ML_GMMTrainer.initialize() does not accept a random generator argument
View options
Closed Issue created
I am trying to fix a bug in bob.bio.gmm!9 (merged) where the rng argument was given at a wrong position: bob.bio.gmm!9 (diffs, comment 17651)
But fixing that revealed that ML_GMMTrainer.initialize() does not accept a random generator argument and the code crashes in the builds https://gitlab.idiap.ch/bob/bob.bio.gmm/-/jobs/80459 :
======================================================================
ERROR: bob.bio.gmm.test.test_scripts.test_ivector_lda_wccn_plda_sequential
----------------------------------------------------------------------
Traceback (most recent call last):
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
self.test(*self.arg)
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/gmm/test/test_scripts.py", line 198, in test_ivector_lda_wccn_plda_sequential
_verify(parameters, test_dir, 'test_ivector_lda_wccn_plda_sequential', ref_modifier='-ivector-lda-wccn-plda')
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/gmm/test/test_scripts.py", line 27, in _verify
executable(parameters)
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/base/script/verify.py", line 457, in main
verify(args, command_line_parameters)
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/base/script/verify.py", line 437, in verify
if not execute(args):
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/base/script/verify.py", line 290, in execute
force = args.force)
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/base/tools/algorithm.py", line 56, in train_projector
algorithm.train_projector(train_features, fs.projector_file)
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/gmm/algorithm/IVector.py", line 139, in train_projector
self.train_ubm(data)
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/bio/gmm/algorithm/GMM.py", line 128, in train_ubm
bob.learn.em.train(self.ubm_trainer, self.ubm, array, self.gmm_training_iterations, self.training_threshold, rng=self.rng)
File "/Users/buildbot/work/gitlab-ci/f6f95b82/0/bob/bob.bio.gmm/build-prefix/lib/python2.7/site-packages/bob/learn/em/train.py", line 49, in train
trainer.initialize(machine, data, rng)
TypeError: function takes at most 2 arguments (3 given)Any idea on how this should be handled? should it ignore rng or should it be changed to accept it?