[OPT FAILURE] Fix gh-4131

flake8 misc/tests/*.py

Make the python code for Gemv also check for the need to initialize the output with beta=0

Skip the gpuarray tests is init_gpu_device == gpu*

fixed rng_mrg int32 overflow, just throw out the error when it overflows

Fix typo and clarify CNMeM config documentation

Make sure imshp and kshp are constant

GpuBatchedDot: streams implementation (WIP)

Fix typo and clarify cuDNN config documentation

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

Improve code misleading tiny issues for Loop of Tutorial.

Fix nan handling in output buffer for CGemv

Catch scipy.sigtools._convolve2d complex warning

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.

Otherwise, if they are symbolic variables, they could be
used during shape inference, which could introduce dependencies
on variables outside of the graph.

Force older conda version on travis

Make BatchedDot don't try to use c code if there is no blas.