提交 · b77082b2742597b2cb2ecd57a20fe52bd395dcba · testgroup / pytensor

28 10月, 2011 11 次提交
- Merge pull request #167 from nouiz/flags · b77082b2
  由 James Bergstra 提交于 10月 28, 2011
```
config flag for max_use_ratio in Equilibrium optimizer
```
  b77082b2
- pep8 fix. · e94955a7
  由 Frederic 提交于 10月 28, 2011
  
  e94955a7
- Add a theano flags that allow to raise the equilibrium max_use_ratio. · 43cb38b8
  由 Frederic 提交于 10月 28, 2011
  
  43cb38b8
- Merge pull request #166 from pascanur/scan_grad_nsteps · f4b29ab6
  由 nouiz 提交于 10月 28, 2011
```
Fix a bug of grad reported by Michael Forbes
```
  f4b29ab6
- fix formatting of the new introduced test · 2abc014b
  由 Razvan Pascanu 提交于 10月 28, 2011
  
  2abc014b
- making scan.py PEP8 compatible · fc22aaf9
  由 Razvan Pascanu 提交于 10月 28, 2011
  
  fc22aaf9
- Fix a bug of grad reported by Michael Forbes · 80f150b2
  由 Razvan Pascanu 提交于 10月 28, 2011
```
The bug is that we do not clip the sequences to n_steps, meaning that when
we revert them for computing the gradient we mess up the gradient.
```
  80f150b2
- Merge pull request #164 from nouiz/pep8 · 5dd4c64a
  由 Razvan Pascanu 提交于 10月 27, 2011
```
Pep8
I've looked over it and everything seems fine.
```
  5dd4c64a
- more pep8 fix. · 660251a9
  由 Frederic 提交于 10月 27, 2011
  
  660251a9
- Move how pydot is imported and detected to respect pep8. · 863677d0
  由 Frederic 提交于 10月 27, 2011
  
  863677d0
- fix pep8 · 7266073f
  由 Frederic 提交于 10月 27, 2011
  
  7266073f
27 10月, 2011 2 次提交
- Merge pull request #163 from pascanur/test_matrix_dot · 2a97ffff
  由 nouiz 提交于 10月 27, 2011
```
Added test for matrix_dot
```
  2a97ffff
- Added test for matrix_dot · 142a7694
  由 Razvan Pascanu 提交于 10月 27, 2011
  
  142a7694
26 10月, 2011 14 次提交
- Merge branch 'nouiz-error' · 32fa17ac
  由 Olivier Delalleau 提交于 10月 26, 2011
  
  32fa17ac
- Fixed some typos · 346400c1
  由 Olivier Delalleau 提交于 10月 26, 2011
  
  346400c1
- Add comment. · d6934160
  由 Frederic 提交于 10月 26, 2011
  
  d6934160
- Don't allow to change at run time the tensor.cmp_sloppy flags value as this don't have any effect. · a0cf1c8a
  由 Frederic 提交于 10月 26, 2011
  
  a0cf1c8a
- Don't implement TensorVariable.__len__ to give better error message as this cause other bug. · 4219e91d
  由 Frederic 提交于 10月 26, 2011
  
  4219e91d
- Merge pull request #161 from dwf/xrange_fixes · 05aaa81b
  由 Olivier Delalleau 提交于 10月 26, 2011
```
Use xrange() instead of range() when iterating
```
  05aaa81b
- Replace [i for i in xrange()] with range() · 3ec9465a
  由 David Warde-Farley 提交于 10月 26, 2011
  
  3ec9465a
- PEP8/readability. · b438341d
  由 David Warde-Farley 提交于 10月 26, 2011
  
  b438341d
- in range() -> in xrange() where possible · 19f7768f
  由 David Warde-Farley 提交于 10月 25, 2011
  
  19f7768f
- small upgrade following code review. · 94349257
  由 Frederic 提交于 10月 25, 2011
  
  94349257
- some pep8 fix. · 0ab95847
  由 Frederic 提交于 10月 25, 2011
  
  0ab95847
- Better error message. · 6ddf8bec
  由 Frederic 提交于 10月 25, 2011
  
  6ddf8bec
- Merge pull request #159 from nouiz/doc · a68974e9
  由 Olivier Delalleau 提交于 10月 25, 2011
```
Doc
```
  a68974e9
- In the tensor.sparse doc, tell that we must import it manually. · 515b2ba8
  由 Frederic 提交于 10月 25, 2011
  
  515b2ba8
25 10月, 2011 13 次提交
- Rename the Interactive Debugger section to Using Test Values. · dc753291
  由 Frederic 提交于 10月 25, 2011
  
  dc753291
- Merge pull request #158 from dwf/matrix_inv_take2 · 749c474f
  由 nouiz 提交于 10月 25, 2011
```
Matrix inv take2
```
  749c474f
- Merge pull request #127 from abergeron/back_shared · 7145ca0b
  由 nouiz 提交于 10月 25, 2011
```
Remove the shared pickle changes
```
  7145ca0b
- matrix_inverse test that triggers LinAlgError. · 1e81c4d8
  由 David Warde-Farley 提交于 10月 24, 2011
  
  1e81c4d8
- Correct import in sandbox/linalg/__init__.py · c9e6bec7
  由 David Warde-Farley 提交于 10月 24, 2011
  
  c9e6bec7
- Use logging infrastructure in place of print · b257a8ad
  由 David Warde-Farley 提交于 10月 24, 2011
  
  b257a8ad
- Rename PSD_hint -> psd and added docstring. · ccaaf770
  由 David Warde-Farley 提交于 10月 24, 2011
```
Pursuant to discussion at https://github.com/Theano/Theano/pull/153/files#r187252 I've renamed the function. I also added a docstring.

Conflicts:

	theano/sandbox/linalg/ops.py
```
  ccaaf770
- Use the _allclose function to decide if arrays are close. · bb6261c4
  由 Razvan Pascanu 提交于 10月 24, 2011
```
Conflicts:

	theano/sandbox/linalg/tests/test_linalg.py
```
  bb6261c4
- Use utt.get_seed() to give a seed for every random generator. · 1d8ef85d
  由 Razvan Pascanu 提交于 10月 24, 2011
```
Conflicts:

	theano/sandbox/linalg/tests/test_linalg.py
```
  1d8ef85d
- remove useless lines · 1a0d4a49
  由 Razvan Pascanu 提交于 10月 24, 2011
  
  1a0d4a49
- Small doc update with direct review by David. · 476eed11
  由 Frederic 提交于 10月 24, 2011
  
  476eed11
- rop for matrix inverse plus test · 7fd2e38f
  由 Razvan Pascanu 提交于 10月 24, 2011
  
  7fd2e38f
- new documentation for the grad method · ab1c7862
  由 Razvan Pascanu 提交于 10月 24, 2011
  
  ab1c7862