提交 · 7cbb30e11b0f5459e8662bc69f8d56d345d63238 · testgroup / pytensor

16 3月, 2011 15 次提交
- Fix borrow flag, after code review. · 7cbb30e1
  由 Pascal Lamblin 提交于 3月 16, 2011
  
  7cbb30e1
- merge · 4bbac540
  由 Olivier Breuleux 提交于 3月 15, 2011
  
  4bbac540
- closes #644 elemwise perform should use numpy vectorized implementations; fixed… · 59561698
  由 Olivier Breuleux 提交于 3月 15, 2011
```
closes #644 elemwise perform should use numpy vectorized implementations; fixed typos in scan and added a warning
```
  59561698
- Actually use an inplace function in the test. · 8ea860cd
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  8ea860cd
- Do not use .value in the Theano tutorials (and associated tests) · 9f67a227
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  9f67a227
- Whitespace · 5f08407d
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  5f08407d
- Update tests to use .get_value() and .set_value() on shared variables. · f5ed1817
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  f5ed1817
- Whitespace. · 27084539
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  27084539
- Use get_value and set_value, not .value, in the library itself. · 5d6a2ac4
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  5d6a2ac4
- Add deprecation warning for .value property of shared variables. · 875424b8
  由 Pascal Lamblin 提交于 3月 15, 2011
```
Also update inline documentation.
```
  875424b8
- In the reshape op, put the output broadcast flag correctly when we receive an… · bb4e32e1
  由 Frederic Bastien 提交于 3月 15, 2011
```
In the reshape op, put the output broadcast flag correctly when we receive an int 1 as the new shape.
```
  bb4e32e1
- white space fix. · e57bfe5f
  由 Frederic Bastien 提交于 3月 15, 2011
  
  e57bfe5f
- work around a crash of gcc due to a bug in gcc. · 49c2fb0b
  由 Frederic Bastien 提交于 3月 15, 2011
  
  49c2fb0b
- 'optimize var[int1::][:int2] to only 1 subtensor.' · e846ef90
  由 Frederic Bastien 提交于 3月 15, 2011
  
  e846ef90
- small refactoring. · addbc07c
  由 Frederic Bastien 提交于 3月 15, 2011
  
  addbc07c
15 3月, 2011 5 次提交
- Added more merge optimization of subtensor of subtensor. case var[::-1][:int]. · 20bc7320
  由 Frederic Bastien 提交于 3月 15, 2011
  
  20bc7320
- 'Merge more subtensor of subtensor(the case var[::-1][int])' · 560af0e6
  由 Frederic Bastien 提交于 3月 15, 2011
  
  560af0e6
- added some test of the not optimized function. · 49e7b915
  由 Frederic Bastien 提交于 3月 15, 2011
  
  49e7b915
- Indentation. · c0b57d75
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  c0b57d75
- Reorder imports according to our convention. · d8cee62e
  由 Pascal Lamblin 提交于 3月 15, 2011
  
  d8cee62e
14 3月, 2011 1 次提交
- better printing of scan profiling. · b629d1cf
  由 Frederic Bastien 提交于 3月 14, 2011
  
  b629d1cf
12 3月, 2011 8 次提交
- Fix a case in local_subtensor_lift where the optimization was failing. · 499334e6
  由 Pascal Lamblin 提交于 3月 11, 2011
```
A broadcasting pattern of '' was used, instead of [].
```
  499334e6
- Also register the optimizations in the specialize phase. · 6dacc650
  由 Pascal Lamblin 提交于 3月 11, 2011
  
  6dacc650
- review commit. Remove debug print. Test a missed case. · 23817993
  由 Frederic Bastien 提交于 3月 11, 2011
  
  23817993
- Fix subtensor merge when the first subtensor is empty. Test it. · 2d4b514d
  由 Pascal Lamblin 提交于 3月 11, 2011
  
  2d4b514d
- Add optimization removing transfers between scalars and scalar tensors · 4357224c
  由 Pascal Lamblin 提交于 3月 11, 2011
  
  4357224c
- Added a Theano flag to disable profiling of memory used by Theano function. · 8cf01ede
  由 Frederic Bastien 提交于 3月 11, 2011
  
  8cf01ede
- small fix when DeepCopyOp don't implement c_code. · 5791145d
  由 Frederic Bastien 提交于 3月 11, 2011
  
  5791145d
- Move test for __main__ at the end of the file. · c0430357
  由 Pascal Lamblin 提交于 3月 11, 2011
  
  c0430357
11 3月, 2011 4 次提交
- Implement TensorFromScalar.__str__ · c74a10e8
  由 Pascal Lamblin 提交于 3月 10, 2011
  
  c74a10e8
- Auto whitespace fix · d5b59a26
  由 Pascal Lamblin 提交于 3月 10, 2011
  
  d5b59a26
- In check_chain, always return a boolean, not a Variable, nor None · f614fd40
  由 Pascal Lamblin 提交于 3月 10, 2011
  
  f614fd40
- error messages · 5c99fda1
  由 Philippe Hamel 提交于 3月 10, 2011
  
  5c99fda1
10 3月, 2011 7 次提交
- 'merge 2 subtensor in this case: var[int:][-1]' · d50c460b
  由 Frederic Bastien 提交于 3月 09, 2011
  
  d50c460b
- Remove the "mode" argument of pydotprint. · e146275a
  由 Pascal Lamblin 提交于 3月 09, 2011
```
It does not really make sense to specify something else than the default.
```
  e146275a
- small profile mode output update. · fbc0927d
  由 Frederic Bastien 提交于 3月 09, 2011
  
  fbc0927d
- duplicate some printing in profile mode to make the reading more easier. · 8e606c7a
  由 Frederic Bastien 提交于 3月 09, 2011
  
  8e606c7a
- small fix to have the info in all case. · 6ede3910
  由 Frederic Bastien 提交于 3月 09, 2011
  
  6ede3910
- in pydotprint, go take the mode by default in the function automatically. · de2768b9
  由 Frederic Bastien 提交于 3月 09, 2011
  
  de2768b9
- Remove -u arguments in nvcc's command line. · 821f69e5
  由 Pascal Lamblin 提交于 3月 09, 2011
  
  821f69e5