* Modify optimizer in new API to support more usecase

* Modify CMake to include only modified examples

* initial commit

* modify rnn_encoder_docoder example

* modify old machine translation

* modify new_api machine translation

* add create lodtensor from list

* modify book example

* Update label role implementation to use new API

* Try trainable embedding layer

* Fix feed order

* Add infer call

* add lod_tensor utility python module

* add lod_tensor test code

* add more lod tensor tests

* modify word2vec example code using new api

* add comment