Commits · 67ca29f48ff5bae70b4d306b4d38ded85bfdec33 · medai / software / mednet

Aug 17, 2023
- [scripts.experiment] Reflect changes from evaluation; closes #44 after... · 67ca29f4
  André Anjos authored 1 year ago
  
  [scripts.experiment] Reflect changes from evaluation; closes #44 after noticing train-analysis is performed
  67ca29f4
Aug 15, 2023
- [doc] Documentation fixes · 50148b8c
  André Anjos authored 1 year ago
  
  50148b8c
Jul 27, 2023
- [scripts.train] Fix overlapping short options for datamodule and device · 180f39d5
  André Anjos authored 1 year ago
  
  180f39d5
Jul 26, 2023
- Added model_transforms in models · 63f51a74
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  63f51a74
- Save and restore normalizer from checkpoint · edadec6d
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  edadec6d
- [ptbench.engine.trainer] Implement type hints · 2fbbd899
  André Anjos authored 1 year ago
  
  2fbbd899
- [ptbench.engine.device] Move AcceleratorProcessor to engine submodule, revamp... · ddb4c028
  André Anjos authored 1 year ago
  
  [ptbench.engine.device] Move AcceleratorProcessor to engine submodule, revamp it, and reflect changes in trainer
  ddb4c028
- [ptbench.scripts] Improved docs, adapt changes from weight-balancing strategy · 0284fbba
  André Anjos authored 1 year ago
  
  0284fbba
- Change 'no set_normalizer' logger info to warning · da18adfa
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  da18adfa
- Check if set_normalizer method is defined in model · 9eb1e023
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  9eb1e023
- Make augmentation transforms part of the model · 5bb7387f
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  5bb7387f
- Updates to the data module and models · cbfbf85d
  André Anjos authored 1 year ago
  
  cbfbf85d
- Improved normalizer, updated pasa model · 32b32762
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  32b32762
- Created common DataModule for Shenzhen dataset · 9c8674d8
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  9c8674d8
- Refactored raw data loading process · 1b2d43a1
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  1b2d43a1
- DataModule instanciation directly inside config file · 7f86dd1f
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  7f86dd1f
- Removed reliance on make_dataset and added method to cache samples for default shenzhen · bd79cd2f
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  bd79cd2f
- Grouped configs and data for default Shenzhen dataset · dbf15e3a
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  Created base class for every LightningDataModule, that can be inherited from. For each protocol, the user should create a new class inheriting from BaseDataModule and implement the setup() function, in which the dataset is defined.
  dbf15e3a
- Moved DataModule to its own script · a321b3b3
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  DataModule is generic and not tied to a specific dataset.
  a321b3b3
- Added DataModule in shenzhen, started support for extra-valid datasets · 4b512c5f
  Daniel CARRON authored 1 year ago and André Anjos committed 1 year ago
  
  4b512c5f
May 10, 2023
- Updated docstrings · f3cc3c55
  Daniel CARRON authored 1 year ago
  
  f3cc3c55
May 09, 2023
- Replaced imports of pytorch_lightning py lightning.pytorch · 7e55b4be
  Daniel CARRON authored 1 year ago
  
  7e55b4be
May 08, 2023
- Fix saving and loading model hyperparameters · 99356e82
  Daniel CARRON authored 1 year ago
  
  99356e82
Apr 12, 2023
- Get epochs from checkpointer and moved checkpointer selection code · bf25f806
  Daniel CARRON authored 2 years ago
  
  bf25f806
- Changed default accelerator · a4e335c1
  Daniel CARRON authored 2 years ago
  
  a4e335c1
- Re-added device/accelerator selection · 8852fa7f
  Daniel CARRON authored 2 years ago
  
  8852fa7f
- Removed unused arguments · c2c4286d
  Daniel CARRON authored 2 years ago
  
  c2c4286d
- Resume training from checkpoints · b42b73d7
  Daniel CARRON authored 2 years ago
  
  b42b73d7
Apr 11, 2023
- Moved train.py to lightning · 9d36fa1f
  Daniel CARRON authored 2 years ago
  
  9d36fa1f
Mar 16, 2023
- removed weight loading functionality · 09e50ca4
  ogueler@idiap.ch authored 2 years ago
  
  09e50ca4
- gave precedence to checkpoints over weights · 010a0147
  ogueler@idiap.ch authored 2 years ago
  
  010a0147
- re-added weight functionality + updated tests · 2381830d
  ogueler@idiap.ch authored 2 years ago
  
  2381830d
Mar 08, 2023
- updated documentation · 458c855b
  Oezgueracar GUELER authored 2 years ago
  
  458c855b
Mar 07, 2023
- removed already used short form -d from --device · 155715d5
  Oezgueracar GUELER authored 2 years ago
  
  155715d5
- fixed checkpointing issue with epochs · c4753bd3
  Oezgueracar GUELER authored 2 years ago
  
  c4753bd3
Mar 01, 2023
- Use clapper (instead of clapp) · 5d4efc36
  André Anjos authored 2 years ago
  
  5d4efc36
Feb 13, 2023
- [qa] Update pre-commit and style · b33248be
  André Anjos authored 2 years ago
  
  b33248be
- Use software/clapp instead of bob/exposed · b6a1b077
  André Anjos authored 2 years ago
  
  b6a1b077
Jan 30, 2023
- [tests] Fix docformatting · eb3ae3d2
  André Anjos authored 2 years ago
  
  eb3ae3d2
Jan 23, 2023
- Initial commit · 0dbdb2f0
  André Anjos authored 2 years ago
  
  0dbdb2f0