提交 · 7b50faa0af51a5359fc0906ed57f1e41bdd4fc38 · testgroup / pytensor

25 1月, 2010 1 次提交
- disabling floatX.py. using config.floatX to adjust the dtype default in tensor.matrix,vector,etc. · 7b50faa0
  由 James Bergstra 提交于 1月 24, 2010
  
  7b50faa0
23 1月, 2010 10 次提交
- merge · ac95733a
  由 James Bergstra 提交于 1月 22, 2010
  
  ac95733a
- removed theano/config.py · 040493f9
  由 James Bergstra 提交于 1月 22, 2010
  
  040493f9
- config overhaul · ed2e752d
  由 James Bergstra 提交于 1月 22, 2010
  
  ed2e752d
- remove debug line. · 21822e88
  由 Frederic Bastien 提交于 1月 22, 2010
  
  21822e88
- removed print · f4d8f92f
  由 Frederic Bastien 提交于 1月 22, 2010
  
  f4d8f92f
- fixed bug in conv2d · 3d65fadf
  由 gdesjardins 提交于 1月 22, 2010
  
  3d65fadf
- fix test with gpu to the new interface. · f2887ce2
  由 Frederic Bastien 提交于 1月 22, 2010
  
  f2887ce2
- don't put the input pattern in the GpuElemwise as we don't do it and this cause not needed copy. · 35d29d38
  由 Frederic Bastien 提交于 1月 22, 2010
  
  35d29d38
- Merge. · 5b6e5783
  由 Pierre-Antoine Manzagol 提交于 1月 22, 2010
  
  5b6e5783
- added test for gpu reshape · e1e774ce
  由 Frederic Bastien 提交于 1月 22, 2010
  
  e1e774ce
22 1月, 2010 29 次提交
- merge · 298f3491
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  298f3491
- More on loading and saving. · 22bb7496
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  22bb7496
- Make library reference point to example, instead of duplicating the guide. · 56566c25
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  56566c25
- merge · 596f9bca
  由 James Bergstra 提交于 1月 21, 2010
  
  596f9bca
- revising libdoc of tensor.basic · ffee86c7
  由 James Bergstra 提交于 1月 21, 2010
  
  ffee86c7
- added a few more tensor3- and tensor4-creation routines · 1f9586a9
  由 James Bergstra 提交于 1月 21, 2010
  
  1f9586a9
- adding script to generate tensor type table · 46820e72
  由 James Bergstra 提交于 1月 21, 2010
  
  46820e72
- Added max_pool2D function. Takes as input an N-d tensor and creates/calls · 0581942d
  由 gdesjardins 提交于 1月 21, 2010
```
DownsampleFactorMax to perform max pooling over the trailing two dimensions.
```
  0581942d
- Added new Prod op which multiplies the contents of a tensor along a given axis. · 8d11831d
  由 gdesjardins 提交于 1月 21, 2010
```
Added tensor.prod function to go with it.
```
  8d11831d
- merge · 03375db6
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  03375db6
- Clean up unused code (left over from previous version). · 2acc5734
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  2acc5734
- Added comments. · 734b9d6f
  由 Pierre-Antoine Manzagol 提交于 1月 21, 2010
  
  734b9d6f
- now if THEANO_GPU is defined to something else then "", theano will automatically import cuda! · 40feefca
  由 Frederic Bastien 提交于 1月 21, 2010
  
  40feefca
- fix gpu convop stuff since when we don't give the shape to ConvOp. · 1f344114
  由 Frederic Bastien 提交于 1月 21, 2010
  
  1f344114
- deleted commented blocks from tensor.opt · a0bf5ef0
  由 James Bergstra 提交于 1月 21, 2010
  
  a0bf5ef0
- tightened index wording · 502e02ab
  由 James Bergstra 提交于 1月 21, 2010
  
  502e02ab
- finished drafting the optimizations.txt file · c64de627
  由 James Bergstra 提交于 1月 21, 2010
  
  c64de627
- small bug fix. · ee3ca99b
  由 Frederic Bastien 提交于 1月 21, 2010
  
  ee3ca99b
- fix test_lenet_108 and test_lenet_256. · abf452d2
  由 Frederic Bastien 提交于 1月 21, 2010
  
  abf452d2
- merge · 09fb5eba
  由 gdesjardins 提交于 1月 21, 2010
  
  09fb5eba
- importing theano.tensor now auto-imports nnet · a24b845c
  由 gdesjardins 提交于 1月 21, 2010
  
  a24b845c
- Document the 'no_default_update' keyword of theano.function in the online doc. · fa39a62d
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  fa39a62d
- renamed conv.py to conv2d and changed interface slightly · c45d0df4
  由 gdesjardins 提交于 1月 21, 2010
  
  c45d0df4
- merge. · 7f17dde5
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  7f17dde5
- Update the tutorial example on random numbers to account for the default_update… · 7c1e4862
  由 Pascal Lamblin 提交于 1月 21, 2010
```
Update the tutorial example on random numbers to account for the default_update mechanism. Update the corresponding test.
```
  7c1e4862
- adjusted test_log1p to only override FAST_COMPILE mode · af04bbfe
  由 James Bergstra 提交于 1月 21, 2010
  
  af04bbfe
- added some code comments to pfunc · f8dfe6a9
  由 James Bergstra 提交于 1月 21, 2010
  
  f8dfe6a9
- Test the default_update behavior of shared random streams. · 539550b7
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  539550b7
- Add default automatic update of shared random streams · e11e3472
  由 Pascal Lamblin 提交于 1月 21, 2010
  
  e11e3472