* fix some enforce

* remove compatible_type to avoid compile error

* remove shared_ptr

* fix tensor error msg

* check INFINITY in cross_entropy

* fix error

* use onehot_cross_entropy without GPU kernel

* add support_gpu

* fix allclose

* fix name error and symplify code

* add support_gpu

* fix allclose

* fix name error and symplify code

* put everything in paddle wheel

* update

* update

* fix unitest

* with platform specs

* finish forward debug

* init cross entropy graident

* add crossentropy grad op

* remove details

* fix static compile

* move namespace rnn to a directory

* User must install Paddle python package before unittest.
* Or use docker to build Paddle