提交 · ee947ae85bc338c781898fd622fccd514f6b208f · testgroup / pytensor

01 11月, 2012 31 次提交
- fixed error in structured text · ee947ae8
  由 Razvan Pascanu 提交于 10月 16, 2012
  
  ee947ae8
- new test for numeric shared · 902ff4d2
  由 Razvan Pascanu 提交于 10月 16, 2012
  
  902ff4d2
- correcting rop2 test · 15a77a59
  由 Razvan Pascanu 提交于 10月 16, 2012
  
  15a77a59
- handling disconnected and undiff cases · b0f5fb53
  由 Razvan Pascanu 提交于 10月 16, 2012
  
  b0f5fb53
- correct number of steps · e8f5ae18
  由 Razvan Pascanu 提交于 10月 16, 2012
  
  e8f5ae18
- correction of grad method for RandomState · c9856df8
  由 Razvan Pascanu 提交于 10月 15, 2012
  
  c9856df8
- correction of grad method for DimShuffle · 5520a058
  由 Razvan Pascanu 提交于 10月 15, 2012
  
  5520a058
- shared variables become sit_sot if are of tensor type · 40012bce
  由 Razvan Pascanu 提交于 10月 15, 2012
  
  40012bce
- change names to more proper things · ee763488
  由 Razvan Pascanu 提交于 10月 15, 2012
  
  ee763488
- fixed error 3 · 98c1a8b8
  由 Razvan Pascanu 提交于 10月 10, 2012
  
  98c1a8b8
- fixed error 2 · d8521d17
  由 Razvan Pascanu 提交于 10月 10, 2012
  
  d8521d17
- fixed error 1 · bc691c11
  由 Razvan Pascanu 提交于 10月 10, 2012
  
  bc691c11
- new implementation of grad method · 3a85ea97
  由 Razvan Pascanu 提交于 10月 10, 2012
  
  3a85ea97
- temporary fix of test (until new pr) · 25d8089a
  由 Razvan Pascanu 提交于 10月 10, 2012
  
  25d8089a
- new test to check for dtype issues of grad · 81728268
  由 Razvan Pascanu 提交于 10月 10, 2012
  
  81728268
- function required to have fast gradient computations · 417067e1
  由 Razvan Pascanu 提交于 10月 10, 2012
  
  417067e1
- Merge pull request #1042 from jlowin/cutils_ext-import-fix · 05e72586
  由 nouiz 提交于 10月 31, 2012
```
cutils_ext import fix -- allow theano to run on picloud servers (not CVM)
```
  05e72586
- Merge pull request #1043 from goodfeli/fix_incsubtensor_opt · e6e6a433
  由 nouiz 提交于 10月 31, 2012
```
fix bug in gpu inc subtensor optimization and add a unit test
```
  e6e6a433
- improved comment explaining why these lines are here · a2a6d7cd
  由 Jeremiah Lowin 提交于 10月 31, 2012
  
  a2a6d7cd
- Merge pull request #1041 from goodfeli/fix_grad · 93f3b868
  由 nouiz 提交于 10月 31, 2012
```
fix bug in gradient when some outputs of a node have different connectio...
```
  93f3b868
- Update comment to explain __init__ touch · 090d80f9
  由 Jeremiah Lowin 提交于 10月 31, 2012
  
  090d80f9
- fix bug introduced by earlier commit (I tried squashing the commits and · e3a886eb
  由 Ian Goodfellow 提交于 10月 31, 2012
```
git just got rid of this edit)
```
  e3a886eb
- add (redundent) __init__ touch to get picloud to work with c|py linker · a7b286ae
  由 Jeremiah Lowin 提交于 10月 31, 2012
  
  a7b286ae
- updated comment · ab1a1c26
  由 Ian Goodfellow 提交于 10月 31, 2012
  
  ab1a1c26
- Merge pull request #1040 from nouiz/misc · b6db0244
  由 lamblin 提交于 10月 31, 2012
```
Misc
```
  b6db0244
- fix bug in gpu inc subtensor optimization and add a unit test · 8f0a7890
  由 Ian Goodfellow 提交于 10月 31, 2012
  
  8f0a7890
- fix bug in gradient when some outputs of a node have different connection patterns · 2484dde0
  由 Ian Goodfellow 提交于 10月 31, 2012
```
(Razvan's branch will introduce a test that catches this bug)
```
  2484dde0
- Sync the timming when using the GPU. Otherwise, we end up with 0s! · a9286c50
  由 Frederic 提交于 10月 31, 2012
  
  a9286c50
- use the nvcc path found by theano. · 5de2bd22
  由 Frederic 提交于 10月 31, 2012
  
  5de2bd22
- call nvcc only if we init this device. · a261e5dc
  由 Frederic 提交于 10月 31, 2012
  
  a261e5dc
- ignore ds_store · 47f7ee3b
  由 Jeremiah Lowin 提交于 10月 31, 2012
  
  47f7ee3b
31 10月, 2012 9 次提交
- Merge pull request #1034 from nouiz/docs · 6f242eec
  由 lamblin 提交于 10月 31, 2012
```
Add erfinv and erfcinv in the doc.
```
  6f242eec
- Merge pull request #1039 from lamblin/test_softmaxgrad_flatten · 9465d81a
  由 nouiz 提交于 10月 31, 2012
```
Add test for the "flatten" vector case
```
  9465d81a
- Fix some pep8/pyflakes violations · 9f3eb6db
  由 Pascal Lamblin 提交于 10月 30, 2012
  
  9f3eb6db
- Make tests more accurate · a885bfb0
  由 Pascal Lamblin 提交于 10月 30, 2012
  
  a885bfb0
- Add test for the "flatten" vector case · b5d38d68
  由 Pascal Lamblin 提交于 9月 19, 2012
```
This case is similar to what is used in DeepLearningBenchmarks
```
  b5d38d68
- add the check_blas.py timing for C2075 with cuda 4.2 · fc8ea431
  由 Frederic 提交于 10月 30, 2012
  
  fc8ea431
- make check_blas.py print the nvcc version. · ebc5feb1
  由 Frederic 提交于 10月 30, 2012
  
  ebc5feb1
- Merge pull request #1037 from lamblin/fix_flatten_infershape · 5d568c17
  由 nouiz 提交于 10月 30, 2012
```
Fix bug in infer_shape of Flatten
```
  5d568c17
- Merge pull request #1038 from lamblin/check_pydot · a591f4bd
  由 nouiz 提交于 10月 30, 2012
```
Check the graphviz executable are there.
```
  a591f4bd