- 10 3月, 2016 6 次提交
-
-
由 Frédéric Bastien 提交于
Kmap
-
由 Frédéric Bastien 提交于
Fix typo in docs
-
由 Wojciech Głogowski 提交于
-
由 Pascal Lamblin 提交于
Impossible to compile a COp
-
由 Samira Shabanian 提交于
-
由 Samira Shabanian 提交于
-
- 09 3月, 2016 13 次提交
-
-
由 abergeron 提交于
[OPT FAILURE] Fix gh-4131
-
由 Francesco 提交于
flake8 misc/tests/*.py
-
-
由 Frédéric Bastien 提交于
Make the python code for Gemv also check for the need to initialize the output with beta=0
-
由 Frédéric Bastien 提交于
Skip the gpuarray tests is init_gpu_device == gpu*
-
由 Taesup (TS) Kim 提交于
-
由 Xavier Bouthillier 提交于
fixed rng_mrg int32 overflow, just throw out the error when it overflows
-
由 Pascal Lamblin 提交于
Fix typo and clarify CNMeM config documentation
-
由 Arnaud Bergeron 提交于
-
由 abergeron 提交于
Make sure imshp and kshp are constant
-
由 abergeron 提交于
GpuBatchedDot: streams implementation (WIP)
-
由 Frédéric Bastien 提交于
Fix typo and clarify cuDNN config documentation
-
由 Li Yao 提交于
-
- 08 3月, 2016 21 次提交
-
-
由 Li Yao 提交于
comments in test pep8 bug fix, int32 to int64, for failure in test_multinomial.py:test_n_samples_compatibility fixed some bugs casting problem in c code fixed fixed more bugs and added into both old and new backend fixed more bugs and added into both old and new backend bug fix pep8 minor fix pep8fix pep8 fix too much memroy?, try something smaller refactored tests refactored tests and pep8 fail on --select=E121
-
由 Frederic Bastien 提交于
-
由 Frédéric Bastien 提交于
Improve code misleading tiny issues for Loop of Tutorial.
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frederic Bastien 提交于
-
由 Frédéric Bastien 提交于
Fix nan handling in output buffer for CGemv
-
由 Frédéric Bastien 提交于
Catch scipy.sigtools._convolve2d complex warning
-
由 Adithya Ganesh 提交于
-
由 Adithya Ganesh 提交于
-
由 Pascal Lamblin 提交于
We are not calling `as_tensor_variable` to avoid transfers when the inputs are on GPU, but the the cuda/numpy arrays should still be converted to Constant variables.
-
由 Pascal Lamblin 提交于
-
由 Pascal Lamblin 提交于
Otherwise, if they are symbolic variables, they could be used during shape inference, which could introduce dependencies on variables outside of the graph.
-
由 Pascal Lamblin 提交于
Force older conda version on travis
-
由 Arnaud Bergeron 提交于
-
由 Arnaud Bergeron 提交于
-