提交 · ae614647ff0fa754f87e572bd92581acacccd0cd · testgroup / pytensor

19 2月, 2016 1 次提交
- Merge pull request #4089 from nouiz/gemv_broadcast · ae614647
  由 abergeron 提交于 2月 18, 2016
```
Fix crash with gemv and some broadcast pattern
```
  ae614647
18 2月, 2016 12 次提交
- Fix crash with gemv and some broadcast pattern · feaf7d42
  由 Frederic Bastien 提交于 2月 18, 2016
  
  feaf7d42
- Merge pull request #4084 from matt-graham/cudnn_paths_fix · b69b276e
  由 Frédéric Bastien 提交于 2月 18, 2016
```
cuDNN header / library paths fix
```
  b69b276e
- Merge pull request #4083 from caglar/fix_cnmemprint_msg · bbb8f719
  由 Frédéric Bastien 提交于 2月 18, 2016
```
Fix the cnmem print percent message
```
  bbb8f719
- Merge pull request #4065 from abergeron/lift_dot22scalar · 6b98e587
  由 Frédéric Bastien 提交于 2月 18, 2016
```
Also lift Dot22Scalar.
```
  6b98e587
- Also changing libs + lib_dirs in gof/cmodule.py for consistency. · 6da87264
  由 Matt Graham 提交于 2月 17, 2016
  
  6da87264
- Making order of libs consistent with include_dirs and lib_dirs. · 9d456b30
  由 Matt Graham 提交于 2月 17, 2016
  
  9d456b30
- Addition of user specified cuDNN paths to Op compile commands. · b4ff2ff2
  由 Matt Graham 提交于 2月 17, 2016
  
  b4ff2ff2
- Add infer_shape tests for the blas ops. · deb1f28e
  由 Arnaud Bergeron 提交于 2月 17, 2016
  
  deb1f28e
- Added the latest changes. · 7fc3d047
  由 Caglar 提交于 2月 17, 2016
  
  7fc3d047
- Test the lifted function for Dot22Scalar. · 0bc7d0df
  由 Arnaud Bergeron 提交于 2月 17, 2016
  
  0bc7d0df
- Flake8 · dfe4049d
  由 Arnaud Bergeron 提交于 2月 17, 2016
  
  dfe4049d
- Merge pull request #4067 from abergeron/debugmode_empty · 8215457d
  由 Frédéric Bastien 提交于 2月 17, 2016
```
Make DebugMode handle a special version of perform.
```
  8215457d
17 2月, 2016 10 次提交
- Merge pull request #4070 from abergeron/fix_buildbot2 · 189069be
  由 Frédéric Bastien 提交于 2月 17, 2016
```
Fix the LogSoftmax tests in DebugMode.
```
  189069be
- Merge pull request #4030 from hantek/fixtesterr · 6b02f8ca
  由 Frédéric Bastien 提交于 2月 17, 2016
```
solving doc test failures
```
  6b02f8ca
- Update opts using values_eq_approx to the new interface. · 68880f84
  由 Arnaud Bergeron 提交于 2月 16, 2016
  
  68880f84
- Fix typo in infer_shape methods. · 0ec6362a
  由 Arnaud Bergeron 提交于 2月 16, 2016
  
  0ec6362a
- Merge pull request #4042 from adbrebs/warn_flags · 3adaa82f
  由 abergeron 提交于 2月 16, 2016
```
[WIP] Move flags declarations and warn user if unknown flags
```
  3adaa82f
- Fix typos and wording. · d3530b07
  由 Arnaud Bergeron 提交于 2月 16, 2016
  
  d3530b07
- Fix the LogSoftmax tests in DebugMode. · 2f739edc
  由 Arnaud Bergeron 提交于 2月 16, 2016
  
  2f739edc
- Add infer_shape to the blas ops. · 62a7e19a
  由 Arnaud Bergeron 提交于 2月 16, 2016
  
  62a7e19a
- Replace Dot22Scalar by a GpuGemm instead and add a test. · 480ca3ad
  由 Arnaud Bergeron 提交于 2月 16, 2016
  
  480ca3ad
- Document debug_perform(). · adb4f3b7
  由 Arnaud Bergeron 提交于 2月 16, 2016
  
  adb4f3b7
16 2月, 2016 14 次提交
- python 3 compatibility · 2c1f7d80
  由 hantek 提交于 2月 15, 2016
  
  2c1f7d80
- 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
- Remove unused imports · 9b0c950a
  由 AdeB 提交于 2月 15, 2016
  
  9b0c950a
- modify test config to check that flags are removed · 325ce519
  由 AdeB 提交于 2月 15, 2016
  
  325ce519
- Make AllocEmpty return constant-valued memory under DebugMode so that it won't trip up checks. · f2e3c8d7
  由 Arnaud Bergeron 提交于 2月 15, 2016
  
  f2e3c8d7
- Allow nodes to have a 'debug' behaviour for DebugMode. · 27ee5379
  由 Arnaud Bergeron 提交于 2月 15, 2016
  
  27ee5379
- Fix calls to as_gpuarray_variable. · 622f1d51
  由 Arnaud Bergeron 提交于 2月 15, 2016
  
  622f1d51
- Also lift Dot22Scalar. · 90d31932
  由 Arnaud Bergeron 提交于 2月 15, 2016
  
  90d31932
- 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
- Print warnings when flags are not recognised · 494ff73b
  由 AdeB 提交于 2月 15, 2016
  
  494ff73b
- Remove default value when flag is not found in the flag dictionary · 556500bb
  由 AdeB 提交于 2月 15, 2016
  
  556500bb
15 2月, 2016 3 次提交
- Re-arrange order of imports in theano/__init__ · 3be031f8
  由 AdeB 提交于 2月 15, 2016
  
  3be031f8
- fix local_bitwidth and python_int_bitwidth imports · a2fb4f62
  由 AdeB 提交于 2月 15, 2016
  
  a2fb4f62
- 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