提交 · e9afaaab0779f48b74abfeac1cf91bc14ebe95b6 · testgroup / pytensor

14 4月, 2016 8 次提交
- Chagne CuDNN to cuDNN at all other places · e9afaaab
  由 Frederic Bastien 提交于 4月 12, 2016
```
Conflicts:
	NEWS.txt
```
  e9afaaab
- Print the old version available. · 006aac92
  由 Frederic Bastien 提交于 4月 12, 2016
  
  006aac92
- Use cuDNN instead of CuDNN. · 3ae18f61
  由 Frederic Bastien 提交于 4月 12, 2016
  
  3ae18f61
- Small user message update. · 5855468f
  由 Frederic Bastien 提交于 4月 12, 2016
  
  5855468f
- Correctly skip test when cudnn isn't available. · d1ca9a09
  由 Frederic Bastien 提交于 4月 12, 2016
  
  d1ca9a09
- Skip some test when cudnn isn't available and make tests pass when device=gpu. · 56bd5b80
  由 Frederic Bastien 提交于 4月 12, 2016
  
  56bd5b80
- Raise min cudnn version to v4. Better report of the problem when not used. · 5639c0f0
  由 Frederic Bastien 提交于 4月 12, 2016
  
  5639c0f0
- Merge pull request #4344 from mgermain/cudnn5 · ef0fc56e
  由 Frédéric Bastien 提交于 4月 12, 2016
```
Added cuDNN v5 support & Added optional dependencies to setup.py
Conflicts:
	.travis.yml
	doc/install.txt
```
  ef0fc56e
13 4月, 2016 10 次提交
- Update gpu pool tensor variable test · 02453383
  由 Harm de Vries 提交于 3月 18, 2016
  
  02453383
- add test · afad1e87
  由 Harm de Vries 提交于 3月 18, 2016
  
  afad1e87
- new gpu backend accepting tensor variables · ef4c9e9b
  由 Harm de Vries 提交于 3月 17, 2016
  
  ef4c9e9b
- Fix problem with python 3.5. · fe0b1477
  由 Pascal Lamblin 提交于 3月 22, 2016
```
It would raise a SystemError wrapping the IndexError already raised.
```
  fe0b1477
- Convert an C assert to a Python error. A user got that. · ffcad649
  由 Frederic Bastien 提交于 3月 23, 2016
  
  ffcad649
- Fix the cnmem memory usage printed · 34e45a0c
  由 Frederic Bastien 提交于 3月 23, 2016
  
  34e45a0c
- Add more direct tests for the scalar and the elemwise Ops · 0a029e57
  由 Pascal Lamblin 提交于 3月 23, 2016
  
  0a029e57
- Make sure the input of Neg.impl is not a numpy.bool_ · 8d3412b5
  由 Pascal Lamblin 提交于 3月 23, 2016
```
Otherwise, -False is True, and -True is False.
```
  8d3412b5
- adding clang-omp++ as an alternative compiler to clang++ as it supports OpenMP · b3d953d5
  由 Kevin Keraudren 提交于 3月 29, 2016
  
  b3d953d5
- Pass dims as intp to PyArray_SimpleNew · 521c44a5
  由 Pascal Lamblin 提交于 4月 05, 2016
  
  521c44a5
30 3月, 2016 4 次提交
- Modifications for 0.8.1 release · feeb0792
  由 Pascal Lamblin 提交于 3月 29, 2016
  
  feeb0792
- FIX: strip accents · 94a6cd57
  由 Gilles Louppe 提交于 3月 23, 2016
  
  94a6cd57
- Don't pass arguments with a space. · 8021da1e
  由 Arnaud Bergeron 提交于 3月 22, 2016
  
  8021da1e
- Only use -target-cpu as native flag on mac. · 5e400500
  由 Arnaud Bergeron 提交于 3月 22, 2016
  
  5e400500
22 3月, 2016 1 次提交
- Modifications for 0.8.0 release · 4b763483
  由 Frederic Bastien 提交于 3月 21, 2016
  
  4b763483
21 3月, 2016 1 次提交
- Merge pull request #4265 from mangwang/master · 16e251d3
  由 abergeron 提交于 3月 21, 2016
```
Repair the -rpath linker option in Windows System
```
  16e251d3
20 3月, 2016 1 次提交
- Ignore the '-Wl,-rpath' preargs in Microsoft linker · 247a9875
  由 Mang Wang 提交于 3月 19, 2016
  
  247a9875
19 3月, 2016 5 次提交
- repair the bug caused by the -rpath option is not understood by the Microsoft linker · a94cf43c
  由 Mang Wang 提交于 3月 18, 2016
  
  a94cf43c
- Merge pull request #4258 from lamblin/fix_debugmode_dnn · 260fc9a7
  由 Frédéric Bastien 提交于 3月 18, 2016
```
Fixes for debugmode in dnn and gpuarray
```
  260fc9a7
- Fixes for gpuarray in DebugMode · 32645f15
  由 Pascal Lamblin 提交于 3月 18, 2016
  
  32645f15
- Apply same workaround for other descriptors · 6328000d
  由 Pascal Lamblin 提交于 3月 17, 2016
  
  6328000d
- Do not check for non-finite values when we have -inf · 045eff2e
  由 Pascal Lamblin 提交于 3月 17, 2016
  
  045eff2e
18 3月, 2016 7 次提交
- Merge pull request #4251 from sygi/sparse-tutorial-typo · 153b0dfb
  由 Frédéric Bastien 提交于 3月 17, 2016
```
(nit) tutorial typos
```
  153b0dfb
- Merge pull request #4253 from lamblin/fix_debugmode_dnn · f04672b4
  由 Frédéric Bastien 提交于 3月 17, 2016
```
Fix tests in DebugMode with DNN convolutions
```
  f04672b4
- Merge pull request #4252 from nouiz/mrg_python · 1567d91d
  由 Frédéric Bastien 提交于 3月 17, 2016
```
Add the same restriction to the perform to return the same error and …
```
  1567d91d
- Fix tests in DebugMode with DNN convolutions · 0f25beee
  由 Pascal Lamblin 提交于 3月 17, 2016
  
  0f25beee
- (nit) tutorial typos · 90399d7e
  由 Jakub Sygnowski 提交于 3月 17, 2016
  
  90399d7e
- Add the same restriction to the perform to return the same error and have test… · d32810b9
  由 Frederic Bastien 提交于 3月 17, 2016
```
Add the same restriction to the perform to return the same error and have test don't use more then 9G of RAM.
```
  d32810b9
- Merge pull request #4249 from lamblin/fix_windows · 22b943c3
  由 Frédéric Bastien 提交于 3月 17, 2016
```
Fix windows test failures
```
  22b943c3
17 3月, 2016 3 次提交
- Merge pull request #4239 from sygi/sparse-tutorial-typo · 08963819
  由 Frédéric Bastien 提交于 3月 17, 2016
```
(nit) sparse tutorial typos
```
  08963819
- Skip another pickle test on Windows · ab6a8441
  由 Pascal Lamblin 提交于 3月 16, 2016
  
  ab6a8441
- Make isnan return 1 instead of -1 on some Windows · d7b03001
  由 Pascal Lamblin 提交于 3月 16, 2016
  
  d7b03001