246666666666666 | 0.0039424877776022575 |
+| Epoch_9.pt | 0.9236666666666669 | 0.0038313199221259313 |
+| Epoch_8_batch_5999.pt | 0.9235 | 0.004327988156347179 |
+| Epoch_7_batch_5999.pt | 0.9233333333333335 | 0.004029214303215277 |
+| Epoch_6_batch_5999.pt | 0.9225 | 0.004844813951249548 |
+| Epoch_6_batch_2999.pt | 0.9213333333333334 | 0.005344894875913643 |
+| Epoch_6.pt | 0.9211666666666666 | 0.003751954223311792 |
+| Epoch_7_batch_2999.pt | 0.9166666666666666 | 0.0040061680838488775 |
+| Epoch_5_batch_2999.pt | 0.9164999999999999 | 0.0037798197095942365 |
+| Epoch_4_batch_5999.pt | 0.9156666666666666 | 0.005295001136641719 |
+| Epoch_5.pt | 0.915 | 0.0038888888888888866 |
+| Epoch_8.pt | 0.9146666666666666 | 0.004042978977480054 |
+| Epoch_3_batch_5999.pt | 0.9145 | 0.004944444444444446 |
+| Epoch_4.pt | 0.9143333333333334 | 0.005068664323536468 |
+| Epoch_4_batch_2999.pt | 0.9126666666666668 | 0.004283185614976973 |
+| Epoch_3_batch_2999.pt | 0.9125 | 0.005517928131481733 |
+| Epoch_7.pt | 0.9099999999999999 | 0.004513354669242204 |
+| Epoch_2.pt | 0.908 | 0.004633479880442902 |
+| Epoch_2_batch_5999.pt | 0.907 | 0.005344894875913641 |
+| Epoch_3.pt | 0.9055 | 0.005483139138742954 |
+| Epoch_2_batch_2999.pt | 0.9036666666666665 | 0.004415181441401225 |
+| Epoch_1.pt | 0.8941666666666667 | 0.005248456563247551 |
+| Epoch_1_batch_5999.pt | 0.8936666666666667 | 0.005333333333333333 |
+| Epoch_1_batch_2999.pt | 0.8833333333333334 | 0.004753166585823874 |
+| Epoch_0.pt | 0.8473333333333333 | 0.0033259176771323995 |
+| Epoch_0_batch_5999.pt | 0.8423333333333334 | 0.0046693114198779325 |
+| Epoch_0_batch_2999.pt | 0.7418333333333333 | 0.0066083093969955585 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_cplfw.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_cplfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..4a7ee3056eedf8e3d742fa5583767320c3f04701
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_cplfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.8380000000000001 | 0.005162782291328419 |
+| Epoch_17.pt | 0.8378333333333334 | 0.004044887033170536 |
+| Epoch_17_batch_2999.pt | 0.8375 | 0.0052602046620278675 |
+| Epoch_10.pt | 0.8356666666666668 | 0.00572626610076672 |
+| Epoch_15_batch_2999.pt | 0.8346666666666666 | 0.005925462944877056 |
+| Epoch_13_batch_2999.pt | 0.8341666666666668 | 0.004995368225036441 |
+| Epoch_12_batch_5999.pt | 0.8338333333333334 | 0.005398044913567211 |
+| Epoch_16_batch_5999.pt | 0.8338333333333333 | 0.005241395064505465 |
+| Epoch_14_batch_2999.pt | 0.8334999999999999 | 0.005902761437885888 |
+| Epoch_12.pt | 0.8334999999999999 | 0.005607809809916662 |
+| Epoch_17_batch_5999.pt | 0.8334999999999999 | 0.005907987896708518 |
+| Epoch_11_batch_5999.pt | 0.8328333333333333 | 0.0063296773073210294 |
+| Epoch_16_batch_2999.pt | 0.8325000000000001 | 0.005796177965896646 |
+| Epoch_14_batch_5999.pt | 0.8323333333333334 | 0.006256848100133819 |
+| Epoch_13.pt | 0.8320000000000001 | 0.0058097406499561625 |
+| Epoch_14.pt | 0.8320000000000001 | 0.006753142844010111 |
+| Epoch_16.pt | 0.8318333333333333 | 0.0058818091615034235 |
+| Epoch_15_batch_5999.pt | 0.8310000000000001 | 0.005979388465366094 |
+| Epoch_12_batch_2999.pt | 0.8303333333333333 | 0.005239922721134044 |
+| Epoch_10_batch_5999.pt | 0.8295 | 0.005253158955556718 |
+| Epoch_13_batch_5999.pt | 0.8291666666666668 | 0.0062075720388316035 |
+| Epoch_11_batch_2999.pt | 0.8291666666666666 | 0.006868473332984994 |
+| Epoch_11.pt | 0.8281666666666666 | 0.004550471417716722 |
+| Epoch_10_batch_2999.pt | 0.8254999999999999 | 0.006925752622113785 |
+| Epoch_9_batch_5999.pt | 0.8099999999999999 | 0.006703601390665933 |
+| Epoch_9.pt | 0.8068333333333333 | 0.006677999626560759 |
+| Epoch_9_batch_2999.pt | 0.8058333333333334 | 0.0055680415246074704 |
+| Epoch_7_batch_5999.pt | 0.8015000000000001 | 0.007675719293112974 |
+| Epoch_6_batch_5999.pt | 0.8013333333333333 | 0.006492160514662914 |
+| Epoch_5_batch_5999.pt | 0.8013333333333333 | 0.007057086448089071 |
+| Epoch_8_batch_2999.pt | 0.8008333333333333 | 0.007869498862423129 |
+| Epoch_8_batch_5999.pt | 0.8006666666666666 | 0.009409084131706488 |
+| Epoch_6_batch_2999.pt | 0.8001666666666667 | 0.008554292836110181 |
+| Epoch_7_batch_2999.pt | 0.7986666666666666 | 0.006969067458029654 |
+| Epoch_5_batch_2999.pt | 0.7941666666666666 | 0.007278625904779583 |
+| Epoch_6.pt | 0.7933333333333333 | 0.008534606386520673 |
+| Epoch_4.pt | 0.7918333333333334 | 0.008611827927154233 |
+| Epoch_4_batch_5999.pt | 0.7886666666666666 | 0.008909005015733823 |
+| Epoch_7.pt | 0.7883333333333333 | 0.0097182531580755 |
+| Epoch_3_batch_5999.pt | 0.7873333333333333 | 0.008473633762194496 |
+| Epoch_4_batch_2999.pt | 0.7871666666666667 | 0.005905897868201721 |
+| Epoch_8.pt | 0.7865 | 0.008208133575309518 |
+| Epoch_5.pt | 0.7865 | 0.006307207321590978 |
+| Epoch_3.pt | 0.7816666666666666 | 0.006464528092350608 |
+| Epoch_3_batch_2999.pt | 0.7811666666666668 | 0.007286254605271452 |
+| Epoch_2_batch_5999.pt | 0.7801666666666666 | 0.006457123542603502 |
+| Epoch_2.pt | 0.779 | 0.00794347313409561 |
+| Epoch_2_batch_2999.pt | 0.7735000000000001 | 0.006253147355680327 |
+| Epoch_1_batch_5999.pt | 0.7708333333333333 | 0.00792421667274079 |
+| Epoch_1.pt | 0.7583333333333333 | 0.0078134104407774 |
+| Epoch_1_batch_2999.pt | 0.744 | 0.00818836874573009 |
+| Epoch_0.pt | 0.7035000000000001 | 0.006026945667091587 |
+| Epoch_0_batch_5999.pt | 0.6981666666666666 | 0.007998649577380449 |
+| Epoch_0_batch_2999.pt | 0.6375 | 0.010497648178708914 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_lfw.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_lfw.txt
new file mode 100644
index 0000000000000000000000000000000000000000..32a3f2f067ee122d52e4ff8686b483ff6b8533f7
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_lfw.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14_batch_5999.pt | 0.9955 | 0.0012184284555256317 |
+| Epoch_10.pt | 0.9954999999999998 | 0.0011928283640879939 |
+| Epoch_17_batch_5999.pt | 0.9954999999999998 | 0.0010555555555555552 |
+| Epoch_13_batch_5999.pt | 0.9953333333333333 | 0.001160034056545619 |
+| Epoch_17_batch_2999.pt | 0.9953333333333333 | 0.000987577157479508 |
+| Epoch_16_batch_5999.pt | 0.9953333333333333 | 0.0010482201257840662 |
+| Epoch_16_batch_2999.pt | 0.9953333333333333 | 0.001105541596785131 |
+| Epoch_14.pt | 0.9953333333333332 | 0.0010772621905369623 |
+| Epoch_13.pt | 0.9951666666666668 | 0.001007686508178724 |
+| Epoch_14_batch_2999.pt | 0.9951666666666666 | 0.0012031337682059844 |
+| Epoch_15_batch_5999.pt | 0.9951666666666666 | 0.001150684176511556 |
+| Epoch_12_batch_2999.pt | 0.9951666666666666 | 0.0011772011166898376 |
+| Epoch_15_batch_2999.pt | 0.9951666666666666 | 0.001150684176511556 |
+| Epoch_11.pt | 0.9950000000000001 | 0.001111111111111109 |
+| Epoch_10_batch_5999.pt | 0.9950000000000001 | 0.0011915339216404008 |
+| Epoch_17.pt | 0.9949999999999999 | 0.00121716123890037 |
+| Epoch_10_batch_2999.pt | 0.9948333333333335 | 0.001228519132638664 |
+| Epoch_11_batch_2999.pt | 0.9948333333333332 | 0.0012285191326386685 |
+| Epoch_12.pt | 0.9948333333333332 | 0.0010671873729054778 |
+| Epoch_13_batch_2999.pt | 0.9948333333333332 | 0.0012031337682059844 |
+| Epoch_16.pt | 0.9948333333333332 | 0.0008766518798921935 |
+| Epoch_12_batch_5999.pt | 0.9946666666666666 | 0.0011863420280034786 |
+| Epoch_9_batch_2999.pt | 0.9945 | 0.0017042068500197737 |
+| Epoch_11_batch_5999.pt | 0.9945 | 0.0011928283640879936 |
+| Epoch_8_batch_2999.pt | 0.9941666666666669 | 0.0014540280364780426 |
+| Epoch_15.pt | 0.9941666666666666 | 0.0011180339887498947 |
+| Epoch_9_batch_5999.pt | 0.9938333333333332 | 0.0015525765124980138 |
+| Epoch_6_batch_2999.pt | 0.9933333333333334 | 0.001531560972454468 |
+| Epoch_7_batch_5999.pt | 0.9931666666666666 | 0.001520436909267114 |
+| Epoch_8.pt | 0.9931666666666666 | 0.0014999999999999983 |
+| Epoch_8_batch_5999.pt | 0.993 | 0.0014229164972073044 |
+| Epoch_7.pt | 0.993 | 0.0014444444444444446 |
+| Epoch_4_batch_2999.pt | 0.9926666666666668 | 0.0014098419489388396 |
+| Epoch_6_batch_5999.pt | 0.9926666666666666 | 0.0012222222222222228 |
+| Epoch_7_batch_2999.pt | 0.9926666666666666 | 0.0014740554623801831 |
+| Epoch_4.pt | 0.9926666666666666 | 0.0015752718754175393 |
+| Epoch_5_batch_2999.pt | 0.9925 | 0.0014326441064697365 |
+| Epoch_4_batch_5999.pt | 0.9923333333333334 | 0.0016887426837300758 |
+| Epoch_3_batch_5999.pt | 0.9923333333333334 | 0.0014529663145135632 |
+| Epoch_5_batch_5999.pt | 0.9923333333333332 | 0.001613982116259329 |
+| Epoch_9.pt | 0.9921666666666666 | 0.001025899184034405 |
+| Epoch_3_batch_2999.pt | 0.9916666666666668 | 0.0017033010796395466 |
+| Epoch_5.pt | 0.9906666666666666 | 0.001724908299584452 |
+| Epoch_3.pt | 0.9906666666666666 | 0.0011166528467912136 |
+| Epoch_2.pt | 0.99 | 0.0010829771494232185 |
+| Epoch_6.pt | 0.9893333333333333 | 0.0017249082995844474 |
+| Epoch_2_batch_5999.pt | 0.9888333333333332 | 0.0012680791345014873 |
+| Epoch_2_batch_2999.pt | 0.9883333333333333 | 0.0009622504486493771 |
+| Epoch_1_batch_5999.pt | 0.9879999999999999 | 0.0012619796324000619 |
+| Epoch_1.pt | 0.9871666666666666 | 0.001025899184034411 |
+| Epoch_1_batch_2999.pt | 0.9843333333333334 | 0.0016517854163687258 |
+| Epoch_0.pt | 0.9726666666666667 | 0.0014740554623801794 |
+| Epoch_0_batch_5999.pt | 0.9685 | 0.002551325000712227 |
+| Epoch_0_batch_2999.pt | 0.9446666666666668 | 0.003581502546952482 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_African.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_African.txt
new file mode 100644
index 0000000000000000000000000000000000000000..86b610904565335dd2d4abc54f2a3027cee222d4
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_African.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_14.pt | 0.8808333333333334 | 0.004901813723284251 |
+| Epoch_14_batch_2999.pt | 0.8801666666666665 | 0.005313911395665874 |
+| Epoch_17_batch_2999.pt | 0.8795 | 0.004707821178306753 |
+| Epoch_15_batch_5999.pt | 0.8793333333333333 | 0.004609437447264788 |
+| Epoch_17_batch_5999.pt | 0.8793333333333333 | 0.005438772555786852 |
+| Epoch_13_batch_2999.pt | 0.8781666666666667 | 0.00495940309032707 |
+| Epoch_16_batch_2999.pt | 0.8780000000000001 | 0.004552844721899525 |
+| Epoch_14_batch_5999.pt | 0.8779999999999999 | 0.005691665988829194 |
+| Epoch_15_batch_2999.pt | 0.8778333333333335 | 0.005134307266916454 |
+| Epoch_17.pt | 0.8773333333333333 | 0.004926622061643417 |
+| Epoch_16_batch_5999.pt | 0.8768333333333332 | 0.005039657542745151 |
+| Epoch_12_batch_5999.pt | 0.8765000000000001 | 0.004032659876435446 |
+| Epoch_13_batch_5999.pt | 0.8763333333333334 | 0.0048035480713816215 |
+| Epoch_16.pt | 0.8763333333333332 | 0.004573136806057042 |
+| Epoch_15.pt | 0.876 | 0.005324066022538193 |
+| Epoch_12_batch_2999.pt | 0.8753333333333332 | 0.004725815626252606 |
+| Epoch_12.pt | 0.874 | 0.004275973645531965 |
+| Epoch_11_batch_5999.pt | 0.8733333333333334 | 0.00528800179301813 |
+| Epoch_11.pt | 0.8726666666666667 | 0.004007708621526985 |
+| Epoch_10.pt | 0.8723333333333333 | 0.005277485372016861 |
+| Epoch_10_batch_5999.pt | 0.8721666666666665 | 0.004714372560794843 |
+| Epoch_13.pt | 0.8718333333333333 | 0.004736579481565278 |
+| Epoch_11_batch_2999.pt | 0.8713333333333333 | 0.004316205466567784 |
+| Epoch_10_batch_2999.pt | 0.8701666666666666 | 0.005313911395665875 |
+| Epoch_8_batch_2999.pt | 0.8451666666666666 | 0.005166666666666669 |
+| Epoch_7_batch_5999.pt | 0.8445 | 0.005566932794703092 |
+| Epoch_8_batch_5999.pt | 0.8425 | 0.005218970618904431 |
+| Epoch_9_batch_2999.pt | 0.8421666666666667 | 0.005583540626334325 |
+| Epoch_6_batch_5999.pt | 0.8383333333333335 | 0.006348908723182954 |
+| Epoch_9_batch_5999.pt | 0.8381666666666667 | 0.006248209620106401 |
+| Epoch_6_batch_2999.pt | 0.8358333333333332 | 0.004297932319409212 |
+| Epoch_7_batch_2999.pt | 0.8348333333333333 | 0.005100532530899631 |
+| Epoch_5_batch_5999.pt | 0.8341666666666667 | 0.005007710104811097 |
+| Epoch_5.pt | 0.8325000000000001 | 0.006579288490181677 |
+| Epoch_7.pt | 0.8315000000000001 | 0.004978038187632841 |
+| Epoch_9.pt | 0.8303333333333333 | 0.005453507196359123 |
+| Epoch_6.pt | 0.8301666666666666 | 0.006719466836580209 |
+| Epoch_5_batch_2999.pt | 0.8274999999999999 | 0.005623199300253403 |
+| Epoch_8.pt | 0.8265 | 0.005273097339852128 |
+| Epoch_4_batch_5999.pt | 0.826 | 0.005693834655697315 |
+| Epoch_4.pt | 0.8254999999999999 | 0.004798726682962655 |
+| Epoch_3_batch_5999.pt | 0.8243333333333334 | 0.005838357624630959 |
+| Epoch_4_batch_2999.pt | 0.8233333333333333 | 0.005998971105196937 |
+| Epoch_3_batch_2999.pt | 0.8206666666666667 | 0.004232443767720201 |
+| Epoch_3.pt | 0.8181666666666667 | 0.00525550857390323 |
+| Epoch_2_batch_5999.pt | 0.8049999999999999 | 0.005649210586171622 |
+| Epoch_2_batch_2999.pt | 0.7998333333333332 | 0.006006426599387291 |
+| Epoch_2.pt | 0.7966666666666666 | 0.007498971122843086 |
+| Epoch_1_batch_5999.pt | 0.7838333333333334 | 0.004253176010094052 |
+| Epoch_1.pt | 0.7788333333333333 | 0.005170249653689992 |
+| Epoch_1_batch_2999.pt | 0.7693333333333333 | 0.004459696053419883 |
+| Epoch_0.pt | 0.7171666666666666 | 0.003469888104361597 |
+| Epoch_0_batch_5999.pt | 0.7051666666666666 | 0.002622904807580643 |
+| Epoch_0_batch_2999.pt | 0.6256666666666667 | 0.006531972647421807 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Asian.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Asian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..36943cd6ef8197926fed53482787b011f7e086aa
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Asian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_13_batch_5999.pt | 0.882 | 0.0036328406053937425 |
+| Epoch_14_batch_5999.pt | 0.882 | 0.004244095389952066 |
+| Epoch_15_batch_2999.pt | 0.881 | 0.00434044999658324 |
+| Epoch_17_batch_2999.pt | 0.8805 | 0.0046017307533064095 |
+| Epoch_14_batch_2999.pt | 0.8795 | 0.0032683480177961616 |
+| Epoch_16_batch_2999.pt | 0.8790000000000001 | 0.0040612593072215765 |
+| Epoch_13.pt | 0.8786666666666667 | 0.003624334762288911 |
+| Epoch_16_batch_5999.pt | 0.8781666666666667 | 0.0033467323292953395 |
+| Epoch_13_batch_2999.pt | 0.8780000000000001 | 0.0036243347622889137 |
+| Epoch_15_batch_5999.pt | 0.8779999999999999 | 0.0034587516480607547 |
+| Epoch_17.pt | 0.8778333333333332 | 0.004112987559751023 |
+| Epoch_14.pt | 0.8773333333333333 | 0.002768874620972698 |
+| Epoch_17_batch_5999.pt | 0.877 | 0.0038151743807532056 |
+| Epoch_15.pt | 0.8766666666666667 | 0.0036260375271290482 |
+| Epoch_12.pt | 0.8763333333333334 | 0.003284832333132106 |
+| Epoch_10.pt | 0.876 | 0.0018459164139817978 |
+| Epoch_11.pt | 0.8755000000000001 | 0.003012832636212506 |
+| Epoch_11_batch_5999.pt | 0.8748333333333334 | 0.0030676528205063414 |
+| Epoch_10_batch_5999.pt | 0.8741666666666668 | 0.0032034896096307776 |
+| Epoch_12_batch_2999.pt | 0.8736666666666666 | 0.0032659863237108973 |
+| Epoch_12_batch_5999.pt | 0.8734999999999999 | 0.003374285475346721 |
+| Epoch_16.pt | 0.8728333333333333 | 0.0036687704402442356 |
+| Epoch_11_batch_2999.pt | 0.8708333333333333 | 0.003576759689515528 |
+| Epoch_10_batch_2999.pt | 0.8683333333333334 | 0.002496911672693804 |
+| Epoch_6_batch_5999.pt | 0.8470000000000001 | 0.0057510063531589035 |
+| Epoch_7_batch_5999.pt | 0.8465 | 0.004946940693218602 |
+| Epoch_8_batch_5999.pt | 0.8460000000000001 | 0.0019436506316150984 |
+| Epoch_9_batch_5999.pt | 0.8455 | 0.0036687704402442343 |
+| Epoch_8_batch_2999.pt | 0.8441666666666666 | 0.0041518254203407505 |
+| Epoch_7_batch_2999.pt | 0.8401666666666665 | 0.005524636181530118 |
+| Epoch_9_batch_2999.pt | 0.8396666666666667 | 0.003170076137263799 |
+| Epoch_5_batch_5999.pt | 0.8368333333333332 | 0.006047395113131019 |
+| Epoch_4.pt | 0.8353333333333334 | 0.00508447163939891 |
+| Epoch_6_batch_2999.pt | 0.8343333333333334 | 0.003961231882216864 |
+| Epoch_8.pt | 0.8329999999999999 | 0.0032659863237109077 |
+| Epoch_6.pt | 0.8321666666666667 | 0.005311587612283448 |
+| Epoch_5_batch_2999.pt | 0.8316666666666667 | 0.0047074933698477515 |
+| Epoch_4_batch_5999.pt | 0.8291666666666666 | 0.0053705300103986595 |
+| Epoch_3_batch_5999.pt | 0.8288333333333334 | 0.005616608967722097 |
+| Epoch_9.pt | 0.8271666666666666 | 0.0032015621187164254 |
+| Epoch_3_batch_2999.pt | 0.8248333333333333 | 0.005749127684771184 |
+| Epoch_7.pt | 0.8240000000000001 | 0.0046361435655611315 |
+| Epoch_5.pt | 0.8230000000000001 | 0.005571089394098522 |
+| Epoch_3.pt | 0.8166666666666667 | 0.005085685551709242 |
+| Epoch_2_batch_5999.pt | 0.8156666666666667 | 0.006182412330330469 |
+| Epoch_4_batch_2999.pt | 0.8155000000000001 | 0.00471437256079484 |
+| Epoch_2.pt | 0.8063333333333335 | 0.005526591162420398 |
+| Epoch_2_batch_2999.pt | 0.8048333333333334 | 0.005711424548251543 |
+| Epoch_1.pt | 0.7868333333333333 | 0.005507850738090717 |
+| Epoch_1_batch_5999.pt | 0.7828333333333333 | 0.005511211916640798 |
+| Epoch_1_batch_2999.pt | 0.769 | 0.006676844083366044 |
+| Epoch_0.pt | 0.741 | 0.00404603143467403 |
+| Epoch_0_batch_5999.pt | 0.7308333333333333 | 0.004121982622566566 |
+| Epoch_0_batch_2999.pt | 0.6881666666666666 | 0.005296458168984487 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Caucasian.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Caucasian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..dfa15cd750816a08f90009f08da51238319f335e
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Caucasian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15_batch_2999.pt | 0.9546666666666667 | 0.00253128572219878 |
+| Epoch_16_batch_5999.pt | 0.9546666666666667 | 0.0022194427061597967 |
+| Epoch_16.pt | 0.9545 | 0.0023313483620397003 |
+| Epoch_15.pt | 0.9540000000000001 | 0.002462057756240042 |
+| Epoch_13_batch_2999.pt | 0.9540000000000001 | 0.0027464904654018385 |
+| Epoch_16_batch_2999.pt | 0.9538333333333334 | 0.0019883922409081383 |
+| Epoch_17_batch_2999.pt | 0.9538333333333332 | 0.002064544908451343 |
+| Epoch_15_batch_5999.pt | 0.9536666666666666 | 0.0021052550357218264 |
+| Epoch_14_batch_2999.pt | 0.9535 | 0.0022145701586660845 |
+| Epoch_14_batch_5999.pt | 0.9535 | 0.0023366378716459008 |
+| Epoch_17_batch_5999.pt | 0.9531666666666668 | 0.002242270674512282 |
+| Epoch_12_batch_5999.pt | 0.9531666666666666 | 0.0029860788111948163 |
+| Epoch_17.pt | 0.9531666666666665 | 0.0023366378716459008 |
+| Epoch_12_batch_2999.pt | 0.953 | 0.0024317854031377013 |
+| Epoch_11.pt | 0.9521666666666666 | 0.002485141027371678 |
+| Epoch_14.pt | 0.9520000000000002 | 0.002627020092785978 |
+| Epoch_13_batch_5999.pt | 0.9516666666666665 | 0.00230404903925864 |
+| Epoch_11_batch_5999.pt | 0.9511666666666667 | 0.002485141027371675 |
+| Epoch_13.pt | 0.9506666666666665 | 0.0028781852993308103 |
+| Epoch_12.pt | 0.9498333333333333 | 0.0023759338644825695 |
+| Epoch_11_batch_2999.pt | 0.9493333333333333 | 0.002920764317354431 |
+| Epoch_10_batch_2999.pt | 0.9488333333333335 | 0.002653323096946567 |
+| Epoch_10_batch_5999.pt | 0.9484999999999999 | 0.002669558617051996 |
+| Epoch_10.pt | 0.9456666666666669 | 0.0026081543542901126 |
+| Epoch_9_batch_5999.pt | 0.9256666666666666 | 0.003307305792474934 |
+| Epoch_9_batch_2999.pt | 0.9241666666666667 | 0.004069231128273327 |
+| Epoch_7_batch_5999.pt | 0.922 | 0.003081205471969341 |
+| Epoch_7_batch_2999.pt | 0.9215 | 0.0026229048075806422 |
+| Epoch_8_batch_5999.pt | 0.9208333333333332 | 0.003636661744234035 |
+| Epoch_8_batch_2999.pt | 0.9205 | 0.003857412297075547 |
+| Epoch_6_batch_2999.pt | 0.9196666666666667 | 0.0035986966090448148 |
+| Epoch_9.pt | 0.9186666666666665 | 0.003071172213574502 |
+| Epoch_6_batch_5999.pt | 0.9178333333333335 | 0.0026063786901644238 |
+| Epoch_5_batch_5999.pt | 0.9148333333333334 | 0.0034016154477425928 |
+| Epoch_5_batch_2999.pt | 0.9135 | 0.0032437023504433365 |
+| Epoch_4_batch_5999.pt | 0.9126666666666667 | 0.0013877773329774245 |
+| Epoch_4.pt | 0.9119999999999999 | 0.004401178293408523 |
+| Epoch_3_batch_2999.pt | 0.9111666666666667 | 0.0031234872881934685 |
+| Epoch_3_batch_5999.pt | 0.9106666666666665 | 0.0024241582476968214 |
+| Epoch_8.pt | 0.9101666666666667 | 0.005249632556218446 |
+| Epoch_3.pt | 0.9093333333333332 | 0.0030550504633038984 |
+| Epoch_7.pt | 0.9066666666666666 | 0.0038005847503304554 |
+| Epoch_5.pt | 0.9065000000000001 | 0.004468338548355086 |
+| Epoch_4_batch_2999.pt | 0.9065 | 0.003328236844611147 |
+| Epoch_6.pt | 0.9043333333333334 | 0.004662696724079402 |
+| Epoch_2_batch_5999.pt | 0.9021666666666667 | 0.004430882084797872 |
+| Epoch_2.pt | 0.8983333333333334 | 0.0028760398012321704 |
+| Epoch_2_batch_2999.pt | 0.8944999999999999 | 0.002653323096946561 |
+| Epoch_1.pt | 0.8865000000000001 | 0.004115988098220321 |
+| Epoch_1_batch_5999.pt | 0.8805 | 0.003496470883902125 |
+| Epoch_1_batch_2999.pt | 0.8636666666666667 | 0.0034228715112776323 |
+| Epoch_0_batch_5999.pt | 0.8295 | 0.004830778379628527 |
+| Epoch_0.pt | 0.8251666666666667 | 0.005136711250317248 |
+| Epoch_0_batch_2999.pt | 0.7656666666666666 | 0.005147215476400207 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Indian.txt b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Indian.txt
new file mode 100644
index 0000000000000000000000000000000000000000..778fa2ec414ab9733024962afbc1481ac880a037
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/accu_files/accu_rfw_Indian.txt
@@ -0,0 +1,58 @@
++------------------------+--------------------+-----------------------+
+| model_name | mean accuracy | standard error |
++------------------------+--------------------+-----------------------+
+| Epoch_15.pt | 0.9103333333333333 | 0.00279770629155871 |
+| Epoch_14_batch_5999.pt | 0.9103333333333333 | 0.0027644122899166063 |
+| Epoch_16_batch_5999.pt | 0.9101666666666667 | 0.0030066798061632962 |
+| Epoch_16_batch_2999.pt | 0.9099999999999999 | 0.003162277660168382 |
+| Epoch_15_batch_2999.pt | 0.9093333333333333 | 0.002572408200620055 |
+| Epoch_17.pt | 0.909 | 0.003222222222222227 |
+| Epoch_17_batch_5999.pt | 0.909 | 0.003173968190463487 |
+| Epoch_14_batch_2999.pt | 0.9088333333333333 | 0.003333796264150557 |
+| Epoch_15_batch_5999.pt | 0.9083333333333332 | 0.003220305943597656 |
+| Epoch_13.pt | 0.9081666666666667 | 0.003517592470728158 |
+| Epoch_13_batch_5999.pt | 0.9078333333333335 | 0.003230353724468143 |
+| Epoch_14.pt | 0.9078333333333333 | 0.003296556377588175 |
+| Epoch_11.pt | 0.9076666666666666 | 0.0024241582476968214 |
+| Epoch_13_batch_2999.pt | 0.9076666666666666 | 0.002867441755680875 |
+| Epoch_17_batch_2999.pt | 0.9075 | 0.003125462928674486 |
+| Epoch_16.pt | 0.9071666666666666 | 0.003239894069294071 |
+| Epoch_12_batch_5999.pt | 0.9048333333333334 | 0.0035525160619440327 |
+| Epoch_12_batch_2999.pt | 0.9046666666666668 | 0.003130889511912304 |
+| Epoch_11_batch_5999.pt | 0.9046666666666667 | 0.003091206165165239 |
+| Epoch_10.pt | 0.9041666666666666 | 0.0032322640461752996 |
+| Epoch_12.pt | 0.9040000000000001 | 0.004121608220220313 |
+| Epoch_10_batch_2999.pt | 0.9031666666666668 | 0.0027605018330677457 |
+| Epoch_10_batch_5999.pt | 0.9031666666666668 | 0.0022005891467042626 |
+| Epoch_11_batch_2999.pt | 0.9021666666666667 | 0.002582586510926545 |
+| Epoch_6_batch_5999.pt | 0.8789999999999999 | 0.004487293445846368 |
+| Epoch_9_batch_5999.pt | 0.8786666666666667 | 0.004751867728965139 |
+| Epoch_7_batch_5999.pt | 0.8775000000000001 | 0.004521894610027696 |
+| Epoch_5_batch_5999.pt | 0.8761666666666666 | 0.004824385075636545 |
+| Epoch_7_batch_2999.pt | 0.8756666666666666 | 0.0032603112780269362 |
+| Epoch_9_batch_2999.pt | 0.8748333333333334 | 0.0032721231828060516 |
+| Epoch_8_batch_2999.pt | 0.8741666666666668 | 0.005590169943749468 |
+| Epoch_9.pt | 0.8725000000000002 | 0.0033077723659120485 |
+| Epoch_6_batch_2999.pt | 0.8720000000000001 | 0.0035642255405212114 |
+| Epoch_6.pt | 0.8698333333333335 | 0.0034645470728153077 |
+| Epoch_5_batch_2999.pt | 0.8696666666666667 | 0.003431876713662335 |
+| Epoch_8_batch_5999.pt | 0.8694999999999998 | 0.00398337594894296 |
+| Epoch_3_batch_5999.pt | 0.8643333333333333 | 0.004283185614976967 |
+| Epoch_8.pt | 0.8606666666666667 | 0.004144012488660335 |
+| Epoch_5.pt | 0.86 | 0.004621474298463427 |
+| Epoch_4.pt | 0.8598333333333332 | 0.003836552593471007 |
+| Epoch_4_batch_5999.pt | 0.859 | 0.004015402444353921 |
+| Epoch_4_batch_2999.pt | 0.8560000000000001 | 0.006056320033294067 |
+| Epoch_3_batch_2999.pt | 0.8553333333333333 | 0.005935871284085815 |
+| Epoch_7.pt | 0.8550000000000001 | 0.005561108336107644 |
+| Epoch_3.pt | 0.8513333333333332 | 0.005084471639398918 |
+| Epoch_2_batch_5999.pt | 0.8441666666666666 | 0.00510537117307006 |
+| Epoch_2.pt | 0.8426666666666666 | 0.004642796092394706 |
+| Epoch_2_batch_2999.pt | 0.8413333333333334 | 0.00588888888888889 |
+| Epoch_1_batch_5999.pt | 0.8358333333333334 | 0.0034716666222151106 |
+| Epoch_1.pt | 0.8283333333333334 | 0.004013864859597433 |
+| Epoch_1_batch_2999.pt | 0.8185 | 0.004447568346583426 |
+| Epoch_0.pt | 0.784 | 0.003929942040850532 |
+| Epoch_0_batch_5999.pt | 0.7765 | 0.00459098687580646 |
+| Epoch_0_batch_2999.pt | 0.7211666666666667 | 0.0054379212622399165 |
++------------------------+--------------------+-----------------------+
diff --git a/bob/bio/facexzoo/models/heads/NPCFace/log.log b/bob/bio/facexzoo/models/heads/NPCFace/log.log
new file mode 100644
index 0000000000000000000000000000000000000000..0d06d0a7bbd2445d935e0cddb97ddcd252c69844
--- /dev/null
+++ b/bob/bio/facexzoo/models/heads/NPCFace/log.log
@@ -0,0 +1,657 @@
+INFO 2020-11-25 15:09:54 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/Grammar.txt
+INFO 2020-11-25 15:09:54 driver.py: 124] Generating grammar tables from /usr/lib/python3.6/lib2to3/PatternGrammar.txt
+INFO 2020-11-25 15:09:54 train.py: 172] Start optimization.
+INFO 2020-11-25 15:09:54 train.py: 173] Namespace(backbone_conf_file='../backbone_conf.yaml', backbone_type='Mobilefacenets', batch_size=512, data_root='/home/wangjun492/wj_data/facex-zoo/deepglint/msra_crop', epoches=18, head_conf_file='../head_conf.yaml', head_type='NPCFace', log_dir='log', lr=0.1, milestones=[10, 13, 16], momentum=0.9, out_dir='out_dir', pretrain_model='mv_epoch_8.pt', print_freq=200, resume=False, save_freq=3000, step='10, 13, 16', tensorboardx_logdir='npc-mobile', train_file='/home/wangjun492/wj_data/facex-zoo/deepglint/msceleb_deepglint_train_file.txt', writer=<tensorboardX.writer.SummaryWriter object at 0x7efbf375c390>)
+backbone param:
+{'feat_dim': 512, 'out_h': 7, 'out_w': 7}
+head param:
+{'feat_dim': 512, 'num_class': 72778, 'margin': 0.35, 'scale': 32}
+INFO 2020-11-25 15:10:19 train.py: 74] Epoch 0, iter 0/6416, lr 0.100000, loss 24.927702
+INFO 2020-11-25 15:11:49 train.py: 74] Epoch 0, iter 200/6416, lr 0.100000, loss 25.481443
+INFO 2020-11-25 15:13:18 train.py: 74] Epoch 0, iter 400/6416, lr 0.100000, loss 26.659269
+INFO 2020-11-25 15:14:48 train.py: 74] Epoch 0, iter 600/6416, lr 0.100000, loss 27.631736
+INFO 2020-11-25 15:16:17 train.py: 74] Epoch 0, iter 800/6416, lr 0.100000, loss 28.146001
+INFO 2020-11-25 15:17:47 train.py: 74] Epoch 0, iter 1000/6416, lr 0.100000, loss 28.272338
+INFO 2020-11-25 15:19:17 train.py: 74] Epoch 0, iter 1200/6416, lr 0.100000, loss 28.236835
+INFO 2020-11-25 15:20:46 train.py: 74] Epoch 0, iter 1400/6416, lr 0.100000, loss 28.099825
+INFO 2020-11-25 15:22:16 train.py: 74] Epoch 0, iter 1600/6416, lr 0.100000, loss 27.929273
+INFO 2020-11-25 15:23:45 train.py: 74] Epoch 0, iter 1800/6416, lr 0.100000, loss 27.690363
+INFO 2020-11-25 15:25:15 train.py: 74] Epoch 0, iter 2000/6416, lr 0.100000, loss 27.420151
+INFO 2020-11-25 15:26:44 train.py: 74] Epoch 0, iter 2200/6416, lr 0.100000, loss 27.146129
+INFO 2020-11-25 15:28:14 train.py: 74] Epoch 0, iter 2400/6416, lr 0.100000, loss 26.847339
+INFO 2020-11-25 15:29:44 train.py: 74] Epoch 0, iter 2600/6416, lr 0.100000, loss 26.526757
+INFO 2020-11-25 15:31:13 train.py: 74] Epoch 0, iter 2800/6416, lr 0.100000, loss 26.193765
+INFO 2020-11-25 15:32:43 train.py: 87] Save checkpoint Epoch_0_batch_2999.pt to disk.
+INFO 2020-11-25 15:32:43 train.py: 74] Epoch 0, iter 3000/6416, lr 0.100000, loss 25.847690
+INFO 2020-11-25 15:34:14 train.py: 74] Epoch 0, iter 3200/6416, lr 0.100000, loss 25.490689
+INFO 2020-11-25 15:35:44 train.py: 74] Epoch 0, iter 3400/6416, lr 0.100000, loss 25.152195
+INFO 2020-11-25 15:37:15 train.py: 74] Epoch 0, iter 3600/6416, lr 0.100000, loss 24.809073
+INFO 2020-11-25 15:38:46 train.py: 74] Epoch 0, iter 3800/6416, lr 0.100000, loss 24.446493
+INFO 2020-11-25 15:40:16 train.py: 74] Epoch 0, iter 4000/6416, lr 0.100000, loss 24.098312
+INFO 2020-11-25 15:41:47 train.py: 74] Epoch 0, iter 4200/6416, lr 0.100000, loss 23.741335
+INFO 2020-11-25 15:43:17 train.py: 74] Epoch 0, iter 4400/6416, lr 0.100000, loss 23.359606
+INFO 2020-11-25 15:44:47 train.py: 74] Epoch 0, iter 4600/6416, lr 0.100000, loss 23.006592
+INFO 2020-11-25 15:46:18 train.py: 74] Epoch 0, iter 4800/6416, lr 0.100000, loss 22.597682
+INFO 2020-11-25 15:47:49 train.py: 74] Epoch 0, iter 5000/6416, lr 0.100000, loss 22.221911
+INFO 2020-11-25 15:49:19 train.py: 74] Epoch 0, iter 5200/6416, lr 0.100000, loss 21.765906
+INFO 2020-11-25 15:50:50 train.py: 74] Epoch 0, iter 5400/6416, lr 0.100000, loss 21.355992
+INFO 2020-11-25 15:52:20 train.py: 74] Epoch 0, iter 5600/6416, lr 0.100000, loss 20.924678
+INFO 2020-11-25 15:53:50 train.py: 74] Epoch 0, iter 5800/6416, lr 0.100000, loss 20.485877
+INFO 2020-11-25 15:55:21 train.py: 87] Save checkpoint Epoch_0_batch_5999.pt to disk.
+INFO 2020-11-25 15:55:21 train.py: 74] Epoch 0, iter 6000/6416, lr 0.100000, loss 20.034333
+INFO 2020-11-25 15:56:52 train.py: 74] Epoch 0, iter 6200/6416, lr 0.100000, loss 19.576644
+INFO 2020-11-25 15:58:22 train.py: 74] Epoch 0, iter 6400/6416, lr 0.100000, loss 19.127555
+INFO 2020-11-25 15:58:29 train.py: 92] Save checkpoint Epoch_0.pt to disk...
+INFO 2020-11-25 15:58:31 train.py: 74] Epoch 1, iter 0/6416, lr 0.100000, loss 18.911504
+INFO 2020-11-25 16:00:01 train.py: 74] Epoch 1, iter 200/6416, lr 0.100000, loss 18.296700
+INFO 2020-11-25 16:01:31 train.py: 74] Epoch 1, iter 400/6416, lr 0.100000, loss 17.933004
+INFO 2020-11-25 16:03:01 train.py: 74] Epoch 1, iter 600/6416, lr 0.100000, loss 17.637282
+INFO 2020-11-25 16:04:31 train.py: 74] Epoch 1, iter 800/6416, lr 0.100000, loss 17.380947
+INFO 2020-11-25 16:06:01 train.py: 74] Epoch 1, iter 1000/6416, lr 0.100000, loss 17.136755
+INFO 2020-11-25 16:07:31 train.py: 74] Epoch 1, iter 1200/6416, lr 0.100000, loss 16.809908
+INFO 2020-11-25 16:09:02 train.py: 74] Epoch 1, iter 1400/6416, lr 0.100000, loss 16.530462
+INFO 2020-11-25 16:10:32 train.py: 74] Epoch 1, iter 1600/6416, lr 0.100000, loss 16.341601
+INFO 2020-11-25 16:12:01 train.py: 74] Epoch 1, iter 1800/6416, lr 0.100000, loss 16.097688
+INFO 2020-11-25 16:13:32 train.py: 74] Epoch 1, iter 2000/6416, lr 0.100000, loss 15.917136
+INFO 2020-11-25 16:15:02 train.py: 74] Epoch 1, iter 2200/6416, lr 0.100000, loss 15.705214
+INFO 2020-11-25 16:16:32 train.py: 74] Epoch 1, iter 2400/6416, lr 0.100000, loss 15.517589
+INFO 2020-11-25 16:18:02 train.py: 74] Epoch 1, iter 2600/6416, lr 0.100000, loss 15.342113
+INFO 2020-11-25 16:19:32 train.py: 74] Epoch 1, iter 2800/6416, lr 0.100000, loss 15.124372
+INFO 2020-11-25 16:21:02 train.py: 87] Save checkpoint Epoch_1_batch_2999.pt to disk.
+INFO 2020-11-25 16:21:03 train.py: 74] Epoch 1, iter 3000/6416, lr 0.100000, loss 14.998705
+INFO 2020-11-25 16:22:33 train.py: 74] Epoch 1, iter 3200/6416, lr 0.100000, loss 14.836050
+INFO 2020-11-25 16:24:03 train.py: 74] Epoch 1, iter 3400/6416, lr 0.100000, loss 14.713484
+INFO 2020-11-25 16:25:33 train.py: 74] Epoch 1, iter 3600/6416, lr 0.100000, loss 14.528275
+INFO 2020-11-25 16:27:03 train.py: 74] Epoch 1, iter 3800/6416, lr 0.100000, loss 14.510679
+INFO 2020-11-25 16:28:33 train.py: 74] Epoch 1, iter 4000/6416, lr 0.100000, loss 14.391305
+INFO 2020-11-25 16:30:03 train.py: 74] Epoch 1, iter 4200/6416, lr 0.100000, loss 14.260403
+INFO 2020-11-25 16:31:33 train.py: 74] Epoch 1, iter 4400/6416, lr 0.100000, loss 14.123975
+INFO 2020-11-25 16:33:03 train.py: 74] Epoch 1, iter 4600/6416, lr 0.100000, loss 14.033432
+INFO 2020-11-25 16:34:33 train.py: 74] Epoch 1, iter 4800/6416, lr 0.100000, loss 13.982409
+INFO 2020-11-25 16:36:03 train.py: 74] Epoch 1, iter 5000/6416, lr 0.100000, loss 13.882295
+INFO 2020-11-25 16:37:33 train.py: 74] Epoch 1, iter 5200/6416, lr 0.100000, loss 13.707344
+INFO 2020-11-25 16:39:02 train.py: 74] Epoch 1, iter 5400/6416, lr 0.100000, loss 13.685294
+INFO 2020-11-25 16:40:33 train.py: 74] Epoch 1, iter 5600/6416, lr 0.100000, loss 13.617356
+INFO 2020-11-25 16:42:03 train.py: 74] Epoch 1, iter 5800/6416, lr 0.100000, loss 13.505162
+INFO 2020-11-25 16:43:32 train.py: 87] Save checkpoint Epoch_1_batch_5999.pt to disk.
+INFO 2020-11-25 16:43:33 train.py: 74] Epoch 1, iter 6000/6416, lr 0.100000, loss 13.465484
+INFO 2020-11-25 16:45:03 train.py: 74] Epoch 1, iter 6200/6416, lr 0.100000, loss 13.385129
+INFO 2020-11-25 16:46:34 train.py: 74] Epoch 1, iter 6400/6416, lr 0.100000, loss 13.262779
+INFO 2020-11-25 16:46:41 train.py: 92] Save checkpoint Epoch_1.pt to disk...
+INFO 2020-11-25 16:46:43 train.py: 74] Epoch 2, iter 0/6416, lr 0.100000, loss 13.185356
+INFO 2020-11-25 16:48:14 train.py: 74] Epoch 2, iter 200/6416, lr 0.100000, loss 12.468792
+INFO 2020-11-25 16:49:45 train.py: 74] Epoch 2, iter 400/6416, lr 0.100000, loss 12.518446
+INFO 2020-11-25 16:51:16 train.py: 74] Epoch 2, iter 600/6416, lr 0.100000, loss 12.622364
+INFO 2020-11-25 16:52:47 train.py: 74] Epoch 2, iter 800/6416, lr 0.100000, loss 12.651303
+INFO 2020-11-25 16:54:18 train.py: 74] Epoch 2, iter 1000/6416, lr 0.100000, loss 12.685659
+INFO 2020-11-25 16:55:49 train.py: 74] Epoch 2, iter 1200/6416, lr 0.100000, loss 12.678821
+INFO 2020-11-25 16:57:20 train.py: 74] Epoch 2, iter 1400/6416, lr 0.100000, loss 12.694099
+INFO 2020-11-25 16:58:51 train.py: 74] Epoch 2, iter 1600/6416, lr 0.100000, loss 12.685656
+INFO 2020-11-25 17:00:22 train.py: 74] Epoch 2, iter 1800/6416, lr 0.100000, loss 12.673783
+INFO 2020-11-25 17:01:53 train.py: 74] Epoch 2, iter 2000/6416, lr 0.100000, loss 12.593529
+INFO 2020-11-25 17:03:25 train.py: 74] Epoch 2, iter 2200/6416, lr 0.100000, loss 12.576401
+INFO 2020-11-25 17:04:56 train.py: 74] Epoch 2, iter 2400/6416, lr 0.100000, loss 12.516576
+INFO 2020-11-25 17:06:27 train.py: 74] Epoch 2, iter 2600/6416, lr 0.100000, loss 12.537273
+INFO 2020-11-25 17:07:59 train.py: 74] Epoch 2, iter 2800/6416, lr 0.100000, loss 12.431063
+INFO 2020-11-25 17:09:30 train.py: 87] Save checkpoint Epoch_2_batch_2999.pt to disk.
+INFO 2020-11-25 17:09:30 train.py: 74] Epoch 2, iter 3000/6416, lr 0.100000, loss 12.405576
+INFO 2020-11-25 17:11:02 train.py: 74] Epoch 2, iter 3200/6416, lr 0.100000, loss 12.372717
+INFO 2020-11-25 17:12:34 train.py: 74] Epoch 2, iter 3400/6416, lr 0.100000, loss 12.407423
+INFO 2020-11-25 17:14:05 train.py: 74] Epoch 2, iter 3600/6416, lr 0.100000, loss 12.299461
+INFO 2020-11-25 17:15:37 train.py: 74] Epoch 2, iter 3800/6416, lr 0.100000, loss 12.250163
+INFO 2020-11-25 17:17:09 train.py: 74] Epoch 2, iter 4000/6416, lr 0.100000, loss 12.273306
+INFO 2020-11-25 17:18:41 train.py: 74] Epoch 2, iter 4200/6416, lr 0.100000, loss 12.175709
+INFO 2020-11-25 17:20:13 train.py: 74] Epoch 2, iter 4400/6416, lr 0.100000, loss 12.185333
+INFO 2020-11-25 17:21:44 train.py: 74] Epoch 2, iter 4600/6416, lr 0.100000, loss 12.133005
+INFO 2020-11-25 17:23:16 train.py: 74] Epoch 2, iter 4800/6416, lr 0.100000, loss 12.055537
+INFO 2020-11-25 17:24:48 train.py: 74] Epoch 2, iter 5000/6416, lr 0.100000, loss 12.051655
+INFO 2020-11-25 17:26:20 train.py: 74] Epoch 2, iter 5200/6416, lr 0.100000, loss 12.050668
+INFO 2020-11-25 17:27:51 train.py: 74] Epoch 2, iter 5400/6416, lr 0.100000, loss 12.024552
+INFO 2020-11-25 17:29:23 train.py: 74] Epoch 2, iter 5600/6416, lr 0.100000, loss 11.949274
+INFO 2020-11-25 17:30:55 train.py: 74] Epoch 2, iter 5800/6416, lr 0.100000, loss 11.900883
+INFO 2020-11-25 17:32:27 train.py: 87] Save checkpoint Epoch_2_batch_5999.pt to disk.
+INFO 2020-11-25 17:32:27 train.py: 74] Epoch 2, iter 6000/6416, lr 0.100000, loss 11.891491
+INFO 2020-11-25 17:33:58 train.py: 74] Epoch 2, iter 6200/6416, lr 0.100000, loss 11.839428
+INFO 2020-11-25 17:35:29 train.py: 74] Epoch 2, iter 6400/6416, lr 0.100000, loss 11.792616
+INFO 2020-11-25 17:35:36 train.py: 92] Save checkpoint Epoch_2.pt to disk...
+INFO 2020-11-25 17:35:38 train.py: 74] Epoch 3, iter 0/6416, lr 0.100000, loss 11.534243
+INFO 2020-11-25 17:37:09 train.py: 74] Epoch 3, iter 200/6416, lr 0.100000, loss 11.044183
+INFO 2020-11-25 17:38:39 train.py: 74] Epoch 3, iter 400/6416, lr 0.100000, loss 11.048038
+INFO 2020-11-25 17:40:10 train.py: 74] Epoch 3, iter 600/6416, lr 0.100000, loss 11.197148
+INFO 2020-11-25 17:41:40 train.py: 74] Epoch 3, iter 800/6416, lr 0.100000, loss 11.270086
+INFO 2020-11-25 17:43:10 train.py: 74] Epoch 3, iter 1000/6416, lr 0.100000, loss 11.389561
+INFO 2020-11-25 17:44:41 train.py: 74] Epoch 3, iter 1200/6416, lr 0.100000, loss 11.395745
+INFO 2020-11-25 17:46:11 train.py: 74] Epoch 3, iter 1400/6416, lr 0.100000, loss 11.422172
+INFO 2020-11-25 17:47:42 train.py: 74] Epoch 3, iter 1600/6416, lr 0.100000, loss 11.410723
+INFO 2020-11-25 17:49:12 train.py: 74] Epoch 3, iter 1800/6416, lr 0.100000, loss 11.483555
+INFO 2020-11-25 17:50:43 train.py: 74] Epoch 3, iter 2000/6416, lr 0.100000, loss 11.449949
+INFO 2020-11-25 17:52:13 train.py: 74] Epoch 3, iter 2200/6416, lr 0.100000, loss 11.391050
+INFO 2020-11-25 17:53:44 train.py: 74] Epoch 3, iter 2400/6416, lr 0.100000, loss 11.413115
+INFO 2020-11-25 17:55:15 train.py: 74] Epoch 3, iter 2600/6416, lr 0.100000, loss 11.387014
+INFO 2020-11-25 17:56:45 train.py: 74] Epoch 3, iter 2800/6416, lr 0.100000, loss 11.391482
+INFO 2020-11-25 17:58:16 train.py: 87] Save checkpoint Epoch_3_batch_2999.pt to disk.
+INFO 2020-11-25 17:58:16 train.py: 74] Epoch 3, iter 3000/6416, lr 0.100000, loss 11.411233
+INFO 2020-11-25 17:59:48 train.py: 74] Epoch 3, iter 3200/6416, lr 0.100000, loss 11.315576
+INFO 2020-11-25 18:01:19 train.py: 74] Epoch 3, iter 3400/6416, lr 0.100000, loss 11.334390
+INFO 2020-11-25 18:02:51 train.py: 74] Epoch 3, iter 3600/6416, lr 0.100000, loss 11.312477
+INFO 2020-11-25 18:04:23 train.py: 74] Epoch 3, iter 3800/6416, lr 0.100000, loss 11.266192
+INFO 2020-11-25 18:05:54 train.py: 74] Epoch 3, iter 4000/6416, lr 0.100000, loss 11.283333
+INFO 2020-11-25 18:07:26 train.py: 74] Epoch 3, iter 4200/6416, lr 0.100000, loss 11.224462
+INFO 2020-11-25 18:08:58 train.py: 74] Epoch 3, iter 4400/6416, lr 0.100000, loss 11.209081
+INFO 2020-11-25 18:10:30 train.py: 74] Epoch 3, iter 4600/6416, lr 0.100000, loss 11.209695
+INFO 2020-11-25 18:12:02 train.py: 74] Epoch 3, iter 4800/6416, lr 0.100000, loss 11.182206
+INFO 2020-11-25 18:13:34 train.py: 74] Epoch 3, iter 5000/6416, lr 0.100000, loss 11.173278
+INFO 2020-11-25 18:15:05 train.py: 74] Epoch 3, iter 5200/6416, lr 0.100000, loss 11.109613
+INFO 2020-11-25 18:16:37 train.py: 74] Epoch 3, iter 5400/6416, lr 0.100000, loss 11.166652
+INFO 2020-11-25 18:18:09 train.py: 74] Epoch 3, iter 5600/6416, lr 0.100000, loss 11.093884
+INFO 2020-11-25 18:19:41 train.py: 74] Epoch 3, iter 5800/6416, lr 0.100000, loss 11.159175
+INFO 2020-11-25 18:21:13 train.py: 87] Save checkpoint Epoch_3_batch_5999.pt to disk.
+INFO 2020-11-25 18:21:13 train.py: 74] Epoch 3, iter 6000/6416, lr 0.100000, loss 11.038569
+INFO 2020-11-25 18:22:45 train.py: 74] Epoch 3, iter 6200/6416, lr 0.100000, loss 11.029128
+INFO 2020-11-25 18:24:16 train.py: 74] Epoch 3, iter 6400/6416, lr 0.100000, loss 11.063368
+INFO 2020-11-25 18:24:23 train.py: 92] Save checkpoint Epoch_3.pt to disk...
+INFO 2020-11-25 18:24:25 train.py: 74] Epoch 4, iter 0/6416, lr 0.100000, loss 11.057874
+INFO 2020-11-25 18:25:56 train.py: 74] Epoch 4, iter 200/6416, lr 0.100000, loss 10.321902
+INFO 2020-11-25 18:27:27 train.py: 74] Epoch 4, iter 400/6416, lr 0.100000, loss 10.310155
+INFO 2020-11-25 18:28:58 train.py: 74] Epoch 4, iter 600/6416, lr 0.100000, loss 10.422067
+INFO 2020-11-25 18:30:29 train.py: 74] Epoch 4, iter 800/6416, lr 0.100000, loss 10.577450
+INFO 2020-11-25 18:32:00 train.py: 74] Epoch 4, iter 1000/6416, lr 0.100000, loss 10.692825
+INFO 2020-11-25 18:33:32 train.py: 74] Epoch 4, iter 1200/6416, lr 0.100000, loss 10.633665
+INFO 2020-11-25 18:35:03 train.py: 74] Epoch 4, iter 1400/6416, lr 0.100000, loss 10.675439
+INFO 2020-11-25 18:36:34 train.py: 74] Epoch 4, iter 1600/6416, lr 0.100000, loss 10.794083
+INFO 2020-11-25 18:38:05 train.py: 74] Epoch 4, iter 1800/6416, lr 0.100000, loss 10.744724
+INFO 2020-11-25 18:39:36 train.py: 74] Epoch 4, iter 2000/6416, lr 0.100000, loss 10.746419
+INFO 2020-11-25 18:41:08 train.py: 74] Epoch 4, iter 2200/6416, lr 0.100000, loss 10.760121
+INFO 2020-11-25 18:42:39 train.py: 74] Epoch 4, iter 2400/6416, lr 0.100000, loss 10.732277
+INFO 2020-11-25 18:44:10 train.py: 74] Epoch 4, iter 2600/6416, lr 0.100000, loss 10.786279
+INFO 2020-11-25 18:45:42 train.py: 74] Epoch 4, iter 2800/6416, lr 0.100000, loss 10.683985
+INFO 2020-11-25 18:47:13 train.py: 87] Save checkpoint Epoch_4_batch_2999.pt to disk.
+INFO 2020-11-25 18:47:13 train.py: 74] Epoch 4, iter 3000/6416, lr 0.100000, loss 10.722752
+INFO 2020-11-25 18:48:45 train.py: 74] Epoch 4, iter 3200/6416, lr 0.100000, loss 10.674939
+INFO 2020-11-25 18:50:17 train.py: 74] Epoch 4, iter 3400/6416, lr 0.100000, loss 10.775858
+INFO 2020-11-25 18:51:48 train.py: 74] Epoch 4, iter 3600/6416, lr 0.100000, loss 10.730112
+INFO 2020-11-25 18:53:20 train.py: 74] Epoch 4, iter 3800/6416, lr 0.100000, loss 10.684429
+INFO 2020-11-25 18:54:52 train.py: 74] Epoch 4, iter 4000/6416, lr 0.100000, loss 10.720501
+INFO 2020-11-25 18:56:24 train.py: 74] Epoch 4, iter 4200/6416, lr 0.100000, loss 10.638486
+INFO 2020-11-25 18:57:56 train.py: 74] Epoch 4, iter 4400/6416, lr 0.100000, loss 10.686998
+INFO 2020-11-25 18:59:28 train.py: 74] Epoch 4, iter 4600/6416, lr 0.100000, loss 10.717062
+INFO 2020-11-25 19:01:00 train.py: 74] Epoch 4, iter 4800/6416, lr 0.100000, loss 10.634711
+INFO 2020-11-25 19:02:31 train.py: 74] Epoch 4, iter 5000/6416, lr 0.100000, loss 10.602083
+INFO 2020-11-25 19:04:03 train.py: 74] Epoch 4, iter 5200/6416, lr 0.100000, loss 10.600415
+INFO 2020-11-25 19:05:35 train.py: 74] Epoch 4, iter 5400/6416, lr 0.100000, loss 10.600393
+INFO 2020-11-25 19:07:07 train.py: 74] Epoch 4, iter 5600/6416, lr 0.100000, loss 10.605459
+INFO 2020-11-25 19:08:39 train.py: 74] Epoch 4, iter 5800/6416, lr 0.100000, loss 10.620802
+INFO 2020-11-25 19:10:10 train.py: 87] Save checkpoint Epoch_4_batch_5999.pt to disk.
+INFO 2020-11-25 19:10:11 train.py: 74] Epoch 4, iter 6000/6416, lr 0.100000, loss 10.568606
+INFO 2020-11-25 19:11:42 train.py: 74] Epoch 4, iter 6200/6416, lr 0.100000, loss 10.526867
+INFO 2020-11-25 19:13:14 train.py: 74] Epoch 4, iter 6400/6416, lr 0.100000, loss 10.546365
+INFO 2020-11-25 19:13:22 train.py: 92] Save checkpoint Epoch_4.pt to disk...
+INFO 2020-11-25 19:13:23 train.py: 74] Epoch 5, iter 0/6416, lr 0.100000, loss 10.762732
+INFO 2020-11-25 19:14:54 train.py: 74] Epoch 5, iter 200/6416, lr 0.100000, loss 9.803253
+INFO 2020-11-25 19:16:25 train.py: 74] Epoch 5, iter 400/6416, lr 0.100000, loss 9.821998
+INFO 2020-11-25 19:17:57 train.py: 74] Epoch 5, iter 600/6416, lr 0.100000, loss 9.986957
+INFO 2020-11-25 19:19:28 train.py: 74] Epoch 5, iter 800/6416, lr 0.100000, loss 10.062394
+INFO 2020-11-25 19:20:59 train.py: 74] Epoch 5, iter 1000/6416, lr 0.100000, loss 10.162395
+INFO 2020-11-25 19:22:30 train.py: 74] Epoch 5, iter 1200/6416, lr 0.100000, loss 10.249884
+INFO 2020-11-25 19:24:01 train.py: 74] Epoch 5, iter 1400/6416, lr 0.100000, loss 10.294340
+INFO 2020-11-25 19:25:32 train.py: 74] Epoch 5, iter 1600/6416, lr 0.100000, loss 10.268831
+INFO 2020-11-25 19:27:04 train.py: 74] Epoch 5, iter 1800/6416, lr 0.100000, loss 10.333934
+INFO 2020-11-25 19:28:35 train.py: 74] Epoch 5, iter 2000/6416, lr 0.100000, loss 10.364582
+INFO 2020-11-25 19:30:06 train.py: 74] Epoch 5, iter 2200/6416, lr 0.100000, loss 10.292656
+INFO 2020-11-25 19:31:37 train.py: 74] Epoch 5, iter 2400/6416, lr 0.100000, loss 10.364636
+INFO 2020-11-25 19:33:09 train.py: 74] Epoch 5, iter 2600/6416, lr 0.100000, loss 10.346629
+INFO 2020-11-25 19:34:40 train.py: 74] Epoch 5, iter 2800/6416, lr 0.100000, loss 10.319169
+INFO 2020-11-25 19:36:11 train.py: 87] Save checkpoint Epoch_5_batch_2999.pt to disk.
+INFO 2020-11-25 19:36:12 train.py: 74] Epoch 5, iter 3000/6416, lr 0.100000, loss 10.363147
+INFO 2020-11-25 19:37:43 train.py: 74] Epoch 5, iter 3200/6416, lr 0.100000, loss 10.326473
+INFO 2020-11-25 19:39:14 train.py: 74] Epoch 5, iter 3400/6416, lr 0.100000, loss 10.361666
+INFO 2020-11-25 19:40:46 train.py: 74] Epoch 5, iter 3600/6416, lr 0.100000, loss 10.352291
+INFO 2020-11-25 19:42:17 train.py: 74] Epoch 5, iter 3800/6416, lr 0.100000, loss 10.302245
+INFO 2020-11-25 19:43:49 train.py: 74] Epoch 5, iter 4000/6416, lr 0.100000, loss 10.206425
+INFO 2020-11-25 19:45:21 train.py: 74] Epoch 5, iter 4200/6416, lr 0.100000, loss 10.297817
+INFO 2020-11-25 19:46:52 train.py: 74] Epoch 5, iter 4400/6416, lr 0.100000, loss 10.294239
+INFO 2020-11-25 19:48:24 train.py: 74] Epoch 5, iter 4600/6416, lr 0.100000, loss 10.335441
+INFO 2020-11-25 19:49:56 train.py: 74] Epoch 5, iter 4800/6416, lr 0.100000, loss 10.272924
+INFO 2020-11-25 19:51:27 train.py: 74] Epoch 5, iter 5000/6416, lr 0.100000, loss 10.262506
+INFO 2020-11-25 19:52:59 train.py: 74] Epoch 5, iter 5200/6416, lr 0.100000, loss 10.278853
+INFO 2020-11-25 19:54:31 train.py: 74] Epoch 5, iter 5400/6416, lr 0.100000, loss 10.283840
+INFO 2020-11-25 19:56:02 train.py: 74] Epoch 5, iter 5600/6416, lr 0.100000, loss 10.228839
+INFO 2020-11-25 19:57:34 train.py: 74] Epoch 5, iter 5800/6416, lr 0.100000, loss 10.191768
+INFO 2020-11-25 19:59:05 train.py: 87] Save checkpoint Epoch_5_batch_5999.pt to disk.
+INFO 2020-11-25 19:59:06 train.py: 74] Epoch 5, iter 6000/6416, lr 0.100000, loss 10.200151
+INFO 2020-11-25 20:00:37 train.py: 74] Epoch 5, iter 6200/6416, lr 0.100000, loss 10.242199
+INFO 2020-11-25 20:02:08 train.py: 74] Epoch 5, iter 6400/6416, lr 0.100000, loss 10.184401
+INFO 2020-11-25 20:02:15 train.py: 92] Save checkpoint Epoch_5.pt to disk...
+INFO 2020-11-25 20:02:17 train.py: 74] Epoch 6, iter 0/6416, lr 0.100000, loss 10.106567
+INFO 2020-11-25 20:03:48 train.py: 74] Epoch 6, iter 200/6416, lr 0.100000, loss 9.496195
+INFO 2020-11-25 20:05:18 train.py: 74] Epoch 6, iter 400/6416, lr 0.100000, loss 9.510301
+INFO 2020-11-25 20:06:49 train.py: 74] Epoch 6, iter 600/6416, lr 0.100000, loss 9.583626
+INFO 2020-11-25 20:08:19 train.py: 74] Epoch 6, iter 800/6416, lr 0.100000, loss 9.736390
+INFO 2020-11-25 20:09:50 train.py: 74] Epoch 6, iter 1000/6416, lr 0.100000, loss 9.819257
+INFO 2020-11-25 20:11:20 train.py: 74] Epoch 6, iter 1200/6416, lr 0.100000, loss 9.921078
+INFO 2020-11-25 20:12:50 train.py: 74] Epoch 6, iter 1400/6416, lr 0.100000, loss 9.965668
+INFO 2020-11-25 20:14:21 train.py: 74] Epoch 6, iter 1600/6416, lr 0.100000, loss 9.946194
+INFO 2020-11-25 20:15:52 train.py: 74] Epoch 6, iter 1800/6416, lr 0.100000, loss 9.993823
+INFO 2020-11-25 20:17:22 train.py: 74] Epoch 6, iter 2000/6416, lr 0.100000, loss 10.049934
+INFO 2020-11-25 20:18:53 train.py: 74] Epoch 6, iter 2200/6416, lr 0.100000, loss 10.080067
+INFO 2020-11-25 20:20:24 train.py: 74] Epoch 6, iter 2400/6416, lr 0.100000, loss 10.078383
+INFO 2020-11-25 20:21:54 train.py: 74] Epoch 6, iter 2600/6416, lr 0.100000, loss 10.061219
+INFO 2020-11-25 20:23:25 train.py: 74] Epoch 6, iter 2800/6416, lr 0.100000, loss 10.055866
+INFO 2020-11-25 20:24:56 train.py: 87] Save checkpoint Epoch_6_batch_2999.pt to disk.
+INFO 2020-11-25 20:24:56 train.py: 74] Epoch 6, iter 3000/6416, lr 0.100000, loss 10.033821
+INFO 2020-11-25 20:26:28 train.py: 74] Epoch 6, iter 3200/6416, lr 0.100000, loss 10.000279
+INFO 2020-11-25 20:28:00 train.py: 74] Epoch 6, iter 3400/6416, lr 0.100000, loss 10.007907
+INFO 2020-11-25 20:29:31 train.py: 74] Epoch 6, iter 3600/6416, lr 0.100000, loss 10.016073
+INFO 2020-11-25 20:31:03 train.py: 74] Epoch 6, iter 3800/6416, lr 0.100000, loss 10.019750
+INFO 2020-11-25 20:32:35 train.py: 74] Epoch 6, iter 4000/6416, lr 0.100000, loss 10.066343
+INFO 2020-11-25 20:34:06 train.py: 74] Epoch 6, iter 4200/6416, lr 0.100000, loss 10.063196
+INFO 2020-11-25 20:35:38 train.py: 74] Epoch 6, iter 4400/6416, lr 0.100000, loss 10.046612
+INFO 2020-11-25 20:37:10 train.py: 74] Epoch 6, iter 4600/6416, lr 0.100000, loss 9.980377
+INFO 2020-11-25 20:38:42 train.py: 74] Epoch 6, iter 4800/6416, lr 0.100000, loss 9.971444
+INFO 2020-11-25 20:40:13 train.py: 74] Epoch 6, iter 5000/6416, lr 0.100000, loss 9.951295
+INFO 2020-11-25 20:41:45 train.py: 74] Epoch 6, iter 5200/6416, lr 0.100000, loss 9.939071
+INFO 2020-11-25 20:43:17 train.py: 74] Epoch 6, iter 5400/6416, lr 0.100000, loss 10.011326
+INFO 2020-11-25 20:44:49 train.py: 74] Epoch 6, iter 5600/6416, lr 0.100000, loss 9.938532
+INFO 2020-11-25 20:46:21 train.py: 74] Epoch 6, iter 5800/6416, lr 0.100000, loss 9.879780
+INFO 2020-11-25 20:47:52 train.py: 87] Save checkpoint Epoch_6_batch_5999.pt to disk.
+INFO 2020-11-25 20:47:53 train.py: 74] Epoch 6, iter 6000/6416, lr 0.100000, loss 9.991037
+INFO 2020-11-25 20:49:25 train.py: 74] Epoch 6, iter 6200/6416, lr 0.100000, loss 9.971898
+INFO 2020-11-25 20:50:56 train.py: 74] Epoch 6, iter 6400/6416, lr 0.100000, loss 9.964046
+INFO 2020-11-25 20:51:03 train.py: 92] Save checkpoint Epoch_6.pt to disk...
+INFO 2020-11-25 20:51:05 train.py: 74] Epoch 7, iter 0/6416, lr 0.100000, loss 9.899056
+INFO 2020-11-25 20:52:36 train.py: 74] Epoch 7, iter 200/6416, lr 0.100000, loss 9.271046
+INFO 2020-11-25 20:54:08 train.py: 74] Epoch 7, iter 400/6416, lr 0.100000, loss 9.260704
+INFO 2020-11-25 20:55:39 train.py: 74] Epoch 7, iter 600/6416, lr 0.100000, loss 9.393151
+INFO 2020-11-25 20:57:10 train.py: 74] Epoch 7, iter 800/6416, lr 0.100000, loss 9.512562
+INFO 2020-11-25 20:58:41 train.py: 74] Epoch 7, iter 1000/6416, lr 0.100000, loss 9.601516
+INFO 2020-11-25 21:00:12 train.py: 74] Epoch 7, iter 1200/6416, lr 0.100000, loss 9.694700
+INFO 2020-11-25 21:01:43 train.py: 74] Epoch 7, iter 1400/6416, lr 0.100000, loss 9.691284
+INFO 2020-11-25 21:03:14 train.py: 74] Epoch 7, iter 1600/6416, lr 0.100000, loss 9.716006
+INFO 2020-11-25 21:04:45 train.py: 74] Epoch 7, iter 1800/6416, lr 0.100000, loss 9.743239
+INFO 2020-11-25 21:06:17 train.py: 74] Epoch 7, iter 2000/6416, lr 0.100000, loss 9.776830
+INFO 2020-11-25 21:07:48 train.py: 74] Epoch 7, iter 2200/6416, lr 0.100000, loss 9.818703
+INFO 2020-11-25 21:09:19 train.py: 74] Epoch 7, iter 2400/6416, lr 0.100000, loss 9.841673
+INFO 2020-11-25 21:10:50 train.py: 74] Epoch 7, iter 2600/6416, lr 0.100000, loss 9.735108
+INFO 2020-11-25 21:12:22 train.py: 74] Epoch 7, iter 2800/6416, lr 0.100000, loss 9.783545
+INFO 2020-11-25 21:13:53 train.py: 87] Save checkpoint Epoch_7_batch_2999.pt to disk.
+INFO 2020-11-25 21:13:53 train.py: 74] Epoch 7, iter 3000/6416, lr 0.100000, loss 9.817886
+INFO 2020-11-25 21:15:24 train.py: 74] Epoch 7, iter 3200/6416, lr 0.100000, loss 9.787124
+INFO 2020-11-25 21:16:55 train.py: 74] Epoch 7, iter 3400/6416, lr 0.100000, loss 9.839368
+INFO 2020-11-25 21:18:26 train.py: 74] Epoch 7, iter 3600/6416, lr 0.100000, loss 9.797990
+INFO 2020-11-25 21:19:57 train.py: 74] Epoch 7, iter 3800/6416, lr 0.100000, loss 9.798788
+INFO 2020-11-25 21:21:28 train.py: 74] Epoch 7, iter 4000/6416, lr 0.100000, loss 9.754932
+INFO 2020-11-25 21:22:59 train.py: 74] Epoch 7, iter 4200/6416, lr 0.100000, loss 9.789546
+INFO 2020-11-25 21:24:30 train.py: 74] Epoch 7, iter 4400/6416, lr 0.100000, loss 9.799263
+INFO 2020-11-25 21:26:01 train.py: 74] Epoch 7, iter 4600/6416, lr 0.100000, loss 9.776393
+INFO 2020-11-25 21:27:32 train.py: 74] Epoch 7, iter 4800/6416, lr 0.100000, loss 9.781977
+INFO 2020-11-25 21:29:04 train.py: 74] Epoch 7, iter 5000/6416, lr 0.100000, loss 9.782919
+INFO 2020-11-25 21:30:35 train.py: 74] Epoch 7, iter 5200/6416, lr 0.100000, loss 9.738867
+INFO 2020-11-25 21:32:06 train.py: 74] Epoch 7, iter 5400/6416, lr 0.100000, loss 9.789223
+INFO 2020-11-25 21:33:37 train.py: 74] Epoch 7, iter 5600/6416, lr 0.100000, loss 9.723247
+INFO 2020-11-25 21:35:09 train.py: 74] Epoch 7, iter 5800/6416, lr 0.100000, loss 9.735522
+INFO 2020-11-25 21:36:40 train.py: 87] Save checkpoint Epoch_7_batch_5999.pt to disk.
+INFO 2020-11-25 21:36:40 train.py: 74] Epoch 7, iter 6000/6416, lr 0.100000, loss 9.717222
+INFO 2020-11-25 21:38:12 train.py: 74] Epoch 7, iter 6200/6416, lr 0.100000, loss 9.714263
+INFO 2020-11-25 21:39:44 train.py: 74] Epoch 7, iter 6400/6416, lr 0.100000, loss 9.681192
+INFO 2020-11-25 21:39:51 train.py: 92] Save checkpoint Epoch_7.pt to disk...
+INFO 2020-11-25 21:39:52 train.py: 74] Epoch 8, iter 0/6416, lr 0.100000, loss 9.711530
+INFO 2020-11-25 21:41:24 train.py: 74] Epoch 8, iter 200/6416, lr 0.100000, loss 8.984945
+INFO 2020-11-25 21:42:55 train.py: 74] Epoch 8, iter 400/6416, lr 0.100000, loss 9.039628
+INFO 2020-11-25 21:44:26 train.py: 74] Epoch 8, iter 600/6416, lr 0.100000, loss 9.191416
+INFO 2020-11-25 21:45:57 train.py: 74] Epoch 8, iter 800/6416, lr 0.100000, loss 9.297601
+INFO 2020-11-25 21:47:28 train.py: 74] Epoch 8, iter 1000/6416, lr 0.100000, loss 9.395964
+INFO 2020-11-25 21:48:59 train.py: 74] Epoch 8, iter 1200/6416, lr 0.100000, loss 9.485150
+INFO 2020-11-25 21:50:30 train.py: 74] Epoch 8, iter 1400/6416, lr 0.100000, loss 9.544018
+INFO 2020-11-25 21:52:01 train.py: 74] Epoch 8, iter 1600/6416, lr 0.100000, loss 9.548304
+INFO 2020-11-25 21:53:33 train.py: 74] Epoch 8, iter 1800/6416, lr 0.100000, loss 9.511209
+INFO 2020-11-25 21:55:04 train.py: 74] Epoch 8, iter 2000/6416, lr 0.100000, loss 9.597686
+INFO 2020-11-25 21:56:35 train.py: 74] Epoch 8, iter 2200/6416, lr 0.100000, loss 9.614700
+INFO 2020-11-25 21:58:06 train.py: 74] Epoch 8, iter 2400/6416, lr 0.100000, loss 9.607562
+INFO 2020-11-25 21:59:38 train.py: 74] Epoch 8, iter 2600/6416, lr 0.100000, loss 9.592671
+INFO 2020-11-25 22:01:09 train.py: 74] Epoch 8, iter 2800/6416, lr 0.100000, loss 9.603088
+INFO 2020-11-25 22:02:40 train.py: 87] Save checkpoint Epoch_8_batch_2999.pt to disk.
+INFO 2020-11-25 22:02:41 train.py: 74] Epoch 8, iter 3000/6416, lr 0.100000, loss 9.601038
+INFO 2020-11-25 22:04:12 train.py: 74] Epoch 8, iter 3200/6416, lr 0.100000, loss 9.618175
+INFO 2020-11-25 22:05:43 train.py: 74] Epoch 8, iter 3400/6416, lr 0.100000, loss 9.614943
+INFO 2020-11-25 22:07:14 train.py: 74] Epoch 8, iter 3600/6416, lr 0.100000, loss 9.585135
+INFO 2020-11-25 22:08:45 train.py: 74] Epoch 8, iter 3800/6416, lr 0.100000, loss 9.630282
+INFO 2020-11-25 22:10:16 train.py: 74] Epoch 8, iter 4000/6416, lr 0.100000, loss 9.642341
+INFO 2020-11-25 22:11:47 train.py: 74] Epoch 8, iter 4200/6416, lr 0.100000, loss 9.637504
+INFO 2020-11-25 22:13:19 train.py: 74] Epoch 8, iter 4400/6416, lr 0.100000, loss 9.628689
+INFO 2020-11-25 22:14:50 train.py: 74] Epoch 8, iter 4600/6416, lr 0.100000, loss 9.571567
+INFO 2020-11-25 22:16:22 train.py: 74] Epoch 8, iter 4800/6416, lr 0.100000, loss 9.589931
+INFO 2020-11-25 22:17:53 train.py: 74] Epoch 8, iter 5000/6416, lr 0.100000, loss 9.562254
+INFO 2020-11-25 22:19:24 train.py: 74] Epoch 8, iter 5200/6416, lr 0.100000, loss 9.574488
+INFO 2020-11-25 22:20:55 train.py: 74] Epoch 8, iter 5400/6416, lr 0.100000, loss 9.592067
+INFO 2020-11-25 22:22:26 train.py: 74] Epoch 8, iter 5600/6416, lr 0.100000, loss 9.515037
+INFO 2020-11-25 22:23:58 train.py: 74] Epoch 8, iter 5800/6416, lr 0.100000, loss 9.562857
+INFO 2020-11-25 22:25:29 train.py: 87] Save checkpoint Epoch_8_batch_5999.pt to disk.
+INFO 2020-11-25 22:25:29 train.py: 74] Epoch 8, iter 6000/6416, lr 0.100000, loss 9.547824
+INFO 2020-11-25 22:27:01 train.py: 74] Epoch 8, iter 6200/6416, lr 0.100000, loss 9.551801
+INFO 2020-11-25 22:28:32 train.py: 74] Epoch 8, iter 6400/6416, lr 0.100000, loss 9.552668
+INFO 2020-11-25 22:28:40 train.py: 92] Save checkpoint Epoch_8.pt to disk...
+INFO 2020-11-25 22:28:41 train.py: 74] Epoch 9, iter 0/6416, lr 0.100000, loss 9.565202
+INFO 2020-11-25 22:30:13 train.py: 74] Epoch 9, iter 200/6416, lr 0.100000, loss 8.796665
+INFO 2020-11-25 22:31:44 train.py: 74] Epoch 9, iter 400/6416, lr 0.100000, loss 8.832894
+INFO 2020-11-25 22:33:15 train.py: 74] Epoch 9, iter 600/6416, lr 0.100000, loss 9.049687
+INFO 2020-11-25 22:34:46 train.py: 74] Epoch 9, iter 800/6416, lr 0.100000, loss 9.127557
+INFO 2020-11-25 22:36:17 train.py: 74] Epoch 9, iter 1000/6416, lr 0.100000, loss 9.240772
+INFO 2020-11-25 22:37:48 train.py: 74] Epoch 9, iter 1200/6416, lr 0.100000, loss 9.282510
+INFO 2020-11-25 22:39:19 train.py: 74] Epoch 9, iter 1400/6416, lr 0.100000, loss 9.333620
+INFO 2020-11-25 22:40:50 train.py: 74] Epoch 9, iter 1600/6416, lr 0.100000, loss 9.344223
+INFO 2020-11-25 22:42:21 train.py: 74] Epoch 9, iter 1800/6416, lr 0.100000, loss 9.437895
+INFO 2020-11-25 22:43:52 train.py: 74] Epoch 9, iter 2000/6416, lr 0.100000, loss 9.447560
+INFO 2020-11-25 22:45:23 train.py: 74] Epoch 9, iter 2200/6416, lr 0.100000, loss 9.441592
+INFO 2020-11-25 22:46:55 train.py: 74] Epoch 9, iter 2400/6416, lr 0.100000, loss 9.390768
+INFO 2020-11-25 22:48:26 train.py: 74] Epoch 9, iter 2600/6416, lr 0.100000, loss 9.464535
+INFO 2020-11-25 22:49:57 train.py: 74] Epoch 9, iter 2800/6416, lr 0.100000, loss 9.425357
+INFO 2020-11-25 22:51:28 train.py: 87] Save checkpoint Epoch_9_batch_2999.pt to disk.
+INFO 2020-11-25 22:51:29 train.py: 74] Epoch 9, iter 3000/6416, lr 0.100000, loss 9.436382
+INFO 2020-11-25 22:53:00 train.py: 74] Epoch 9, iter 3200/6416, lr 0.100000, loss 9.466057
+INFO 2020-11-25 22:54:32 train.py: 74] Epoch 9, iter 3400/6416, lr 0.100000, loss 9.507903
+INFO 2020-11-25 22:56:04 train.py: 74] Epoch 9, iter 3600/6416, lr 0.100000, loss 9.458973
+INFO 2020-11-25 22:57:36 train.py: 74] Epoch 9, iter 3800/6416, lr 0.100000, loss 9.452028
+INFO 2020-11-25 22:59:07 train.py: 74] Epoch 9, iter 4000/6416, lr 0.100000, loss 9.442735
+INFO 2020-11-25 23:00:39 train.py: 74] Epoch 9, iter 4200/6416, lr 0.100000, loss 9.365954
+INFO 2020-11-25 23:02:11 train.py: 74] Epoch 9, iter 4400/6416, lr 0.100000, loss 9.399442
+INFO 2020-11-25 23:03:43 train.py: 74] Epoch 9, iter 4600/6416, lr 0.100000, loss 9.447410
+INFO 2020-11-25 23:05:14 train.py: 74] Epoch 9, iter 4800/6416, lr 0.100000, loss 9.446903
+INFO 2020-11-25 23:06:46 train.py: 74] Epoch 9, iter 5000/6416, lr 0.100000, loss 9.442265
+INFO 2020-11-25 23:08:18 train.py: 74] Epoch 9, iter 5200/6416, lr 0.100000, loss 9.402736
+INFO 2020-11-25 23:09:50 train.py: 74] Epoch 9, iter 5400/6416, lr 0.100000, loss 9.413441
+INFO 2020-11-25 23:11:22 train.py: 74] Epoch 9, iter 5600/6416, lr 0.100000, loss 9.411291
+INFO 2020-11-25 23:12:54 train.py: 74] Epoch 9, iter 5800/6416, lr 0.100000, loss 9.388273
+INFO 2020-11-25 23:14:25 train.py: 87] Save checkpoint Epoch_9_batch_5999.pt to disk.
+INFO 2020-11-25 23:14:25 train.py: 74] Epoch 9, iter 6000/6416, lr 0.100000, loss 9.437387
+INFO 2020-11-25 23:15:56 train.py: 74] Epoch 9, iter 6200/6416, lr 0.100000, loss 9.430422
+INFO 2020-11-25 23:17:28 train.py: 74] Epoch 9, iter 6400/6416, lr 0.100000, loss 9.384275
+INFO 2020-11-25 23:17:35 train.py: 92] Save checkpoint Epoch_9.pt to disk...
+INFO 2020-11-25 23:17:36 train.py: 74] Epoch 10, iter 0/6416, lr 0.010000, loss 9.310372
+INFO 2020-11-25 23:19:08 train.py: 74] Epoch 10, iter 200/6416, lr 0.010000, loss 7.696374
+INFO 2020-11-25 23:20:39 train.py: 74] Epoch 10, iter 400/6416, lr 0.010000, loss 7.366358
+INFO 2020-11-25 23:22:10 train.py: 74] Epoch 10, iter 600/6416, lr 0.010000, loss 7.185824
+INFO 2020-11-25 23:23:42 train.py: 74] Epoch 10, iter 800/6416, lr 0.010000, loss 7.078823
+INFO 2020-11-25 23:25:13 train.py: 74] Epoch 10, iter 1000/6416, lr 0.010000, loss 7.020524
+INFO 2020-11-25 23:26:44 train.py: 74] Epoch 10, iter 1200/6416, lr 0.010000, loss 6.939969
+INFO 2020-11-25 23:28:15 train.py: 74] Epoch 10, iter 1400/6416, lr 0.010000, loss 6.877689
+INFO 2020-11-25 23:29:46 train.py: 74] Epoch 10, iter 1600/6416, lr 0.010000, loss 6.806614
+INFO 2020-11-25 23:31:18 train.py: 74] Epoch 10, iter 1800/6416, lr 0.010000, loss 6.783090
+INFO 2020-11-25 23:32:49 train.py: 74] Epoch 10, iter 2000/6416, lr 0.010000, loss 6.713297
+INFO 2020-11-25 23:34:21 train.py: 74] Epoch 10, iter 2200/6416, lr 0.010000, loss 6.676922
+INFO 2020-11-25 23:35:52 train.py: 74] Epoch 10, iter 2400/6416, lr 0.010000, loss 6.684383
+INFO 2020-11-25 23:37:23 train.py: 74] Epoch 10, iter 2600/6416, lr 0.010000, loss 6.622241
+INFO 2020-11-25 23:38:55 train.py: 74] Epoch 10, iter 2800/6416, lr 0.010000, loss 6.637861
+INFO 2020-11-25 23:40:26 train.py: 87] Save checkpoint Epoch_10_batch_2999.pt to disk.
+INFO 2020-11-25 23:40:26 train.py: 74] Epoch 10, iter 3000/6416, lr 0.010000, loss 6.497284
+INFO 2020-11-25 23:41:58 train.py: 74] Epoch 10, iter 3200/6416, lr 0.010000, loss 6.502413
+INFO 2020-11-25 23:43:29 train.py: 74] Epoch 10, iter 3400/6416, lr 0.010000, loss 6.496943
+INFO 2020-11-25 23:45:01 train.py: 74] Epoch 10, iter 3600/6416, lr 0.010000, loss 6.439115
+INFO 2020-11-25 23:46:32 train.py: 74] Epoch 10, iter 3800/6416, lr 0.010000, loss 6.475040
+INFO 2020-11-25 23:48:04 train.py: 74] Epoch 10, iter 4000/6416, lr 0.010000, loss 6.430689
+INFO 2020-11-25 23:49:36 train.py: 74] Epoch 10, iter 4200/6416, lr 0.010000, loss 6.405270
+INFO 2020-11-25 23:51:07 train.py: 74] Epoch 10, iter 4400/6416, lr 0.010000, loss 6.358830
+INFO 2020-11-25 23:52:39 train.py: 74] Epoch 10, iter 4600/6416, lr 0.010000, loss 6.321556
+INFO 2020-11-25 23:54:10 train.py: 74] Epoch 10, iter 4800/6416, lr 0.010000, loss 6.335605
+INFO 2020-11-25 23:55:42 train.py: 74] Epoch 10, iter 5000/6416, lr 0.010000, loss 6.270074
+INFO 2020-11-25 23:57:14 train.py: 74] Epoch 10, iter 5200/6416, lr 0.010000, loss 6.288233
+INFO 2020-11-25 23:58:45 train.py: 74] Epoch 10, iter 5400/6416, lr 0.010000, loss 6.265128
+INFO 2020-11-26 00:00:17 train.py: 74] Epoch 10, iter 5600/6416, lr 0.010000, loss 6.260416
+INFO 2020-11-26 00:01:49 train.py: 74] Epoch 10, iter 5800/6416, lr 0.010000, loss 6.222077
+INFO 2020-11-26 00:03:21 train.py: 87] Save checkpoint Epoch_10_batch_5999.pt to disk.
+INFO 2020-11-26 00:03:21 train.py: 74] Epoch 10, iter 6000/6416, lr 0.010000, loss 6.218251
+INFO 2020-11-26 00:04:53 train.py: 74] Epoch 10, iter 6200/6416, lr 0.010000, loss 6.230017
+INFO 2020-11-26 00:06:24 train.py: 74] Epoch 10, iter 6400/6416, lr 0.010000, loss 6.198102
+INFO 2020-11-26 00:06:32 train.py: 92] Save checkpoint Epoch_10.pt to disk...
+INFO 2020-11-26 00:06:33 train.py: 74] Epoch 11, iter 0/6416, lr 0.010000, loss 6.262739
+INFO 2020-11-26 00:08:04 train.py: 74] Epoch 11, iter 200/6416, lr 0.010000, loss 5.667192
+INFO 2020-11-26 00:09:34 train.py: 74] Epoch 11, iter 400/6416, lr 0.010000, loss 5.679188
+INFO 2020-11-26 00:11:05 train.py: 74] Epoch 11, iter 600/6416, lr 0.010000, loss 5.656450
+INFO 2020-11-26 00:12:36 train.py: 74] Epoch 11, iter 800/6416, lr 0.010000, loss 5.663527
+INFO 2020-11-26 00:14:06 train.py: 74] Epoch 11, iter 1000/6416, lr 0.010000, loss 5.653868
+INFO 2020-11-26 00:15:36 train.py: 74] Epoch 11, iter 1200/6416, lr 0.010000, loss 5.687948
+INFO 2020-11-26 00:17:07 train.py: 74] Epoch 11, iter 1400/6416, lr 0.010000, loss 5.711788
+INFO 2020-11-26 00:18:37 train.py: 74] Epoch 11, iter 1600/6416, lr 0.010000, loss 5.718742
+INFO 2020-11-26 00:20:08 train.py: 74] Epoch 11, iter 1800/6416, lr 0.010000, loss 5.676516
+INFO 2020-11-26 00:21:38 train.py: 74] Epoch 11, iter 2000/6416, lr 0.010000, loss 5.680991
+INFO 2020-11-26 00:23:09 train.py: 74] Epoch 11, iter 2200/6416, lr 0.010000, loss 5.694247
+INFO 2020-11-26 00:24:40 train.py: 74] Epoch 11, iter 2400/6416, lr 0.010000, loss 5.728337
+INFO 2020-11-26 00:26:11 train.py: 74] Epoch 11, iter 2600/6416, lr 0.010000, loss 5.665781
+INFO 2020-11-26 00:27:41 train.py: 74] Epoch 11, iter 2800/6416, lr 0.010000, loss 5.723630
+INFO 2020-11-26 00:29:12 train.py: 87] Save checkpoint Epoch_11_batch_2999.pt to disk.
+INFO 2020-11-26 00:29:12 train.py: 74] Epoch 11, iter 3000/6416, lr 0.010000, loss 5.702121
+INFO 2020-11-26 00:30:44 train.py: 74] Epoch 11, iter 3200/6416, lr 0.010000, loss 5.718015
+INFO 2020-11-26 00:32:16 train.py: 74] Epoch 11, iter 3400/6416, lr 0.010000, loss 5.713120
+INFO 2020-11-26 00:33:48 train.py: 74] Epoch 11, iter 3600/6416, lr 0.010000, loss 5.737073
+INFO 2020-11-26 00:35:20 train.py: 74] Epoch 11, iter 3800/6416, lr 0.010000, loss 5.725788
+INFO 2020-11-26 00:36:52 train.py: 74] Epoch 11, iter 4000/6416, lr 0.010000, loss 5.769676
+INFO 2020-11-26 00:38:23 train.py: 74] Epoch 11, iter 4200/6416, lr 0.010000, loss 5.774718
+INFO 2020-11-26 00:39:55 train.py: 74] Epoch 11, iter 4400/6416, lr 0.010000, loss 5.744242
+INFO 2020-11-26 00:41:27 train.py: 74] Epoch 11, iter 4600/6416, lr 0.010000, loss 5.738924
+INFO 2020-11-26 00:42:59 train.py: 74] Epoch 11, iter 4800/6416, lr 0.010000, loss 5.739225
+INFO 2020-11-26 00:44:31 train.py: 74] Epoch 11, iter 5000/6416, lr 0.010000, loss 5.750110
+INFO 2020-11-26 00:46:02 train.py: 74] Epoch 11, iter 5200/6416, lr 0.010000, loss 5.784163
+INFO 2020-11-26 00:47:35 train.py: 74] Epoch 11, iter 5400/6416, lr 0.010000, loss 5.804254
+INFO 2020-11-26 00:49:06 train.py: 74] Epoch 11, iter 5600/6416, lr 0.010000, loss 5.775759
+INFO 2020-11-26 00:50:38 train.py: 74] Epoch 11, iter 5800/6416, lr 0.010000, loss 5.750277
+INFO 2020-11-26 00:52:10 train.py: 87] Save checkpoint Epoch_11_batch_5999.pt to disk.
+INFO 2020-11-26 00:52:10 train.py: 74] Epoch 11, iter 6000/6416, lr 0.010000, loss 5.818751
+INFO 2020-11-26 00:53:42 train.py: 74] Epoch 11, iter 6200/6416, lr 0.010000, loss 5.786312
+INFO 2020-11-26 00:55:14 train.py: 74] Epoch 11, iter 6400/6416, lr 0.010000, loss 5.753539
+INFO 2020-11-26 00:55:21 train.py: 92] Save checkpoint Epoch_11.pt to disk...
+INFO 2020-11-26 00:55:23 train.py: 74] Epoch 12, iter 0/6416, lr 0.010000, loss 5.729906
+INFO 2020-11-26 00:56:54 train.py: 74] Epoch 12, iter 200/6416, lr 0.010000, loss 5.263418
+INFO 2020-11-26 00:58:25 train.py: 74] Epoch 12, iter 400/6416, lr 0.010000, loss 5.272457
+INFO 2020-11-26 00:59:57 train.py: 74] Epoch 12, iter 600/6416, lr 0.010000, loss 5.262656
+INFO 2020-11-26 01:01:28 train.py: 74] Epoch 12, iter 800/6416, lr 0.010000, loss 5.286423
+INFO 2020-11-26 01:02:59 train.py: 74] Epoch 12, iter 1000/6416, lr 0.010000, loss 5.301856
+INFO 2020-11-26 01:04:30 train.py: 74] Epoch 12, iter 1200/6416, lr 0.010000, loss 5.297711
+INFO 2020-11-26 01:06:01 train.py: 74] Epoch 12, iter 1400/6416, lr 0.010000, loss 5.357860
+INFO 2020-11-26 01:07:32 train.py: 74] Epoch 12, iter 1600/6416, lr 0.010000, loss 5.400959
+INFO 2020-11-26 01:09:03 train.py: 74] Epoch 12, iter 1800/6416, lr 0.010000, loss 5.393822
+INFO 2020-11-26 01:10:35 train.py: 74] Epoch 12, iter 2000/6416, lr 0.010000, loss 5.396180
+INFO 2020-11-26 01:12:06 train.py: 74] Epoch 12, iter 2200/6416, lr 0.010000, loss 5.426024
+INFO 2020-11-26 01:13:37 train.py: 74] Epoch 12, iter 2400/6416, lr 0.010000, loss 5.446758
+INFO 2020-11-26 01:15:09 train.py: 74] Epoch 12, iter 2600/6416, lr 0.010000, loss 5.489548
+INFO 2020-11-26 01:16:40 train.py: 74] Epoch 12, iter 2800/6416, lr 0.010000, loss 5.476563
+INFO 2020-11-26 01:18:11 train.py: 87] Save checkpoint Epoch_12_batch_2999.pt to disk.
+INFO 2020-11-26 01:18:12 train.py: 74] Epoch 12, iter 3000/6416, lr 0.010000, loss 5.500869
+INFO 2020-11-26 01:19:43 train.py: 74] Epoch 12, iter 3200/6416, lr 0.010000, loss 5.496167
+INFO 2020-11-26 01:21:13 train.py: 74] Epoch 12, iter 3400/6416, lr 0.010000, loss 5.512305
+INFO 2020-11-26 01:22:44 train.py: 74] Epoch 12, iter 3600/6416, lr 0.010000, loss 5.553058
+INFO 2020-11-26 01:24:15 train.py: 74] Epoch 12, iter 3800/6416, lr 0.010000, loss 5.511114
+INFO 2020-11-26 01:25:46 train.py: 74] Epoch 12, iter 4000/6416, lr 0.010000, loss 5.577352
+INFO 2020-11-26 01:27:18 train.py: 74] Epoch 12, iter 4200/6416, lr 0.010000, loss 5.553564
+INFO 2020-11-26 01:28:49 train.py: 74] Epoch 12, iter 4400/6416, lr 0.010000, loss 5.552995
+INFO 2020-11-26 01:30:20 train.py: 74] Epoch 12, iter 4600/6416, lr 0.010000, loss 5.593556
+INFO 2020-11-26 01:31:51 train.py: 74] Epoch 12, iter 4800/6416, lr 0.010000, loss 5.614996
+INFO 2020-11-26 01:33:22 train.py: 74] Epoch 12, iter 5000/6416, lr 0.010000, loss 5.628087
+INFO 2020-11-26 01:34:53 train.py: 74] Epoch 12, iter 5200/6416, lr 0.010000, loss 5.578542
+INFO 2020-11-26 01:36:24 train.py: 74] Epoch 12, iter 5400/6416, lr 0.010000, loss 5.590160
+INFO 2020-11-26 01:37:56 train.py: 74] Epoch 12, iter 5600/6416, lr 0.010000, loss 5.600044
+INFO 2020-11-26 01:39:27 train.py: 74] Epoch 12, iter 5800/6416, lr 0.010000, loss 5.624286
+INFO 2020-11-26 01:40:58 train.py: 87] Save checkpoint Epoch_12_batch_5999.pt to disk.
+INFO 2020-11-26 01:40:58 train.py: 74] Epoch 12, iter 6000/6416, lr 0.010000, loss 5.683837
+INFO 2020-11-26 01:42:30 train.py: 74] Epoch 12, iter 6200/6416, lr 0.010000, loss 5.618696
+INFO 2020-11-26 01:44:02 train.py: 74] Epoch 12, iter 6400/6416, lr 0.010000, loss 5.659023
+INFO 2020-11-26 01:44:09 train.py: 92] Save checkpoint Epoch_12.pt to disk...
+INFO 2020-11-26 01:44:10 train.py: 74] Epoch 13, iter 0/6416, lr 0.001000, loss 5.618520
+INFO 2020-11-26 01:45:42 train.py: 74] Epoch 13, iter 200/6416, lr 0.001000, loss 5.012071
+INFO 2020-11-26 01:47:13 train.py: 74] Epoch 13, iter 400/6416, lr 0.001000, loss 4.988137
+INFO 2020-11-26 01:48:44 train.py: 74] Epoch 13, iter 600/6416, lr 0.001000, loss 4.895365
+INFO 2020-11-26 01:50:15 train.py: 74] Epoch 13, iter 800/6416, lr 0.001000, loss 4.930635
+INFO 2020-11-26 01:51:46 train.py: 74] Epoch 13, iter 1000/6416, lr 0.001000, loss 4.891228
+INFO 2020-11-26 01:53:17 train.py: 74] Epoch 13, iter 1200/6416, lr 0.001000, loss 4.912731
+INFO 2020-11-26 01:54:49 train.py: 74] Epoch 13, iter 1400/6416, lr 0.001000, loss 4.867206
+INFO 2020-11-26 01:56:20 train.py: 74] Epoch 13, iter 1600/6416, lr 0.001000, loss 4.909080
+INFO 2020-11-26 01:57:51 train.py: 74] Epoch 13, iter 1800/6416, lr 0.001000, loss 4.910253
+INFO 2020-11-26 01:59:22 train.py: 74] Epoch 13, iter 2000/6416, lr 0.001000, loss 4.914638
+INFO 2020-11-26 02:00:53 train.py: 74] Epoch 13, iter 2200/6416, lr 0.001000, loss 4.908720
+INFO 2020-11-26 02:02:25 train.py: 74] Epoch 13, iter 2400/6416, lr 0.001000, loss 4.907465
+INFO 2020-11-26 02:03:57 train.py: 74] Epoch 13, iter 2600/6416, lr 0.001000, loss 4.931486
+INFO 2020-11-26 02:05:28 train.py: 74] Epoch 13, iter 2800/6416, lr 0.001000, loss 4.895203
+INFO 2020-11-26 02:06:59 train.py: 87] Save checkpoint Epoch_13_batch_2999.pt to disk.
+INFO 2020-11-26 02:07:00 train.py: 74] Epoch 13, iter 3000/6416, lr 0.001000, loss 4.927824
+INFO 2020-11-26 02:08:31 train.py: 74] Epoch 13, iter 3200/6416, lr 0.001000, loss 4.893639
+INFO 2020-11-26 02:10:03 train.py: 74] Epoch 13, iter 3400/6416, lr 0.001000, loss 4.898834
+INFO 2020-11-26 02:11:35 train.py: 74] Epoch 13, iter 3600/6416, lr 0.001000, loss 4.902920
+INFO 2020-11-26 02:13:06 train.py: 74] Epoch 13, iter 3800/6416, lr 0.001000, loss 4.898564
+INFO 2020-11-26 02:14:38 train.py: 74] Epoch 13, iter 4000/6416, lr 0.001000, loss 4.884068
+INFO 2020-11-26 02:16:10 train.py: 74] Epoch 13, iter 4200/6416, lr 0.001000, loss 4.923888
+INFO 2020-11-26 02:17:42 train.py: 74] Epoch 13, iter 4400/6416, lr 0.001000, loss 4.896753
+INFO 2020-11-26 02:19:14 train.py: 74] Epoch 13, iter 4600/6416, lr 0.001000, loss 4.901687
+INFO 2020-11-26 02:20:46 train.py: 74] Epoch 13, iter 4800/6416, lr 0.001000, loss 4.914706
+INFO 2020-11-26 02:22:18 train.py: 74] Epoch 13, iter 5000/6416, lr 0.001000, loss 4.938848
+INFO 2020-11-26 02:23:50 train.py: 74] Epoch 13, iter 5200/6416, lr 0.001000, loss 4.907446
+INFO 2020-11-26 02:25:22 train.py: 74] Epoch 13, iter 5400/6416, lr 0.001000, loss 4.953235
+INFO 2020-11-26 02:26:54 train.py: 74] Epoch 13, iter 5600/6416, lr 0.001000, loss 4.876094
+INFO 2020-11-26 02:28:26 train.py: 74] Epoch 13, iter 5800/6416, lr 0.001000, loss 4.937435
+INFO 2020-11-26 02:29:57 train.py: 87] Save checkpoint Epoch_13_batch_5999.pt to disk.
+INFO 2020-11-26 02:29:58 train.py: 74] Epoch 13, iter 6000/6416, lr 0.001000, loss 4.932445
+INFO 2020-11-26 02:31:29 train.py: 74] Epoch 13, iter 6200/6416, lr 0.001000, loss 4.922741
+INFO 2020-11-26 02:33:00 train.py: 74] Epoch 13, iter 6400/6416, lr 0.001000, loss 4.929702
+INFO 2020-11-26 02:33:07 train.py: 92] Save checkpoint Epoch_13.pt to disk...
+INFO 2020-11-26 02:33:09 train.py: 74] Epoch 14, iter 0/6416, lr 0.001000, loss 4.960059
+INFO 2020-11-26 02:34:40 train.py: 74] Epoch 14, iter 200/6416, lr 0.001000, loss 4.811613
+INFO 2020-11-26 02:36:11 train.py: 74] Epoch 14, iter 400/6416, lr 0.001000, loss 4.833166
+INFO 2020-11-26 02:37:42 train.py: 74] Epoch 14, iter 600/6416, lr 0.001000, loss 4.833502
+INFO 2020-11-26 02:39:14 train.py: 74] Epoch 14, iter 800/6416, lr 0.001000, loss 4.820003
+INFO 2020-11-26 02:40:45 train.py: 74] Epoch 14, iter 1000/6416, lr 0.001000, loss 4.843599
+INFO 2020-11-26 02:42:16 train.py: 74] Epoch 14, iter 1200/6416, lr 0.001000, loss 4.782953
+INFO 2020-11-26 02:43:47 train.py: 74] Epoch 14, iter 1400/6416, lr 0.001000, loss 4.838921
+INFO 2020-11-26 02:45:18 train.py: 74] Epoch 14, iter 1600/6416, lr 0.001000, loss 4.815551
+INFO 2020-11-26 02:46:49 train.py: 74] Epoch 14, iter 1800/6416, lr 0.001000, loss 4.845832
+INFO 2020-11-26 02:48:21 train.py: 74] Epoch 14, iter 2000/6416, lr 0.001000, loss 4.849728
+INFO 2020-11-26 02:49:52 train.py: 74] Epoch 14, iter 2200/6416, lr 0.001000, loss 4.817868
+INFO 2020-11-26 02:51:23 train.py: 74] Epoch 14, iter 2400/6416, lr 0.001000, loss 4.877688
+INFO 2020-11-26 02:52:55 train.py: 74] Epoch 14, iter 2600/6416, lr 0.001000, loss 4.875107
+INFO 2020-11-26 02:54:26 train.py: 74] Epoch 14, iter 2800/6416, lr 0.001000, loss 4.841876
+INFO 2020-11-26 02:55:58 train.py: 87] Save checkpoint Epoch_14_batch_2999.pt to disk.
+INFO 2020-11-26 02:55:58 train.py: 74] Epoch 14, iter 3000/6416, lr 0.001000, loss 4.905459
+INFO 2020-11-26 02:57:30 train.py: 74] Epoch 14, iter 3200/6416, lr 0.001000, loss 4.877974
+INFO 2020-11-26 02:59:02 train.py: 74] Epoch 14, iter 3400/6416, lr 0.001000, loss 4.868736
+INFO 2020-11-26 03:00:34 train.py: 74] Epoch 14, iter 3600/6416, lr 0.001000, loss 4.853268
+INFO 2020-11-26 03:02:05 train.py: 74] Epoch 14, iter 3800/6416, lr 0.001000, loss 4.889563
+INFO 2020-11-26 03:03:37 train.py: 74] Epoch 14, iter 4000/6416, lr 0.001000, loss 4.812276
+INFO 2020-11-26 03:05:09 train.py: 74] Epoch 14, iter 4200/6416, lr 0.001000, loss 4.878132
+INFO 2020-11-26 03:06:40 train.py: 74] Epoch 14, iter 4400/6416, lr 0.001000, loss 4.882167
+INFO 2020-11-26 03:08:12 train.py: 74] Epoch 14, iter 4600/6416, lr 0.001000, loss 4.912398
+INFO 2020-11-26 03:09:44 train.py: 74] Epoch 14, iter 4800/6416, lr 0.001000, loss 4.898185
+INFO 2020-11-26 03:11:15 train.py: 74] Epoch 14, iter 5000/6416, lr 0.001000, loss 4.898757
+INFO 2020-11-26 03:12:47 train.py: 74] Epoch 14, iter 5200/6416, lr 0.001000, loss 4.893168
+INFO 2020-11-26 03:14:19 train.py: 74] Epoch 14, iter 5400/6416, lr 0.001000, loss 4.873687
+INFO 2020-11-26 03:15:51 train.py: 74] Epoch 14, iter 5600/6416, lr 0.001000, loss 4.900874
+INFO 2020-11-26 03:17:23 train.py: 74] Epoch 14, iter 5800/6416, lr 0.001000, loss 4.869009
+INFO 2020-11-26 03:18:55 train.py: 87] Save checkpoint Epoch_14_batch_5999.pt to disk.
+INFO 2020-11-26 03:18:55 train.py: 74] Epoch 14, iter 6000/6416, lr 0.001000, loss 4.878964
+INFO 2020-11-26 03:20:26 train.py: 74] Epoch 14, iter 6200/6416, lr 0.001000, loss 4.910442
+INFO 2020-11-26 03:21:58 train.py: 74] Epoch 14, iter 6400/6416, lr 0.001000, loss 4.951179
+INFO 2020-11-26 03:22:05 train.py: 92] Save checkpoint Epoch_14.pt to disk...
+INFO 2020-11-26 03:22:07 train.py: 74] Epoch 15, iter 0/6416, lr 0.001000, loss 4.787207
+INFO 2020-11-26 03:23:38 train.py: 74] Epoch 15, iter 200/6416, lr 0.001000, loss 4.822606
+INFO 2020-11-26 03:25:08 train.py: 74] Epoch 15, iter 400/6416, lr 0.001000, loss 4.815923
+INFO 2020-11-26 03:26:39 train.py: 74] Epoch 15, iter 600/6416, lr 0.001000, loss 4.775978
+INFO 2020-11-26 03:28:10 train.py: 74] Epoch 15, iter 800/6416, lr 0.001000, loss 4.816824
+INFO 2020-11-26 03:29:40 train.py: 74] Epoch 15, iter 1000/6416, lr 0.001000, loss 4.787887
+INFO 2020-11-26 03:31:11 train.py: 74] Epoch 15, iter 1200/6416, lr 0.001000, loss 4.853176
+INFO 2020-11-26 03:32:41 train.py: 74] Epoch 15, iter 1400/6416, lr 0.001000, loss 4.808012
+INFO 2020-11-26 03:34:12 train.py: 74] Epoch 15, iter 1600/6416, lr 0.001000, loss 4.829766
+INFO 2020-11-26 03:35:42 train.py: 74] Epoch 15, iter 1800/6416, lr 0.001000, loss 4.828946
+INFO 2020-11-26 03:37:13 train.py: 74] Epoch 15, iter 2000/6416, lr 0.001000, loss 4.802634
+INFO 2020-11-26 03:38:44 train.py: 74] Epoch 15, iter 2200/6416, lr 0.001000, loss 4.822742
+INFO 2020-11-26 03:40:14 train.py: 74] Epoch 15, iter 2400/6416, lr 0.001000, loss 4.829261
+INFO 2020-11-26 03:41:45 train.py: 74] Epoch 15, iter 2600/6416, lr 0.001000, loss 4.871683
+INFO 2020-11-26 03:43:16 train.py: 74] Epoch 15, iter 2800/6416, lr 0.001000, loss 4.848487
+INFO 2020-11-26 03:44:46 train.py: 87] Save checkpoint Epoch_15_batch_2999.pt to disk.
+INFO 2020-11-26 03:44:47 train.py: 74] Epoch 15, iter 3000/6416, lr 0.001000, loss 4.800260
+INFO 2020-11-26 03:46:18 train.py: 74] Epoch 15, iter 3200/6416, lr 0.001000, loss 4.793696
+INFO 2020-11-26 03:47:50 train.py: 74] Epoch 15, iter 3400/6416, lr 0.001000, loss 4.843006
+INFO 2020-11-26 03:49:22 train.py: 74] Epoch 15, iter 3600/6416, lr 0.001000, loss 4.820165
+INFO 2020-11-26 03:50:54 train.py: 74] Epoch 15, iter 3800/6416, lr 0.001000, loss 4.817658
+INFO 2020-11-26 03:52:25 train.py: 74] Epoch 15, iter 4000/6416, lr 0.001000, loss 4.854495
+INFO 2020-11-26 03:53:57 train.py: 74] Epoch 15, iter 4200/6416, lr 0.001000, loss 4.886664
+INFO 2020-11-26 03:55:29 train.py: 74] Epoch 15, iter 4400/6416, lr 0.001000, loss 4.833145
+INFO 2020-11-26 03:57:01 train.py: 74] Epoch 15, iter 4600/6416, lr 0.001000, loss 4.812392
+INFO 2020-11-26 03:58:33 train.py: 74] Epoch 15, iter 4800/6416, lr 0.001000, loss 4.821527
+INFO 2020-11-26 04:00:05 train.py: 74] Epoch 15, iter 5000/6416, lr 0.001000, loss 4.860501
+INFO 2020-11-26 04:01:37 train.py: 74] Epoch 15, iter 5200/6416, lr 0.001000, loss 4.891657
+INFO 2020-11-26 04:03:09 train.py: 74] Epoch 15, iter 5400/6416, lr 0.001000, loss 4.856526
+INFO 2020-11-26 04:04:40 train.py: 74] Epoch 15, iter 5600/6416, lr 0.001000, loss 4.841456
+INFO 2020-11-26 04:06:12 train.py: 74] Epoch 15, iter 5800/6416, lr 0.001000, loss 4.833061
+INFO 2020-11-26 04:07:44 train.py: 87] Save checkpoint Epoch_15_batch_5999.pt to disk.
+INFO 2020-11-26 04:07:44 train.py: 74] Epoch 15, iter 6000/6416, lr 0.001000, loss 4.862416
+INFO 2020-11-26 04:09:16 train.py: 74] Epoch 15, iter 6200/6416, lr 0.001000, loss 4.901599
+INFO 2020-11-26 04:10:48 train.py: 74] Epoch 15, iter 6400/6416, lr 0.001000, loss 4.867302
+INFO 2020-11-26 04:10:55 train.py: 92] Save checkpoint Epoch_15.pt to disk...
+INFO 2020-11-26 04:10:56 train.py: 74] Epoch 16, iter 0/6416, lr 0.000100, loss 4.768134
+INFO 2020-11-26 04:12:28 train.py: 74] Epoch 16, iter 200/6416, lr 0.000100, loss 4.826395
+INFO 2020-11-26 04:13:59 train.py: 74] Epoch 16, iter 400/6416, lr 0.000100, loss 4.771744
+INFO 2020-11-26 04:15:30 train.py: 74] Epoch 16, iter 600/6416, lr 0.000100, loss 4.783202
+INFO 2020-11-26 04:17:01 train.py: 74] Epoch 16, iter 800/6416, lr 0.000100, loss 4.756491
+INFO 2020-11-26 04:18:33 train.py: 74] Epoch 16, iter 1000/6416, lr 0.000100, loss 4.764306
+INFO 2020-11-26 04:20:04 train.py: 74] Epoch 16, iter 1200/6416, lr 0.000100, loss 4.745123
+INFO 2020-11-26 04:21:35 train.py: 74] Epoch 16, iter 1400/6416, lr 0.000100, loss 4.782097
+INFO 2020-11-26 04:23:06 train.py: 74] Epoch 16, iter 1600/6416, lr 0.000100, loss 4.756188
+INFO 2020-11-26 04:24:38 train.py: 74] Epoch 16, iter 1800/6416, lr 0.000100, loss 4.779078
+INFO 2020-11-26 04:26:09 train.py: 74] Epoch 16, iter 2000/6416, lr 0.000100, loss 4.722829
+INFO 2020-11-26 04:27:40 train.py: 74] Epoch 16, iter 2200/6416, lr 0.000100, loss 4.744510
+INFO 2020-11-26 04:29:11 train.py: 74] Epoch 16, iter 2400/6416, lr 0.000100, loss 4.742054
+INFO 2020-11-26 04:30:43 train.py: 74] Epoch 16, iter 2600/6416, lr 0.000100, loss 4.757106
+INFO 2020-11-26 04:32:14 train.py: 74] Epoch 16, iter 2800/6416, lr 0.000100, loss 4.760223
+INFO 2020-11-26 04:33:46 train.py: 87] Save checkpoint Epoch_16_batch_2999.pt to disk.
+INFO 2020-11-26 04:33:46 train.py: 74] Epoch 16, iter 3000/6416, lr 0.000100, loss 4.736575
+INFO 2020-11-26 04:35:18 train.py: 74] Epoch 16, iter 3200/6416, lr 0.000100, loss 4.758691
+INFO 2020-11-26 04:36:49 train.py: 74] Epoch 16, iter 3400/6416, lr 0.000100, loss 4.744581
+INFO 2020-11-26 04:38:21 train.py: 74] Epoch 16, iter 3600/6416, lr 0.000100, loss 4.749971
+INFO 2020-11-26 04:39:53 train.py: 74] Epoch 16, iter 3800/6416, lr 0.000100, loss 4.747253
+INFO 2020-11-26 04:41:24 train.py: 74] Epoch 16, iter 4000/6416, lr 0.000100, loss 4.776212
+INFO 2020-11-26 04:42:56 train.py: 74] Epoch 16, iter 4200/6416, lr 0.000100, loss 4.804538
+INFO 2020-11-26 04:44:28 train.py: 74] Epoch 16, iter 4400/6416, lr 0.000100, loss 4.757755
+INFO 2020-11-26 04:45:59 train.py: 74] Epoch 16, iter 4600/6416, lr 0.000100, loss 4.765916
+INFO 2020-11-26 04:47:31 train.py: 74] Epoch 16, iter 4800/6416, lr 0.000100, loss 4.778348
+INFO 2020-11-26 04:49:03 train.py: 74] Epoch 16, iter 5000/6416, lr 0.000100, loss 4.752482
+INFO 2020-11-26 04:50:34 train.py: 74] Epoch 16, iter 5200/6416, lr 0.000100, loss 4.748292
+INFO 2020-11-26 04:52:06 train.py: 74] Epoch 16, iter 5400/6416, lr 0.000100, loss 4.722651
+INFO 2020-11-26 04:53:38 train.py: 74] Epoch 16, iter 5600/6416, lr 0.000100, loss 4.734725
+INFO 2020-11-26 04:55:10 train.py: 74] Epoch 16, iter 5800/6416, lr 0.000100, loss 4.737980
+INFO 2020-11-26 04:56:41 train.py: 87] Save checkpoint Epoch_16_batch_5999.pt to disk.
+INFO 2020-11-26 04:56:42 train.py: 74] Epoch 16, iter 6000/6416, lr 0.000100, loss 4.798186
+INFO 2020-11-26 04:58:14 train.py: 74] Epoch 16, iter 6200/6416, lr 0.000100, loss 4.723335
+INFO 2020-11-26 04:59:46 train.py: 74] Epoch 16, iter 6400/6416, lr 0.000100, loss 4.773817
+INFO 2020-11-26 04:59:53 train.py: 92] Save checkpoint Epoch_16.pt to disk...
+INFO 2020-11-26 04:59:54 train.py: 74] Epoch 17, iter 0/6416, lr 0.000100, loss 4.723282
+INFO 2020-11-26 05:01:26 train.py: 74] Epoch 17, iter 200/6416, lr 0.000100, loss 4.732407
+INFO 2020-11-26 05:02:57 train.py: 74] Epoch 17, iter 400/6416, lr 0.000100, loss 4.744728
+INFO 2020-11-26 05:04:28 train.py: 74] Epoch 17, iter 600/6416, lr 0.000100, loss 4.758975
+INFO 2020-11-26 05:06:00 train.py: 74] Epoch 17, iter 800/6416, lr 0.000100, loss 4.766218
+INFO 2020-11-26 05:07:31 train.py: 74] Epoch 17, iter 1000/6416, lr 0.000100, loss 4.787281
+INFO 2020-11-26 05:09:02 train.py: 74] Epoch 17, iter 1200/6416, lr 0.000100, loss 4.777162
+INFO 2020-11-26 05:10:33 train.py: 74] Epoch 17, iter 1400/6416, lr 0.000100, loss 4.758057
+INFO 2020-11-26 05:12:05 train.py: 74] Epoch 17, iter 1600/6416, lr 0.000100, loss 4.710682
+INFO 2020-11-26 05:13:36 train.py: 74] Epoch 17, iter 1800/6416, lr 0.000100, loss 4.720905
+INFO 2020-11-26 05:15:07 train.py: 74] Epoch 17, iter 2000/6416, lr 0.000100, loss 4.766242
+INFO 2020-11-26 05:16:39 train.py: 74] Epoch 17, iter 2200/6416, lr 0.000100, loss 4.755566
+INFO 2020-11-26 05:18:10 train.py: 74] Epoch 17, iter 2400/6416, lr 0.000100, loss 4.757743
+INFO 2020-11-26 05:19:42 train.py: 74] Epoch 17, iter 2600/6416, lr 0.000100, loss 4.750778
+INFO 2020-11-26 05:21:13 train.py: 74] Epoch 17, iter 2800/6416, lr 0.000100, loss 4.754477
+INFO 2020-11-26 05:22:44 train.py: 87] Save checkpoint Epoch_17_batch_2999.pt to disk.
+INFO 2020-11-26 05:22:45 train.py: 74] Epoch 17, iter 3000/6416, lr 0.000100, loss 4.729516
+INFO 2020-11-26 05:24:16 train.py: 74] Epoch 17, iter 3200/6416, lr 0.000100, loss 4.749175
+INFO 2020-11-26 05:25:48 train.py: 74] Epoch 17, iter 3400/6416, lr 0.000100, loss 4.752513
+INFO 2020-11-26 05:27:19 train.py: 74] Epoch 17, iter 3600/6416, lr 0.000100, loss 4.744433
+INFO 2020-11-26 05:28:51 train.py: 74] Epoch 17, iter 3800/6416, lr 0.000100, loss 4.760126
+INFO 2020-11-26 05:30:22 train.py: 74] Epoch 17, iter 4000/6416, lr 0.000100, loss 4.775823
+INFO 2020-11-26 05:31:54 train.py: 74] Epoch 17, iter 4200/6416, lr 0.000100, loss 4.786304
+INFO 2020-11-26 05:33:26 train.py: 74] Epoch 17, iter 4400/6416, lr 0.000100, loss 4.764655
+INFO 2020-11-26 05:34:57 train.py: 74] Epoch 17, iter 4600/6416, lr 0.000100, loss 4.771241
+INFO 2020-11-26 05:36:29 train.py: 74] Epoch 17, iter 4800/6416, lr 0.000100, loss 4.738297
+INFO 2020-11-26 05:38:01 train.py: 74] Epoch 17, iter 5000/6416, lr 0.000100, loss 4.785628
+INFO 2020-11-26 05:39:32 train.py: 74] Epoch 17, iter 5200/6416, lr 0.000100, loss 4.737215
+INFO 2020-11-26 05:41:04 train.py: 74] Epoch 17, iter 5400/6416, lr 0.000100, loss 4.759377
+INFO 2020-11-26 05:42:35 train.py: 74] Epoch 17, iter 5600/6416, lr 0.000100, loss 4.791402
+INFO 2020-11-26 05:44:07 train.py: 74] Epoch 17, iter 5800/6416, lr 0.000100, loss 4.769801
+INFO 2020-11-26 05:45:39 train.py: 87] Save checkpoint Epoch_17_batch_5999.pt to disk.
+INFO 2020-11-26 05:45:39 train.py: 74] Epoch 17, iter 6000/6416, lr 0.000100, loss 4.763365
+INFO 2020-11-26 05:47:10 train.py: 74] Epoch 17, iter 6200/6416, lr 0.000100, loss 4.753882
+INFO 2020-11-26 05:48:42 train.py: 74] Epoch 17, iter 6400/6416, lr 0.000100, loss 4.780777
+INFO 2020-11-26 05:48:49 train.py: 92] Save checkpoint Epoch_17.pt to disk...
+INFO 2020-11-26 05:48:50 train.py: 175] Optimization done!
diff --git a/bob/bio/facexzoo/transformers/pytorch.py b/bob/bio/facexzoo/transformers/pytorch.py
new file mode 100644
index 0000000000000000000000000000000000000000..d2b8ee1c1d34639424ac290d2e0618606f96c823
--- /dev/null
+++ b/bob/bio/facexzoo/transformers/pytorch.py
@@ -0,0 +1,1085 @@
+#!/usr/bin/env python
+# vim: set fileencoding=utf-8 :
+# Hatef Otroshi <hatef.otroshi@idiap.ch>
+
+import imp
+import os
+import torch
+import numpy as np
+from bob.bio.base.algorithm import Distance
+# from bob.bio.base.pipelines.vanilla_biometrics import VanillaBiometricsPipeline
+from bob.bio.base.pipelines import PipelineSimple
+from bob.bio.face.utils import dnn_default_cropping
+from bob.bio.face.utils import embedding_transformer
+# from bob.bio.face.utils import cropped_positions_arcface
+from bob.bio.facexzoo.utils import cropped_positions_arcface
+from bob.extension.download import get_file
+from sklearn.base import BaseEstimator
+from sklearn.base import TransformerMixin
+from sklearn.utils import check_array
+# from bob.bio.face.annotator import BobIpMTCNN
+from bob.bio.face.annotator import MTCNN
+
+# from bob.learn.pytorch.architectures.facexzoo import FaceXZooModelFactory
+from bob.bio.facexzoo.backbones import FaceXZooModelFactory
+
+# from bob.bio.face.embeddings.pytorch import PyTorchModel
+################# PyTorchModel from bob.bio.face.embeddings.pytorch
+from sklearn.base import BaseEstimator
+from sklearn.base import TransformerMixin
+class PyTorchModel(TransformerMixin, BaseEstimator):
+ """
+ Base Transformer using pytorch models
+
+
+ Parameters
+ ----------
+
+ checkpoint_path: str
+ Path containing the checkpoint
+
+ config:
+ Path containing some configuration file (e.g. .json, .prototxt)
+
+ preprocessor:
+ A function that will transform the data right before forward. The default transformation is `X/255`
+
+ """
+
+ def __init__(
+ self,
+ checkpoint_path=None,
+ config=None,
+ preprocessor=lambda x: x / 255,
+ memory_demanding=False,
+ device=None,
+ **kwargs,
+ ):
+
+ super().__init__(**kwargs)
+ self.checkpoint_path = checkpoint_path
+ self.config = config
+ self.model = None
+ self.preprocessor = preprocessor
+ self.memory_demanding = memory_demanding
+ self.device = device
+
+ def transform(self, X):
+ """__call__(image) -> feature
+
+ Extracts the features from the given image.
+
+ **Parameters:**
+
+ image : 2D :py:class:`numpy.ndarray` (floats)
+ The image to extract the features from.
+
+ **Returns:**
+
+ feature : 2D or 3D :py:class:`numpy.ndarray` (floats)
+ The list of features extracted from the image.
+ """
+ import torch
+
+ if self.model is None:
+ self._load_model()
+ X = check_array(X, allow_nd=True)
+ X = torch.Tensor(X)
+ with torch.no_grad():
+ X = self.preprocessor(X)
+
+ def _transform(X):
+ with torch.no_grad():
+ return self.model(X.to(self.device)).cpu().detach().numpy()
+
+ if self.memory_demanding:
+ features = np.array([_transform(x[None, ...]) for x in X])
+
+ # If we ndim is > than 3. We should stack them all
+ # The enroll_features can come from a source where there are `N` samples containing
+ # nxd samples
+ if features.ndim >= 3:
+ features = np.vstack(features)
+
+ return features
+
+ else:
+ return _transform(X)
+
+ def __getstate__(self):
+ # Handling unpicklable objects
+
+ d = self.__dict__.copy()
+ d["model"] = None
+ return d
+
+ def _more_tags(self):
+ return {"stateless": True, "requires_fit": False}
+
+ def place_model_on_device(self, device=None):
+ import torch
+
+ if device is None:
+ device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+ self.device = device
+
+ if self.model is not None:
+ self.model.to(device)
+################# PyTorchModel
+
+
+# from bob.bio.face.embeddings.pytorch import FaceXZooModel
+class FaceXZooModel(PyTorchModel):
+ """
+ FaceXZoo models
+ """
+
+ def __init__(
+ self,
+ preprocessor=lambda x: (x - 127.5) / 128.0,
+ memory_demanding=False,
+ device=None,
+ arch="MobileFaceNet",
+ head='MV-Softmax',
+ **kwargs,
+ ):
+
+ self.arch = arch
+ self.head = head
+ _model = FaceXZooModelFactory(self.arch, self.head)
+ filename = _model.get_facexzoo_file()
+ checkpoint_name = _model.get_checkpoint_name()
+ config = None
+ path = os.path.dirname(filename)
+ checkpoint_path = filename#os.path.join(path, self.arch + ".pt")
+
+ super(FaceXZooModel, self).__init__(
+ checkpoint_path,
+ config,
+ memory_demanding=memory_demanding,
+ preprocessor=preprocessor,
+ device=device,
+ **kwargs,
+ )
+
+ def _load_model(self):
+
+ _model = FaceXZooModelFactory(self.arch, self.head)
+ self.model = _model.get_model()
+
+ model_dict = self.model.state_dict()
+
+ pretrained_dict = torch.load(
+ self.checkpoint_path, map_location=torch.device("cpu")
+ )["state_dict"]
+
+ pretrained_dict_keys = pretrained_dict.keys()
+ model_dict_keys = model_dict.keys()
+
+ new_pretrained_dict = {}
+ for k in model_dict:
+ new_pretrained_dict[k] = pretrained_dict["backbone." + k]
+ model_dict.update(new_pretrained_dict)
+ self.model.load_state_dict(model_dict)
+
+ self.model.eval()
+ self.place_model_on_device()
+
+ def _more_tags(self):
+ return {"stateless": True, "requires_fit": False}
+
+ def fit(self, X, y=None):
+ return self
+
+
+
+# from bob.bio.face.embeddings.pytorch import iresnet_template
+# iresnet_template is replaced with pipeline_template
+def pipeline_template(embedding, annotation_type, fixed_positions=None):
+ # DEFINE CROPPING
+ cropped_image_size = (112, 112)
+ if annotation_type == "eyes-center" or annotation_type == "bounding-box":
+ # Hard coding eye positions for backward consistency
+ # cropped_positions = {
+ cropped_positions = cropped_positions_arcface()
+ if annotation_type == "bounding-box":
+ # This will allow us to use `BoundingBoxAnnotatorCrop`
+ cropped_positions.update(
+ {"topleft": (0, 0), "bottomright": cropped_image_size}
+ )
+
+ else:
+ cropped_positions = dnn_default_cropping(cropped_image_size, annotation_type)
+
+ annotator = MTCNN(min_size=40, factor=0.709, thresholds=(0.1, 0.2, 0.2))
+ transformer = embedding_transformer(
+ cropped_image_size=cropped_image_size,
+ embedding=embedding,
+ cropped_positions=cropped_positions,
+ fixed_positions=fixed_positions,
+ color_channel="rgb",
+ annotator=annotator,
+ )
+
+ algorithm = Distance()
+
+ return PipelineSimple(transformer, algorithm)
+
+
+
+def MobileFaceNet(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the MobileFaceNet pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`MobileFaceNet` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(
+ arch="MobileFaceNet", memory_demanding=memory_demanding
+ ),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def ResNet50_ir(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the ResNet50_ir pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`ResNet50_ir` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="ResNet50_ir", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+def ResNet152_irse(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the ResNet152_irse pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`ResNet152_irse` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="ResNet152_irse", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def HRNet(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the HRNet pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`HRNet` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="HRNet", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+def EfficientNet_B0(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the EfficientNet_B0 pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`EfficientNet_B0` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="EfficientNet_B0", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def TF_NAS_A(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the TF_NAS_A pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`TF_NAS_A` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="TF_NAS_A", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+def LightCNN29(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the LightCNN29 pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`LightCNN29` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="LightCNN29", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+def GhostNet(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the GhostNet pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`GhostNet` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="GhostNet", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+def AttentionNet56(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the AttentionNet pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`AttentionNet` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="AttentionNet56", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def AttentionNet92(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the AttentionNet pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`AttentionNet` to extract the features
+
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="AttentionNet92", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+def ResNeSt50(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the ResNeSt50 pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`ResNeSt50` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="ResNeSt50", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+def ReXNet_1(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the ReXNet_1 pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`ReXNet_1` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="ReXNet_1", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def RepVGG_A0(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the RepVGG_A0 pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`RepVGG_A0` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="RepVGG_A0", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def RepVGG_B0(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the RepVGG_B0 pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`RepVGG_B0` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="RepVGG_B0", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def RepVGG_B1(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the RepVGG_B1 pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`RepVGG_B1` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(arch="RepVGG_B1", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+def SwinTransformer_preprocessor(X):
+ preprocessor_=lambda x: (x - 127.5) / 128.0
+
+ X = preprocessor_(X)
+ if X.size(2) != 224:
+ X = torch.nn.functional.interpolate(X, mode='bilinear', size=(224, 224), align_corners=False)
+
+ return X
+
+def SwinTransformer_S(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the SwinTransformer_S pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`RepVGG_B1` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(preprocessor=SwinTransformer_preprocessor,
+ arch="SwinTransformer_S", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+def SwinTransformer_T(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the SwinTransformer_T pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`RepVGG_B1` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(preprocessor=SwinTransformer_preprocessor,
+ arch="SwinTransformer_T", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+############# Heads
+
+
+def AM_Softmax(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the AM_Softmax pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`AM_Softmax` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="AM-Softmax", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def AdaM_Softmax(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the AdaM_Softmax pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`AdaM_Softmax` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="AdaM-Softmax", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+def AdaCos(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the AdaCos pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`AdaCos` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="AdaCos", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+
+def ArcFace(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the ArcFace pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`ArcFace` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="ArcFace", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+
+def MV_Softmax(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the MV_Softmax pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`MV_Softmax` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="MV-Softmax", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+
+def CurricularFace(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the CurricularFace pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`CurricularFace` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="CurricularFace", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+
+def CircleLoss(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the CircleLoss pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`CircleLoss` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="CircleLoss", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+def NPCFace(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the NPCFace pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`NPCFace` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="NPCFace", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
+
+
+
+
+def MagFace(annotation_type, fixed_positions=None, memory_demanding=False):
+ """
+ Get the MagFace pipeline which will crop the face :math:`112 \\times 112` and
+ use the :py:class:`MagFace` to extract the features
+
+ .. warning::
+
+ If you are at Idiap, please use the option `-l sge-gpu` while running the `vanilla-biometrics` pipeline.
+
+
+
+ Parameters
+ ----------
+
+ annotation_type: str
+ Type of the annotations (e.g. `eyes-center')
+
+ fixed_positions: dict
+ Set it if in your face images are registered to a fixed position in the image
+
+ memory_demanding: bool
+
+ """
+
+ return pipeline_template(
+ embedding=FaceXZooModel(head="MagFace", memory_demanding=memory_demanding),
+ annotation_type=annotation_type,
+ fixed_positions=fixed_positions,
+ )
\ No newline at end of file
diff --git a/bob/bio/facexzoo/utils.py b/bob/bio/facexzoo/utils.py
new file mode 100644
index 0000000000000000000000000000000000000000..535ec98fb079a9a8520a5eef41f148d6d8be875e
--- /dev/null
+++ b/bob/bio/facexzoo/utils.py
@@ -0,0 +1,42 @@
+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+# SPDX-FileContributor: Hatef OTROSHI <hatef.otroshi@idiap.ch>
+# SPDX-License-Identifier: MIT
+def cropped_positions_arcface(annotation_type="eyes-center"):
+ """
+ Returns the 112 x 112 crop used in iResnet based models
+ The crop follows the following rule:
+
+ - In X --> (112/2)-1
+ - In Y, leye --> 16+(112/2) --> 72
+ - In Y, reye --> (112/2)-16 --> 40
+
+ This will leave 16 pixels between left eye and left border and right eye and right border
+
+ For reference, https://github.com/deepinsight/insightface/blob/master/recognition/arcface_mxnet/common/face_align.py
+ contains the cropping code for training the original ArcFace-InsightFace model. Due to this code not being very explicit,
+ we choose to pick our own default cropped positions. They have been tested to provide good evaluation performance
+ on the Mobio dataset.
+
+ For sensitive applications, you can use custom cropped position that you optimize for your specific dataset,
+ such as is done in https://gitlab.idiap.ch/bob/bob.bio.face/-/blob/master/notebooks/50-shades-of-face.ipynb
+
+ """
+
+ if isinstance(annotation_type, list):
+ return [cropped_positions_arcface(item) for item in annotation_type]
+
+
+ if annotation_type == "eyes-center":
+ cropped_positions = {
+ "leye": (51.5014, 73.5318), #"leye": (55, 72),
+ "reye": (51.6963, 38.2946) #"reye": (55, 40),
+ }
+ elif annotation_type == "left-profile":
+
+ cropped_positions = {"leye": (40, 30), "mouth": (85, 30)}
+ elif annotation_type == "right-profile":
+ return {"reye": (40, 82), "mouth": (85, 82)}
+ else:
+ raise ValueError(f"Annotations of the type `{annotation_type}` not supported")
+
+ return cropped_positions
\ No newline at end of file
diff --git a/bob/bio/invert/__init__.py b/bob/bio/invert/__init__.py
new file mode 100644
index 0000000000000000000000000000000000000000..e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
diff --git a/bob/bio/invert/invertibility_pipeline.py b/bob/bio/invert/invertibility_pipeline.py
new file mode 100644
index 0000000000000000000000000000000000000000..39cba4e38dec0770b4179045f05284e17f69b2af
--- /dev/null
+++ b/bob/bio/invert/invertibility_pipeline.py
@@ -0,0 +1,179 @@
+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+# SPDX-FileContributor: Hatef OTROSHI <hatef.otroshi@idiap.ch>
+# SPDX-License-Identifier: MIT
+
+"""
+Implementation of the Vanilla Biometrics pipeline using Dask :ref:`bob.bio.base.struct_bio_rec_sys`_
+
+This file contains simple processing blocks meant to be used
+for bob.bio experiments
+"""
+
+import logging
+import tempfile
+
+from sklearn.base import BaseEstimator
+from sklearn.pipeline import Pipeline
+
+from bob.bio.base.pipelines import BioAlgorithm
+from bob.bio.base.pipelines.pipelines import PipelineSimple
+from bob.bio.base.pipelines.pipelines import check_valid_pipeline
+from bob.bio.base.pipelines.score_writers import FourColumnsScoreWriter
+
+logger = logging.getLogger(__name__)
+
+
+class InvertBiometricsPipeline(PipelineSimple):
+ """
+ Invert Biometrics Pipeline
+
+ This is the backbone of most biometric recognition systems.
+ It implements three subpipelines and they are the following:
+
+ - :py:class:`PipelineSimple.train_background_model`: Initializes or trains your transformer.
+ It will run :py:meth:`sklearn.base.BaseEstimator.fit`
+
+ - :py:class:`PipelineSimple.enroll_templates`: Creates enrollment templates
+ It will run :py:meth:`sklearn.base.BaseEstimator.transform` followed by a sequence of
+ :py:meth:`bob.bio.base.pipelines.abstract_classes.BioAlgorithm.create_templates`
+
+ - :py:class:`PipelineSimple.probe_templates`: Creates probe templates
+ It will run :py:meth:`sklearn.base.BaseEstimator.transform` followed by a sequence of
+ :py:meth:`bob.bio.base.pipelines.abstract_classes.BioAlgorithm.create_templates`
+
+ - :py:class:`PipelineSimple.compute_scores`: Computes scores
+ It will run :py:meth:`bob.bio.base.pipelines.abstract_classes.BioAlgorithm.compare`
+
+
+ Example
+ -------
+ >>> from sklearn.preprocessing import FunctionTransformer
+ >>> from sklearn.pipeline import make_pipeline
+ >>> from bob.bio.base.algorithm import Distance
+ >>> from bob.bio.base.pipelines import PipelineSimple
+ >>> from bob.pipelines import wrap
+ >>> import numpy
+ >>> linearize = lambda samples: [numpy.reshape(x, (-1,)) for x in samples]
+ >>> transformer = wrap(["sample"], FunctionTransformer(linearize))
+ >>> transformer_pipeline = make_pipeline(transformer)
+ >>> biometric_algorithm = Distance()
+ >>> pipeline = PipelineSimple(transformer_pipeline, biometric_algorithm)
+ >>> pipeline(samples_for_training_back_ground_model, samplesets_for_enroll, samplesets_for_scoring) # doctest: +SKIP
+
+
+ To run this pipeline using Dask, used the function
+ :py:func:`dask_bio_pipeline`.
+
+ Example
+ -------
+ >>> from base.pipelines import dask_bio_pipeline
+ >>> pipeline = InvertBiometricsPipeline(transformer, biometric_algoritm)
+ >>> pipeline = dask_bio_pipeline(pipeline)
+ >>> pipeline(samples_for_training_back_ground_model, samplesets_for_enroll, samplesets_for_scoring).compute() # doctest: +SKIP
+
+
+ Parameters
+ ----------
+
+ transformer: :py:class`sklearn.pipeline.Pipeline` or a `sklearn.base.BaseEstimator`
+ Transformer that will preprocess your data
+
+ biometric_algorithm: :py:class:`bob.bio.base.pipelines.vanilla_biometrics.abstract_classes.BioAlgorithm`
+ Biometrics algorithm object that implements the methods `enroll` and `score` methods
+
+ score_writer: :any:`bob.bio.base.pipelines.vanilla_biometrics.ScoreWriter`
+ Format to write scores. Default to :any:`bob.bio.base.pipelines.vanilla_biometrics.FourColumnsScoreWriter`
+
+ """
+
+ def __init__(
+ self,
+ transformer,
+ inversionAttack_transformer,
+ biometric_algorithm,
+ score_writer=None,
+ ):
+ self.transformer = transformer
+ self.inversionAttack_transformer = inversionAttack_transformer
+ self.biometric_algorithm = biometric_algorithm
+ self.score_writer = score_writer
+ if self.score_writer is None:
+ tempdir = tempfile.TemporaryDirectory()
+ self.score_writer = FourColumnsScoreWriter(tempdir.name)
+
+ check_valid_pipeline(self)
+
+ def __call__(
+ self,
+ background_model_samples,
+ biometric_reference_samples,
+ probe_samples,
+ invert_references_samples,
+ allow_scoring_with_all_biometric_references=True,
+ return_templates=False,
+ ):
+ logger.info(
+ f" >> Vanilla Biometrics: Training background model with pipeline {self.transformer}"
+ )
+
+ # Training background model (fit will return even if samples is ``None``,
+ # in which case we suppose the algorithm is not trainable in any way)
+ self.transformer = self.train_background_model(background_model_samples)
+
+ logger.info(
+ f" >> Creating biometric references with the biometric algorithm {self.biometric_algorithm}"
+ )
+
+ # Create biometric samples
+ biometric_references = self.enroll_templates(
+ biometric_reference_samples
+ )
+
+ logger.info(" >> PipelineSimple: Creating probe templates")
+ probe_templates = self.probe_templates(probe_samples)
+
+ logger.info(
+ f" >> Computing scores with the biometric algorithm {self.biometric_algorithm}"
+ )
+
+ # Scores all probes
+ scores_probes = self.compute_scores(
+ probe_templates,
+ biometric_references,
+ allow_scoring_with_all_biometric_references,
+ )
+
+ # Scores all inverted references
+ scores_inversionAttack = self.compute_scores_invertedReferences(
+ invert_references_samples,
+ biometric_references,
+ allow_scoring_with_all_biometric_references,
+ )
+
+ return scores_probes, scores_inversionAttack
+
+ def compute_scores_invertedReferences(
+ self,
+ invert_references_samples,
+ biometric_references,
+ allow_scoring_with_all_biometric_references=True,
+ ):
+ # probes is a list of SampleSets
+ invert_reference_features = self.inversionAttack_transformer.transform(
+ invert_references_samples
+ )
+
+ invert_reference_templates = (
+ self.biometric_algorithm.create_templates_from_samplesets(
+ invert_reference_features, enroll=False
+ )
+ )
+
+ scores = self.biometric_algorithm.score_sample_templates(
+ invert_reference_templates,
+ biometric_references,
+ allow_scoring_with_all_biometric_references,
+ )
+
+ # scores is a list of Samples
+ return scores
diff --git a/bob/bio/invert/pipeline.py b/bob/bio/invert/pipeline.py
new file mode 100644
index 0000000000000000000000000000000000000000..08be78dc1e8b7ae490240f860fc1d9c22966131c
--- /dev/null
+++ b/bob/bio/invert/pipeline.py
@@ -0,0 +1,259 @@
+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+# SPDX-FileContributor: Hatef OTROSHI <hatef.otroshi@idiap.ch>
+# SPDX-License-Identifier: MIT
+import copy
+import logging
+import os
+
+import dask.bag
+
+from dask.delayed import Delayed
+
+import bob.pipelines
+
+from bob.bio.base.pipelines import BioAlgDaskWrapper as BioAlgorithmDaskWrapper
+from bob.bio.base.pipelines import CSVScoreWriter
+from bob.bio.base.pipelines import FourColumnsScoreWriter
+from bob.bio.base.pipelines import PipelineScoreNorm
+from bob.bio.base.pipelines import TNormScores
+from bob.bio.base.pipelines import ZNormScores
+from bob.bio.base.pipelines import checkpoint_pipeline_simple
+from bob.bio.base.pipelines import dask_bio_pipeline
+from bob.bio.base.pipelines import is_biopipeline_checkpointed
+from bob.pipelines import estimator_requires_fit
+from bob.pipelines import is_instance_nested
+from bob.pipelines.distributed import dask_get_partition_size
+from bob.pipelines.utils import flatten_samplesets
+
+logger = logging.getLogger(__name__)
+
+
+def compute_scores(result, dask_client):
+ if isinstance(result, Delayed) or isinstance(result, dask.bag.Bag):
+ if dask_client is not None:
+ result = result.compute(scheduler=dask_client)
+ else:
+ logger.warning(
+ "`dask_client` not set. Your pipeline will run locally"
+ )
+ result = result.compute(scheduler="single-threaded")
+ return result
+
+
+def post_process_scores(pipeline, scores, path):
+ written_scores = pipeline.write_scores(scores)
+ return pipeline.post_process(written_scores, path)
+
+
+def get_inverted_references(biometric_references):
+ inverted_references = copy.deepcopy(biometric_references)
+ references = [r.reference_id for r in inverted_references]
+
+ # breakdown sampleset
+ inverted_references = flatten_samplesets(inverted_references)
+
+ for sampleset in inverted_references:
+ # sampleset.references = [sampleset.reference_id]
+ sampleset.references = copy.deepcopy(references)
+ for sample in sampleset:
+ sample.key = sample.key + "-inverted"
+
+ sampleset.key = sampleset.key + "-inverted"
+
+ return inverted_references
+
+
+def execute_inverted_simple_biometrics(
+ pipeline,
+ database,
+ dask_client,
+ groups,
+ output,
+ write_metadata_scores,
+ checkpoint,
+ dask_partition_size,
+ dask_n_workers,
+ **kwargs,
+):
+ """
+ Function that executes the PipelineSimple.
+
+ This is called when using the ``bob bio pipeline simple``
+ command.
+
+ This is also callable from a script without fear of interrupting the running
+ Dask instance, allowing chaining multiple experiments while keeping the
+ workers alive.
+
+ When using Dask, something to keep in mind is that we want to split our data and
+ processing time on multiple workers. There is no recipe to make everything work on
+ any system. So if you encounter some balancing error (a few of all the available
+ workers actually working while the rest waits, or the scheduler being overloaded
+ trying to organise millions of tiny tasks), you can specify ``dask_n_partitions``
+ or ``dask_partition_size``.
+ The first will try to split any set of data into a number of chunks (ideally, we
+ would want one per worker), and the second creates similar-sized partitions in each
+ set.
+ If the memory on the workers is not sufficient, try reducing the size of the
+ partitions (or increasing the number of partitions).
+
+ Parameters
+ ----------
+
+ pipeline: Instance of :py:class:`bob.bio.base.pipelines.PipelineSimple`
+ A constructed PipelineSimple object.
+
+ database: Instance of :py:class:`bob.bio.base.pipelines.abstract_class.Database`
+ A database interface instance
+
+ dask_client: instance of :py:class:`dask.distributed.Client` or ``None``
+ A Dask client instance used to run the experiment in parallel on multiple
+ machines, or locally.
+ Basic configs can be found in ``bob.pipelines.config.distributed``.
+
+ dask_n_partitions: int or None
+ Specifies a number of partitions to split the data into.
+
+ dask_partition_size: int or None
+ Specifies a data partition size when using dask. Ignored when dask_n_partitions
+ is set.
+
+ dask_n_workers: int or None
+ Sets the starting number of Dask workers. Does not prevent Dask from requesting
+ more or releasing workers depending on load.
+
+ groups: list of str
+ Groups of the dataset that will be requested from the database interface.
+
+ output: str
+ Path where the scores will be saved.
+
+ write_metadata_scores: bool
+ Use the CSVScoreWriter instead of the FourColumnScoreWriter when True.
+
+ checkpoint: bool
+ Whether checkpoint files will be created for every step of the pipelines.
+
+ checkpoint_dir: str
+ If `checkpoint` is set, this path will be used to save the checkpoints.
+ If `None`, the content of `output` will be used.
+
+ force: bool
+ If set, it will force generate all the checkpoints of an experiment. This option doesn't work if `--memory` is set
+ """
+
+ if not os.path.exists(output):
+ os.makedirs(output, exist_ok=True)
+
+ if write_metadata_scores:
+ pipeline.score_writer = CSVScoreWriter(os.path.join(output, "./tmp"))
+ else:
+ pipeline.score_writer = FourColumnsScoreWriter(
+ os.path.join(output, "./tmp")
+ )
+
+ # Check if it's already checkpointed
+ if checkpoint and not is_biopipeline_checkpointed(pipeline):
+ hash_fn = database.hash_fn if hasattr(database, "hash_fn") else None
+ pipeline = checkpoint_pipeline_simple(pipeline, output, hash_fn=hash_fn)
+
+ # Here we have to checkpoint the `inversionAttack_transformer`
+ # inversionAttack_transformer
+
+ pipeline.inversionAttack_transformer = bob.pipelines.wrap(
+ ["checkpoint"],
+ pipeline.inversionAttack_transformer,
+ features_dir=output,
+ hash_fn=hash_fn,
+ )
+
+ pass
+
+ # Load the background model samples only if the transformer requires fitting
+ if estimator_requires_fit(pipeline.transformer):
+ background_model_samples = database.background_model_samples()
+ else:
+ background_model_samples = []
+
+ for group in groups:
+
+ score_probes_file_name = os.path.join(
+ output,
+ f"scores-{group}" + (".csv" if write_metadata_scores else ""),
+ )
+ score_inversionAttack_file_name = os.path.join(
+ output,
+ f"scores_inversion-{group}"
+ + (".csv" if write_metadata_scores else ""),
+ )
+
+ biometric_references = database.references(group=group)
+ probes = database.probes(group=group)
+ inverted_references = get_inverted_references(biometric_references)
+
+ # If there's no data to be processed, continue
+ if len(biometric_references) == 0 or len(probes) == 0:
+ logger.warning(
+ f"Current dataset ({database}) does not have `{group}` set. The experiment will not be executed."
+ )
+ continue
+
+ if dask_client is not None and not is_instance_nested(
+ pipeline.biometric_algorithm,
+ "biometric_algorithm",
+ BioAlgorithmDaskWrapper,
+ ):
+ # Scaling up
+ if dask_n_workers is not None and not isinstance(dask_client, str):
+ dask_client.cluster.scale(dask_n_workers)
+
+ n_objects = max(
+ len(background_model_samples),
+ len(biometric_references),
+ len(probes),
+ )
+ partition_size = None
+ if not isinstance(dask_client, str):
+ partition_size = dask_get_partition_size(
+ dask_client.cluster, n_objects
+ )
+ if dask_partition_size is not None:
+ partition_size = dask_partition_size
+
+ pipeline = dask_bio_pipeline(
+ pipeline,
+ partition_size=partition_size,
+ )
+
+ # Here we have to dask the `inversionAttack_transformer`
+ # inversionAttack_transformer
+
+ pipeline.inversionAttack_transformer = bob.pipelines.wrap(
+ ["dask"],
+ pipeline.inversionAttack_transformer,
+ partition_size=partition_size,
+ )
+
+ logger.info(f"Running vanilla biometrics for group {group}")
+ allow_scoring_with_all_biometric_references = (
+ database.allow_scoring_with_all_biometric_references
+ if hasattr(database, "allow_scoring_with_all_biometric_references")
+ else False
+ )
+ scores_probes, scores_inversionAttack = pipeline(
+ background_model_samples,
+ biometric_references,
+ probes,
+ inverted_references,
+ allow_scoring_with_all_biometric_references=allow_scoring_with_all_biometric_references,
+ )
+
+ post_processed_scores = post_process_scores(
+ pipeline, scores_probes, score_probes_file_name
+ )
+ _ = compute_scores(post_processed_scores, dask_client)
+
+ post_processed_scores = post_process_scores(
+ pipeline, scores_inversionAttack, score_inversionAttack_file_name
+ )
+ _ = compute_scores(post_processed_scores, dask_client)
diff --git a/bob/bio/invert/transformers.py b/bob/bio/invert/transformers.py
new file mode 100644
index 0000000000000000000000000000000000000000..b472c22f420bf72374b60410dba210bf5e552aa9
--- /dev/null
+++ b/bob/bio/invert/transformers.py
@@ -0,0 +1,80 @@
+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+# SPDX-FileContributor: Hatef OTROSHI <hatef.otroshi@idiap.ch>
+# SPDX-License-Identifier: MIT
+import numpy as np
+import torch
+
+from sklearn.base import BaseEstimator
+from sklearn.base import TransformerMixin
+
+from bob.pipelines import Sample
+from bob.pipelines import SampleBatch
+from bob.pipelines import SampleSet
+
+from .networks import Generator
+
+
+class InversionTransformer(TransformerMixin, BaseEstimator):
+ """
+ Transforms any :math:`\mathbb{R}^n` into an image :math:`\mathbb{R}^{h \\times w \\times c}`.
+
+ Parameters
+ ----------
+
+ checkpoint: str
+ Checkpoint of the image generator
+
+ generator:
+ instance of the generator network
+
+ """
+
+ def __init__(self, checkpoint, generator=None):
+
+ self.device = torch.device(
+ "cuda" if torch.cuda.is_available() else "cpu"
+ )
+
+ self.generator = Generator() if generator is None else generator
+
+ # TODO: use the checkpoint variable here
+ self.generator.load_state_dict(
+ torch.load(
+ checkpoint,
+ map_location=self.device,
+ )
+ )
+ self.generator.eval()
+ self.generator.to(self.device)
+ self.checkpoint = checkpoint
+
+ def _more_tags(self):
+ return {"stateless": True, "requires_fit": False}
+
+ def fit(self, X, y=None):
+ return self
+
+ def transform(self, samples):
+ def _transform(data):
+ data = data.flatten()
+ data = np.reshape(data, (1, data.shape[0], 1, 1))
+ embedding = torch.Tensor(data).to(self.device)
+ reconstructed_img = self.generator(embedding)[0]
+ return reconstructed_img.cpu().detach().numpy() * 255.0
+
+ if isinstance(samples[0], SampleSet):
+ return [
+ SampleSet(
+ self.transform(sset.samples),
+ parent=sset,
+ )
+ for sset in samples
+ ]
+ else:
+ return [
+ Sample(
+ _transform(sample.data),
+ parent=sample,
+ )
+ for sample in samples
+ ]
diff --git a/bob/bio/invert/wrappers.py b/bob/bio/invert/wrappers.py
new file mode 100644
index 0000000000000000000000000000000000000000..0f712bb0ba47858d5805744494b69525b6270af7
--- /dev/null
+++ b/bob/bio/invert/wrappers.py
@@ -0,0 +1,17 @@
+# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
+# SPDX-FileContributor: Hatef OTROSHI <hatef.otroshi@idiap.ch>
+# SPDX-License-Identifier: MIT
+from sklearn.pipeline import Pipeline
+
+
+def get_invert_pipeline(FR_transformer, inv_transformer, feature_extractor):
+
+ pipeline = Pipeline(
+ FR_transformer.steps
+ + [
+ ("inverted-samples", inv_transformer),
+ ("inverted-features", feature_extractor),
+ ]
+ )
+
+ return pipeline
diff --git a/buildout.cfg b/buildout.cfg
new file mode 100644
index 0000000000000000000000000000000000000000..db012640b714db1aafc5c2c5cdb19d0a440cf7a1
--- /dev/null
+++ b/buildout.cfg
@@ -0,0 +1,47 @@
+; -*- coding: utf-8 -*-
+; Wed Feb 12 13:37:08 2020
+
+[buildout]
+parts = scripts
+
+develop = src/bob.bio.base
+ src/bob.bio.face
+ src/bob.bio.invert
+ src/bob.learn.tensorflow
+ src/bob.pipelines
+ src/timm
+ .
+
+
+
+
+eggs = bob.bio.facexzoo
+ bob.bio.invert
+ bob.bio.base
+ bob.bio.face
+ bob.bio.invert
+ bob.learn.tensorflow
+ bob.pipelines
+ timm
+
+
+extensions = bob.buildout
+ mr.developer
+
+newest = false
+verbose = true
+auto-checkout = *
+
+
+[sources]
+
+bob.bio.base = git git@gitlab.idiap.ch:bob/bob.bio.base rev=8d70e55c15e3d2cdcafcfd92b11f138e2f30f5bb
+bob.bio.face = git git@gitlab.idiap.ch:bob/bob.bio.face rev=3567e990d0e523ceb5d3f9598054d8a27d7f7000
+bob.learn.tensorflow = git git@gitlab.idiap.ch:bob/bob.learn.tensorflow rev=f420d1b322762c81b79f59fa103c4ad07713fd79
+bob.pipelines = git git@gitlab.idiap.ch:bob/bob.pipelines rev=d8162ffc4fa072a14a8a4d7ac3b558de464a56ef
+timm = git git@github.com:rwightman/pytorch-image-models rev=1d01c2b68c90619b173d4675d9a283be2bd44a33 ;v0.3.3
+
+
+[scripts]
+recipe = bob.buildout:scripts
+dependent-scripts = true
\ No newline at end of file
diff --git a/pyproject.toml b/pyproject.toml
new file mode 100644
index 0000000000000000000000000000000000000000..77711a8aa25547d56404e668e69d3fb09278f6e7
--- /dev/null
+++ b/pyproject.toml
@@ -0,0 +1,17 @@
+[build-system]
+ requires = ["setuptools", "wheel", "bob.extension"]
+ build-backend = "setuptools.build_meta"
+
+[tool.isort]
+ profile = "black"
+ line_length = 80
+ order_by_type = true
+ lines_between_types = 1
+
+[tool.black]
+ line-length = 80
+
+[tool.pytest.ini_options]
+markers = [
+ "slow: marks tests as slow (deselect with '-m \"not slow\"')",
+]
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000000000000000000000000000000000000..089f1abe678a25e3bdd9f06da16152d3b03da86f
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,3 @@
+setuptools
+numpy
+bob.extension
\ No newline at end of file
diff --git a/setup.py b/setup.py
new file mode 100644
index 0000000000000000000000000000000000000000..294a5a334fa539e08c34d22bf70716d94776df94
--- /dev/null
+++ b/setup.py
@@ -0,0 +1,57 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+from setuptools import setup, dist
+dist.Distribution(dict(setup_requires=['bob.extension']))
+
+from bob.extension.utils import find_packages
+from bob.extension.utils import load_requirements
+
+install_requires = load_requirements()
+
+
+setup(
+
+ name='bob.paper.neurips2023_face_ti',
+ version=open("version.txt").read().rstrip(),
+ description='New package',
+
+ url='https://gitlab.idiap.ch/bob/bob.paper.neurips2023_face_ti',
+ license='BSD-3',
+
+ # there may be multiple authors (separate entries by comma)
+ author='Hatef OTROSHI',
+ author_email='hatef.otroshi@idiap.ch',
+
+ # you may add more keywords separating those by commas (a, b, c, ...)
+ keywords = "bob",
+
+ long_description=open('README.md').read(),
+
+ # leave this here, it is pretty standard
+ packages=find_packages(),
+ include_package_data=True,
+ zip_safe = False,
+
+ install_requires=install_requires,
+
+ entry_points={
+ # add entry points (scripts, bob resources here, if any)
+ },
+
+ # check classifiers, add and remove as you see fit
+ # full list here: https://pypi.org/classifiers/
+ # don't remove the Bob framework unless it's not a bob package
+ classifiers = [
+ 'Framework :: Bob',
+ 'Development Status :: 4 - Beta',
+ 'Intended Audience :: Science/Research',
+ 'License :: OSI Approved :: BSD-3',
+ 'Natural Language :: English',
+ 'Programming Language :: Python',
+ 'Programming Language :: Python :: 3',
+ 'Topic :: Scientific/Engineering :: Artificial Intelligence',
+ 'Topic :: Software Development :: Libraries :: Python Modules',
+ ],
+
+)
\ No newline at end of file
diff --git a/version.txt b/version.txt
new file mode 100644
index 0000000000000000000000000000000000000000..79d48fc4015cc8015db8c42188f074088e94ee37
--- /dev/null
+++ b/version.txt
@@ -0,0 +1 @@
+0.0.1b0
\ No newline at end of file