Add default GpuKernelBase.get_params() that retrieves GPU context
from first node input type, and then factorize get_params() for all
sub-classes that retrieve context by this way.

Method get_params() must be overridden for all subclasses that get
context by different way.

Extend GpuKernelBase interface to allow sub-classes
to use parameters types other than gpu_context_type
(especially ParamsType).

Remove bug in optimization

[BUG, CRASH] Fixes in DebugMode for GPU

The addition gave non-sensical results.

It also fixes a race condition.

CPU wraps around, but GPU returns max_int16, so DebugMode was
complaining.

Pass the right variable when not working inplace.

The corresponding CPU test does the same

[BUG,CRASH] Fix crash in GPU advanced indexing with no arrays

minor fix to test_nnet.py

Add sigmoid_binary_crossentropy function

Add magma gpu svd and matrix inverse

Update libgpuarray version in the doc