Currently use `Operator("fc", X="x", W='w1', B='b1')` as operator
creation method.

Fix #3198

Rename operator.py to op.py because it is conflict with protobuf

* because it uses `numpy.flip` in `test_image.py`

Initialize Gradient Checker

Save log prob

cpplint code style for *.h and *.cu

Add unittest for GradOpBuilder

Simplify building process of gradient operators

* move net_op to operators

* init softmax grad op

* add compute code

* export Backward to python

* update test ,export op.type to python

* update python test, fix compute bug

* update unit test

* use eigen

* optimize eigen code

* add gpu test

* register softmax_grad GPU kernel and fix test bug

* typo

* follow comments

Move pybind.cc/tensor_bind.h to paddle::framework

* Do not use `virtualenv` in Docker
* Split unittest into many target, make them parallel

Add @EMPTY@ to default no_grad_vars in `Backward`

Fix CPU compile

error fix on the front page

update MKLDNN cmakes