提交 · b2a1e9011e8a9bce0b4e16f582eb3937b418a85e · testgroup / pytensor

03 12月, 2014 1 次提交
- Fix typo. · b2a1e901
  由 Arnaud Bergeron 提交于 12月 02, 2014
  
  b2a1e901
02 12月, 2014 3 次提交
- Add mention of compile.* values in the config doc. · 2e01a29b
  由 Arnaud Bergeron 提交于 12月 01, 2014
  
  2e01a29b
- Set the compile.timeout default relative to compile.wait · 715353c4
  由 Arnaud Bergeron 提交于 12月 01, 2014
  
  715353c4
- Update documentation based on review · 0398f4fc
  由 Arnaud Bergeron 提交于 12月 01, 2014
  
  0398f4fc
29 11月, 2014 9 次提交
- Add configuration variables for the compile lock times. · acf5e176
  由 Arnaud Bergeron 提交于 11月 28, 2014
  
  acf5e176
- Try to use the full compiler path instead of a relative one. · cc703505
  由 Arnaud Bergeron 提交于 11月 28, 2014
  
  cc703505
- Merge pull request #2307 from nouiz/dnn_pool · 604d4d2b
  由 Frédéric Bastien 提交于 11月 28, 2014
```
CRASH/OPT fix (quick to review)
```
  604d4d2b
- Merge pull request #2308 from dmitriy-serdyuk/none_checks · f5d4ed30
  由 Frédéric Bastien 提交于 11月 28, 2014
```
Removed comparisons with None
```
  f5d4ed30
- Fixed typo in comment · 0d99dd17
  由 serdyuk 提交于 11月 28, 2014
  
  0d99dd17
- small fix following code review · 1dce9aeb
  由 Frederic 提交于 11月 28, 2014
  
  1dce9aeb
- Better hack to skip broadcast check on the GPU. · 092835c1
  由 Frederic 提交于 11月 28, 2014
  
  092835c1
- Make all dnn op do not check the broadcast flags. · 4c326058
  由 Frederic 提交于 11月 28, 2014
  
  4c326058
- Removed comparisons with None · 08bd312a
  由 serdyuk 提交于 11月 28, 2014
  
  08bd312a
28 11月, 2014 6 次提交
- Fix test following fix to dnn pool opt · 574753c0
  由 Frederic 提交于 11月 28, 2014
  
  574753c0
- Fix crash in python 3, due to difference in comprenhension list. · 5f92bf25
  由 Frederic 提交于 11月 28, 2014
  
  5f92bf25
- Enable local_pool_dnn opt. It was disabled by mistake in a rebase. · b8fe4eb6
  由 Frederic 提交于 11月 28, 2014
  
  b8fe4eb6
- Merge pull request #2298 from nouiz/convdnn_broadcast · 5c9e0b31
  由 abergeron 提交于 11月 27, 2014
```
Fix broadcast pattern of dnn conv op.
```
  5c9e0b31
- Merge pull request #2305 from nouiz/revert · ad4b5528
  由 abergeron 提交于 11月 27, 2014
```
Revert change that break too many test in floatX=float32
```
  ad4b5528
- Revert change that break too many test in floatX=float32 · d81faed2
  由 Frederic 提交于 11月 27, 2014
  
  d81faed2
27 11月, 2014 15 次提交
- remove warning from code check tool about hiding name · ac315f60
  由 Frederic 提交于 11月 27, 2014
  
  ac315f60
- [CRASH] we where returning a python int while this was reused in… · ec787896
  由 Frederic 提交于 11月 27, 2014
```
[CRASH] we where returning a python int while this was reused in Elemwise.perform that expect ndarray.
```
  ec787896
- lower the number of GPU op compiled · 567b2db8
  由 Frederic 提交于 11月 27, 2014
  
  567b2db8
- Merge pull request #2304 from 317070/patch-1 · 5bf33866
  由 Frédéric Bastien 提交于 11月 27, 2014
```
language fix
```
  5bf33866
- language fix · 81063980
  由 Jonas Degrave 提交于 11月 27, 2014
  
  81063980
- Merge pull request #2303 from abergeron/fix_scan_args_cond · dd9a8860
  由 Frédéric Bastien 提交于 11月 27, 2014
```
Fix scan_args for the 'as_while' case properly.
```
  dd9a8860
- Merge pull request #2294 from nouiz/mixed · 3ff7e92b
  由 abergeron 提交于 11月 27, 2014
```
Fix crashes related to sparse
```
  3ff7e92b
- Test dnn conv outputs broadcastable information. · 22471e39
  由 Frederic 提交于 11月 26, 2014
  
  22471e39
- crash fix to the new tests that will be in the next commit. · 75405be7
  由 Frederic 提交于 11月 26, 2014
```
Make subtensor.make_node infer more broadcastable. This remove crash from local_subtensor_merge that was loosing broadcastable dimensions.
```
  75405be7
- Merge pull request #2302 from abergeron/try_flags_cleanup · f392dbf4
  由 Frédéric Bastien 提交于 11月 26, 2014
```
Make sure we remove the exe_path even if try_run is False in GCC_compiler.try_compile_tmp().
```
  f392dbf4
- Fix scan_args for the 'as_while' case properly. · 157ad811
  由 Arnaud Bergeron 提交于 11月 26, 2014
  
  157ad811
- Make sure we remove the exe_path even if try_run is False. · 70bf5e3a
  由 Arnaud Bergeron 提交于 11月 26, 2014
  
  70bf5e3a
- Generate less GpuContiguous implementation · 2d73c170
  由 Frederic 提交于 11月 26, 2014
  
  2d73c170
- remove import not used · 0592f247
  由 Frederic 提交于 11月 26, 2014
  
  0592f247
- hack to lower the number of compilation on the GPU. · 55a1fe82
  由 Frederic 提交于 11月 26, 2014
  
  55a1fe82
26 11月, 2014 3 次提交
- Merge pull request #2275 from carriepl/TypedList_example · f5c05820
  由 Frédéric Bastien 提交于 11月 26, 2014
```
Fixed bug with TypedList example
```
  f5c05820
- Fix broadcast pattern of dnn conv op. · 01187ffe
  由 Frederic 提交于 11月 25, 2014
  
  01187ffe
- Merge pull request #2297 from f0k/metaopt-again · aef67878
  由 Frédéric Bastien 提交于 11月 25, 2014
```
conv2d meta-optimizer missing commits
```
  aef67878
25 11月, 2014 3 次提交
- Documented convolution meta-optimizer in doc/library/tensor/nnet/conv.txt · 30e8007d
  由 f0k 提交于 11月 24, 2014
  
  30e8007d
- Removed output borrowing from meta-optimization function compilation · 07dd0f15
  由 f0k 提交于 11月 24, 2014
  
  07dd0f15
- Rename config flag conv_meta.verbose to metaopt.verbose · 6798e88c
  由 f0k 提交于 11月 24, 2014
  
  6798e88c