推荐系统中文文档

improve demo/mnist dataProvider speed

[Sparse Bug]  Test and sparse_remote_update can not co-exsit, crash trainer if necessary

Check the input size for convolution.

Change-Id: I2b58c8d854aa31096a6b6e49c1c120f7acec622b

add python api_predict for quick start

Remove do_average_in_cpu, max_average_window, average_window from optimizers.py

Add an extra parameter label to chunk_evaluator.

update chinese catalog

Make Travis-CI pre-commit check more readable.

add contribute_to_paddle_cn.md

add chinese doc for gpu-profiling

also, fix another code reference error in en doc.

resolve #834

In order to keep consistent with other evaluators, an extra parameter
label is add to chunk_evaluator. Document and demos are modified
accordingly.

These three parameters have already been moved to model_average. Leaving
them here will cause duplicate keys in kwargs

translation about to chinese by jiangfeng

add chinese doc for gpu-profiling, and fix the out-dated lineNos in en doc referenced to a code file.

resolve #834

fix a index bug

Remove custom glog-like and gflags-like macros

fix a few bugs related to frame size.