* documents will be ready afterwards

not ready now

* Update index_en.rst

* Update index_cn.rst

* Update index_en.rst

* Update index_cn.rst

* Update index_en.rst

* fleet api

* fleet api

* fleet api

* refine inference api

test=develop

* fix typo

* fix typo

* fix comments

* fix typo

* update faq by sheqi

* fix format errors

* renew links; fix format issues

* improve the doc of metrics introduction

* tweak the doc of metrics introduction

* test=develop, update install doc

* add setuptools dependency

* Update compile_Ubuntu.md

update cudnn version

* Update install_Ubuntu.md

update cudnn version

* Update install_CentOS.md

update cudnn version

* test=develop, refine ubuntu cuda support

* test=develop, refine ubuntu cuda support

* test=develop, refine python support on python3.5

* fix python2 in ubuntu16.04 need gcc 5.4

* add nccl related doc

* refine doc

* add gpu fast install doc

* refine command of pip

* add avx off for mac complie

* refine code

* refine doc

* update memory_optimize.rst, test=develop

* follow comment, fix typo, test=develop

* Create quick_start.rst

* Update quick_start.rst

* test=develop, update install doc

* add setuptools dependency

* Update compile_Ubuntu.md

update cudnn version

* Update install_Ubuntu.md

update cudnn version

* Update install_CentOS.md

update cudnn version

* test=develop, refine ubuntu cuda support

* test=develop, refine ubuntu cuda support

* test=develop, refine python support on python3.5

* fix python2 in ubuntu16.04 need gcc 5.4

* add nccl related doc

* refine doc

* add gpu fast install doc

* refine command of pip

* add avx off for mac complie

* Modify errors in VisualDL usage manual
Including:
	line break errors;
	punctuation errors;
	Add explanation for each pictures.

* Modify errors in VisualDL usage manual
Both Chinese and English Documents are modified.
Including :
	line break errors.
	spell errors

* Modify the line break error in visualdl_usage_en
The Line break errors in visualdl_usage_en will make the demo program crash into errors.
I have fixed these errors.

* Modify the line break error in visualdl_usage_en.md
The errors have been fixed.

* test=develop, update install doc

* add setuptools dependency

* Update compile_Ubuntu.md

update cudnn version

* Update install_Ubuntu.md

update cudnn version

* Update install_CentOS.md

update cudnn version

* test=develop, refine ubuntu cuda support

* test=develop, refine ubuntu cuda support

* test=develop, refine python support on python3.5

* fix python2 in ubuntu16.04 need gcc 5.4

* add nccl related doc

* refine doc

* add gpu fast install doc

* refine command of pip

* Add fluid benchmark document

Compare some models' performance as reference for users.
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Refine doc.
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Refine document
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* Refine doc
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* test=develop, update install doc

* add setuptools dependency

* Update compile_Ubuntu.md

update cudnn version

* Update install_Ubuntu.md

update cudnn version

* Update install_CentOS.md

update cudnn version

* test=develop, refine ubuntu cuda support

* test=develop, refine ubuntu cuda support

* test=develop, refine python support on python3.5

* fix python2 in ubuntu16.04 need gcc 5.4

* add nccl related doc

* refine doc

* add gpu fast install doc

* Modify errors in VisualDL usage manual
Including:
	line break errors;
	punctuation errors;
	Add explanation for each pictures.

* Modify errors in VisualDL usage manual
Both Chinese and English Documents are modified.
Including :
	line break errors.
	spell errors

* update compile_CentOS.md
* update user guide to install virtualenv

* update compile_Ubuntu.md
* update user guide to install virtualenv

* test=develop, update install doc

* add setuptools dependency

* Update compile_Ubuntu.md

update cudnn version

* Update install_Ubuntu.md

update cudnn version

* Update install_CentOS.md

update cudnn version

* test=develop, refine ubuntu cuda support

* test=develop, refine ubuntu cuda support

* test=develop, refine python support on python3.5

* fix python2 in ubuntu16.04 need gcc 5.4

* add nccl related doc

* refine doc

* update dygraph user guide to fit new api

* refine doc

* remove pylayer support doc

* refine words

* refine sample code

* Add usage manual for VisualDL
The usage manual will be helpful for beginners to make use of VisualDL.

Components of VisualDL are introduced in details, including:
	* Definition of Functions
	* Parameters explanation
	* Demo program for each component

* add_index

* refine inference api

test=develop

* fix typo

* fix typo

* fix comments

* modify write_docs_cn, write_docs_en, include specifications about python 2.7.15 and creating  virtual env with conda

* gitignore update

* revert .gitignore;  update write_docs

* icafe DLTP-1328, stress  exe.close()

* DLTP-1383 不可迭代更正

* DLTP-1404 修复动态图文档中每段代码的格式和大小都不一样

* DLTP-1390 Pyreader中文显示格式问题

* DLTP-1393 无效链接

* DLTP-1421 paddle.fluid.layers.reduce_max中文翻译有问题

* DyGraph 调整标签结构

* add .swp file

* sync from Book; GoogLeNet

* TRT预测库仅支持GPU编译

DLTP-739 完善tensorRT使用文档 TRT需要GPU编译

* DLTP-965 预测模型的保存和加载文档优化

* Windows_installation_English

* DLTP-1417 受训改为训练

* modify write_docs_cn, write_docs_en, include specifications about python 2.7.15 and creating  virtual env with conda

* gitignore update

* revert .gitignore;  update write_docs

* icafe DLTP-1328, stress  exe.close()

* DLTP-1383 不可迭代更正

* DLTP-1404 修复动态图文档中每段代码的格式和大小都不一样

* DLTP-1390 Pyreader中文显示格式问题

* DLTP-1393 无效链接

* DLTP-1421 paddle.fluid.layers.reduce_max中文翻译有问题

* DyGraph 调整标签结构

* add flag files

* add flags file

* delete space

* modify write_docs_cn, write_docs_en, include specifications about python 2.7.15 and creating  virtual env with conda

* gitignore update

* revert .gitignore;  update write_docs

* add python inference api doc

* Polish executor doc

* routine_commmitid_update0616

* routine_commitid_update0616

* split layers,test=develop

* split nn layers

* remove nn header, test=develop