提交 · cbc68829483a9bf35a2a61d8db5085bb705ada03 · testgroup / pytensor

19 4月, 2016 2 次提交
- Fix gh-4399. Add CpuContiguous.grad · cbc68829
  由 Frederic Bastien 提交于 4月 19, 2016
  
  cbc68829
- Add opt that move CpuContiguous to the GPU. fix gh-4400 · 9c8561fc
  由 Frederic Bastien 提交于 4月 19, 2016
  
  9c8561fc
16 4月, 2016 3 次提交
- Merge pull request #4381 from nouiz/set_subtensor1_crash · 189b1352
  由 Frédéric Bastien 提交于 4月 15, 2016
```
Fix crash in GpuAdvanvedIncSubtensor1 when unbroadcasting the value.
```
  189b1352
- Merge pull request #4263 from shabanian/opt · 7927f4a0
  由 Frédéric Bastien 提交于 4月 15, 2016
```
sumsqr2dot added to the opt.py
```
  7927f4a0
- Fix crash in GpuAdvanvedIncSubtensor1 when unbroadcasting the value. · f68c8fd5
  由 Frederic Bastien 提交于 4月 15, 2016
  
  f68c8fd5
15 4月, 2016 5 次提交
- Merge pull request #4375 from nouiz/work_aroung_nvcc · d0f73df4
  由 Pascal Lamblin 提交于 4月 15, 2016
```
Fix crashes probably due to nvcc not handling unsigned int as we expect
```
  d0f73df4
- Merge pull request #4374 from slefrancois/edit_release · 15e22007
  由 Frédéric Bastien 提交于 4月 14, 2016
```
Changed release doc to say new releases several times a year, instead…
```
  15e22007
- Merge pull request #4290 from adbrebs/stack_trace_opt · ee47964c
  由 Pascal Lamblin 提交于 4月 14, 2016
```
[WIP] Helper function that check stack traces
```
  ee47964c
- Merge pull request #4373 from fabioperez/master · 630c4e20
  由 Frédéric Bastien 提交于 4月 14, 2016
```
Close single quote
```
  630c4e20
- Fix crashes probably due to nvcc not handling unsigned int as we expect · ef10cd2e
  由 Frederic Bastien 提交于 4月 14, 2016
  
  ef10cd2e
14 4月, 2016 7 次提交
- Updated release number policy · 7a8a599c
  由 slefrancois 提交于 4月 14, 2016
  
  7a8a599c
- Changed release doc to say new releases several times a year, instead of monthly · bba57ae4
  由 slefrancois 提交于 4月 14, 2016
  
  bba57ae4
- Add new arguments/options to the argument of the function · 2f6c0084
  由 AdeB 提交于 4月 14, 2016
  
  2f6c0084
- Allows the first argument of check_stack_trace to be a fgraph · 812c1513
  由 AdeB 提交于 3月 27, 2016
  
  812c1513
- Add helper function check_stack_trace to check if a stack trace is in a graph or not · 8bedd35e
  由 AdeB 提交于 3月 23, 2016
  
  8bedd35e
- Move copy_stack_trace to gof.opt · 8c7511bf
  由 AdeB 提交于 3月 23, 2016
  
  8c7511bf
- Close single quote · 8f0c7d2e
  由 Fábio Perez 提交于 4月 14, 2016
  
  8f0c7d2e
13 4月, 2016 8 次提交
- Merge pull request #4366 from nouiz/cudnn_init · c10aa585
  由 Frédéric Bastien 提交于 4月 13, 2016
```
cudnn v5 cleanup init, small test fix.
```
  c10aa585
- Small update to error message · 5f872965
  由 Frederic Bastien 提交于 4月 12, 2016
  
  5f872965
- Make code more readable · 3c015784
  由 Frederic Bastien 提交于 4月 12, 2016
  
  3c015784
- Chagne CuDNN to cuDNN at all other places · 783d83e7
  由 Frederic Bastien 提交于 4月 12, 2016
  
  783d83e7
- Print the old version available. · fc4b5c18
  由 Frederic Bastien 提交于 4月 12, 2016
  
  fc4b5c18
- Use cuDNN instead of CuDNN. · 24a9ca84
  由 Frederic Bastien 提交于 4月 12, 2016
  
  24a9ca84
- Small user message update. · 3bdd23f6
  由 Frederic Bastien 提交于 4月 12, 2016
  
  3bdd23f6
- Merge pull request #4362 from tete1030/master · 7444fdd6
  由 Frédéric Bastien 提交于 4月 12, 2016
```
Fix PersistentNdarrayLoad problem restoring same object multiple times
```
  7444fdd6
12 4月, 2016 15 次提交
- changed f_test · 92956aec
  由 Samira Shabanian 提交于 4月 12, 2016
  
  92956aec
- Merge pull request #4363 from mafg/patch-1 · c6b6b0aa
  由 Frédéric Bastien 提交于 4月 12, 2016
```
Update elemwise_openmp_speedup.py
```
  c6b6b0aa
- Update doc about cudnn version supported · 35673bc1
  由 Frederic Bastien 提交于 4月 12, 2016
  
  35673bc1
- Correctly skip test when cudnn isn't available. · a117388e
  由 Frederic Bastien 提交于 4月 12, 2016
  
  a117388e
- Skip some test when cudnn isn't available and make tests pass when device=gpu. · 47445fc1
  由 Frederic Bastien 提交于 4月 12, 2016
  
  47445fc1
- Raise min cudnn version to v4. Better report of the problem when not used. · 9de5c0d5
  由 Frederic Bastien 提交于 4月 12, 2016
  
  9de5c0d5
- Merge pull request #4344 from mgermain/cudnn5 · f191575f
  由 Frédéric Bastien 提交于 4月 12, 2016
```
Added cuDNN v5 support & Added optional dependencies to setup.py
```
  f191575f
- Rollback wrong fix · 1a8ec990
  由 texot 提交于 4月 12, 2016
  
  1a8ec990
- Update elemwise_openmp_speedup.py · 94bd0d40
  由 Marco 提交于 4月 12, 2016
```
Removed duplicate print code for "Fast op".
```
  94bd0d40
- Fix PersistentNdarrayLoad problem restoring same object twice · 9c57a03c
  由 texot 提交于 4月 12, 2016
  
  9c57a03c
- updated comment · b424cc1d
  由 Mathieu Germain 提交于 4月 11, 2016
  
  b424cc1d
- Merge pull request #4335 from fvisin/fix_elemwise · 1efb1539
  由 Pascal Lamblin 提交于 4月 11, 2016
```
Check RST files in '/theano/doc/' to follow numpy's docstring + Fix elemwise
```
  1efb1539
- Merge pull request #4330 from nouiz/cuda_inplace_true_div · 2034dfb8
  由 Pascal Lamblin 提交于 4月 11, 2016
```
fix gh-4328 This was done for python3, but not python2.
```
  2034dfb8
- Merge pull request #4324 from yobibyte/fix · 9d544fb4
  由 Frédéric Bastien 提交于 4月 11, 2016
```
Moved theano.sandbox.softsign to tensor/nnet/nnet.py. Added test. #4314
```
  9d544fb4
- Merge pull request #4340 from lamblin/fix_broadcast · 6fb0300f
  由 Frédéric Bastien 提交于 4月 11, 2016
```
Add broadcastable dimensions in opt where needed
```
  6fb0300f