提交 · 666b86b3d685cf15304f58e6ebd91cf19f35734c · testgroup / pytensor

16 2月, 2016 5 次提交
- Merge pull request #4041 from abergeron/fix_buildbot2 · 666b86b3
  由 Frédéric Bastien 提交于 2月 16, 2016
```
Fix the output shape in python to match the C code for DownsampleFactorMaxGradGrad.
```
  666b86b3
- Fix imports in tests. · b05017a8
  由 Arnaud Bergeron 提交于 2月 15, 2016
  
  b05017a8
- Explain more how dummy is there for backward compat and cleanup a bit. · c386dd5d
  由 Arnaud Bergeron 提交于 2月 15, 2016
  
  c386dd5d
- Merge pull request #4061 from nouiz/param · e088e2a8
  由 Pascal Lamblin 提交于 2月 15, 2016
```
Correct In parameter name following Param removal
```
  e088e2a8
- Correct In parameter name following Param removal · fdd3150b
  由 Frederic Bastien 提交于 2月 15, 2016
  
  fdd3150b
15 2月, 2016 1 次提交
- Merge pull request #4040 from caglar/fix_fload_multi_bug · 50e06772
  由 Frédéric Bastien 提交于 2月 15, 2016
```
Fixes the multinomial grad dtype bug
```
  50e06772
13 2月, 2016 6 次提交
- Merge pull request #4045 from nouiz/corrmm · 9f46b692
  由 Frédéric Bastien 提交于 2月 12, 2016
```
Don't use CpuCorrMM when there is no CPU BLAS
```
  9f46b692
- Merge pull request #3942 from shabanian/tensor_signal_pep8 · 5d8526de
  由 Frédéric Bastien 提交于 2月 12, 2016
```
change shape dimensions int32 to int64
```
  5d8526de
- flake8 · 57aae87f
  由 Caglar 提交于 2月 12, 2016
  
  57aae87f
- Add a comment explaining the dummy param. · 7145695f
  由 Arnaud Bergeron 提交于 2月 12, 2016
  
  7145695f
- Move the backward compat where it should be. · 028aaef8
  由 Arnaud Bergeron 提交于 2月 12, 2016
  
  028aaef8
- Remove name that was deprecated in 2010. This is super old. · 51d97ce2
  由 Arnaud Bergeron 提交于 2月 12, 2016
  
  51d97ce2
12 2月, 2016 18 次提交
- Make sure the docstring is attached to the right object and reuse objects. · a262b3b0
  由 Frederic Bastien 提交于 2月 12, 2016
  
  a262b3b0
- Merge pull request #4009 from AnishShah/elu · 1d017d25
  由 carriepl 提交于 2月 12, 2016
```
Exponential Linear Units (ELUs)
```
  1d017d25
- Fix check · 26d0f24a
  由 Frederic Bastien 提交于 2月 12, 2016
  
  26d0f24a
- Merge pull request #4043 from diogo149/master · d9e0d1c5
  由 Frédéric Bastien 提交于 2月 12, 2016
```
Correct warning re: cuDNN and pool_2d's ignore_border argument
```
  d9e0d1c5
- Don't use CpuCorrMM when there is no CPU BLAS · 435caa76
  由 Frederic Bastien 提交于 2月 12, 2016
  
  435caa76
- Correct warning re: cuDNN and pool_2d's ignore_border argument · c0371515
  由 Diogo Moitinho de Almeida 提交于 2月 11, 2016
  
  c0371515
- Exponential Linear Units (ELUs) · 2055feec
  由 Anish Shah 提交于 2月 12, 2016
  
  2055feec
- fixed the undefined variable problem. · 6010cadd
  由 Caglar 提交于 2月 11, 2016
  
  6010cadd
- Remove pool.py from the whitelist (it has no errors). · 2c192649
  由 Arnaud Bergeron 提交于 2月 11, 2016
  
  2c192649
- Remove the bad compat for pickled DownsampleFactorMaxGrad and use a better one. · 28b99277
  由 Arnaud Bergeron 提交于 2月 11, 2016
  
  28b99277
- Fix the output shape in python to match the C code for DownsampleFactorMaxGradGrad. · 2d29bb3c
  由 Arnaud Bergeron 提交于 2月 11, 2016
  
  2d29bb3c
- Merge pull request #4033 from nouiz/py3_repr · 8a736df5
  由 Frédéric Bastien 提交于 2月 11, 2016
```
Fix CudaNdarray repr in python 3 and a crash.
```
  8a736df5
- Merge pull request #3795 from aalmah/rand_weighted_select · 931dfec1
  由 carriepl 提交于 2月 11, 2016
```
Sample multinomial with replacement
```
  931dfec1
- fixed the multi grad bug. · 8f442150
  由 Caglar 提交于 2月 11, 2016
  
  8f442150
- Add test for the previous fix · 20609f04
  由 Frederic Bastien 提交于 2月 11, 2016
  
  20609f04
- fix gh-4036. We need the shape and we don't compute it. It wasn't timmed, so… · 332601b4
  由 Frederic Bastien 提交于 2月 11, 2016
```
fix gh-4036. We need the shape and we don't compute it. It wasn't timmed, so don't try this optimization.
```
  332601b4
- Better error message · dbb49e4b
  由 Frederic Bastien 提交于 2月 11, 2016
  
  dbb49e4b
- Make DebugMode raise the good error when no code exist for an op · 12c4530d
  由 Frederic Bastien 提交于 2月 11, 2016
  
  12c4530d
11 2月, 2016 10 次提交
- Merge pull request #3964 from memimo/3929 · b2362664
  由 carriepl 提交于 2月 11, 2016
```
Cleaunup profileMode deprecation in docs
```
  b2362664
- Merge pull request #4032 from sygi/dot-a-file-in-ldlibrary · 4bec565c
  由 Frédéric Bastien 提交于 2月 11, 2016
```
(#4024) cutting the *.a extension in ldlibrary file
```
  4bec565c
- Fix CudaNdarray repr in python 3. This fix daily buildbot error in python 3 · 3835a616
  由 Frederic Bastien 提交于 2月 11, 2016
  
  3835a616
- Merge pull request #4000 from nouiz/cudnn_version · e9425be8
  由 Frédéric Bastien 提交于 2月 11, 2016
```
Cudnn version, print it, warn when too new, make mandatory in the new back-end.
```
  e9425be8
- (#4024) cutting the *.a extension in ldlibrary file · 455db13e
  由 Jakub Sygnowski 提交于 2月 11, 2016
  
  455db13e
- Make cnmem print the start size to make it more clear for user. · ef41de9c
  由 Frederic Bastien 提交于 2月 10, 2016
  
  ef41de9c
- Fix typo in warning · 290631f3
  由 Frederic Bastien 提交于 2月 10, 2016
  
  290631f3
- Fix crash during opt when we where looking for cudnn version. · 2c13f91b
  由 Frederic Bastien 提交于 2月 10, 2016
  
  2c13f91b
- If cudnn isn't there and we use cuda back-end, print it. · 89be3bcf
  由 Frederic Bastien 提交于 2月 10, 2016
  
  89be3bcf
- Only raise error if convolution or pooling are used in the new back-end. · b7f0c300
  由 Frederic Bastien 提交于 2月 10, 2016
  
  b7f0c300