提交 · b0b4f13db12f6eed3ca718cd70a9b26d3cdc65f1 · testgroup / pytensor

03 2月, 2016 1 次提交
- small fix following code review · b0b4f13d
  由 Frederic Bastien 提交于 2月 02, 2016
  
  b0b4f13d
02 2月, 2016 6 次提交
- small doc update · 49b79e5e
  由 Frederic Bastien 提交于 2月 01, 2016
  
  49b79e5e
- Doc dnn.enable flag · 52790dda
  由 Frederic Bastien 提交于 2月 01, 2016
  
  52790dda
- small code change · 9d1344fa
  由 Frederic Bastien 提交于 2月 01, 2016
  
  9d1344fa
- Use config.dnn.enabled in gpuarray · 08c6bbc3
  由 Frederic Bastien 提交于 2月 01, 2016
  
  08c6bbc3
- Add config flag dnn.enabled={auto,True,False}. · b7559702
  由 Frederic Bastien 提交于 2月 01, 2016
```
fix gh-3962
```
  b7559702
- Merge pull request #3936 from nouiz/32bit · 9627228c
  由 Pascal Lamblin 提交于 2月 01, 2016
```
Mostly tests fix for 32bit python
```
  9627228c
01 2月, 2016 9 次提交
- Merge pull request #3947 from Sentient07/issue-#3944 · d6b3dff4
  由 Frédéric Bastien 提交于 2月 01, 2016
```
Renamed BatchedDotOp to GpuBatchedDot
```
  d6b3dff4
- Fix a test in 32bit. We can't make min/max reduction on 0 sized tensor. · 319e0aae
  由 Frederic Bastien 提交于 2月 01, 2016
  
  319e0aae
- Fix code review comment · 1ed8895e
  由 Frederic Bastien 提交于 2月 01, 2016
  
  1ed8895e
- added BatchedDotOp · ed3034a3
  由 Ramana.S 提交于 1月 29, 2016
  
  ed3034a3
- Merge pull request #3955 from pra85/patch-2 · aecc4921
  由 Frédéric Bastien 提交于 1月 31, 2016
```
Fix typo in docs
```
  aecc4921
- Merge pull request #3883 from skaae/bessel · 47278f72
  由 Frédéric Bastien 提交于 1月 31, 2016
```
[WIP] add bessel0
```
  47278f72
- Merge pull request #3952 from nouiz/doc · dfce3aee
  由 Frédéric Bastien 提交于 1月 31, 2016
```
Fix doc gen
```
  dfce3aee
- add j0 and j1 to if else check · 14fdf308
  由 skaae 提交于 1月 31, 2016
  
  14fdf308
- remove __eq__ and __hash__ · 6f5082fe
  由 skaae 提交于 1月 31, 2016
  
  6f5082fe
31 1月, 2016 1 次提交
- Fix typo in docs · 61c71ab6
  由 Prayag Verma 提交于 1月 31, 2016
```
debuging > debugging
```
  61c71ab6
30 1月, 2016 3 次提交
- Fix docgen as now warning are treated as error · dac4da86
  由 Frederic Bastien 提交于 1月 29, 2016
  
  dac4da86
- Better test update for python 32bit fix · b2f99581
  由 Frederic Bastien 提交于 1月 29, 2016
  
  b2f99581
- Fix comments and move code to detect those type of errors · bdfe90b7
  由 Frederic Bastien 提交于 1月 29, 2016
  
  bdfe90b7
29 1月, 2016 6 次提交
- Renamed BatchedDotOp to GpuBatchedDot · 72d3a280
  由 Ramana.S 提交于 1月 29, 2016
  
  72d3a280
- Merge pull request #3943 from lamblin/fix_batcheddotop_import · 18319b8f
  由 carriepl 提交于 1月 28, 2016
```
Fix path for BatchedDot
```
  18319b8f
- Fix path for BatchedDot · 73fbb214
  由 Pascal Lamblin 提交于 1月 28, 2016
  
  73fbb214
- Merge pull request #3508 from cooijmanstim/batched_gemm · f87853c5
  由 carriepl 提交于 1月 28, 2016
```
WIP: BatchedDotOp
```
  f87853c5
- Make test in python 32bit. sum, average cause small difference result ~1e-17 · 55655580
  由 Frederic Bastien 提交于 1月 28, 2016
  
  55655580
- Fix test with python 32bit · 5c022322
  由 Frederic Bastien 提交于 1月 28, 2016
  
  5c022322
28 1月, 2016 4 次提交
- Merge pull request #3940 from abergeron/fix_buildbot · 4800a51c
  由 Frédéric Bastien 提交于 1月 27, 2016
```
Fix most of the remaining DeprecationWarnings in the tests.
```
  4800a51c
- Don't use floating-point axes. · 7fc2bcbf
  由 Arnaud Bergeron 提交于 1月 27, 2016
  
  7fc2bcbf
- Remove confusing sys.stdout.flush() which displaces the printing of messages. · 1f984665
  由 Arnaud Bergeron 提交于 1月 27, 2016
  
  1f984665
- Merge pull request #3938 from carriepl/half_padding · 6e25ecf9
  由 carriepl 提交于 1月 27, 2016
```
Add 'half' padding to GpuArray dnn convolution
```
  6e25ecf9
27 1月, 2016 7 次提交
- Add tests for border_mode half · 0c17be82
  由 carriepl 提交于 1月 27, 2016
  
  0c17be82
- Add border_mode 'half' to dnn convolution · 3ee95b87
  由 carriepl 提交于 1月 27, 2016
  
  3ee95b87
- Fix test in 32 bit python · 1908d995
  由 Frederic Bastien 提交于 1月 26, 2016
  
  1908d995
- remove bincount uint64 restriction, it was a left over from the numpy implementation · 49785840
  由 Frederic Bastien 提交于 1月 26, 2016
  
  49785840
- Make repeat test pass on 32bit python. We use the numpy code only for vector, not scalar · 817ff62b
  由 Frederic Bastien 提交于 1月 26, 2016
  
  817ff62b
- Make test of bincount pass with 32bit python. We don't have the numpy limit anymore · 45c7b96a
  由 Frederic Bastien 提交于 1月 26, 2016
  
  45c7b96a
- Better error message · 2bbcab02
  由 Frederic Bastien 提交于 1月 26, 2016
  
  2bbcab02
26 1月, 2016 3 次提交
- Merge pull request #3918 from nouiz/tx1 · e891fb3f
  由 Pascal Lamblin 提交于 1月 25, 2016
```
Doc installation on Tegra TX1 board.
```
  e891fb3f
- Merge pull request #3845 from f0k/add-cudnn-half-padding · 45efd371
  由 Pascal Lamblin 提交于 1月 25, 2016
```
Add border_mode="half" to cuDNN convolutions
```
  45efd371
- Merge pull request #3890 from mohammadpz/remove_param_class · 62d5cbeb
  由 Pascal Lamblin 提交于 1月 25, 2016
```
Param class in now replaced by In
```
  62d5cbeb