add cluster train for quick_start

Rename math.py to layer_math.py

* Fix #903

fix bug: gradient_clipping_threshold should be allowed to set with parameter-grain

Gan tutorial

add prefetch and onPassEnd to PaddleApi.h

 * gradient_clipping_threshold should be allowed to set with parameter-grain

Remove bazel from dockerfile and docs

try to connect again if refused error found

Create index_cn.md on image classification

Bug fix for BatchNorm when testing more than models in test mode.

A tiny fix in PyDataProvider2

Remove utils/CommandLineParser.h

Change float to real in NormLayer.h

按照review建议进行了修改

add rnn_cn.md

Fix data provider bug in srl demo

删除了关于train.sh文件路径的说明

Fix typo in PyDataProvider2.py

Add const to GradientMachine::eval

Adding namespace in timing macros

Once encoutering a single sample whose size is larger than batch size,
the provider will yield empty batch and terminate the current pass unexpectedly
if can_over_batch_size=False.

priorbox layer for Single Shot Multibox Detection Network

* hidden decorator kwargs in DataProvider.__init__
* also add unit test for this.