* upload dygraph seq2seq for v1.6.2

* update README

* upgrade to version 1.7

* remove unused line

remove unused line

* fix base_output dirname

* update run.sh to be the same as static graph

* update run.sh to be the same as static graph

* update Readme

* update Readme

* update README.md

change version number to 1.7

* update transformer support remove build once; test=develop

* fix optimizer; test=develop

* add new model, bmn dygraph model for video temporal proposal generation

* matching new API, decoupling model and  train module

* fix train reader usage error

* fix cpu num set error

* fixed eval/test func

* change pred/eval to 1 thread or 1 gpu card
Co-authored-by: NChen Weihang <sunny_cwh@163.com>

cleanup readme

* EMNLP2019-MAL

* EMNLP2019-MAL, update README.md

* EMNLP2019-MAL, update README.md

* EMNLP2019-MAL, update license

test=develop

* Optimizer init with parameters, test=develop

* fix bug for eval, test=develop

* update ptb to support remove build once; test=develop

* update orc; test=develop

* polish code, test=develop
Co-authored-by: Nzhongpu <2013000149@qq.com>

* add compress ce

* add ce for word2vec

* fix mnist multi gpu save error; test=develop

* fix inference error; test=develop

* only run inference on rankid == 0; test=develop

* fix elementwise

* fix_pix2pix

* fix reader

* fix se_resnet model for Optimizer and Linear upgrade, test=develop

* polish code style for se_resnet, test=develop

test=develop

* deepfm dygraph

* deepfm dygraph fix readme

* adapt dygraph dev api and fix readme

* fix eval and readme

* deepfm dygraph add infer.py, format train output

* fix bug of last commit

* add more log in train and infer. remove is_sparse

* update build_once for cycle_gan
test=develop

* add input_channel arg
 for cycle_gan
test=develop