- 01 10月, 2014 1 次提交
-
-
由 Frederic 提交于
This is because if we exclude the 'gpu' tag, we don't want those opt to also be enabled.
-
- 30 9月, 2014 5 次提交
- 27 9月, 2014 13 次提交
- 26 9月, 2014 6 次提交
-
-
由 Frédéric Bastien 提交于
Include and library flags in nvcc.flags in config not passed to nvcc correctly
-
由 Frédéric Bastien 提交于
Fix crash when numpy.distutils.__config__ is present but None
-
由 Yaroslav Bulatov 提交于
-
由 Geoffrey French [UEA] 提交于
Added -I and -L to the list of argument patterns from nvcc.flags that are not presented to nvcc in comma-separated form.
-
由 Arnaud Bergeron 提交于
-
由 Frédéric Bastien 提交于
Implemented cuDNN softmax but it is not currently passing tests.
-
- 25 9月, 2014 4 次提交
-
-
由 Dustin Webb 提交于
-
由 Dustin Webb 提交于
Incremeneted GpuDnnSoftmax version and retested as an extra sanity test of my recent refactoring.
-
由 Dustin Webb 提交于
-
由 abergeron 提交于
[MRG, BUG] Scan Opt fixes, enhencements
-
- 24 9月, 2014 8 次提交
-
-
由 Dustin Webb 提交于
-
由 Dustin Webb 提交于
-
由 Dustin Webb 提交于
-
由 Dustin Webb 提交于
-
由 Dustin Webb 提交于
-
由 Dustin Webb 提交于
Fixed problem with cudnn softmax tests. This fix should be generalized for use of the class. Also there are two cases it cannot handle.
-
由 Dustin Webb 提交于
Fixed eq and hash functions.
-
由 Dustin Webb 提交于
-
- 23 9月, 2014 3 次提交
-
-
由 Frédéric Bastien 提交于
Add support for subsampling and gradient to GpuDnnConv
-
由 Frederic 提交于
-
由 Frederic Bastien 提交于
-