提交 · rel-0.8.2 · testgroup / pytensor

21 4月, 2016 1 次提交
- Modifications for 0.8.2 release · 127aa70c
  由 Frederic Bastien 提交于 4月 21, 2016
  
  127aa70c
20 4月, 2016 1 次提交
- Fix https://github.com/fchollet/keras/issues/2339 and fix… · 63ed6b58
  由 Frederic Bastien 提交于 4月 15, 2016
```
Fix https://github.com/fchollet/keras/issues/2339 and fix https://github.com/fchollet/keras/issues/2345
```
  63ed6b58
15 4月, 2016 4 次提交
- Update elemwise_openmp_speedup.py · 5db700bc
  由 Marco 提交于 4月 12, 2016
```
Removed duplicate print code for "Fast op".
```
  5db700bc
- Rollback wrong fix · 9e1c5052
  由 texot 提交于 4月 12, 2016
  
  9e1c5052
- Fix PersistentNdarrayLoad problem restoring same object twice · 126ef599
  由 texot 提交于 4月 12, 2016
  
  126ef599
- Fix crashes probably due to nvcc not handling unsigned int as we expect · 4d641a3e
  由 Frederic Bastien 提交于 4月 14, 2016
  
  4d641a3e
14 4月, 2016 11 次提交
- fix other tests · d878af4f
  由 Harm de Vries 提交于 3月 23, 2016
  
  d878af4f
- Small update to error message · 26d47b89
  由 Frederic Bastien 提交于 4月 12, 2016
  
  26d47b89
- Make code more readable · e4dd0b3a
  由 Frederic Bastien 提交于 4月 12, 2016
  
  e4dd0b3a
- 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 1 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