提交 · 54c3b53ad1bb7a491556f6edbf603adc08d7e438 · testgroup / pytensor

03 7月, 2010 1 次提交
- A somewhat more reasonable/basic setup.py using only distutils. Seems to work. · 54c3b53a
  由 David Warde Farley 提交于 7月 02, 2010
  
  54c3b53a
17 6月, 2010 3 次提交
- added eq and hash for tensor.Dot · 0265894e
  由 James Bergstra 提交于 6月 16, 2010
  
  0265894e
- merge · fefdf2b1
  由 Simon Lemieux 提交于 6月 16, 2010
  
  fefdf2b1
- neighbours seems to work on gpu · 8da1f256
  由 Simon Lemieux 提交于 6月 16, 2010
  
  8da1f256
15 6月, 2010 3 次提交
- remove not used stuff. · 81ae8ae9
  由 Frederic Bastien 提交于 6月 14, 2010
  
  81ae8ae9
- small change. · 0d268386
  由 Frederic Bastien 提交于 6月 14, 2010
  
  0d268386
- adding gpu code for neighbours.py · 21d76325
  由 Simon Lemieux 提交于 6月 14, 2010
  
  21d76325
14 6月, 2010 2 次提交
- make code versionable. · 946a8fa5
  由 Frederic Bastien 提交于 6月 14, 2010
  
  946a8fa5
- make test more complete, but the new test fail! · 790462b6
  由 Frederic Bastien 提交于 6月 14, 2010
  
  790462b6
12 6月, 2010 3 次提交
- merge · 4ba6a10a
  由 Razvan Pascanu 提交于 6月 11, 2010
  
  4ba6a10a
- aliasing bug reported by Ian fixed .. · 655ecd92
  由 Razvan Pascanu 提交于 6月 11, 2010
  
  655ecd92
- adding simplified neighbours op for cpu · dfc04612
  由 Simon Lemieux 提交于 6月 11, 2010
  
  dfc04612
11 6月, 2010 21 次提交
- made the gpu optimize phase earlier to have better performance in some case. · 56f5d44a
  由 Frederic Bastien 提交于 6月 11, 2010
  
  56f5d44a
- make Reshape.infer_shape work in more case. · 6b2b5132
  由 Frederic Bastien 提交于 6月 11, 2010
  
  6b2b5132
- make the optimizer local_inplace_setsubtensor work for GpuIncSubtensor too. · 2da94091
  由 Frederic Bastien 提交于 6月 11, 2010
  
  2da94091
- make the GpuAlloc work in more case. · b8a75301
  由 Frederic Bastien 提交于 6月 11, 2010
  
  b8a75301
- small clean up that make CudaNdarray_setitem it a little faster. · 5cc01161
  由 Frederic Bastien 提交于 6月 11, 2010
  
  5cc01161
- Add sanity check on seed in rng_mrg, to ensure the results' validity · 12fa8f65
  由 Pascal Lamblin 提交于 6月 10, 2010
  
  12fa8f65
- Remove one of the last references to weave · f3fedb46
  由 Pascal Lamblin 提交于 6月 10, 2010
  
  f3fedb46
- Added free(newdims) when exiting before the end of the function. · 70965d0d
  由 Pascal Lamblin 提交于 6月 10, 2010
  
  70965d0d
- add __str__ to Join and GpuJoin · 6b25596e
  由 Frederic Bastien 提交于 6月 10, 2010
  
  6b25596e
- merge · 303885e5
  由 James Bergstra 提交于 6月 10, 2010
  
  303885e5
- cmodule - reversed order of includes to make a nonstandard numpy installation work as well · e35a57c3
  由 James Bergstra 提交于 6月 10, 2010
  
  e35a57c3
- optimize Alloc to GpuAlloc more often. · c6d5ee22
  由 Frederic Bastien 提交于 6月 10, 2010
  
  c6d5ee22
- added GpuAlloc and opt to put them into the graph and test. · 3c51a2f1
  由 Frederic Bastien 提交于 6月 10, 2010
  
  3c51a2f1
- pfunc - fix to helper that guessed at which node is responsible for a wrong dtype in an update · a163846d
  由 James Bergstra 提交于 6月 10, 2010
  
  a163846d
- merge · e6d987fd
  由 James Bergstra 提交于 6月 10, 2010
  
  e6d987fd
- cuda_ndarray removed buggy EqualAndIgnore from setitem · b74f3871
  由 James Bergstra 提交于 6月 10, 2010
  
  b74f3871
- better printing in fprint_CudaNdarray · bbfdaab7
  由 Frederic Bastien 提交于 6月 10, 2010
  
  bbfdaab7
- small optimization. · 5b5097e2
  由 Frederic Bastien 提交于 6月 10, 2010
  
  5b5097e2
- make test use less memory · bf5bd3d5
  由 Frederic Bastien 提交于 6月 10, 2010
  
  bf5bd3d5
- more test for CudaNdarray setitem. · 98a6fbb9
  由 Frederic Bastien 提交于 6月 10, 2010
  
  98a6fbb9
- added multinomial for MRG · 45c5780f
  由 Simon Lemieux 提交于 6月 10, 2010
  
  45c5780f
10 6月, 2010 7 次提交
- fix to CudaNdarray setitem when we unbroadcast. · 0016c4d6
  由 Frederic Bastien 提交于 6月 09, 2010
  
  0016c4d6
- make mrg test_normal0 don't execute gpu test in fast compile mode. · 274bb747
  由 Frederic Bastien 提交于 6月 09, 2010
  
  274bb747
- make mrg test faster · ab65e9c5
  由 Frederic Bastien 提交于 6月 09, 2010
  
  ab65e9c5
- make rng test on gpu fully on gpu · 7dab9f38
  由 Frederic Bastien 提交于 6月 09, 2010
  
  7dab9f38
- fix mrg(use_cuda=True).binomial(dtype=float32) to work on the gpu event if floatX is not float32 · 4b00bd50
  由 Frederic Bastien 提交于 6月 09, 2010
  
  4b00bd50
- added assert. · a51132f6
  由 Frederic Bastien 提交于 6月 09, 2010
  
  a51132f6
- small fix · a7074259
  由 Frederic Bastien 提交于 6月 09, 2010
  
  a7074259