* follow comments to fix bugs

Remove main function in some unittest.

Add analystic scripts to documentation

* reduce trainer count for unit test on MAC OSX

* Fix lots of trainer_config_helpers bug, and complete unittest for `layers.py`

* Correctly handling multiple inputs and integer inputs for recurrent_group

* Fix ScatterAgentLayer for generation

* Revert sequence_(nest)_rnn.conf

* Add `device` parameter to ExtraAttr in trainer_config_helpers.
* add unittest for it.

* add type compatible check for ParamAttr

Local training with "sparse_update = True" parameter triggers kSgdSparseCpuTraining mode, fix bugs under it.

fix cudnn conv bug which occurs in image classfication demo in GTX GPU

* Also refine unittest to multiple iteration to prevent luckily random number.

* fix bug in dotmul_operator's api and anotation
* update rnn document
* remove redundant info of projection and operator in layers.py

* it makes unit test failed.

* using map to replace unordered_map on Mac

* remove unnecessary field set in ParameterConfig, Evaluators, etc

update beam_search and seqToseq config, and add ExpActivation api