Wed, 04 Jun 2008 19:04:40 -0400 |
James Bergstra |
deprecating __getitem__[fieldname] syntax
|
Wed, 04 Jun 2008 18:48:50 -0400 |
James Bergstra |
reverting dataset
|
Wed, 04 Jun 2008 17:49:28 -0400 |
James Bergstra |
mergec
|
Wed, 04 Jun 2008 17:49:09 -0400 |
James Bergstra |
initial commit of amat
|
Wed, 04 Jun 2008 17:34:01 -0400 |
Thierry Bertin-Mahieux |
commented junk in the default test (main function) of mlp_factory_approach so the test still works
|
Wed, 04 Jun 2008 17:00:44 -0400 |
Thierry Bertin-Mahieux |
we can now load and save in a file, see test class in the file for an example, but basically it's model1.save(filename) or learn_algo(filename) to load
|
Tue, 03 Jun 2008 21:34:40 -0400 |
Yoshua Bengio |
Automated merge with http://lgcm.iro.umontreal.ca/hg/pylearn
|
Tue, 03 Jun 2008 21:34:24 -0400 |
Yoshua Bengio |
Split LearningAlgorithm into OfflineLearningAlgorithm and OnlineLearningAlgorithm
|
Tue, 03 Jun 2008 16:52:06 -0400 |
Thierry Bertin-Mahieux |
small debugging with dummytests
|
Tue, 03 Jun 2008 16:45:23 -0400 |
Thierry Bertin-Mahieux |
small debugging with dummytests
|
Tue, 03 Jun 2008 16:41:55 -0400 |
Thierry Bertin-Mahieux |
created 'dummytests', tests that checks consistency of new weird datasets, where we can't compare with actual values in a matrix, for instance. Useful as a first debugging when creating a dataset
|
Tue, 03 Jun 2008 16:06:21 -0400 |
James Bergstra |
merged
|
Tue, 03 Jun 2008 16:05:28 -0400 |
James Bergstra |
beginning to hack on #20, fixing for Thierry
|
Tue, 03 Jun 2008 13:25:40 -0400 |
Frederic Bastien |
Automated merge with ssh://projects@lgcm.iro.umontreal.ca/hg/pylearn
|
Tue, 03 Jun 2008 13:25:33 -0400 |
Frederic Bastien |
need to make multiple test do to cached dataset
|
Tue, 03 Jun 2008 13:24:41 -0400 |
Frederic Bastien |
optimication in CachedDataSet.minibatches_nowrap
|
Tue, 03 Jun 2008 13:23:28 -0400 |
Frederic Bastien |
code clean up
|
Tue, 03 Jun 2008 13:22:45 -0400 |
Frederic Bastien |
implemented CachedDataSet.__iter__ as an optimization
|
Tue, 03 Jun 2008 12:25:53 -0400 |
Frederic Bastien |
optimization of CachedDataSet__getitem__
|
Tue, 03 Jun 2008 12:18:22 -0400 |
Frederic Bastien |
make test_speed generic
|
Tue, 03 Jun 2008 13:18:33 -0400 |
James Bergstra |
merged
|
Tue, 03 Jun 2008 13:14:45 -0400 |
James Bergstra |
polished filetensor a little
|
Tue, 03 Jun 2008 12:14:03 -0400 |
Frederic Bastien |
Automated merge with ssh://projects@lgcm.iro.umontreal.ca/hg/pylearn
|
Tue, 03 Jun 2008 12:13:17 -0400 |
Frederic Bastien |
better function name in test_speed
|
Mon, 02 Jun 2008 17:09:58 -0400 |
James Bergstra |
merged
|
Mon, 02 Jun 2008 17:08:17 -0400 |
James Bergstra |
mlp_factory_approach.py, updated and un-deprecated by popular demand
|
Mon, 02 Jun 2008 11:59:41 -0400 |
Frederic Bastien |
Optimisation in ArrayDataSet::__getitem__
|
Fri, 30 May 2008 10:19:16 -0400 |
delallea |
Hmm... that was committed by Fred I think, I got lost by Mercurial I think
|
Fri, 30 May 2008 10:14:46 -0400 |
delallea |
If I understand properly, the length of an unbounded stream is sys.maxint
|
Thu, 29 May 2008 10:42:29 -0400 |
Frederic Bastien |
more test. Are we wanting to limit the size of minibatch to the size of the dataset?
|
Thu, 29 May 2008 10:41:35 -0400 |
Frederic Bastien |
more general test
|
Thu, 29 May 2008 10:30:59 -0400 |
Frederic Bastien |
optimization in ArrayDataSet::__iter__
|
Wed, 28 May 2008 14:16:32 -0400 |
Frederic Bastien |
Automated merge with ssh://projects@lgcm.iro.umontreal.ca/hg/pylearn
|
Wed, 28 May 2008 14:16:28 -0400 |
Frederic Bastien |
code cleanup
|
Wed, 28 May 2008 14:09:19 -0400 |
James Bergstra |
re-enabling old ArrayDataSet indexing
|
Tue, 27 May 2008 15:59:24 -0400 |
Frederic Bastien |
Automated merge with ssh://projects@lgcm.iro.umontreal.ca/hg/pylearn
|
Tue, 27 May 2008 15:59:02 -0400 |
Frederic Bastien |
moved the function from misc.py that have dependency on theano in misc_theano.py
|
Tue, 27 May 2008 15:49:09 -0400 |
delallea |
Fixed by James
|
Tue, 27 May 2008 13:46:03 -0400 |
Frederic Bastien |
Automated merge with ssh://projects@lgcm.iro.umontreal.ca/hg/pylearn
|
Fri, 23 May 2008 10:21:52 -0400 |
Frederic Bastien |
more test speed
|
Fri, 16 May 2008 16:40:26 -0400 |
Frederic Bastien |
Added speed test for ArraDataSet
|
Fri, 16 May 2008 16:38:07 -0400 |
Frederic Bastien |
optimisation of ArrayDataSet
|
Tue, 27 May 2008 13:23:05 -0400 |
James Bergstra |
reinstating changeset 216 clobbered accidentally by 218
|
Fri, 23 May 2008 17:12:12 -0400 |
James Bergstra |
merged
|
Fri, 23 May 2008 17:11:39 -0400 |
James Bergstra |
deprecation note for mlp
|
Fri, 23 May 2008 17:10:03 -0400 |
James Bergstra |
added LookupList to pylearn namespace
|
Fri, 23 May 2008 16:01:01 -0400 |
Thierry Bertin-Mahieux |
should have solved the problem with minibatches not handling subsets of fieldnames, although maybe not super efficient
|
Fri, 23 May 2008 14:16:54 -0400 |
Thierry Bertin-Mahieux |
redo what James had done, so invert node and name
|
Fri, 23 May 2008 14:07:53 -0400 |
Thierry Bertin-Mahieux |
tester on iterator consistency now triggers a bug in dataset, linked to the combination of minibatch and slicing
|
Fri, 23 May 2008 13:44:25 -0400 |
Thierry Bertin-Mahieux |
test on simple slicing works
|
Fri, 23 May 2008 13:16:42 -0400 |
Thierry Bertin-Mahieux |
added a class to test any new dataset
|
Fri, 23 May 2008 12:22:54 -0400 |
Thierry Bertin-Mahieux |
small debugging
|
Thu, 22 May 2008 19:08:46 -0400 |
James Bergstra |
harmless typo in ScalarSoftplus.c_code
|
Thu, 22 May 2008 19:07:51 -0400 |
James Bergstra |
Modified ArrayDataSet's handling of column fields.
|
Thu, 22 May 2008 17:41:14 -0400 |
James Bergstra |
tlearn->pylearn in mlp
|
Thu, 22 May 2008 16:31:07 -0400 |
Frederic Bastien |
small modif for test
|
Wed, 21 May 2008 17:43:56 -0400 |
Thierry Bertin-Mahieux |
previous commit was supposed to concern only one file, dataset.py, try to undo my other changes with this commit (nothing was broken though, just useless debugging prints)
|
Wed, 21 May 2008 17:42:20 -0400 |
Thierry Bertin-Mahieux |
previous commit was supposed to concern only one file, dataset.py, try to undo my other changes with this commit (nothing was broken though, just useless debugging prints)
|
Wed, 21 May 2008 17:39:30 -0400 |
Thierry Bertin-Mahieux |
in __get__, problem if the i.stop was None, i being the slice, added one line replacing None by the len(self)
|
Sat, 17 May 2008 00:01:47 -0400 |
Yoshua Bengio |
work in progress denoising auto-encoder
|