提交 · 5708a1a00d00ea3572d6502ea0762d047582caa9 · testgroup / pytensor

30 9月, 2015 4 次提交
- Merge pull request #3470 from lamblin/fix_cast_args · 5708a1a0
  由 Frédéric Bastien 提交于 9月 29, 2015
```
Use make_new_inplace mechanism to fix optimization with cast
```
  5708a1a0
- Use make_new_inplace mechanism to fix optimization with cast · 95c17c54
  由 Pascal Lamblin 提交于 9月 29, 2015
  
  95c17c54
- Merge pull request #3453 from bouthilx/stack_doc · 8e17dcd6
  由 Frédéric Bastien 提交于 9月 29, 2015
```
Fix stack(tensors, axis=0) documentation
```
  8e17dcd6
- Merge pull request #3356 from abergeron/gpuarray_cudnnv3 · 1ef9be9d
  由 Pascal Lamblin 提交于 9月 29, 2015
```
cuDNN v3 support for gpuarray
```
  1ef9be9d
29 9月, 2015 9 次提交
- Fix flake8 problem. · 642446c5
  由 Arnaud Bergeron 提交于 9月 29, 2015
  
  642446c5
- Fix error messages. · a0e8767f
  由 Arnaud Bergeron 提交于 9月 28, 2015
  
  a0e8767f
- Rename out to dx for clarity. · c107d35d
  由 Arnaud Bergeron 提交于 9月 28, 2015
  
  c107d35d
- Fix small compile nit. · e237869c
  由 Arnaud Bergeron 提交于 9月 28, 2015
  
  e237869c
- Make alpha and beta respect the type of the inputs. · ef211924
  由 Arnaud Bergeron 提交于 9月 28, 2015
  
  ef211924
- Fix stack(tensors, axis=0) documentation · 09f6c39e
  由 Xavier Bouthillier 提交于 9月 27, 2015
  
  09f6c39e
- Merge pull request #3456 from nouiz/nanguard · e617dc50
  由 abergeron 提交于 9月 28, 2015
```
[CRASH] in nanguardmode with pycobject
```
  e617dc50
- Merge pull request #3163 from mjwillson/GpuCrossentropySoftmax_performance · 545d463f
  由 abergeron 提交于 9月 28, 2015
```
Performance improvement for GpuCrossentropySoftmaxArgmax1HotWithBias
```
  545d463f
- Merge pull request #3445 from nouiz/work_around_dnn_pool_infer_shape · d15bdccf
  由 abergeron 提交于 9月 28, 2015
```
Work around dnn pool infer shape
```
  d15bdccf
28 9月, 2015 9 次提交
- Merge pull request #3460 from Theano/revert-3435-scan_grad_speedup · 8a615a8d
  由 carriepl 提交于 9月 28, 2015
```
Revert "Scan grad speedup"
```
  8a615a8d
- Revert "Scan grad speedup" · 573dd60e
  由 carriepl 提交于 9月 28, 2015
  
  573dd60e
- Merge pull request #3435 from carriepl/scan_grad_speedup · 18fe6205
  由 carriepl 提交于 9月 28, 2015
```
Scan grad speedup
```
  18fe6205
- Small doc refactor for r op. · 66623c32
  由 Frederic 提交于 9月 28, 2015
  
  66623c32
- Merge pull request #3442 from nouiz/numpy_ver · c23a145c
  由 carriepl 提交于 9月 28, 2015
```
update numpy version to 1.7.0
```
  c23a145c
- Merge pull request #3455 from nouiz/bugfix_sum_prod · 41177b4c
  由 carriepl 提交于 9月 28, 2015
```
[BUG] in opt local_op_of_op
```
  41177b4c
- Merge pull request #3448 from eulerreich/patch-1 · a60bf55d
  由 Frédéric Bastien 提交于 9月 28, 2015
```
We can't make users notice unreleased features
```
  a60bf55d
- [CRASH] fix gh-3447 · f78fd294
  由 Frederic 提交于 9月 28, 2015
  
  f78fd294
- [BUG] in opt local_op_of_op fix gh-3451 · c4b90369
  由 Frederic 提交于 9月 28, 2015
  
  c4b90369
27 9月, 2015 1 次提交
- We can't make users notice unreleased features · d90e4312
  由 eulerreich 提交于 9月 26, 2015
  
  d90e4312
26 9月, 2015 17 次提交
- Make _handle apply specific. · 3a422ba3
  由 Arnaud Bergeron 提交于 9月 25, 2015
  
  3a422ba3
- Fix some things in code review. · 7bbbe727
  由 Arnaud Bergeron 提交于 9月 25, 2015
  
  7bbbe727
- Make sure to warn people that tensor_format doesn't work. · fae01b90
  由 Arnaud Bergeron 提交于 9月 25, 2015
  
  fae01b90
- Make sure not to use the descriptor op if not available. · acad83a3
  由 Arnaud Bergeron 提交于 9月 25, 2015
  
  acad83a3
- Merge pull request #3440 from carriepl/scan_push_out_seq_scan · 7576ab6d
  由 abergeron 提交于 9月 25, 2015
```
Scan push out seq scan
```
  7576ab6d
- Update doc and setup.py with the new min requirement · 56ce1fef
  由 Frederic 提交于 9月 25, 2015
  
  56ce1fef
- Better doc · 812244dd
  由 Frederic 提交于 9月 25, 2015
  
  812244dd
- work around gh-3444 · 6dfdace2
  由 Frederic 提交于 9月 25, 2015
  
  6dfdace2
- Try numpy 1.7.1 to work around the useless numpy output · 8b8f5487
  由 Frederic 提交于 9月 25, 2015
  
  8b8f5487
- Try with nose 1.3.0 and numpy 1.7.0 · 2ea9c7f2
  由 Frederic 提交于 9月 25, 2015
  
  2ea9c7f2
- revert numpy version change · 5b80957b
  由 Frederic 提交于 9月 25, 2015
  
  5b80957b
- bump scipy to 0.12 · 593b0f85
  由 Frederic 提交于 9月 25, 2015
  
  593b0f85
- update numpy version · d0250cb3
  由 Harm de Vries 提交于 9月 25, 2015
  
  d0250cb3
- Fix direct calls to GpuDnnConvDesc · 171aa1d7
  由 Arnaud Bergeron 提交于 9月 25, 2015
  
  171aa1d7
- Improve comment · 3cc67599
  由 carriepl 提交于 9月 25, 2015
  
  3cc67599
- Add some missing tests and fix small mistakes. · fa436a06
  由 Arnaud Bergeron 提交于 9月 25, 2015
  
  fa436a06
- Add test for pushout_nonseq_scan · 3e54b38a
  由 carriepl 提交于 9月 25, 2015
  
  3e54b38a