* Implement norm_op by CUDA instead of Eigen.
* Remove the commented code.

* support test using executor without reader

* run imgnet

* update fluid benchmark

* wip

* update

* update all models

* support pyreader

* update

* clean up

* make profile batches contollable

* update API.spec

* update scripts

* clean dockerfile

* update

* clean comments

* add scope argument docstring

* use num_trainers to determine nccl init comms

* add save_inference_model for Trainer

* add comment

* update comment

* fix mac compile

* add unit test

* update API.spec

* revert cpplint-cpp-source

* Add intermediate

* fix flatten/squeeze/unsqueeze

* Considering compatibility issues, we could not fix the origin op

* follow comment

* reset the shape of XShape

* rmsprop optimizer support v1 mode

* typo

* optimize code

* refine code

* optimize unit test

* update test_rmsprop_op.py

* update formula of rmsprop

* optimize document

* update API.spec for RMSPropOptimizer

* add default value to check_output_with_place equal_nan

* reimplement auc in pyton

* reimplement auc in fluid

* add auc unittest


* replace new auc in layers


* add batch Auc in Fluid

* name formated

* Refine fusion lstm op code for better readability

* Enable peephole in fusion lstm op (seq_mode part) and add unit test

* Enable peephole in fused lstop op (batch_mode part)

Set batch_mode as default as well

* Use pre-commit to clean format

* Follow up review comments as well as adding more unit tests for seq mode