- 28 11月, 2014 4 次提交
- 27 11月, 2014 15 次提交
-
-
由 Frederic 提交于
-
由 Frederic 提交于
[CRASH] we where returning a python int while this was reused in Elemwise.perform that expect ndarray.
-
由 Frederic 提交于
-
由 Frédéric Bastien 提交于
language fix
-
由 Jonas Degrave 提交于
-
由 Frédéric Bastien 提交于
Fix scan_args for the 'as_while' case properly.
-
由 abergeron 提交于
Fix crashes related to sparse
-
由 Frederic 提交于
-
由 Frederic 提交于
Make subtensor.make_node infer more broadcastable. This remove crash from local_subtensor_merge that was loosing broadcastable dimensions.
-
由 Frédéric Bastien 提交于
Make sure we remove the exe_path even if try_run is False in GCC_compiler.try_compile_tmp().
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
-
由 Frederic 提交于
-
由 Frederic 提交于
-
由 Frederic 提交于
-
- 26 11月, 2014 3 次提交
-
-
由 Frédéric Bastien 提交于
Fixed bug with TypedList example
-
由 Frederic 提交于
-
由 Frédéric Bastien 提交于
conv2d meta-optimizer missing commits
-
- 25 11月, 2014 18 次提交
-
-
由 f0k 提交于
-
由 f0k 提交于
-
由 f0k 提交于
-
由 Frederic 提交于
-
由 Frederic 提交于
Do not enforce that if the input is sparse, we need the grad to return sparse.
-
由 Frederic 提交于
I do not force the grad to be the same format. This would require conversion. I do not think we always enforce this already, so I do not want to enforce that now.
-
由 Frederic 提交于
Also make theano.grad test the output is good.
-
由 Pascal Lamblin 提交于
Turn off test_value when checking for beta init.
-
由 Frédéric Bastien 提交于
conv2d meta-optimizer
-
由 Frédéric Bastien 提交于
Use clang++ on mac (if present).
-
由 Arnaud Bergeron 提交于
-
由 f0k 提交于
-
由 f0k 提交于
-
由 f0k 提交于
-
由 f0k 提交于
-
由 abergeron 提交于
Enable again dnn pool grad after fixed it.
-
由 Frederic 提交于
bug: The optimization where converting the ignore_border=False to ignore_border=True. crash: The optimization that convert the gpudownsample grad to cuddpoolgrad was swapping the inputs parameter. This is caused in part by cudnn using strange definition of input and output. Now the op use the same inputs order as the other pool grad ops. The only difference is tha the c_code() use the name that cudnn use. There is a comment that tell that.
-
由 abergeron 提交于
Add grad() to GpuDnnConvGrad* ops
-