update detection_map

a606775f · wanghaox · 26f03ea1 · db8da025 · a606775f · a606775f
1000 changed file
--- a/.copyright.hook
+++ b/.copyright.hook
+from __future__ import absolute_import
+from __future__ import print_function
+from __future__ import unicode_literals
+
+import argparse
+import io, re
+import sys, os
+import subprocess
+import platform
+
+COPYRIGHT = '''
+  Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+'''
+
+LANG_COMMENT_MARK = None
+
+NEW_LINE_MARK = None
+
+COPYRIGHT_HEADER = None
+
+if platform.system() == "Windows":
+    NEW_LINE_MARK = "\r\n"
+else:
+    NEW_LINE_MARK = '\n'
+    COPYRIGHT_HEADER = COPYRIGHT.split(NEW_LINE_MARK)[1]
+    p = re.search('(\d{4})', COPYRIGHT_HEADER).group(0)
+    process = subprocess.Popen(["date", "+%Y"], stdout=subprocess.PIPE)
+    date, err = process.communicate()
+    date = date.decode("utf-8").rstrip("\n")
+    COPYRIGHT_HEADER = COPYRIGHT_HEADER.replace(p, date)
+
+
+def generate_copyright(template, lang='C'):
+    if lang == 'Python':
+        LANG_COMMENT_MARK = '#'
+    else:
+        LANG_COMMENT_MARK = "//"
+
+    lines = template.split(NEW_LINE_MARK)
+    BLANK = " "
+    ans = LANG_COMMENT_MARK + BLANK + COPYRIGHT_HEADER + NEW_LINE_MARK
+    for lino, line in enumerate(lines):
+        if lino == 0 or lino == 1 or lino == len(lines) - 1: continue
+        if len(line)  == 0:
+            BLANK = ""
+        else:
+            BLANK = " "
+        ans += LANG_COMMENT_MARK + BLANK + line + NEW_LINE_MARK
+
+    return ans + "\n"
+
+
+def lang_type(filename):
+    if filename.endswith(".py"):
+        return "Python"
+    elif filename.endswith(".h"):
+        return "C"
+    elif filename.endswith(".c"):
+        return "C"
+    elif filename.endswith(".hpp"):
+        return "C"
+    elif filename.endswith(".cc"):
+        return "C"
+    elif filename.endswith(".cpp"):
+        return "C"
+    elif filename.endswith(".cu"):
+        return "C"
+    elif filename.endswith(".cuh"):
+        return "C"
+    elif filename.endswith(".go"):
+        return "C"
+    elif filename.endswith(".proto"):
+        return "C"
+    else:
+        print("Unsupported filetype %s", filename)
+        exit(0)
+
+
+PYTHON_ENCODE = re.compile("^[ \t\v]*#.*?coding[:=][ \t]*([-_.a-zA-Z0-9]+)")
+
+
+def main(argv=None):
+    parser = argparse.ArgumentParser(
+        description='Checker for copyright declaration.')
+    parser.add_argument('filenames', nargs='*', help='Filenames to check')
+    args = parser.parse_args(argv)
+
+    retv = 0
+    for filename in args.filenames:
+        fd = io.open(filename, encoding="utf-8")
+        first_line = fd.readline()
+        second_line = fd.readline()
+        if "COPYRIGHT (C)" in first_line.upper(): continue
+        if first_line.startswith("#!") or PYTHON_ENCODE.match(
+                second_line) != None or PYTHON_ENCODE.match(first_line) != None:
+            continue
+        original_contents = io.open(filename, encoding="utf-8").read()
+        new_contents = generate_copyright(
+            COPYRIGHT, lang_type(filename)) + original_contents
+        print('Auto Insert Copyright Header {}'.format(filename))
+        retv = 1
+        with io.open(filename, 'w') as output_file:
+            output_file.write(new_contents)
+
+    return retv
+
+
+if __name__ == '__main__':
+    exit(main())
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@@ -31,3 +31,11 @@
    -   id: go-fmt
        types:
        - go
+-   repo: local
+    hooks:
+    -   id: copyright_checker
+        name: copyright_checker
+        entry: python ./.copyright.hook
+        language: system
+        files: \.(c|cc|cxx|cpp|cu|h|hpp|hxx|proto|py)$
+        exclude: (?!.*third_party)^.*$ | (?!.*book)^.*$
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -16,14 +16,14 @@ cmake_minimum_required(VERSION 3.0)
 set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_CURRENT_SOURCE_DIR}/cmake")
 set(PADDLE_SOURCE_DIR ${CMAKE_CURRENT_SOURCE_DIR})
 set(PADDLE_BINARY_DIR ${CMAKE_CURRENT_BINARY_DIR})
-SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
-SET(CMAKE_C_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")

 include(system)

 project(paddle CXX C Go)
-message(STATUS "CXX compiler: " ${CMAKE_CXX_COMPILER} ", version: " ${CMAKE_CXX_COMPILER_VERSION})
-message(STATUS "C compiler: " ${CMAKE_C_COMPILER} ", version: " ${CMAKE_C_COMPILER_VERSION})
+message(STATUS "CXX compiler: ${CMAKE_CXX_COMPILER}, version: "
+        "${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION}")
+message(STATUS "C compiler: ${CMAKE_C_COMPILER}, version: "
+        "${CMAKE_C_COMPILER_ID} ${CMAKE_C_COMPILER_VERSION}")

 find_package(Sphinx)
 if(NOT CMAKE_CROSSCOMPILING)
@@ -31,9 +31,6 @@ if(NOT CMAKE_CROSSCOMPILING)
 endif(NOT CMAKE_CROSSCOMPILING)
 find_package(Git REQUIRED)
 find_package(Threads REQUIRED)
-if(NOT ANDROID AND NOT IOS)
-    find_package(Boost QUIET)
-endif()

 include(simd)

@@ -55,6 +52,8 @@ option(WITH_COVERAGE    "Compile PaddlePaddle with code coverage"       OFF)
 option(COVERALLS_UPLOAD "Package code coverage data to coveralls"       OFF)
 option(ON_TRAVIS        "Exclude special unit test on Travis CI"        OFF)
 option(WITH_C_API       "Compile PaddlePaddle with C-API(Prediction)"   OFF)
+# TODO: Only compile PaddlePaddle fluid version by WITH_FLUID option. 
+option(WITH_FLUID       "Compile PaddlePaddle fluid only(TODO)"         ON)
 option(WITH_GOLANG      "Compile PaddlePaddle with GOLANG"              OFF)
 option(GLIDE_INSTALL    "Download and install go dependencies "         ON)
 option(USE_NNPACK       "Compile PaddlePaddle with NNPACK library"      OFF)
@@ -107,6 +106,10 @@ if (WITH_C_API AND WITH_PYTHON)
    "different Python interpreter from compiling.")
 endif()

+if (WITH_C_API)
+  set(WITH_FLUID OFF CACHE STRING "Disable install fluid when compile the C_API" FORCE)
+endif()
+
 if(MOBILE_INFERENCE)
    set(THIRD_PARTY_BUILD_TYPE MinSizeRel)
 else()
@@ -134,6 +137,7 @@ include(external/openblas)  # download, build, install openblas
 include(external/mkldnn)    # download, build, install mkldnn
 include(external/swig)      # download, build, install swig
 include(external/warpctc)   # download, build, install warpctc
+include(external/boost)     # download, build, install boost
 include(external/any)       # download libn::any
 include(external/eigen)     # download eigen3
 include(external/pybind11)  # download pybind11
@@ -158,7 +162,6 @@ include_directories("${PADDLE_SOURCE_DIR}")
 include_directories("${PADDLE_SOURCE_DIR}/paddle/cuda/include")
 include_directories("${CMAKE_CURRENT_BINARY_DIR}/proto")
 include_directories("${CMAKE_CURRENT_BINARY_DIR}/go/pserver/client/c")
-include_directories(${Boost_INCLUDE_DIRS})

 set(EXTERNAL_LIBS
    ${GFLAGS_LIBRARIES}
@@ -201,6 +204,10 @@ if(WITH_GOLANG)
 endif(WITH_GOLANG)

 set(PADDLE_PYTHON_BUILD_DIR "${CMAKE_CURRENT_BINARY_DIR}/python/build")
+
+SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
+SET(CMAKE_C_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
+
 add_subdirectory(paddle)
 if(WITH_PYTHON)
  add_subdirectory(python)

--- a/CODE_OF_CONDUCT.md
+++ b/CODE_OF_CONDUCT.md
+# Contributor Covenant Code of Conduct
+
+## Our Pledge
+
+In the interest of fostering an open and welcoming environment, we as contributors and maintainers pledge to making participation in our project and our community a harassment-free experience for everyone, regardless of age, body size, disability, ethnicity, gender identity and expression, level of experience, nationality, personal appearance, race, religion, or sexual identity and orientation.
+
+## Our Standards
+
+Examples of behavior that contributes to creating a positive environment include:
+
+* Using welcoming and inclusive language
+* Being respectful of differing viewpoints and experiences
+* Gracefully accepting constructive criticism
+* Focusing on what is best for the community
+* Showing empathy towards other community members
+
+Examples of unacceptable behavior by participants include:
+
+* The use of sexualized language or imagery and unwelcome sexual attention or advances
+* Trolling, insulting/derogatory comments, and personal or political attacks
+* Public or private harassment
+* Publishing others' private information, such as a physical or electronic address, without explicit permission
+* Other conduct which could reasonably be considered inappropriate in a professional setting
+
+## Our Responsibilities
+
+Project maintainers are responsible for clarifying the standards of acceptable behavior and are expected to take appropriate and fair corrective action in response to any instances of unacceptable behavior.
+
+Project maintainers have the right and responsibility to remove, edit, or reject comments, commits, code, wiki edits, issues, and other contributions that are not aligned to this Code of Conduct, or to ban temporarily or permanently any contributor for other behaviors that they deem inappropriate, threatening, offensive, or harmful.
+
+## Scope
+
+This Code of Conduct applies both within project spaces and in public spaces when an individual is representing the project or its community. Examples of representing a project or community include using an official project e-mail address, posting via an official social media account, or acting as an appointed representative at an online or offline event. Representation of a project may be further defined and clarified by project maintainers.
+
+## Enforcement
+
+Instances of abusive, harassing, or otherwise unacceptable behavior may be reported by contacting the project team at paddle-dev@baidu.com. The project team will review and investigate all complaints, and will respond in a way that it deems appropriate to the circumstances. The project team is obligated to maintain confidentiality with regard to the reporter of an incident. Further details of specific enforcement policies may be posted separately.
+
+Project maintainers who do not follow or enforce the Code of Conduct in good faith may face temporary or permanent repercussions as determined by other members of the project's leadership.
+
+## Attribution
+
+This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4, available at [http://contributor-covenant.org/version/1/4][version]
+
+[homepage]: http://contributor-covenant.org
+[version]: http://contributor-covenant.org/version/1/4/
--- a/CODE_OF_CONDUCT_cn.md
+++ b/CODE_OF_CONDUCT_cn.md
+# 参与者公约
+
+## 我们的保证
+
+为了促进一个开放透明且友好的环境，我们作为贡献者和维护者保证：无论年龄、种族、民族、性别认同和表达（方式）、体型、身体健全与否、经验水平、国籍、个人表现、宗教或性别取向，参与者在我们项目和社区中都免于骚扰。
+
+## 我们的标准
+
+有助于创造正面环境的行为包括但不限于：
+* 使用友好和包容性语言
+* 尊重不同的观点和经历
+* 耐心地接受建设性批评
+* 关注对社区最有利的事情
+* 友善对待其他社区成员
+
+身为参与者不能接受的行为包括但不限于：
+* 使用与性有关的言语或是图像，以及不受欢迎的性骚扰
+* 捣乱/煽动/造谣的行为或进行侮辱/贬损的评论，人身攻击及政治攻击
+* 公开或私下的骚扰
+* 未经许可地发布他人的个人资料，例如住址或是电子地址
+* 其他可以被合理地认定为不恰当或者违反职业操守的行为
+
+## 我们的责任
+
+项目维护者有责任为「可接受的行为」标准做出诠释，以及对已发生的不被接受的行为采取恰当且公平的纠正措施。
+
+项目维护者有权利及责任去删除、编辑、拒绝与本行为标准有所违背的评论(comments)、提交(commits)、代码、wiki 编辑、问题(issues)和其他贡献，以及项目维护者可暂时或永久性的禁止任何他们认为有不适当、威胁、冒犯、有害行为的贡献者。
+
+## 使用范围
+
+当一个人代表该项目或是其社区时，本行为标准适用于其项目平台和公共平台。
+
+代表项目或是社区的情况，举例来说包括使用官方项目的电子邮件地址、通过官方的社区媒体账号发布或线上或线下事件中担任指定代表。
+
+该项目的呈现方式可由其项目维护者进行进一步的定义及解释。
+
+## 强制执行
+
+可以通过paddle-dev@baidu.com，来联系项目团队来举报滥用、骚扰或其他不被接受的行为。
+
+任何维护团队认为有必要且适合的所有投诉都将进行审查及调查，并做出相对应的回应。项目小组有对事件回报者有保密的义务。具体执行的方针近一步细节可能会单独公布。
+
+没有切实地遵守或是执行本行为标准的项目维护人员，可能会因项目领导人或是其他成员的决定，暂时或是永久地取消其参与资格。
+
+## 来源
+
+本行为标准改编自[贡献者公约][主页]，版本 1.4
+可在此观看https://www.contributor-covenant.org/zh-cn/version/1/4/code-of-conduct.html
+
+[主页]: https://www.contributor-covenant.org
--- a/Dockerfile
+++ b/Dockerfile
@@ -27,7 +27,7 @@ RUN apt-get update && \
    curl sed grep graphviz libjpeg-dev zlib1g-dev  \
    python-matplotlib gcc-4.8 g++-4.8 \
    automake locales clang-format swig doxygen cmake  \
-    liblapack-dev liblapacke-dev libboost-dev \
+    liblapack-dev liblapacke-dev \
    clang-3.8 llvm-3.8 libclang-3.8-dev \
    net-tools libtool && \
    apt-get clean -y

--- a/README.md
+++ b/README.md
@@ -2,8 +2,8 @@


 [![Build Status](https://travis-ci.org/PaddlePaddle/Paddle.svg?branch=develop)](https://travis-ci.org/PaddlePaddle/Paddle)
-[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://doc.paddlepaddle.org/develop/doc/)
-[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://doc.paddlepaddle.org/develop/doc_cn/)
+[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/index_en.html)
+[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/index_cn.html)
 [![Coverage Status](https://coveralls.io/repos/github/PaddlePaddle/Paddle/badge.svg?branch=develop)](https://coveralls.io/github/PaddlePaddle/Paddle?branch=develop)
 [![Release](https://img.shields.io/github/release/PaddlePaddle/Paddle.svg)](https://github.com/PaddlePaddle/Paddle/releases)
 [![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)
@@ -37,6 +37,7 @@ Please refer to our [release announcement](https://github.com/PaddlePaddle/Paddl

      - Optimized math operations through SSE/AVX intrinsics, BLAS libraries
      (e.g. MKL, OpenBLAS, cuBLAS) or customized CPU/GPU kernels.
+      - Optimized CNN networks through MKL-DNN library.
      - Highly optimized recurrent networks which can handle **variable-length**
      sequence without padding.
      - Optimized local and distributed training for models with high dimensional
@@ -61,32 +62,32 @@ Please refer to our [release announcement](https://github.com/PaddlePaddle/Paddl
 ## Installation

 It is recommended to check out the
-[Docker installation guide](http://doc.paddlepaddle.org/develop/doc/getstarted/build_and_install/docker_install_en.html)
+[Docker installation guide](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/docker_install_en.html)
 before looking into the
-[build from source guide](http://doc.paddlepaddle.org/develop/doc/getstarted/build_and_install/build_from_source_en.html).
+[build from source guide](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/build_from_source_en.html).

 ## Documentation

-We provide [English](http://doc.paddlepaddle.org/develop/doc/) and
-[Chinese](http://doc.paddlepaddle.org/doc_cn/) documentation.
+We provide [English](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/index_en.html) and
+[Chinese](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/index_cn.html) documentation.

- [Deep Learning 101](http://book.paddlepaddle.org/index.html)
+- [Deep Learning 101](http://www.paddlepaddle.org/docs/develop/book/01.fit_a_line/index.html)

  You might want to start from this online interactive book that can run in a Jupyter Notebook.

- [Distributed Training](http://doc.paddlepaddle.org/develop/doc/howto/usage/cluster/cluster_train_en.html)
+- [Distributed Training](http://www.paddlepaddle.org/docs/develop/documentation/en/howto/usage/cluster/cluster_train_en.html)

  You can run distributed training jobs on MPI clusters.

- [Distributed Training on Kubernetes](http://doc.paddlepaddle.org/develop/doc/howto/usage/k8s/k8s_en.html)
+- [Distributed Training on Kubernetes](http://www.paddlepaddle.org/docs/develop/documentation/en/howto/usage/cluster/k8s_en.html)

   You can also run distributed training jobs on Kubernetes clusters.

- [Python API](http://doc.paddlepaddle.org/develop/doc/api/index_en.html)
+- [Python API](http://www.paddlepaddle.org/docs/develop/documentation/en/api/index_en.html)

   Our new API enables much shorter programs.

- [How to Contribute](http://doc.paddlepaddle.org/develop/doc/howto/dev/contribute_to_paddle_en.html)
+- [How to Contribute](http://www.paddlepaddle.org/docs/develop/documentation/en/howto/dev/contribute_to_paddle_en.html)

   We appreciate your contributions!


--- a/benchmark/IntelOptimizedPaddle.md
+++ b/benchmark/IntelOptimizedPaddle.md
@@ -7,11 +7,11 @@ Machine:

 System: CentOS release 6.3 (Final), Docker 1.12.1.

-PaddlePaddle: (TODO: will rerun after 0.11.0)
- paddlepaddle/paddle:latest (for MKLML and MKL-DNN)
+PaddlePaddle:
+- paddlepaddle/paddle:0.11.0 (for MKLML and MKL-DNN)
  - MKL-DNN tag v0.11
  - MKLML 2018.0.1.20171007
- paddlepaddle/paddle:latest-openblas (for OpenBLAS)
+- paddlepaddle/paddle:0.11.0-openblas (for OpenBLAS)
  - OpenBLAS v0.2.20
 	 
 On each machine, we will test and compare the performance of training on single node using MKL-DNN / MKLML / OpenBLAS respectively.
@@ -22,6 +22,7 @@ On each machine, we will test and compare the performance of training on single

 #### Training
 Test on batch size 64, 128, 256 on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
+Pay attetion that the speed below includes forward, backward and parameter update time. So we can not directly compare the data with the benchmark of caffe `time` [command](https://github.com/PaddlePaddle/Paddle/blob/develop/benchmark/caffe/image/run.sh#L9), which only contain forward and backward. The updating time of parameter would become very heavy when the weight size are large, especially on alexnet.

 Input image size - 3 * 224 * 224, Time: images/second

@@ -55,33 +56,57 @@ Input image size - 3 * 224 * 224, Time: images/second

 <img src="figs/googlenet-cpu-train.png" width="500">

+- AlexNet
+
+| BatchSize    | 64     | 128    | 256    |
+|--------------|--------| ------ | -------|
+| OpenBLAS     | 45.62  | 72.79  | 107.22 | 
+| MKLML        | 66.37  | 105.60 | 144.04 |
+| MKL-DNN      | 399.00 | 498.94 | 626.53 | 
+
+<img src="figs/alexnet-cpu-train.png" width="500">
+
 #### Inference
 Test on batch size 1, 2, 4, 8, 16 on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
 - VGG-19

 | BatchSize | 1     | 2     | 4     | 8     | 16    |
 |-----------|-------|-------|-------|-------|-------|
-| OpenBLAS  | 1.07  | 1.08  | 1.06  | 0.88  | 0.65  |
+| OpenBLAS  | 1.10  | 1.96  | 3.62  | 3.63  | 2.25  |
 | MKLML     | 5.58  | 9.80  | 15.15 | 21.21 | 28.67 |
 | MKL-DNN   | 75.07 | 88.64 | 82.58 | 92.29 | 96.75 |

+<img src="figs/vgg-cpu-infer.png" width="500">
+
 - ResNet-50

 | BatchSize | 1     | 2      | 4      | 8      | 16     |
 |-----------|-------|--------|--------|--------|--------|
-| OpenBLAS  | 3.35  | 3.19   | 3.09   | 2.55   | 1.96   |
+| OpenBLAS  | 3.31  | 6.72   | 11.59  | 13.17  | 9.27   |
 | MKLML     | 6.33  | 12.02  | 22.88  | 40.53  | 63.09  |
 | MKL-DNN   | 107.83| 148.84 | 177.78 | 189.35 | 217.69 |

+<img src="figs/resnet-cpu-infer.png" width="500">

 - GoogLeNet

 | BatchSize | 1      | 2      | 4      | 8      | 16     |
 |-----------|--------|--------|--------|--------|--------|
-| OpenBLAS  | 12.04  | 11.31  | 10.00  | 9.07   | 4.34   |
+| OpenBLAS  | 12.06  | 23.56  | 34.48  | 36.45  | 23.12  |
 | MKLML     | 22.74  | 41.56  | 81.22  | 133.47 | 210.53 |
 | MKL-DNN   | 175.10 | 272.92 | 450.70 | 512.00 | 600.94 |

+<img src="figs/googlenet-cpu-infer.png" width="500">
+
+- AlexNet
+
+| BatchSize | 1      | 2      | 4      | 8      | 16     |
+|-----------|--------|--------|--------|--------|--------|
+| OpenBLAS  | 3.53   | 6.23   | 15.04  | 26.06  | 31.62  |
+| MKLML     | 21.32  | 36.55  | 73.06  | 131.15 | 192.77 |
+| MKL-DNN   | 442.91 | 656.41 | 719.10 | 847.68 | 850.51 |
+
+<img src="figs/alexnet-cpu-infer.png" width="500">

 ### Laptop
 TBD
--- a/benchmark/cluster/README.md
+++ b/benchmark/cluster/README.md
+# Cluster Training Benchmark
+
+## Setup
+
+- Platform
+  - Kubernetes: v1.6.2
+  - Linux Kernel: v3.10.0
+
+- Resource
+  - CPU: 10 Cores per Pod
+  - Memory: 5GB per Pod
+
+- Docker Image
+
+  We use different base Docker Image to run the benchmark on Kubernetes:
+  - PaddlePaddle v2: paddlepaddle/paddle:0.11.0
+  - PaddlePaddle Fluid: paddlepaddle/paddle:[commit-id]
+  - TensorFlow: tensorflow/tensorflow:1.5.0-rc0
+
+- Model
+  vgg16 is used in this benchmark.
+
+## Cases
+
+- Variable
+  - Batch Size of training data.
+  - PServer count of the training job.
+  - The number of trainers.
+
+- Invariant
+  - The resource of trainer/pserver Pod.
+
+### Measure the Performance for Different Batch Size
+
+- PServer Count: 40
+- Trainer Count: 100
+- Metrics: mini-batch / sec
+
+| Batch Size | 32 | 64 | 128 | 256 |
+| -- | -- | -- | -- | -- |
+| PaddlePaddle Fluid | - | - | - | - |
+| PaddlePaddle v2 | - | - | - | - |
+| TensorFlow | - | - | - | - |
+
+### Measure the Performance for Different PServer Count
+
+- Trainer Count: 100
+- Batch Size: 64
+- Metrics: mini-batch / sec
+
+| PServer Count | 10 | 20 | 40 | 60 |
+| -- | -- | -- | -- | -- |
+| PaddlePaddle Fluid | - | - | - | - |
+| PaddlePaddle v2 | - | - | - | - |
+| TensorFlow | - | - | - | - |
+
+### Measure Parallel Efficiency By Increasing Trainer Count
+
+- PServer Count: 20
+- Batch Size: 64
+- Metrics:
+
+$S = \div(T1, TN)$
+
+which S is the ratio of T1 over TN, training time of 1 and N trainers.
+The parallel efficiency is:
+
+$E = \div(S, N)$
+
+| Trainer Counter | 1 | 10 | 20 | 30 | 40 | 50 | 60 | 70 | 80 | 90 | 100 |
+| -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
+| PaddlePaddle Fluid | - | - | - | - | - | - | - | - | - | - | - |
+| PaddlePaddle v2 | - | - | - | - | - | - | - | - | - | - | - | - |
+| TensorFlow | - | - | - | - | - | - | - | - | - | - | - | - | - |
+
+## Reproduce the benchmark
+
+TODO
--- a/benchmark/figs/alexnet-cpu-infer.png
+++ b/benchmark/figs/alexnet-cpu-infer.png
--- a/benchmark/figs/alexnet-cpu-train.png
+++ b/benchmark/figs/alexnet-cpu-train.png
--- a/benchmark/figs/googlenet-cpu-infer.png
+++ b/benchmark/figs/googlenet-cpu-infer.png
--- a/benchmark/figs/googlenet-cpu-train.png
+++ b/benchmark/figs/googlenet-cpu-train.png
--- a/benchmark/figs/resnet-cpu-infer.png
+++ b/benchmark/figs/resnet-cpu-infer.png
--- a/benchmark/figs/resnet-cpu-train.png
+++ b/benchmark/figs/resnet-cpu-train.png
--- a/benchmark/figs/vgg-cpu-infer.png
+++ b/benchmark/figs/vgg-cpu-infer.png
--- a/benchmark/figs/vgg-cpu-train.png
+++ b/benchmark/figs/vgg-cpu-train.png
--- a/benchmark/paddle/image/alexnet.py
+++ b/benchmark/paddle/image/alexnet.py
-#!/usr/bin/env python
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.

 from paddle.trainer_config_helpers import *

@@ -6,10 +18,24 @@ height = 227
 width = 227
 num_class = 1000
 batch_size = get_config_arg('batch_size', int, 128)
+gp = get_config_arg('layer_num', int, 1)
+is_infer = get_config_arg("is_infer", bool, False)
+num_samples = get_config_arg('num_samples', int, 2560)

-args = {'height': height, 'width': width, 'color': True, 'num_class': num_class}
+args = {
+    'height': height,
+    'width': width,
+    'color': True,
+    'num_class': num_class,
+    'is_infer': is_infer,
+    'num_samples': num_samples
+}
 define_py_data_sources2(
-    "train.list", None, module="provider", obj="process", args=args)
+    "train.list" if not is_infer else None,
+    "test.list" if is_infer else None,
+    module="provider",
+    obj="process",
+    args=args)

 settings(
    batch_size=batch_size,
@@ -31,7 +57,7 @@ net = img_pool_layer(input=net, pool_size=3, stride=2)

 # conv2
 net = img_conv_layer(
-    input=net, filter_size=5, num_filters=256, stride=1, padding=2, groups=1)
+    input=net, filter_size=5, num_filters=256, stride=1, padding=2, groups=gp)
 net = img_cmrnorm_layer(input=net, size=5, scale=0.0001, power=0.75)
 net = img_pool_layer(input=net, pool_size=3, stride=2)

@@ -40,11 +66,11 @@ net = img_conv_layer(
    input=net, filter_size=3, num_filters=384, stride=1, padding=1)
 # conv4
 net = img_conv_layer(
-    input=net, filter_size=3, num_filters=384, stride=1, padding=1, groups=1)
+    input=net, filter_size=3, num_filters=384, stride=1, padding=1, groups=gp)

 # conv5
 net = img_conv_layer(
-    input=net, filter_size=3, num_filters=256, stride=1, padding=1, groups=1)
+    input=net, filter_size=3, num_filters=256, stride=1, padding=1, groups=gp)
 net = img_pool_layer(input=net, pool_size=3, stride=2)

 net = fc_layer(
@@ -59,6 +85,9 @@ net = fc_layer(
    layer_attr=ExtraAttr(drop_rate=0.5))
 net = fc_layer(input=net, size=1000, act=SoftmaxActivation())

-lab = data_layer('label', num_class)
-loss = cross_entropy(input=net, label=lab)
-outputs(loss)
+if is_infer:
+    outputs(net)
+else:
+    lab = data_layer('label', num_class)
+    loss = cross_entropy(input=net, label=lab)
+    outputs(loss)
--- a/benchmark/paddle/image/googlenet.py
+++ b/benchmark/paddle/image/googlenet.py
@@ -7,13 +7,15 @@ num_class = 1000
 batch_size = get_config_arg('batch_size', int, 128)
 use_gpu = get_config_arg('use_gpu', bool, True)
 is_infer = get_config_arg("is_infer", bool, False)
+num_samples = get_config_arg('num_samples', int, 2560)

 args = {
    'height': height,
    'width': width,
    'color': True,
    'num_class': num_class,
-    'is_infer': is_infer
+    'is_infer': is_infer,
+    'num_samples': num_samples
 }
 define_py_data_sources2(
    "train.list" if not is_infer else None,

--- a/benchmark/paddle/image/plotlog.py
+++ b/benchmark/paddle/image/plotlog.py
+# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import sys
+import argparse
+import matplotlib.pyplot as plt
+
+
+def parse_args():
+    parser = argparse.ArgumentParser('Parse Log')
+    parser.add_argument(
+        '--file_path', '-f', type=str, help='the path of the log file')
+    parser.add_argument(
+        '--sample_rate',
+        '-s',
+        type=float,
+        default=1.0,
+        help='the rate to take samples from log')
+    parser.add_argument(
+        '--log_period', '-p', type=int, default=1, help='the period of log')
+
+    args = parser.parse_args()
+    return args
+
+
+def parse_file(file_name):
+    loss = []
+    error = []
+    with open(file_name) as f:
+        for i, line in enumerate(f):
+            line = line.strip()
+            if not line.startswith('pass'):
+                continue
+            line_split = line.split(' ')
+            if len(line_split) != 5:
+                continue
+
+            loss_str = line_split[2][:-1]
+            cur_loss = float(loss_str.split('=')[-1])
+            loss.append(cur_loss)
+
+            err_str = line_split[3][:-1]
+            cur_err = float(err_str.split('=')[-1])
+            error.append(cur_err)
+
+    accuracy = [1.0 - err for err in error]
+
+    return loss, accuracy
+
+
+def sample(metric, sample_rate):
+    interval = int(1.0 / sample_rate)
+    if interval > len(metric):
+        return metric[:1]
+
+    num = len(metric) / interval
+    idx = [interval * i for i in range(num)]
+    metric_sample = [metric[id] for id in idx]
+    return metric_sample
+
+
+def plot_metric(metric,
+                batch_id,
+                graph_title,
+                line_style='b-',
+                line_label='y',
+                line_num=1):
+    plt.figure()
+    plt.title(graph_title)
+    if line_num == 1:
+        plt.plot(batch_id, metric, line_style, label=line_label)
+    else:
+        for i in range(line_num):
+            plt.plot(batch_id, metric[i], line_style[i], label=line_label[i])
+    plt.xlabel('batch')
+    plt.ylabel(graph_title)
+    plt.legend()
+    plt.savefig(graph_title + '.jpg')
+    plt.close()
+
+
+def main():
+    args = parse_args()
+    assert args.sample_rate > 0. and args.sample_rate <= 1.0, "The sample rate should in the range (0, 1]."
+
+    loss, accuracy = parse_file(args.file_path)
+    batch = [args.log_period * i for i in range(len(loss))]
+
+    batch_sample = sample(batch, args.sample_rate)
+    loss_sample = sample(loss, args.sample_rate)
+    accuracy_sample = sample(accuracy, args.sample_rate)
+
+    plot_metric(loss_sample, batch_sample, 'loss', line_label='loss')
+    plot_metric(
+        accuracy_sample,
+        batch_sample,
+        'accuracy',
+        line_style='g-',
+        line_label='accuracy')
+
+
+if __name__ == '__main__':
+    main()
--- a/benchmark/paddle/image/provider.py
+++ b/benchmark/paddle/image/provider.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import io, os
 import random
 import numpy as np
@@ -14,6 +28,7 @@ def initHook(settings, height, width, color, num_class, **kwargs):
    else:
        settings.data_size = settings.height * settings.width
    settings.is_infer = kwargs.get('is_infer', False)
+    settings.num_samples = kwargs.get('num_samples', 2560)
    if settings.is_infer:
        settings.slots = [dense_vector(settings.data_size)]
    else:
@@ -23,7 +38,7 @@ def initHook(settings, height, width, color, num_class, **kwargs):
 @provider(
    init_hook=initHook, min_pool_size=-1, cache=CacheType.CACHE_PASS_IN_MEM)
 def process(settings, file_list):
-    for i in xrange(2560 if settings.is_infer else 1024):
+    for i in xrange(settings.num_samples):
        img = np.random.rand(1, settings.data_size).reshape(-1, 1).flatten()
        if settings.is_infer:
            yield img.astype('float32')

--- a/benchmark/paddle/image/resnet.py
+++ b/benchmark/paddle/image/resnet.py
@@ -7,13 +7,15 @@ num_class = 1000
 batch_size = get_config_arg('batch_size', int, 64)
 layer_num = get_config_arg("layer_num", int, 50)
 is_infer = get_config_arg("is_infer", bool, False)
+num_samples = get_config_arg('num_samples', int, 2560)

 args = {
    'height': height,
    'width': width,
    'color': True,
    'num_class': num_class,
-    'is_infer': is_infer
+    'is_infer': is_infer,
+    'num_samples': num_samples
 }
 define_py_data_sources2(
    "train.list" if not is_infer else None,

--- a/benchmark/paddle/image/run_mkldnn_infer.sh
+++ b/benchmark/paddle/image/run_mkldnn_infer.sh
@@ -37,7 +37,7 @@ function infer() {
      --trainer_count=1 \
      --num_passes=1 \
      --save_dir="models/${topology}-${layer_num}" \
-      --config_args="batch_size=128,layer_num=${layer_num}" \
+      --config_args="batch_size=128,layer_num=${layer_num},num_samples=256" \
      > /dev/null 2>&1
    echo "Done"
  fi
@@ -79,8 +79,9 @@ fi
 # inference benchmark
 for use_mkldnn in True False; do
  for batchsize in 1 2 4 8 16; do
-    infer googlenet v1 $batchsize $use_mkldnn
-    infer resnet 50 $batchsize $use_mkldnn
    infer vgg 19 $batchsize $use_mkldnn
+    infer resnet 50 $batchsize $use_mkldnn
+    infer googlenet v1 $batchsize $use_mkldnn
+    infer alexnet 2 $batchsize $use_mkldnn
  done
 done
--- a/benchmark/paddle/image/run_mkldnn_train.sh
+++ b/benchmark/paddle/image/run_mkldnn_train.sh
@@ -28,6 +28,10 @@ function train() {
    --test_period=100 \
    --config_args=$args \
    2>&1 | tee ${log} 
+
+  avg_time=`tail ${log} -n 1 | awk -F ' ' '{print $8}' | sed 's/avg=//'`
+  fps=`awk 'BEGIN{printf "%.2f",('$bs' / '$avg_time' * 1000)}'`
+  echo "FPS: $fps images/sec" 2>&1 | tee -a ${log}
 }

 if [ ! -f "train.list" ]; then
@@ -43,5 +47,6 @@ for use_mkldnn in True False; do
    train vgg 19 $batchsize $use_mkldnn
    train resnet 50 $batchsize $use_mkldnn
    train googlenet v1 $batchsize $use_mkldnn
+    train alexnet 2 $batchsize $use_mkldnn
  done
 done
--- a/benchmark/paddle/image/run_openblas_infer.sh
+++ b/benchmark/paddle/image/run_openblas_infer.sh
+set -e
+
+function clock_to_seconds() {
+  hours=`echo $1 | awk -F ':' '{print $1}'`
+  mins=`echo $1 | awk -F ':' '{print $2}'`
+  secs=`echo $1 | awk -F ':' '{print $3}'`
+  echo `awk 'BEGIN{printf "%.2f",('$secs' + '$mins' * 60 + '$hours' * 3600)}'`
+}
+
+function infer() {
+  export OPENBLAS_MAIN_FREE=1
+  topology=$1
+  layer_num=$2
+  bs=$3
+  trainers=`nproc`
+  if [ $trainers -gt $bs ]; then
+    trainers=$bs
+  fi
+  log="logs/infer-${topology}-${layer_num}-${trainers}openblas-${bs}.log"
+  threads=$((`nproc` / trainers))
+  if [ $threads -eq 0 ]; then
+    threads=1
+  fi
+  export OPENBLAS_NUM_THREADS=$threads
+
+  models_in="models/${topology}-${layer_num}/pass-00000/"
+  if [ ! -d $models_in ]; then
+    echo "./run_mkl_infer.sh to save the model first"
+    exit 0
+  fi
+  log_period=$((32 / bs))
+  paddle train --job=test \
+    --config="${topology}.py" \
+    --use_mkldnn=False \
+    --use_gpu=False \
+    --trainer_count=$trainers \
+    --log_period=$log_period \
+    --config_args="batch_size=${bs},layer_num=${layer_num},is_infer=True,num_samples=256" \
+    --init_model_path=$models_in \
+    2>&1 | tee ${log}
+
+  # calculate the last 5 logs period time of 160(=32*5) samples,
+  # the time before are burning time.
+  start=`tail ${log} -n 7 | head -n 1 | awk -F ' ' '{print $2}' | xargs`
+  end=`tail ${log} -n 2 | head -n 1 | awk -F ' ' '{print $2}' | xargs`
+  start_sec=`clock_to_seconds $start`
+  end_sec=`clock_to_seconds $end`
+  fps=`awk 'BEGIN{printf "%.2f",(160 / ('$end_sec' - '$start_sec'))}'`
+  echo "Last 160 samples start: ${start}(${start_sec} sec), end: ${end}(${end_sec} sec;" >> ${log}
+  echo "FPS: $fps images/sec" 2>&1 | tee -a ${log}
+}
+
+if [ ! -f "train.list" ]; then
+  echo " " > train.list
+fi
+if [ ! -f "test.list" ]; then
+  echo " " > test.list
+fi
+if [ ! -d "logs" ]; then
+  mkdir logs
+fi
+
+# inference benchmark
+for batchsize in 1 2 4 8 16; do
+  infer vgg 19 $batchsize
+  infer resnet 50 $batchsize 
+  infer googlenet v1 $batchsize
+  infer alexnet 2 $batchsize
+done
--- a/benchmark/paddle/image/run_openblas_train.sh
+++ b/benchmark/paddle/image/run_openblas_train.sh
+set -e
+
+function train() {
+  export OPENBLAS_NUM_THREADS=1
+  topology=$1
+  layer_num=$2
+  bs=$3
+  thread=`nproc`
+  # each trainer_count use only 1 core to avoid conflict
+  log="logs/train-${topology}-${layer_num}-${thread}openblas-${bs}.log"
+  args="batch_size=${bs},layer_num=${layer_num}"
+  config="${topology}.py"
+  paddle train --job=time \
+    --config=$config \
+    --use_mkldnn=False \
+    --use_gpu=False \
+    --trainer_count=$thread \
+    --log_period=3 \
+    --test_period=30 \
+    --config_args=$args \
+    2>&1 | tee ${log} 
+
+  avg_time=`tail ${log} -n 1 | awk -F ' ' '{print $8}' | sed 's/avg=//'`
+  fps=`awk 'BEGIN{printf "%.2f",('$bs' / '$avg_time' * 1000)}'`
+  echo "FPS: $fps images/sec" 2>&1 | tee -a ${log}
+}
+
+if [ ! -f "train.list" ]; then
+  echo " " > train.list
+fi
+if [ ! -d "logs" ]; then
+  mkdir logs
+fi
+
+# training benchmark
+for batchsize in 64 128 256; do
+  train vgg 19 $batchsize
+  train resnet 50 $batchsize
+  train googlenet v1 $batchsize
+  train alexnet 2 $batchsize
+done
--- a/benchmark/paddle/image/vgg.py
+++ b/benchmark/paddle/image/vgg.py
@@ -7,13 +7,15 @@ num_class = 1000
 batch_size = get_config_arg('batch_size', int, 64)
 layer_num = get_config_arg('layer_num', int, 19)
 is_infer = get_config_arg("is_infer", bool, False)
+num_samples = get_config_arg('num_samples', int, 2560)

 args = {
    'height': height,
    'width': width,
    'color': True,
    'num_class': num_class,
-    'is_infer': is_infer
+    'is_infer': is_infer,
+    'num_samples': num_samples
 }
 define_py_data_sources2(
    "train.list" if not is_infer else None,

--- a/benchmark/paddle/rnn/imdb.py
+++ b/benchmark/paddle/rnn/imdb.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from __future__ import print_function
 import six.moves.cPickle as pickle
 import gzip

--- a/benchmark/paddle/rnn/provider.py
+++ b/benchmark/paddle/rnn/provider.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import io, os
 import random
 import numpy as np

--- a/benchmark/tensorflow/image/alexnet.py
+++ b/benchmark/tensorflow/image/alexnet.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from six.moves import xrange  # pylint: disable=redefined-builtin
 from datetime import datetime
 import math

--- a/benchmark/tensorflow/image/alexnet_multi_gpu.py
+++ b/benchmark/tensorflow/image/alexnet_multi_gpu.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from six.moves import xrange  # pylint: disable=redefined-builtin
 from datetime import datetime
 import math

--- a/benchmark/tensorflow/image/googlenet.py
+++ b/benchmark/tensorflow/image/googlenet.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from six.moves import xrange
 from datetime import datetime
 import math

--- a/benchmark/tensorflow/image/googlenet_multi_gpu.py
+++ b/benchmark/tensorflow/image/googlenet_multi_gpu.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from six.moves import xrange  # pylint: disable=redefined-builtin
 from datetime import datetime
 import math

--- a/benchmark/tensorflow/image/smallnet_mnist_cifar.py
+++ b/benchmark/tensorflow/image/smallnet_mnist_cifar.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 from six.moves import xrange  # pylint: disable=redefined-builtin
 from datetime import datetime
 import math

--- a/benchmark/tensorflow/rnn/reader.py
+++ b/benchmark/tensorflow/rnn/reader.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import os.path
 import io
 import numpy as np

--- a/cmake/external/boost.cmake
+++ b/cmake/external/boost.cmake
+# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+include(ExternalProject)
+
+set(BOOST_PROJECT       "extern_boost")
+set(BOOST_VER           "1.41.0")
+set(BOOST_TAR           "boost_1_41_0")
+set(BOOST_URL           "http://sourceforge.net/projects/boost/files/boost/${BOOST_VER}/${BOOST_TAR}.tar.gz")
+set(BOOST_SOURCES_DIR ${THIRD_PARTY_PATH}/boost)
+set(BOOST_DOWNLOAD_DIR  "${BOOST_SOURCES_DIR}/src/${BOOST_PROJECT}")
+set(BOOST_INCLUDE_DIR "${BOOST_DOWNLOAD_DIR}/${BOOST_TAR}" CACHE PATH "boost include directory." FORCE)
+
+include_directories(${BOOST_INCLUDE_DIR})
+
+ExternalProject_Add(
+    ${BOOST_PROJECT}
+    ${EXTERNAL_PROJECT_LOG_ARGS}
+    DOWNLOAD_DIR          ${BOOST_DOWNLOAD_DIR}
+    DOWNLOAD_COMMAND      wget --no-check-certificate ${BOOST_URL} -c -q -O ${BOOST_TAR}.tar.gz
+                          && tar zxf ${BOOST_TAR}.tar.gz
+    DOWNLOAD_NO_PROGRESS  1
+    PREFIX                ${BOOST_SOURCES_DIR}
+    CONFIGURE_COMMAND     ""
+    BUILD_COMMAND         ""
+    INSTALL_COMMAND       ""
+    UPDATE_COMMAND        ""
+)
+
+if (${CMAKE_VERSION} VERSION_LESS "3.3.0")
+    set(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/boost_dummy.c)
+    file(WRITE ${dummyfile} "const char *dummy = \"${dummyfile}\";")
+    add_library(boost STATIC ${dummyfile})
+else()
+    add_library(boost INTERFACE)
+endif()
+
+add_dependencies(boost ${BOOST_PROJECT})
+list(APPEND external_project_dependencies boost)
+set(Boost_INCLUDE_DIR ${BOOST_INCLUDE_DIR})
--- a/cmake/external/eigen.cmake
+++ b/cmake/external/eigen.cmake
 INCLUDE(ExternalProject)

 SET(EIGEN_SOURCE_DIR ${THIRD_PARTY_PATH}/eigen3)
-
-INCLUDE_DIRECTORIES(${EIGEN_SOURCE_DIR}/src/extern_eigen3)
+SET(EIGEN_INCLUDE_DIR ${EIGEN_SOURCE_DIR}/src/extern_eigen3)
+INCLUDE_DIRECTORIES(${EIGEN_INCLUDE_DIR})

 ExternalProject_Add(
    extern_eigen3
@@ -19,7 +19,7 @@ ExternalProject_Add(

 if (${CMAKE_VERSION} VERSION_LESS "3.3.0")
    set(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/eigen3_dummy.c)
-    file(WRITE ${dummyfile} "const char * dummy_eigen3 = \"${dummyfile}\";")
+    file(WRITE ${dummyfile} "const char *dummy_eigen3 = \"${dummyfile}\";")
    add_library(eigen3 STATIC ${dummyfile})
 else()
    add_library(eigen3 INTERFACE)
@@ -28,3 +28,9 @@ endif()
 add_dependencies(eigen3 extern_eigen3)

 LIST(APPEND external_project_dependencies eigen3)
+
+IF(NOT WITH_C_API AND WITH_FLUID)
+    INSTALL(FILES ${EIGEN_INCLUDE_DIR}/Eigen/Core DESTINATION third_party/eigen3/Eigen)
+    INSTALL(DIRECTORY ${EIGEN_INCLUDE_DIR}/Eigen/src DESTINATION third_party/eigen3/Eigen)
+    INSTALL(DIRECTORY ${EIGEN_INCLUDE_DIR}/unsupported/Eigen DESTINATION third_party/eigen3/unsupported)
+ENDIF()
--- a/cmake/external/gflags.cmake
+++ b/cmake/external/gflags.cmake
@@ -52,7 +52,7 @@ ADD_DEPENDENCIES(gflags extern_gflags)

 LIST(APPEND external_project_dependencies gflags)

-IF(WITH_C_API)
+IF(WITH_C_API OR WITH_FLUID)
  INSTALL(DIRECTORY ${GFLAGS_INCLUDE_DIR} DESTINATION third_party/gflags)
  IF(ANDROID)
    INSTALL(FILES ${GFLAGS_LIBRARIES} DESTINATION third_party/gflags/lib/${ANDROID_ABI})

--- a/cmake/external/glog.cmake
+++ b/cmake/external/glog.cmake
@@ -68,7 +68,7 @@ LINK_LIBRARIES(glog gflags)

 LIST(APPEND external_project_dependencies glog)

-IF(WITH_C_API)
+IF(WITH_C_API OR WITH_FLUID)
  INSTALL(DIRECTORY ${GLOG_INCLUDE_DIR} DESTINATION third_party/glog)
  IF(ANDROID)
    INSTALL(FILES ${GLOG_LIBRARIES} DESTINATION third_party/glog/lib/${ANDROID_ABI})

--- a/cmake/external/grpc.cmake
+++ b/cmake/external/grpc.cmake
@@ -33,7 +33,7 @@ ExternalProject_Add(
    extern_grpc
    DEPENDS protobuf zlib
    GIT_REPOSITORY "https://github.com/grpc/grpc.git"
-    GIT_TAG "v1.7.x"
+    GIT_TAG "v1.8.x"
    PREFIX          ${GRPC_SOURCES_DIR}
    UPDATE_COMMAND  ""
    CONFIGURE_COMMAND ""

--- a/cmake/external/mkldnn.cmake
+++ b/cmake/external/mkldnn.cmake
@@ -63,9 +63,30 @@ ExternalProject_Add(
                        -DMKLROOT:PATH=${MKLML_ROOT}
 )

-ADD_LIBRARY(mkldnn SHARED IMPORTED GLOBAL)
-SET_PROPERTY(TARGET mkldnn PROPERTY IMPORTED_LOCATION ${MKLDNN_LIB})
-ADD_DEPENDENCIES(mkldnn ${MKLDNN_PROJECT})
+ADD_LIBRARY(shared_mkldnn SHARED IMPORTED GLOBAL)
+SET_PROPERTY(TARGET shared_mkldnn PROPERTY IMPORTED_LOCATION ${MKLDNN_LIB})
+ADD_DEPENDENCIES(shared_mkldnn ${MKLDNN_PROJECT})
 MESSAGE(STATUS "MKLDNN library: ${MKLDNN_LIB}")
 add_definitions(-DPADDLE_WITH_MKLDNN)
-LIST(APPEND external_project_dependencies mkldnn)
+LIST(APPEND external_project_dependencies shared_mkldnn)
+
+# generate a static dummy target to track mkldnn dependencies
+# for cc_library(xxx SRCS xxx.c DEPS mkldnn)
+SET(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/mkldnn_dummy.c)
+FILE(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
+ADD_LIBRARY(mkldnn STATIC ${dummyfile})
+TARGET_LINK_LIBRARIES(mkldnn ${MKLDNN_LIB} ${MKLML_LIB} ${MKLML_IOMP_LIB})
+ADD_DEPENDENCIES(mkldnn ${MKLDNN_PROJECT})
+
+# copy the real so.0 lib to install dir
+# it can be directly contained in wheel or capi
+SET(MKLDNN_SHARED_LIB ${MKLDNN_INSTALL_DIR}/libmkldnn.so.0)
+ADD_CUSTOM_COMMAND(OUTPUT ${MKLDNN_SHARED_LIB}
+    COMMAND cp ${MKLDNN_LIB} ${MKLDNN_SHARED_LIB}
+    DEPENDS mkldnn)
+ADD_CUSTOM_TARGET(mkldnn_shared_lib ALL DEPENDS ${MKLDNN_SHARED_LIB})
+
+IF(WITH_C_API)
+  INSTALL(FILES ${MKLDNN_SHARED_LIB} DESTINATION lib)
+ENDIF()
+
--- a/cmake/external/mklml.cmake
+++ b/cmake/external/mklml.cmake
@@ -66,3 +66,7 @@ ADD_LIBRARY(mklml SHARED IMPORTED GLOBAL)
 SET_PROPERTY(TARGET mklml PROPERTY IMPORTED_LOCATION ${MKLML_LIB})
 ADD_DEPENDENCIES(mklml ${MKLML_PROJECT})
 LIST(APPEND external_project_dependencies mklml)
+
+IF(WITH_C_API)
+  INSTALL(FILES ${MKLML_LIB} ${MKLML_IOMP_LIB} DESTINATION lib)
+ENDIF()
--- a/cmake/external/openblas.cmake
+++ b/cmake/external/openblas.cmake
@@ -30,23 +30,21 @@ IF(NOT ${CBLAS_FOUND})
        CACHE FILEPATH "openblas library." FORCE)

    SET(OPENBLAS_CC "${CMAKE_C_COMPILER} -Wno-unused-but-set-variable -Wno-unused-variable")
+    SET(OPENBLAS_COMMIT "v0.2.20")

    IF(CMAKE_CROSSCOMPILING)
        SET(OPTIONAL_ARGS HOSTCC=${HOST_C_COMPILER})
        GET_FILENAME_COMPONENT(CROSS_SUFFIX ${CMAKE_C_COMPILER} DIRECTORY)
        SET(CROSS_SUFFIX ${CROSS_SUFFIX}/)
        IF(ANDROID)
-            # arm_soft_fp_abi branch of OpenBLAS to support softfp
-            #   https://github.com/xianyi/OpenBLAS/tree/arm_soft_fp_abi
-            SET(OPENBLAS_COMMIT "b5c96fcfcdc82945502a2303116a64d89985daf5")
            IF(ANDROID_ABI MATCHES "^armeabi(-v7a)?$")
+                # use softfp
                SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV7 ARM_SOFTFP_ABI=1 USE_THREAD=0)
            ELSEIF(ANDROID_ABI STREQUAL "arm64-v8a")
                SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV8 BINARY=64 USE_THREAD=0)
            ENDIF()
        ELSEIF(IOS)
            IF(CMAKE_OSX_ARCHITECTURES MATCHES "arm64")
-                SET(OPENBLAS_COMMIT "b5c96fcfcdc82945502a2303116a64d89985daf5")
                SET(OPENBLAS_CC "${OPENBLAS_CC} ${CMAKE_C_FLAGS} -isysroot ${CMAKE_OSX_SYSROOT}")
                SET(OPENBLAS_CC "${OPENBLAS_CC} -arch arm64")
                SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV8 BINARY=64 USE_THREAD=0 CROSS_SUFFIX=${CROSS_SUFFIX})
@@ -56,14 +54,12 @@ IF(NOT ${CBLAS_FOUND})
            ENDIF()
        ELSEIF(RPI)
            # use hardfp
-            SET(OPENBLAS_COMMIT "v0.2.20")
            SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV7 USE_THREAD=0)
        ENDIF()
    ELSE()
        IF(APPLE)
            SET(OPENBLAS_CC "${CMAKE_C_COMPILER} -isysroot ${CMAKE_OSX_SYSROOT}")
        ENDIF()
-        SET(OPENBLAS_COMMIT "v0.2.20")
        SET(OPTIONAL_ARGS "")
        IF(CMAKE_SYSTEM_PROCESSOR MATCHES "^x86(_64)?$")
            SET(OPTIONAL_ARGS DYNAMIC_ARCH=1 NUM_THREADS=64)
@@ -104,6 +100,11 @@ IF(NOT ${CBLAS_FOUND})
                \"${CBLAS_INSTALL_DIR}/lib -> ${CMAKE_INSTALL_PREFIX}/${TMP_INSTALL_DIR}\"
            )"
        )
+        INSTALL(CODE "execute_process(
+            COMMAND rm -r ${CMAKE_INSTALL_PREFIX}/${TMP_INSTALL_DIR}/cmake
+                    ${CMAKE_INSTALL_PREFIX}/${TMP_INSTALL_DIR}/pkgconfig
+            )"
+        )
    ENDIF()
 ENDIF(NOT ${CBLAS_FOUND})

@@ -113,7 +114,7 @@ INCLUDE_DIRECTORIES(${CBLAS_INC_DIR})
 # FIXME(gangliao): generate cblas target to track all high performance
 # linear algebra libraries for cc_library(xxx SRCS xxx.c DEPS cblas)
 SET(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/cblas_dummy.c)
-FILE(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
+FILE(WRITE ${dummyfile} "const char *dummy_cblas = \"${dummyfile}\";")
 ADD_LIBRARY(cblas STATIC ${dummyfile})
 TARGET_LINK_LIBRARIES(cblas ${CBLAS_LIBRARIES})


--- a/cmake/external/protobuf.cmake
+++ b/cmake/external/protobuf.cmake
@@ -250,12 +250,12 @@ IF(NOT PROTOBUF_FOUND)
    SET(PROTOBUF_PROTOC_LIBRARY ${extern_protobuf_PROTOC_LIBRARY}
        CACHE FILEPATH "protoc library." FORCE)

-    IF(WITH_C_API)
+    IF(WITH_C_API OR WITH_FLUID)
        INSTALL(DIRECTORY ${PROTOBUF_INCLUDE_DIR} DESTINATION third_party/protobuf)
        IF(ANDROID)
-            INSTALL(FILES ${PROTOBUF_LIBRARY} DESTINATION third_party/protobuf/lib/${ANDROID_ABI})
+            INSTALL(FILES ${PROTOBUF_LITE_LIBRARY} DESTINATION third_party/protobuf/lib/${ANDROID_ABI})
        ELSE()
-            INSTALL(FILES ${PROTOBUF_LIBRARY} DESTINATION third_party/protobuf/lib)
+            INSTALL(FILES ${PROTOBUF_LITE_LIBRARY} DESTINATION third_party/protobuf/lib)
        ENDIF()
    ENDIF()


--- a/cmake/external/warpctc.cmake
+++ b/cmake/external/warpctc.cmake
@@ -63,7 +63,7 @@ ExternalProject_Add(
 MESSAGE(STATUS "warp-ctc library: ${WARPCTC_LIBRARIES}")
 INCLUDE_DIRECTORIES(${WARPCTC_INCLUDE_DIR})

-ADD_LIBRARY(warpctc STATIC IMPORTED GLOBAL)
+ADD_LIBRARY(warpctc SHARED IMPORTED GLOBAL)
 SET_PROPERTY(TARGET warpctc PROPERTY IMPORTED_LOCATION ${WARPCTC_LIBRARIES})
 ADD_DEPENDENCIES(warpctc extern_warpctc)


--- a/cmake/generic.cmake
+++ b/cmake/generic.cmake
@@ -120,7 +120,7 @@ function(merge_static_libs TARGET_NAME)
      DEPENDS ${libs})

    # Generate dummy staic lib
-    file(WRITE ${target_SRCS} "const char *dummy = \"${target_SRCS}\";")
+    file(WRITE ${target_SRCS} "const char *dummy_${TARGET_NAME} = \"${target_SRCS}\";")
    add_library(${TARGET_NAME} STATIC ${target_SRCS})
    target_link_libraries(${TARGET_NAME} ${libs_deps})

@@ -160,7 +160,7 @@ function(merge_static_libs TARGET_NAME)
      DEPENDS ${libs} ${target_OBJS})

    # Generate dummy staic lib
-    file(WRITE ${target_SRCS} "const char *dummy = \"${target_SRCS}\";")
+    file(WRITE ${target_SRCS} "const char *dummy_${TARGET_NAME} = \"${target_SRCS}\";")
    add_library(${TARGET_NAME} STATIC ${target_SRCS})
    target_link_libraries(${TARGET_NAME} ${libs_deps})

@@ -324,7 +324,7 @@ function(go_library TARGET_NAME)
    )

  # Add dummy code to support `make target_name` under Terminal Command
-  file(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
+  file(WRITE ${dummyfile} "const char *dummy_${TARGET_NAME} = \"${dummyfile}\";")
  if (go_library_SHARED OR go_library_shared)
    add_library(${TARGET_NAME} SHARED ${dummyfile})
  else()

--- a/cmake/make_resource.py
+++ b/cmake/make_resource.py
+#   Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import os
 import re
 import sys

--- a/doc/api/index_cn.rst
+++ b/doc/api/index_cn.rst
@@ -7,3 +7,4 @@ API
    模型配置 <v2/model_configs.rst>
    数据访问 <v2/data.rst>
    训练与应用 <v2/run_logic.rst>
+    v2/fluid.rst
--- a/doc/api/v1/data_provider/dataprovider_cn.rst
+++ b/doc/api/v1/data_provider/dataprovider_cn.rst
-.. _api_dataprovider:
-
-DataProvider的介绍
-==================
-
-DataProvider是PaddlePaddle负责提供数据的模块。其作用是将数据传入内存或显存，让神经网络可以进行训练或预测。用户可以通过简单使用Python接口 :ref:`api_pydataprovider2` ，来自定义传数据的过程。如果有更复杂的使用，或者需要更高的效率，用户也可以在C++端自定义一个 ``DataProvider`` 。
-
-PaddlePaddle需要用户在网络配置（trainer_config.py）中定义使用哪种DataProvider，并且在DataProvider中实现如何访问训练文件列表（train.list）或测试文件列表（test.list）。
-
- train.list和test.list存放在本地（推荐直接存放到训练目录，以相对路径引用)。一般情况下，两者均为纯文本文件，其中每一行对应一个数据文件地址：
-  
-  - 如果数据文件存于本地磁盘，这个地址则为它的绝对路径或相对路径(相对于PaddlePaddle程序运行时的路径)。
-  - 地址也可以为hdfs文件路径，或者数据库连接路径等。
-  - 由于这个地址会被DataProvider使用，因此，如何解析该地址也是用户自定义DataProvider时需要考虑的地方。
- 如果没有设置test.list，或设置为None，那么在训练过程中不会执行测试操作；否则，会根据命令行参数指定的测试方式，在训练过程中进行测试，从而防止过拟合。
--- a/doc/api/v1/data_provider/dataprovider_en.rst
+++ b/doc/api/v1/data_provider/dataprovider_en.rst
-Introduction
-==============
-DataProvider is a module that loads training or testing data into cpu or gpu
-memory for the following triaining or testing process.
-
-For simple use, users can use Python :code:`PyDataProvider` to dynamically reads
-the original data in any format or in any form, and then transfer them into a
-data format PaddlePaddle requires. The process is extremly flexible and highly
-customized, with sacrificing the efficiency only a little. This is extremly
-useful when you have to dynamically generate certain kinds of data according to,
-for example, the training performance.
-
-Besides, users also can customize a C++ :code:`DataProvider` for a more
-complex usage, or for a higher efficiency.
-
-The following parameters are required to define in the PaddlePaddle network
-configuration file (trainer_config.py): which DataProvider is chosen to used,
-and specific parameters for DataProvider, including training file list
-(train.list) and testing file list (test.list).
-
-Train.list and test.list are simply two plain text files, which defines path
-of training or testing data. It is recommended that directly placing them into
-the training directory, and reference to them by using a relative path (
-relative to the PaddePaddle program).
-
-Testing or evaluating will not be performed during training if the test.list is
-not set or set to None. Otherwise, PaddlePaddle will evaluate the trained model
-by the specified tesing data while training, every testing period (a user
-defined command line parameter in PaddlePaddle) to prevent over-fitting.
-
-Each line of train.list and test.list is an absolute or relative path (relative
-to the PaddePaddle program runtime) of data file. Fascinatingly more, each line
-can also be a HDFS file path or a SQL connection string. As long as the user
-assures how to access each file in DataProvider.
--- a/doc/api/v1/data_provider/pydataprovider2_cn.rst
+++ b/doc/api/v1/data_provider/pydataprovider2_cn.rst
-..  _api_pydataprovider2:
-
-PyDataProvider2的使用
-=====================
-
-PyDataProvider2是PaddlePaddle使用Python提供数据的推荐接口。该接口使用多线程读取数据，并提供了简单的Cache功能；同时可以使用户只关注如何从文件中读取每一条数据，而不用关心数据如何传输，如何存储等等。
-
-..  contents::
-
-MNIST的使用场景
---------------
-
-我们以MNIST手写识别为例，来说明PyDataProvider2的简单使用场景。
-
-样例数据
-++++++++
-
-MNIST是一个包含有70,000张灰度图片的数字分类数据集。样例数据 ``mnist_train.txt`` 如下：
-
-..  literalinclude:: src/mnist_train.txt
-
-其中每行数据代表一张图片，行内使用 ``;`` 分成两部分。第一部分是图片的标签，为0-9中的一个数字；第二部分是28*28的图片像素灰度值。 对应的 ``train.list`` 即为这个数据文件的名字：
-
-..  literalinclude:: src/train.list
-
-dataprovider的使用
-++++++++++++++++++
-
-..  literalinclude:: src/mnist_provider.dict.py
-
- 首先，引入PaddlePaddle的PyDataProvider2包。
- 其次，定义一个Python的 `Decorator <http://www.learnpython.org/en/Decorators>`_ `@provider`_ 。用于将下一行的数据输入函数标记成一个PyDataProvider2，同时设置它的input_types属性。
-  
-  - `input_types`_：设置这个PyDataProvider2返回什么样的数据。本例根据网络配置中 ``data_layer`` 的名字，显式指定返回的是一个28*28维的稠密浮点数向量和一个[0-9]的10维整数标签。
-
-    ..  literalinclude:: src/mnist_config.py
-         :lines: 9-10
-
-  - 注意：如果用户不显示指定返回数据的对应关系，那么PaddlePaddle会根据layer的声明顺序，来确定对应关系。但这个关系可能不正确，所以推荐使用显式指定的方式来设置input_types。
- 最后，实现数据输入函数（如本例的 ``process`` 函数）。
-
-  - 该函数的功能是：打开文本文件，读取每一行，将行中的数据转换成与input_types一致的格式，然后返回给PaddlePaddle进程。注意，
-    
-    - 返回的顺序需要和input_types中定义的顺序一致。
-    - 返回时，必须使用Python关键词 ``yield`` ，相关概念是 ``generator`` 。
-    - 一次yield调用，返回一条完整的样本。如果想为一个数据文件返回多条样本，只需要在函数中调用多次yield即可（本例中使用for循环进行多次调用）。
-  
-  - 该函数具有两个参数：
-  
-    - settings：在本例中没有使用，具体可以参考 `init_hook`_ 中的说明。
-    - filename：为 ``train.list`` 或 ``test.list`` 中的一行，即若干数据文件路径的某一个。
-
-网络配置中的调用
-++++++++++++++++
-
-在网络配置里，只需要一行代码就可以调用这个PyDataProvider2，如，
-
-..  literalinclude:: src/mnist_config.py
-     :lines: 1-7
-
-训练数据是 ``train.list`` ，没有测试数据，调用的PyDataProvider2是 ``mnist_provider`` 模块中的 ``process`` 函数。
-
-小结
-+++++
-
-至此，简单的PyDataProvider2样例就说明完毕了。对用户来说，仅需要知道如何从 **一个文件** 中读取 **一条样本** ，就可以将数据传送给PaddlePaddle了。而PaddlePaddle则会帮用户做以下工作：
-
-* 将数据组合成Batch进行训练
-* 对训练数据进行Shuffle
-* 多线程的数据读取
-* 缓存训练数据到内存(可选)
-* CPU->GPU双缓存
-
-是不是很简单呢？
-
-时序模型的使用场景
------------------
-样例数据
-++++++++
-
-时序模型是指数据的某一维度是一个序列形式，即包含时间步信息。所谓时间步信息，不一定和时间有关系，只是说明数据的顺序是重要的。例如，文本信息就是一个序列数据。
-
-本例采用英文情感分类的数据，即将一段英文文本数据，分类成正面情绪和负面情绪两类(用0和1表示)。样例数据 ``sentimental_train.txt`` 如下：
-
-..  literalinclude:: src/sentimental_train.txt
-
-dataprovider的使用
-++++++++++++++++++
-
-相对MNIST而言，这个dataprovider较复杂，主要原因是增加了初始化机制 `init_hook`_。本例的 ``on_init`` 函数就是根据该机制配置的，它会在dataprovider创建的时候执行。
-
- 其中 ``input_types`` 和在 `@provider`_ 中配置的效果一致。本例中的输入特征是词ID的序列，因此使用 ``integer_value_sequence`` 类型来设置。
- 将 ``dictionary`` 存入settings对象，在 ``process`` 函数中使用。 dictionary是从网络配置中传入的dict对象，即一个将单词字符串映射到单词ID的字典。
-
-..  literalinclude:: src/sentimental_provider.py
-
-网络配置中的调用
-++++++++++++++++
-
-调用这个PyDataProvider2的方法，基本上和MNIST样例一致，除了
-
-* 在配置中需要读取外部字典。
-* 在声明DataProvider的时候传入dictionary作为参数。
-
-..  literalinclude:: src/sentimental_config.py
-     :emphasize-lines: 12-14
-
-参考(Reference)
---------------
-
-@provider
-+++++++++
-
-``@provider`` 是一个Python的 `Decorator`_ ，可以将某一个函数标记成一个PyDataProvider2。如果不了解 `Decorator`_ 是什么也没关系，只需知道这是一个标记属性的方法就可以了。它包含的属性参数如下:
-
-*  input_types：数据输入格式。具体的格式说明，请参考 `input_types`_ 。
-*  should_shuffle：是不是要对数据做Shuffle。训练时默认shuffle，测试时默认不shuffle。
-*  min_pool_size：设置内存中最小暂存的数据条数，也是PaddlePaddle所能够保证的shuffle粒度。如果为-1，则会预先读取全部数据到内存中。
-*  pool_size： 设置内存中暂存的数据条数。如果为-1（默认），则不在乎内存暂存多少条数据。如果设置，则推荐大于训练时batch size的值，并且在内存足够的情况下越大越好。
-*  can_over_batch_size：是否允许暂存略微多余pool_size的数据。由于这样做可以避免很多死锁问题，一般推荐设置成True。
-*  calc_batch_size：可以传入一个函数，用于自定义每条数据的batch size（默认为1）。
-*  cache： 数据缓存的策略，具体请参考 `cache`_ 。
-*  init_hook：初始化时调用的函数，具体请参考 `init_hook`_ 。
-*  check：如果为true，会根据input_types检查数据的合法性。
-*  check_fail_continue：如果为true，那么当check出数据不合法时，会扔到这条数据，继续训练或预测。（对check=false的情况，没有作用）
-
-input_types
-+++++++++++
-
-PaddlePaddle的数据包括四种主要类型，和三种序列模式。
-
-四种数据类型：
-
-* dense_vector：稠密的浮点数向量。
-* sparse_binary_vector：稀疏的01向量，即大部分值为0，但有值的地方必须为1。
-* sparse_float_vector：稀疏的向量，即大部分值为0，但有值的部分可以是任何浮点数。
-* integer：整数标签。
-
-三种序列模式：
-
-* SequenceType.NO_SEQUENCE：不是一条序列
-* SequenceType.SEQUENCE：是一条时间序列
-* SequenceType.SUB_SEQUENCE： 是一条时间序列，且序列的每一个元素还是一个时间序列。
-
-不同的数据类型和序列模式返回的格式不同，列表如下：
-
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-|                      | NO_SEQUENCE         | SEQUENCE                          |  SUB_SEQUENCE                                  |
-+======================+=====================+===================================+================================================+
-| dense_vector         | [f, f, ...]         | [[f, ...], [f, ...], ...]         | [[[f, ...], ...], [[f, ...], ...],...]         |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-| sparse_binary_vector | [i, i, ...]         | [[i, ...], [i, ...], ...]         | [[[i, ...], ...], [[i, ...], ...],...]         |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-| sparse_float_vector  | [(i,f), (i,f), ...] | [[(i,f), ...], [(i,f), ...], ...] | [[[(i,f), ...], ...], [[(i,f), ...], ...],...] |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-| integer_value        |  i                  | [i, i, ...]                       | [[i, ...], [i, ...], ...]                      |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-
-其中，f代表一个浮点数，i代表一个整数。
-
-注意：对sparse_binary_vector和sparse_float_vector，PaddlePaddle存的是有值位置的索引。例如，
-
- 对一个5维非序列的稀疏01向量 ``[0, 1, 1, 0, 0]`` ，类型是sparse_binary_vector，返回的是 ``[1, 2]`` 。
- 对一个5维非序列的稀疏浮点向量 ``[0, 0.5, 0.7, 0, 0]`` ，类型是sparse_float_vector，返回的是 ``[(1, 0.5), (2, 0.7)]`` 。
-
-init_hook
-+++++++++
-
-init_hook可以传入一个函数。该函数在初始化的时候会被调用，其参数如下:
-
-* 第一个参数是settings对象，它和数据传入函数的第一个参数（如本例中 ``process`` 函数的 ``settings`` 参数）必须一致。该对象具有以下两个属性：
-    * settings.input_types：数据输入格式，具体请参考 `input_types`_ 。
-    * settings.logger：一个logging对象。
-* 其他参数使用 ``kwargs`` （key word arguments）传入，包括以下两种：
-    * PaddlePaddle定义的参数: 1）is_train：bool型参数，表示用于训练或预测；2）file_list：所有文件列表。
-    * 用户定义的参数：使用args在网络配置中设置。
-
-注意：PaddlePaddle保留添加参数的权力，因此init_hook尽量使用 ``**kwargs`` 来接受不使用的函数以保证兼容性。
-
-cache
-+++++
-
-PyDataProvider2提供了两种简单的Cache策略：
-
-* CacheType.NO_CACHE：不缓存任何数据，每次都会从python端读取数据
-* CacheType.CACHE_PASS_IN_MEM：第一个pass会从python端读取数据，剩下的pass会直接从内存里
-  读取数据。 
-
-
-注意事项
--------
-
-可能的内存泄露问题
-++++++++++++++++++
-
-PaddlePaddle将train.list中的每一行都传递给process函数，从而生成多个generator。当训练数据非常多时，就会生成非常多的generator。
-
-虽然每个generator在没有调用的时候，是几乎不占内存的；但当调用过一次后，generator便会存下当前的上下文(Context)，而这个Context可能会非常大。并且，generator至少需要调用两次才会知道是否停止。所以，即使process函数里面只有一个yield，也需要两次随机选择到相同generator的时候，才会释放该段内存。
-
-..  code-block:: python
-
-    def func():
-        yield 0
-
-    f = func()  # 创建generator
-    tmp = next(f)  # 调用一次，返回0
-    tmp = next(f)  # 调用第二次的时候，才会Stop Iteration
-
-由于顺序调用这些generator不会出现上述问题，因此有两种解决方案：
-
-1. **最佳推荐**：将样本的地址放入另一个文本文件，train.list写入那个文本文件的地址。即不要将每一个样本都放入train.list。
-2. 在generator的上下文中尽量留下非常少的变量引用，例如
-
-..  code-block:: python
-
-    def real_process(fn):
-        # ... read from fn
-        return result   # 当函数返回的时候，python可以解除掉内部变量的引用。
-
-    def process(fn):
-        yield real_process(fn)
-
-注意：这个问题是PyDataProvider读数据时候的逻辑问题，很难整体修正。
-
-内存不够用的情况
-++++++++++++++++
-
-PyDataProvider2会尽可能多的使用内存。因此，对于内存较小的机器，推荐使用 ``pool_size`` 变量来设置内存中暂存的数据条。具体请参考 `@provider`_ 中的说明。
-
--- a/doc/api/v1/data_provider/pydataprovider2_en.rst
+++ b/doc/api/v1/data_provider/pydataprovider2_en.rst
-..  _api_pydataprovider2:
-
-PyDataProvider2
-===============
-
-We highly recommand users to use PyDataProvider2 to provide training or testing
-data to PaddlePaddle. The user only needs to focus on how to read a single
-sample from the original data file by using PyDataProvider2, leaving all of the
-trivial work, including, transfering data into cpu/gpu memory, shuffle, binary
-serialization to PyDataProvider2. PyDataProvider2 uses multithreading and a
-fanscinating but simple cache strategy to optimize the efficiency of the data
-providing process.
-
-DataProvider for the non-sequential model
-----------------------------------------
-
-Here we use the MNIST handwriting recognition data as an example to illustrate
-how to write a simple PyDataProvider.
-
-MNIST is a handwriting classification data set. It contains 70,000 digital
-grayscale images. Labels of the training sample range from 0 to 9. All the
-images have been size-normalized and centered into images with the same size
-of 28 x 28 pixels.
-
-A small part of the original data as an example is shown as below:
-
-.. literalinclude:: src/mnist_train.txt
-
-Each line of the data contains two parts, separated by :code:`;`. The first part is
-label of an image. The second part contains 28x28 pixel float values.
-
-Just write path of the above data into train.list. It looks like this:
-
-.. literalinclude:: src/train.list
-
-The corresponding dataprovider is shown as below:
-
-.. literalinclude:: src/mnist_provider.dict.py
-
-The first line imports PyDataProvider2 package.
-The main function is the process function, that has two parameters.
-The first parameter is the settings, which is not used in this example.
-The second parameter is the filename, that is exactly each line of train.list.
-This parameter is passed to the process function by PaddlePaddle.
-
-:code:`@provider` is a Python
-`Decorator <http://www.learnpython.org/en/Decorators>`_ .
-It sets some properties to DataProvider, and constructs a real PaddlePaddle
-DataProvider from a very simple user implemented python function. It does not
-matter if you are not familiar with `Decorator`_. You can keep it simple by
-just taking :code:`@provider` as a fixed mark above the provider function you
-implemented.
-
-`input_types`_ defines the data format that a DataProvider returns.
-In this example, it is set to a 28x28-dimensional dense vector and an integer
-scalar, whose value ranges from 0 to 9.
-`input_types`_ can be set to several kinds of input formats, please refer to the
-document of `input_types`_ for more details.
-
-
-The process method is the core part to construct a real DataProvider in
-PaddlePaddle. It implements how to open the text file, how to read one sample
-from the original text file, convert them into `input_types`_, and give them
-back to PaddlePaddle process at line 23.
-Note that data yielded by the process function must follow the same order that
-`input_types`_ are defined.
-
-
-With the help of PyDataProvider2, user can focus on how to generate ONE traning
-sample by using keywords :code:`yield`.
-:code:`yield` is a python keyword, and a concept related to it includes
-:code:`generator`.
-
-Only a few lines of codes need to be added into the training configuration file,
-you can take this as an example.
-
-.. literalinclude:: src/mnist_config.py
-
-Here we specify training data by :code:`train.list`, and no testing data is specified.
-The method which actually provide data is :code:`process`.
-
-User also can use another style to provide data, which defines the
-:code:`data_layer`'s name explicitly when `yield`. For example,
-the :code:`dataprovider` is shown as below.
-
-.. literalinclude:: src/mnist_provider.dict.py
-   :linenos:
-
-If user did't give the :code:`data_layer`'s name, PaddlePaddle will use
-the order of :code:`data_layer` definition roughly to determine which feature to
-which :code:`data_layer`. This order may be not correct, so TO DEFINE THE
-:code:`data_layer`'s NAMES EXPLICITLY IS THE RECOMMANDED WAY TO PROVIDER DATA.
-
-Now, this simple example of using PyDataProvider is finished.
-The only thing that the user should know is how to generte **one sample** from
-**one data file**.
-And PaddlePadle will do all of the rest things\:
-
-* Form a training batch
-* Shuffle the training data
-* Read data with multithreading
-* Cache the training data (Optional)
-* CPU-> GPU double buffering.
-
-Is this cool?
-
-..  _api_pydataprovider2_sequential_model:
-
-DataProvider for the sequential model
-------------------------------------
-A sequence model takes sequences as its input. A sequence is made up of several
-timesteps. The so-called timestep, is not necessary to have something to do
-with time. It can also be explained to that the order of data are taken into
-consideration into model design and training.
-For example, the sentence can be interpreted as a kind of sequence data in NLP
-tasks.
-
-Here is an example on data proivider for English sentiment classification data.
-The original input data are simple English text, labeled into positive or
-negative sentiment (marked by 0 and 1 respectively).
-
-A small part of the original data as an example can be found in the path below:
-
-.. literalinclude:: src/sentimental_train.txt
-
-The corresponding data provider can be found in the path below:
-
-.. literalinclude:: src/sentimental_provider.py
-
-This data provider for sequential model is a little more complex than that
-for MINST dataset.
-A new initialization method is introduced here.
-The method :code:`on_init` is configured to DataProvider by :code:`@provider`'s
-:code:`init_hook` parameter, and it will be invoked once DataProvider is
-initialized. The :code:`on_init` function has the following parameters:
-
-* The first parameter is the settings object.
-* The rest parameters are passed by key word arguments. Some of them are passed
-  by PaddlePaddle, see reference for `init_hook`_.
-  The :code:`dictionary` object is a python dict object passed from the trainer
-  configuration file, and it maps word string to word id.
-
-To pass these parameters into DataProvider, the following lines should be added
-into trainer configuration file.
-
-.. literalinclude:: src/sentimental_config.py
-
-The definition is basically same as MNIST example, except:
-* Load dictionary in this configuration
-* Pass it as a parameter to the DataProvider
-
-The `input_types` is configured in method :code:`on_init`. It has the same
-effect to configure them by :code:`@provider`'s :code:`input_types` parameter.
-However, the :code:`input_types` is set at runtime, so we can set it to
-different types according to the input data. Input of the neural network is a
-sequence of word id, so set :code:`seq_type` to :code:`integer_value_sequence`.
-
-Durning :code:`on_init`, we save :code:`dictionary` variable to
-:code:`settings`, and it will be used in :code:`process`. Note the settings
-parameter for the process function and for the on_init's function are a same
-object.
-
-The basic processing logic is the same as MNIST's :code:`process` method. Each
-sample in the data file is given back to PaddlePaddle process.
-
-Thus, the basic usage of PyDataProvider is here.
-Please refer to the following section reference for details.
-
-Reference
---------
-
-@provider
-+++++++++
-
-.. autofunction:: paddle.trainer.PyDataProvider2.provider
-
-input_types
-+++++++++++
-
-PaddlePaddle has four data types, and three sequence types.
-The four data types are:
-
-* :code:`dense_vector`: dense float vector.
-* :code:`sparse_binary_vector`: sparse binary vector, most of the value is 0, and
-  the non zero elements are fixed to 1.
-* :code:`sparse_float_vector`: sparse float vector, most of the value is 0, and some
-  non zero elements can be any float value. They are given by the user.
-* :code:`integer`: an integer scalar, that is especially used for label or word index.
-
-The three sequence types are:
-
-* :code:`SequenceType.NO_SEQUENCE` means the sample is not a sequence.
-* :code:`SequenceType.SEQUENCE` means the sample is a sequence.
-* :code:`SequenceType.SUB_SEQUENCE` means it is a nested sequence, that each timestep of
-  the input sequence is also a sequence.
-
-Different input type has a defferenct input format. Their formats are shown
-in the above table.
-
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-|                      | NO_SEQUENCE         | SEQUENCE                          |  SUB_SEQUENCE                                  |
-+======================+=====================+===================================+================================================+
-| dense_vector         | [f, f, ...]         | [[f, ...], [f, ...], ...]         | [[[f, ...], ...], [[f, ...], ...],...]         |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-| sparse_binary_vector | [i, i, ...]         | [[i, ...], [i, ...], ...]         | [[[i, ...], ...], [[i, ...], ...],...]         |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-| sparse_float_vector  | [(i,f), (i,f), ...] | [[(i,f), ...], [(i,f), ...], ...] | [[[(i,f), ...], ...], [[(i,f), ...], ...],...] |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-| integer_value        |  i                  | [i, i, ...]                       | [[i, ...], [i, ...], ...]                      |
-+----------------------+---------------------+-----------------------------------+------------------------------------------------+
-
-where f represents a float value, i represents an integer value.
-
-init_hook
-+++++++++
-
-init_hook is a function that is invoked once the data provoder is initialized.
-Its parameters lists as follows:
-
-* The first parameter is a settings object, which is the same to :code:`settings`
-  in :code:`process` method. The object contains several attributes, including:
-
-  * :code:`settings.input_types`: the input types. Reference `input_types`_.
-  * :code:`settings.logger`: a logging object.
-
-* The rest parameters are the key word arguments. It is made up of PaddpePaddle
-  pre-defined parameters and user defined parameters.
-
-  * PaddlePaddle-defined parameters including:
-
-    * :code:`is_train` is a bool parameter that indicates the DataProvider is used in
-      training or testing.
-    * :code:`file_list` is the list of all files.
-
-  * User-defined parameters args can be set in training configuration.
-
-Note, PaddlePaddle reserves the right to add pre-defined parameter, so please
-use :code:`**kwargs` in init_hook to ensure compatibility by accepting the
-parameters which your init_hook does not use.
-
-cache
-+++++
-DataProvider provides two simple cache strategy. They are:
-
-* :code:`CacheType.NO_CACHE` means do not cache any data, then data is read at runtime by
-  the user implemented python module every pass.
-* :code:`CacheType.CACHE_PASS_IN_MEM` means the first pass reads data by the user
-  implemented python module, and the rest passes will directly read data from
-  memory.
--- a/doc/api/v1/data_provider/src/mnist_config.py
+++ b/doc/api/v1/data_provider/src/mnist_config.py
-from paddle.trainer_config_helpers import *
-
-define_py_data_sources2(
-    train_list='train.list',
-    test_list=None,
-    module='mnist_provider',
-    obj='process')
-
-img = data_layer(name='pixel', size=784)
-label = data_layer(name='label', size=10)
--- a/doc/api/v1/data_provider/src/mnist_provider.dict.py
+++ b/doc/api/v1/data_provider/src/mnist_provider.dict.py
-from paddle.trainer.PyDataProvider2 import *
-
-
-# Define a py data provider
-@provider(
-    input_types={'pixel': dense_vector(28 * 28),
-                 'label': integer_value(10)})
-def process(settings, filename):  # settings is not used currently.
-    f = open(filename, 'r')  # open one of training file
-
-    for line in f:  # read each line
-        label, pixel = line.split(';')
-
-        # get features and label
-        pixels_str = pixel.split(' ')
-
-        pixels_float = []
-        for each_pixel_str in pixels_str:
-            pixels_float.append(float(each_pixel_str))
-
-        # give data to paddle.
-        yield {"pixel": pixels_float, 'label': int(label)}
-
-    f.close()  # close file
--- a/doc/api/v1/data_provider/src/mnist_train.txt
+++ b/doc/api/v1/data_provider/src/mnist_train.txt
-5;0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.215686 0.533333 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.67451 0.992157 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.070588 0.886275 0.992157 0 0 0 0 0 0 0 0 0 0 0.192157 0.070588 0 0 0 0 0 0 0 0 0 0 0 0 0 0.670588 0.992157 0.992157 0 0 0 0 0 0 0 0 0 0.117647 0.933333 0.858824 0.313725 0 0 0 0 0 0 0 0 0 0 0 0.090196 0.858824 0.992157 0.831373 0 0 0 0 0 0 0 0 0 0.141176 0.992157 0.992157 0.611765 0.054902 0 0 0 0 0 0 0 0 0 0 0.258824 0.992157 0.992157 0.529412 0 0 0 0 0 0 0 0 0 0.368627 0.992157 0.992157 0.419608 0.003922 0 0 0 0 0 0 0 0 0 0.094118 0.835294 0.992157 0.992157 0.517647 0 0 0 0 0 0 0 0 0 0.603922 0.992157 0.992157 0.992157 0.603922 0.545098 0.043137 0 0 0 0 0 0 0 0.447059 0.992157 0.992157 0.956863 0.062745 0 0 0 0 0 0 0 0 0.011765 0.666667 0.992157 0.992157 0.992157 0.992157 0.992157 0.745098 0.137255 0 0 0 0 0 0.152941 0.866667 0.992157 0.992157 0.521569 0 0 0 0 0 0 0 0 0 0.070588 0.992157 0.992157 0.992157 0.803922 0.352941 0.745098 0.992157 0.945098 0.317647 0 0 0 0 0.580392 0.992157 0.992157 0.764706 0.043137 0 0 0 0 0 0 0 0 0 0.070588 0.992157 0.992157 0.776471 0.043137 0 0.007843 0.27451 0.882353 0.941176 0.176471 0 0 0.180392 0.898039 0.992157 0.992157 0.313725 0 0 0 0 0 0 0 0 0 0 0.070588 0.992157 0.992157 0.713725 0 0 0 0 0.627451 0.992157 0.729412 0.062745 0 0.509804 0.992157 0.992157 0.776471 0.035294 0 0 0 0 0 0 0 0 0 0 0.494118 0.992157 0.992157 0.968627 0.168627 0 0 0 0.423529 0.992157 0.992157 0.364706 0 0.717647 0.992157 0.992157 0.317647 0 0 0 0 0 0 0 0 0 0 0 0.533333 0.992157 0.984314 0.945098 0.603922 0 0 0 0.003922 0.466667 0.992157 0.988235 0.976471 0.992157 0.992157 0.788235 0.007843 0 0 0 0 0 0 0 0 0 0 0 0.686275 0.882353 0.364706 0 0 0 0 0 0 0.098039 0.588235 0.992157 0.992157 0.992157 0.980392 0.305882 0 0 0 0 0 0 0 0 0 0 0 0 0.101961 0.67451 0.321569 0 0 0 0 0 0 0 0.105882 0.733333 0.976471 0.811765 0.713725 0 0 0 0 0 0 0 0 0 0 0 0 0 0.65098 0.992157 0.321569 0 0 0 0 0 0 0 0 0 0.25098 0.007843 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0.94902 0.219608 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.968627 0.764706 0.152941 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.498039 0.25098 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0;
-0;0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.298039 0.333333 0.333333 0.333333 0.337255 0.333333 0.333333 0.109804 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.027451 0.223529 0.776471 0.964706 0.988235 0.988235 0.988235 0.992157 0.988235 0.988235 0.780392 0.098039 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.14902 0.698039 0.988235 0.992157 0.988235 0.901961 0.87451 0.568627 0.882353 0.976471 0.988235 0.988235 0.501961 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.188235 0.647059 0.988235 0.988235 0.745098 0.439216 0.098039 0 0 0 0.572549 0.988235 0.988235 0.988235 0 0 0 0 0 0 0 0 0 0 0 0 0 0.2 0.933333 0.992157 0.941176 0.247059 0 0 0 0 0 0 0.188235 0.898039 0.992157 0.992157 0 0 0 0 0 0 0 0 0 0 0 0.039216 0.639216 0.933333 0.988235 0.913725 0.278431 0 0 0 0 0 0 0 0.113725 0.843137 0.988235 0.988235 0 0 0 0 0 0 0 0 0 0 0 0.235294 0.988235 0.992157 0.988235 0.815686 0.07451 0 0 0 0 0 0 0 0.333333 0.988235 0.988235 0.552941 0 0 0 0 0 0 0 0 0 0 0.211765 0.878431 0.988235 0.992157 0.701961 0.329412 0.109804 0 0 0 0 0 0 0 0.698039 0.988235 0.913725 0.145098 0 0 0 0 0 0 0 0 0 0.188235 0.890196 0.988235 0.988235 0.745098 0.047059 0 0 0 0 0 0 0 0 0 0.882353 0.988235 0.568627 0 0 0 0 0 0 0 0 0 0.2 0.933333 0.992157 0.992157 0.992157 0.447059 0.294118 0 0 0 0 0 0 0 0 0.447059 0.992157 0.768627 0 0 0 0 0 0 0 0 0 0 0.623529 0.988235 0.988235 0.988235 0.988235 0.992157 0.47451 0 0 0 0 0 0 0 0.188235 0.933333 0.87451 0.509804 0 0 0 0 0 0 0 0 0 0 0.992157 0.988235 0.937255 0.792157 0.988235 0.894118 0.082353 0 0 0 0 0 0 0.027451 0.647059 0.992157 0.654902 0 0 0 0 0 0 0 0 0 0 0 0.623529 0.988235 0.913725 0.329412 0.376471 0.184314 0 0 0 0 0 0 0.027451 0.513725 0.988235 0.635294 0.219608 0 0 0 0 0 0 0 0 0 0 0 0.196078 0.929412 0.988235 0.988235 0.741176 0.309804 0 0 0 0 0 0 0.529412 0.988235 0.678431 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.223529 0.992157 0.992157 1 0.992157 0.992157 0.992157 0.992157 1 0.992157 0.992157 0.882353 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.023529 0.478431 0.654902 0.658824 0.952941 0.988235 0.988235 0.988235 0.992157 0.988235 0.729412 0.278431 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.196078 0.647059 0.764706 0.764706 0.768627 0.580392 0.047059 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0;
-4;0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.180392 0.470588 0.623529 0.623529 0.623529 0.588235 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.243137 0.494118 0.862745 0.870588 0.960784 0.996078 0.996078 0.996078 0.996078 0.992157 0.466667 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.317647 0.639216 0.639216 0.639216 0.639216 0.639216 0.470588 0.262745 0.333333 0.929412 0.694118 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.811765 0.694118 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.811765 0.694118 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.811765 0.694118 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.184314 0.992157 0.694118 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.192157 0.996078 0.384314 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.454902 0.980392 0.219608 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.564706 0.941176 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.588235 0.776471 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.945098 0.560784 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.054902 0.952941 0.356863 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.337255 0.917647 0.109804 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.698039 0.701961 0.019608 0.4 0.662745 0.662745 0.662745 0.662745 0.662745 0.662745 0.662745 0.376471 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.090196 0.639216 0.972549 0.945098 0.913725 0.996078 0.996078 0.996078 0.996078 1 0.996078 0.996078 1 0.996078 0 0 0 0 0 0 0 0 0 0 0.007843 0.105882 0.717647 0.776471 0.905882 0.996078 0.996078 0.988235 0.980392 0.862745 0.537255 0.223529 0.223529 0.368627 0.376471 0.6 0.6 0.6 0 0 0 0 0 0 0 0 0.262745 0.470588 0.6 0.996078 0.996078 0.996078 0.996078 0.847059 0.356863 0.156863 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.909804 0.705882 0.823529 0.635294 0.490196 0.219608 0.113725 0.062745 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.152941 0.152941 0.156863 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0;
--- a/doc/api/v1/data_provider/src/sentimental_config.py
+++ b/doc/api/v1/data_provider/src/sentimental_config.py
-from paddle.trainer_config_helpers import *
-
-dictionary = dict()
-...  #  read dictionary from outside
-
-define_py_data_sources2(
-    train_list='train.list',
-    test_list=None,
-    module='sentimental_provider',
-    obj='process',
-    # above codes same as mnist sample.
-    args={  # pass to provider.
-        'dictionary': dictionary
-    })
--- a/doc/api/v1/data_provider/src/sentimental_provider.py
+++ b/doc/api/v1/data_provider/src/sentimental_provider.py
-from paddle.trainer.PyDataProvider2 import *
-
-
-def on_init(settings, dictionary, **kwargs):
-    # on_init will invoke when data provider is initialized. The dictionary
-    # is passed from trainer_config, and is a dict object with type
-    # (word string => word id).
-
-    # set input types in runtime. It will do the same thing as
-    # @provider(input_types) will do, but it is set dynamically during runtime.
-    settings.input_types = {
-        # The text is a sequence of integer values, and each value is a word id.
-        # The whole sequence is the sentences that we want to predict its
-        # sentimental.
-        'data': integer_value_sequence(len(dictionary)),  # text input
-        'label': integer_value(2)  # label positive/negative
-    }
-
-    # save dictionary as settings.dictionary. 
-    # It will be used in process method.
-    settings.dictionary = dictionary
-
-
-@provider(init_hook=on_init)
-def process(settings, filename):
-    f = open(filename, 'r')
-
-    for line in f:  # read each line of file
-        label, sentence = line.split('\t')  # get label and sentence
-        words = sentence.split(' ')  # get words
-
-        # convert word string to word id
-        # the word not in dictionary will be ignored.
-        word_ids = []
-
-        for each_word in words:
-            if each_word in settings.dictionary:
-                word_ids.append(settings.dictionary[each_word])
-
-        # give data to paddle.
-        yield word_ids, int(label)
-
-    f.close()
--- a/doc/api/v1/data_provider/src/sentimental_train.txt
+++ b/doc/api/v1/data_provider/src/sentimental_train.txt
-0       I saw this movie at the AFI Dallas festival . It all takes place at a lake house and it looks wonderful .
-1       This documentary makes you travel all around the globe . It contains rare and stunning sequels from the wilderness .
-...
--- a/doc/api/v1/data_provider/src/train.list
+++ b/doc/api/v1/data_provider/src/train.list
-mnist_train.txt
--- a/doc/api/v1/index_cn.rst
+++ b/doc/api/v1/index_cn.rst
-API中文手册
-============
-
-DataProvider API
----------------
-
-..  toctree::
-    :maxdepth: 1
-
-    data_provider/dataprovider_cn.rst
-    data_provider/pydataprovider2_cn.rst
-
-..  _api_trainer_config:
-
-Model Config API
----------------
-
-..  toctree::
-    :maxdepth: 1
-
-    trainer_config_helpers/optimizers.rst
-    trainer_config_helpers/data_sources.rst
-    trainer_config_helpers/layers.rst
-    trainer_config_helpers/activations.rst
-    trainer_config_helpers/poolings.rst
-    trainer_config_helpers/networks.rst
-    trainer_config_helpers/evaluators.rst
-    trainer_config_helpers/attrs.rst
-
-
-Applications API
----------------
-
-..  toctree::
-    :maxdepth: 1
-
-    predict/swig_py_paddle_cn.rst
--- a/doc/api/v1/index_en.rst
+++ b/doc/api/v1/index_en.rst
-API
-===
-
-DataProvider API
----------------
-
-..  toctree::
-    :maxdepth: 1
-
-    data_provider/dataprovider_en.rst
-    data_provider/pydataprovider2_en.rst
-
-..  _api_trainer_config:
-
-Model Config API
----------------
-
-..  toctree::
-    :maxdepth: 1
-
-    trainer_config_helpers/optimizers.rst
-    trainer_config_helpers/data_sources.rst
-    trainer_config_helpers/layers.rst
-    trainer_config_helpers/activations.rst 
-    trainer_config_helpers/poolings.rst
-    trainer_config_helpers/networks.rst
-    trainer_config_helpers/evaluators.rst
-    trainer_config_helpers/attrs.rst
-
-
-Applications API
----------------
-
-..  toctree::
-    :maxdepth: 1
-
-    predict/swig_py_paddle_en.rst
--- a/doc/api/v1/predict/src/predict_sample.py
+++ b/doc/api/v1/predict/src/predict_sample.py
-# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-from py_paddle import swig_paddle, DataProviderConverter
-from paddle.trainer.PyDataProvider2 import dense_vector
-from paddle.trainer.config_parser import parse_config
-
-TEST_DATA = [[[
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.215686, 0.533333, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.67451, 0.992157, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0.070588, 0.886275, 0.992157, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.192157,
-    0.070588, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.670588, 0.992157,
-    0.992157, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.117647, 0.933333, 0.858824, 0.313725,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.090196, 0.858824, 0.992157, 0.831373, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0.141176, 0.992157, 0.992157, 0.611765, 0.054902, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0.258824, 0.992157, 0.992157, 0.529412, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0.368627, 0.992157, 0.992157, 0.419608, 0.003922, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0.094118, 0.835294, 0.992157, 0.992157, 0.517647, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0.603922, 0.992157, 0.992157, 0.992157, 0.603922,
-    0.545098, 0.043137, 0, 0, 0, 0, 0, 0, 0, 0.447059, 0.992157, 0.992157,
-    0.956863, 0.062745, 0, 0, 0, 0, 0, 0, 0, 0, 0.011765, 0.666667, 0.992157,
-    0.992157, 0.992157, 0.992157, 0.992157, 0.745098, 0.137255, 0, 0, 0, 0, 0,
-    0.152941, 0.866667, 0.992157, 0.992157, 0.521569, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0.070588, 0.992157, 0.992157, 0.992157, 0.803922, 0.352941, 0.745098,
-    0.992157, 0.945098, 0.317647, 0, 0, 0, 0, 0.580392, 0.992157, 0.992157,
-    0.764706, 0.043137, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.070588, 0.992157, 0.992157,
-    0.776471, 0.043137, 0, 0.007843, 0.27451, 0.882353, 0.941176, 0.176471, 0,
-    0, 0.180392, 0.898039, 0.992157, 0.992157, 0.313725, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0.070588, 0.992157, 0.992157, 0.713725, 0, 0, 0, 0, 0.627451,
-    0.992157, 0.729412, 0.062745, 0, 0.509804, 0.992157, 0.992157, 0.776471,
-    0.035294, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.494118, 0.992157, 0.992157,
-    0.968627, 0.168627, 0, 0, 0, 0.423529, 0.992157, 0.992157, 0.364706, 0,
-    0.717647, 0.992157, 0.992157, 0.317647, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0.533333, 0.992157, 0.984314, 0.945098, 0.603922, 0, 0, 0, 0.003922,
-    0.466667, 0.992157, 0.988235, 0.976471, 0.992157, 0.992157, 0.788235,
-    0.007843, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.686275, 0.882353, 0.364706, 0,
-    0, 0, 0, 0, 0, 0.098039, 0.588235, 0.992157, 0.992157, 0.992157, 0.980392,
-    0.305882, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.101961, 0.67451, 0.321569,
-    0, 0, 0, 0, 0, 0, 0, 0.105882, 0.733333, 0.976471, 0.811765, 0.713725, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.65098, 0.992157, 0.321569, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0.25098, 0.007843, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1,
-    0.94902, 0.219608, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0.968627, 0.764706, 0.152941, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.498039, 0.25098, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0
-]], [[
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0.298039, 0.333333, 0.333333, 0.333333, 0.337255,
-    0.333333, 0.333333, 0.109804, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0.027451, 0.223529, 0.776471, 0.964706, 0.988235, 0.988235, 0.988235,
-    0.992157, 0.988235, 0.988235, 0.780392, 0.098039, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0.14902, 0.698039, 0.988235, 0.992157, 0.988235, 0.901961,
-    0.87451, 0.568627, 0.882353, 0.976471, 0.988235, 0.988235, 0.501961, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.188235, 0.647059, 0.988235, 0.988235,
-    0.745098, 0.439216, 0.098039, 0, 0, 0, 0.572549, 0.988235, 0.988235,
-    0.988235, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.2, 0.933333, 0.992157,
-    0.941176, 0.247059, 0, 0, 0, 0, 0, 0, 0.188235, 0.898039, 0.992157,
-    0.992157, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.039216, 0.639216, 0.933333,
-    0.988235, 0.913725, 0.278431, 0, 0, 0, 0, 0, 0, 0, 0.113725, 0.843137,
-    0.988235, 0.988235, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.235294, 0.988235,
-    0.992157, 0.988235, 0.815686, 0.07451, 0, 0, 0, 0, 0, 0, 0, 0.333333,
-    0.988235, 0.988235, 0.552941, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.211765,
-    0.878431, 0.988235, 0.992157, 0.701961, 0.329412, 0.109804, 0, 0, 0, 0, 0,
-    0, 0, 0.698039, 0.988235, 0.913725, 0.145098, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0.188235, 0.890196, 0.988235, 0.988235, 0.745098, 0.047059, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0.882353, 0.988235, 0.568627, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.2,
-    0.933333, 0.992157, 0.992157, 0.992157, 0.447059, 0.294118, 0, 0, 0, 0, 0,
-    0, 0, 0, 0.447059, 0.992157, 0.768627, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0.623529, 0.988235, 0.988235, 0.988235, 0.988235, 0.992157, 0.47451, 0, 0,
-    0, 0, 0, 0, 0, 0.188235, 0.933333, 0.87451, 0.509804, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0.992157, 0.988235, 0.937255, 0.792157, 0.988235, 0.894118,
-    0.082353, 0, 0, 0, 0, 0, 0, 0.027451, 0.647059, 0.992157, 0.654902, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0.623529, 0.988235, 0.913725, 0.329412, 0.376471,
-    0.184314, 0, 0, 0, 0, 0, 0, 0.027451, 0.513725, 0.988235, 0.635294,
-    0.219608, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.196078, 0.929412, 0.988235,
-    0.988235, 0.741176, 0.309804, 0, 0, 0, 0, 0, 0, 0.529412, 0.988235,
-    0.678431, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.223529, 0.992157,
-    0.992157, 1, 0.992157, 0.992157, 0.992157, 0.992157, 1, 0.992157, 0.992157,
-    0.882353, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.023529,
-    0.478431, 0.654902, 0.658824, 0.952941, 0.988235, 0.988235, 0.988235,
-    0.992157, 0.988235, 0.729412, 0.278431, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0.196078, 0.647059, 0.764706, 0.764706, 0.768627,
-    0.580392, 0.047059, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-    0, 0, 0, 0, 0, 0, 0
-]]]
-
-
-def main():
-    conf = parse_config("./mnist_model/trainer_config.py", "")
-    print conf.data_config.load_data_args
-    network = swig_paddle.GradientMachine.createFromConfigProto(
-        conf.model_config)
-    assert isinstance(network, swig_paddle.GradientMachine)  # For code hint.
-    network.loadParameters("./mnist_model/")
-    converter = DataProviderConverter([dense_vector(784)])
-    inArg = converter(TEST_DATA)
-    print network.forwardTest(inArg)
-
-
-if __name__ == '__main__':
-    swig_paddle.initPaddle("--use_gpu=0")
-    main()
--- a/doc/api/v1/predict/swig_py_paddle_cn.rst
+++ b/doc/api/v1/predict/swig_py_paddle_cn.rst
-.. _api_swig_py_paddle:
-
-基于Python的预测
-================
-
-预测流程
--------
-
-PaddlePaddle使用swig对常用的预测接口进行了封装，通过编译会生成py_paddle软件包，安装该软件包就可以在python环境下实现模型预测。可以使用python的 ``help()`` 函数查询软件包相关API说明。
-
-基于Python的模型预测，主要包括以下五个步骤。
-
-1. 初始化PaddlePaddle环境
-
-   在程序开始阶段，通过调用 ``swig_paddle.initPaddle()`` 并传入相应的命令行参数初始化PaddlePaddle。
-
-2. 解析模型配置文件
-   
-   初始化之后，可以通过调用 ``parse_config()`` 解析训练模型时用的配置文件。注意预测数据通常不包含label, 同时预测网络通常直接输出最后一层的结果而不是像训练网络一样再接一层cost layer，所以一般需要对训练用的模型配置文件稍作相应修改才能在预测时使用。
-
-3. 构造paddle.GradientMachine
-  
-   通过调用 ``swig_paddle.GradientMachine.createFromConfigproto()`` 传入上一步解析出来的模型配置就可以创建一个 ``GradientMachine``。
-
-4. 准备预测数据
-  
-   swig_paddle中的预测接口的参数是自定义的C++数据类型，py_paddle里面提供了一个工具类 ``DataProviderConverter`` 可以用于接收和PyDataProvider2一样的输入数据并转换成预测接口所需的数据类型。
-
-5. 模型预测
-  
-   通过调用 ``forwardTest()`` 传入预测数据，直接返回计算结果。
-
-
-预测Demo
--------
-
-如下是一段使用mnist model来实现手写识别的预测代码。完整的代码见 ``src_root/doc/ui/predict/predict_sample.py`` 。mnist model可以通过 ``src_root\demo\mnist`` 目录下的demo训练出来。
-
-..  literalinclude:: src/predict_sample.py
-    :language: python
-    :lines: 15-18,121-136
-
-
-Demo预测输出如下，其中value即为softmax层的输出。由于TEST_DATA包含两条预测数据，所以输出的value包含两个向量 。
-
-..  code-block:: text
-
-    [{'id': None, 'value': array(
-      [[  5.53018653e-09,   1.12194102e-05,   1.96644767e-09,
-          1.43630644e-02,   1.51111044e-13,   9.85625684e-01,
-          2.08823112e-10,   2.32777140e-08,   2.00186201e-09,
-          1.15501715e-08],
-       [  9.99982715e-01,   1.27787406e-10,   1.72296313e-05,
-          1.49316648e-09,   1.36540484e-11,   6.93137714e-10,
-          2.70634608e-08,   3.48565123e-08,   5.25639710e-09,
-          4.48684503e-08]], dtype=float32)}]
-
-
--- a/doc/api/v1/predict/swig_py_paddle_en.rst
+++ b/doc/api/v1/predict/swig_py_paddle_en.rst
-Python Prediction
-==================
-
-PaddlePaddle offers a set of clean prediction interfaces for python with the help of
-SWIG. The main steps of predict values in python are:
-
-* Parse training configurations
-* Construct GradientMachine
-* Prepare data
-* Predict
-
-Here is a sample python script that shows the typical prediction process for the
-MNIST classification problem. A complete sample code could be found at
-:code:`src_root/doc/ui/predict/predict_sample.py`.
-
-..  literalinclude:: src/predict_sample.py
-    :language: python
-    :lines: 15-18,90-100,101-104
-
-The module that does the most of the job is py_paddle.swig_paddle, it's
-generated by SWIG and has complete documents, for more details you can use
-python's :code:`help()` function. Let's walk through the above python script:
-
-* At the beginning, use :code:`swig_paddle.initPaddle()` to initialize
-  PaddlePaddle with command line arguments, for more about command line arguments
-  see :ref:`cmd_detail_introduction` .
-* Parse the configuration file that is used in training with :code:`parse_config()`.
-  Because data to predict with always have no label, and output of prediction work
-  normally is the output layer rather than the cost layer, so you should modify
-  the configuration file accordingly before using it in the prediction work.
-* Create a neural network with
-  :code:`swig_paddle.GradientMachine.createFromConfigproto()`, which takes the
-  parsed configuration :code:`conf.model_config` as argument. Then load the
-  trained parameters from the model with :code:`network.loadParameters()`.
-* Create a data converter object of utility class :code:`DataProviderConverter`.
-    - Note: As swig_paddle can only accept C++ matrices, we offer a utility
-      class DataProviderConverter that can accept the same input data with
-      PyDataProvider2, for more information please refer to document
-      of :ref:`api_pydataprovider2` .
-* Do the prediction with :code:`forwardTest()`, which takes the converted
-  input data and outputs the activations of the output layer.
-
-Here is a typical output:
-
-..  code-block:: text
-
-    [{'id': None, 'value': array([[  5.53018653e-09,   1.12194102e-05,   1.96644767e-09,
-          1.43630644e-02,   1.51111044e-13,   9.85625684e-01,
-          2.08823112e-10,   2.32777140e-08,   2.00186201e-09,
-          1.15501715e-08],
-       [  9.99982715e-01,   1.27787406e-10,   1.72296313e-05,
-          1.49316648e-09,   1.36540484e-11,   6.93137714e-10,
-          2.70634608e-08,   3.48565123e-08,   5.25639710e-09,
-          4.48684503e-08]], dtype=float32)}]
-
-:code:`value` is the output of the output layer, each row represents result of
-the corresponding row in the input data, each element represents activation of
-the corresponding neuron in the output layer.
-
--- a/doc/api/v2/config/layer.rst
+++ b/doc/api/v2/config/layer.rst
@@ -252,6 +252,11 @@ first_seq
 ..  autoclass:: paddle.v2.layer.first_seq
    :noindex:

+sub_seq
+---------
+..  autoclass:: paddle.v2.layer.sub_seq
+    :noindex:
+
 concat
 ------
 ..  autoclass:: paddle.v2.layer.concat
@@ -467,7 +472,7 @@ lambda_cost
    :noindex:

 square_error_cost
--------
+-----------------
 ..  autoclass:: paddle.v2.layer.square_error_cost
    :noindex:

@@ -533,7 +538,7 @@ Miscs
 =====

 dropout
--------------
+--------
 ..  autoclass:: paddle.v2.layer.dropout
    :noindex:


--- a/doc/api/v2/fluid.rst
+++ b/doc/api/v2/fluid.rst
@@ -15,4 +15,4 @@ Fluid
    fluid/param_attr.rst
    fluid/profiler.rst
    fluid/regularizer.rst
-
+    fluid/io.rst
--- a/doc/api/v2/fluid/io.rst
+++ b/doc/api/v2/fluid/io.rst
+===========
+IO
+===========
+
+
+
+is_parameter
+-----------
+..  autofunction:: paddle.v2.fluid.io.is_parameter
+    :noindex:
--- a/doc/api/v2/fluid/layers.rst
+++ b/doc/api/v2/fluid/layers.rst
@@ -18,18 +18,28 @@ dynamic_lstm
 ..  autofunction:: paddle.v2.fluid.layers.dynamic_lstm
    :noindex:

+dynamic_lstmp
+-------------
+..  autofunction:: paddle.v2.fluid.layers.dynamic_lstmp
+    :noindex:
+
+dynamic_gru
+-----------
+..  autofunction:: paddle.v2.fluid.layers.dynamic_gru
+    :noindex:
+
 data
---------
+----
 ..  autofunction:: paddle.v2.fluid.layers.data
    :noindex:

 mean
---------
+----
 ..  autofunction:: paddle.v2.fluid.layers.mean
    :noindex:

 mul
---------
+---
 ..  autofunction:: paddle.v2.fluid.layers.mul
    :noindex:

@@ -38,6 +48,16 @@ elementwise_add
 ..  autofunction:: paddle.v2.fluid.layers.elementwise_add
    :noindex:

+elementwise_sub
+---------------
+..  autofunction:: paddle.v2.fluid.layers.elementwise_sub
+    :noindex:
+
+elementwise_mul
+---------------
+..  autofunction:: paddle.v2.fluid.layers.elementwise_mul
+    :noindex:
+
 elementwise_div
 ---------------
 ..  autofunction:: paddle.v2.fluid.layers.elementwise_div
@@ -45,13 +65,13 @@ elementwise_div


 dropout
---------
+-------
 ..  autofunction:: paddle.v2.fluid.layers.dropout
    :noindex:


 reshape
---------
+--------
 ..  autofunction:: paddle.v2.fluid.layers.reshape
    :noindex:

@@ -68,12 +88,6 @@ scale
    :noindex:


-reshape
---------
-..  autofunction:: paddle.v2.fluid.layers.reshape
-    :noindex:
-
-
 transpose
 ---------
 ..  autofunction:: paddle.v2.fluid.layers.transpose
@@ -81,67 +95,67 @@ transpose


 sigmoid_cross_entropy_with_logits
---------
+---------------------------------
 ..  autofunction:: paddle.v2.fluid.layers.esigmoid_cross_entropy_with_logits
    :noindex:


 cast
---------
+----
 ..  autofunction:: paddle.v2.fluid.layers.cast
    :noindex:


 concat
---------
+-------
 ..  autofunction:: paddle.v2.fluid.layers.concat
    :noindex:


 sums
---------
+----
 ..  autofunction:: paddle.v2.fluid.layers.sums
    :noindex:


 linear_chain_crf
---------
+----------------
 ..  autofunction:: paddle.v2.fluid.layers.linear_chain_crf
    :noindex:


 assign
---------
+-------
 ..  autofunction:: paddle.v2.fluid.layers.embedding
    :noindex:


 split_lod_tensor
---------
+----------------
 ..  autofunction:: paddle.v2.fluid.layers.split_lod_tensor
    :noindex:


 merge_lod_tensor
---------
+----------------
 ..  autofunction:: paddle.v2.fluid.layers.merge_lod_tensor
    :noindex:

 cos_sim
---------
+--------
 ..  autofunction:: paddle.v2.fluid.layers.cos_sim
    :noindex:


 cross_entropy
---------
+-------------
 ..  autofunction:: paddle.v2.fluid.layers.cross_entropy
    :noindex:



 square_error_cost
---------
+-----------------
 ..  autofunction:: paddle.v2.fluid.layers.square_error_cost
    :noindex:

@@ -153,74 +167,80 @@ accuracy


 sequence_conv
---------
+-------------
 ..  autofunction:: paddle.v2.fluid.layers.sequence_conv
    :noindex:


 conv2d
---------
+------
 ..  autofunction:: paddle.v2.fluid.layers.conv2d
    :noindex:


 sequence_pool
---------
+-------------
 ..  autofunction:: paddle.v2.fluid.layers.sequence_pool
    :noindex:


+sequence_first_step
+-------------------
+..  autofunction:: paddle.v2.fluid.layers.sequence_first_step
+    :noindex:
+
+
+sequence_last_step
+------------------
+..  autofunction:: paddle.v2.fluid.layers.sequence_last_step
+    :noindex:
+
+
 pool2d
---------
+------
 ..  autofunction:: paddle.v2.fluid.layers.pool2d
    :noindex:


 batch_norm
---------
+----------
 ..  autofunction:: paddle.v2.fluid.layers.batch_norm
    :noindex:


 beam_search_decode
---------
+------------------
 ..  autofunction:: paddle.v2.fluid.layers.beam_search_decode
    :noindex:


-lstm
---------
-..  autofunction:: paddle.v2.fluid.layers.lstm
-    :noindex:
-
-
 lod_rank_table
---------
+--------------
 ..  autofunction:: paddle.v2.fluid.layers.lod_rank_table
    :noindex:


 max_sequence_len
---------
+----------------
 ..  autofunction:: paddle.v2.fluid.layers.max_sequence_len
    :noindex:


 topk
---------
+-----
 ..  autofunction:: paddle.v2.fluid.layers.topk
    :noindex:


 lod_tensor_to_array
---------
+-------------------
 ..  autofunction:: paddle.v2.fluid.layers.lod_tensor_to_array
    :noindex:



 array_to_lod_tensor
---------
+-------------------
 ..  autofunction:: paddle.v2.fluid.layers.array_to_lod_tensor
    :noindex:

@@ -228,26 +248,26 @@ array_to_lod_tensor


 fill_constant
---------
+-------------
 ..  autofunction:: paddle.v2.fluid.layers.fill_constant
    :noindex:



 fill_constant_batch_size_like
---------
+-----------------------------
 ..  autofunction:: paddle.v2.fluid.layers.fill_constant_batch_size_like
    :noindex:


 ones
---------
+----
 ..  autofunction:: paddle.v2.fluid.layers.ones
    :noindex:


 zeros
---------
+-----
 ..  autofunction:: paddle.v2.fluid.layers.zeros
    :noindex:

@@ -259,14 +279,14 @@ increment


 array_write
---------
+-----------
 ..  autofunction:: paddle.v2.fluid.layers.array_write
    :noindex:



 create_array
---------
+------------
 ..  autofunction:: paddle.v2.fluid.layers.create_array
    :noindex:

@@ -278,25 +298,249 @@ less_than


 array_read
---------
+----------
 ..  autofunction:: paddle.v2.fluid.layers.array_read
    :noindex:


 shrink_memory
---------
+--------------
 ..  autofunction:: paddle.v2.fluid.layers.shrink_memory
    :noindex:


 array_length
---------
+-------------
 ..  autofunction:: paddle.v2.fluid.layers.array_length
    :noindex:


 conv2d_transpose
---------
+----------------
 ..  autofunction:: paddle.v2.fluid.layers.conv2d_transpose
    :noindex:

+
+sequence_expand
+---------------
+..  autofunction:: paddle.v2.fluid.layers.sequence_expand
+    :noindex:
+
+
+gru_unit
+--------
+..  autofunction:: paddle.v2.fluid.layers.gru_unit
+    :noindex:
+
+
+lstm_unit
+---------
+..  autofunction:: paddle.v2.fluid.layers.lstm_unit
+    :noindex:
+
+
+sequence_softmax
+----------------
+..  autofunction:: paddle.v2.fluid.layers.sequence_softmax
+    :noindex:
+
+
+reduce_sum
+----------
+..  autofunction:: paddle.v2.fluid.layers.reduce_sum
+    :noindex:
+
+
+reduce_mean
+-----------
+..  autofunction:: paddle.v2.fluid.layers.reduce_mean
+    :noindex:
+
+
+reduce_max
+----------
+..  autofunction:: paddle.v2.fluid.layers.reduce_max
+    :noindex:
+
+
+reduce_min
+----------
+..  autofunction:: paddle.v2.fluid.layers.reduce_min
+    :noindex:
+
+
+split
+-----
+..  autofunction:: paddle.v2.fluid.layers.split
+    :noindex:
+
+
+matmul
+------
+..  autofunction:: paddle.v2.fluid.layers.matmul
+    :noindex:
+
+logsigmoid
+----------
+..  autofunction:: paddle.v2.fluid.layers.logsigmoid
+    :noindex:
+
+exp
+---
+..  autofunction:: paddle.v2.fluid.layers.exp
+    :noindex:
+
+relu
+----
+..  autofunction:: paddle.v2.fluid.layers.relu
+    :noindex:
+
+tanh
+----
+..  autofunction:: paddle.v2.fluid.layers.tanh
+    :noindex:
+
+tanh_shrink
+-----------
+..  autofunction:: paddle.v2.fluid.layers.tanh_shrink
+    :noindex:
+
+softshrink
+----------
+..  autofunction:: paddle.v2.fluid.layers.softshrink
+    :noindex:
+
+sqrt
+----
+..  autofunction:: paddle.v2.fluid.layers.sqrt
+    :noindex:
+
+abs
+----
+..  autofunction:: paddle.v2.fluid.layers.abs
+    :noindex:
+
+ceil
+----
+..  autofunction:: paddle.v2.fluid.layers.ceil
+    :noindex:
+
+floor
+-----
+..  autofunction:: paddle.v2.fluid.layers.floor
+    :noindex:
+
+round
+-----
+..  autofunction:: paddle.v2.fluid.layers.round
+    :noindex:
+
+reciprocal
+----------
+..  autofunction:: paddle.v2.fluid.layers.reciprocal
+    :noindex:
+
+log
+---
+..  autofunction:: paddle.v2.fluid.layers.log
+    :noindex:
+
+square
+------
+..  autofunction:: paddle.v2.fluid.layers.square
+    :noindex:
+
+softplus
+--------
+..  autofunction:: paddle.v2.fluid.layers.softplus
+    :noindex:
+
+softsign
+---------
+..  autofunction:: paddle.v2.fluid.layers.softsign
+    :noindex:
+
+brelu
+-----
+..  autofunction:: paddle.v2.fluid.layers.brelu
+    :noindex:
+
+leaky_relu
+----------
+..  autofunction:: paddle.v2.fluid.layers.leaky_relu
+    :noindex:
+
+soft_relu
+---------
+..  autofunction:: paddle.v2.fluid.layers.soft_relu
+    :noindex:
+
+elu
+----
+..  autofunction:: paddle.v2.fluid.layers.elu
+    :noindex:
+
+relu6
+-----
+..  autofunction:: paddle.v2.fluid.layers.relu6
+    :noindex:
+
+pow
+----
+..  autofunction:: paddle.v2.fluid.layers.pow
+    :noindex:
+
+hard_shrink
+-----------
+..  autofunction:: paddle.v2.fluid.layers.hard_shrink
+    :noindex:
+
+thresholded_relu
+----------------
+..  autofunction:: paddle.v2.fluid.layers.thresholded_relu
+    :noindex:
+
+hard_sigmoid
+-------------
+..  autofunction:: paddle.v2.fluid.layers.hard_sigmoid
+    :noindex:
+
+swish
+------
+..  autofunction:: paddle.v2.fluid.layers.swish
+    :noindex:
+
+im2sequence
+------
+..  autofunction:: paddle.v2.fluid.layers.im2sequence
+    :noindex:
+
+edit_distance
+---------------
+..  autofunction:: paddle.v2.fluid.layers.edit_distance_error
+    :noindex:
+
+ctc_greedy_decoder
+---------------
+..  autofunction:: paddle.v2.fluid.layers.ctc_greedy_decoder
+    :noindex:
+
+l2_normalize
+------------
+..  autofunction:: paddle.v2.fluid.layers.l2_normalize
+    :noindex:
+
+sequence_reshape
+----------------
+..  autofunction:: paddle.v2.fluid.layers.sequence_reshape
+    :noindex:
+
+row_conv
+--------
+..  autofunction:: paddle.v2.fluid.layers.row_conv
+    :noindex:
+
+multiplex
+---------
+..  autofunction:: paddle.v2.fluid.layers.multiplex
+    :noindex:
--- a/doc/api/v2/fluid/nets.rst
+++ b/doc/api/v2/fluid/nets.rst
@@ -3,20 +3,31 @@ Nets
 ===========

 simple_img_conv_pool
-----------
+--------------------
 ..  autofunction:: paddle.v2.fluid.nets.simple_img_conv_pool
    :noindex:


 img_conv_group
-----------
+---------------
 ..  autofunction:: paddle.v2.fluid.nets.img_conv_group
    :noindex:


 sequence_conv_pool
-----------
+------------------
 ..  autofunction:: paddle.v2.fluid.nets.sequence_conv_pool
    :noindex:


+glu
+---
+..  autofunction:: paddle.v2.fluid.nets.glu
+    :noindex:
+
+
+scaled_dot_product_attention
+----------------------------
+..  autofunction:: paddle.v2.fluid.nets.scaled_dot_product_attention
+    :noindex:
+
--- a/doc/api/v2/fluid/optimizer.rst
+++ b/doc/api/v2/fluid/optimizer.rst
@@ -18,7 +18,7 @@ SGDOptimizer


 MomentumOptimizer
-----------
+-----------------
 ..  automodule:: paddle.v2.fluid.optimizer
    :members: MomentumOptimizer
    :noindex:
@@ -26,14 +26,14 @@ MomentumOptimizer


 AdagradOptimizer
-----------
+----------------
 ..  automodule:: paddle.v2.fluid.optimizer
    :members: AdagradOptimizer
    :noindex:


 AdamOptimizer
-----------
+-------------
 ..  automodule:: paddle.v2.fluid.optimizer
    :members: AdamOptimizer
    :noindex:
@@ -47,7 +47,7 @@ AdamaxOptimizer


 DecayedAdagradOptimizer
-----------
+-----------------------
 ..  automodule:: paddle.v2.fluid.optimizer
    :members: DecayedAdagradOptimizer
    :noindex:

--- a/doc/api/v2/fluid/regularizer.rst
+++ b/doc/api/v2/fluid/regularizer.rst
@@ -3,14 +3,14 @@ Regularizer
 ===========

 WeightDecayRegularizer
-----------
+----------------------
 ..  automodule:: paddle.v2.fluid.regularizer
    :members: WeightDecayRegularizer
    :noindex:


 L2DecayRegularizer
-----------
+------------------
 ..  automodule:: paddle.v2.fluid.regularizer
    :members: L2DecayRegularizer
    :noindex:
@@ -18,7 +18,7 @@ L2DecayRegularizer


 L1DecayRegularizer
-----------
+-------------------
 ..  automodule:: paddle.v2.fluid.regularizer
    :members: L1DecayRegularizer


--- a/doc/design/backward.md
+++ b/doc/design/backward.md
+# Backward Building
+
+## Motivation
+
+In Neural Network, most models are solved by the backpropagation algorithm(known as **BP**) at present. Technically, BP calculates the gradient of the loss function, then propagates it back through the networks following the chain rule. However, when configuring the model structure, users do not need to define the backward part. So a mechanism is required by the framework which can complete the model's backward part automatically according to the given forward part.
+
+When implementing a specific `op`, the developer is also asked to implement its backward version, called `grad_op`. A `grad_op` takes gradients of its corresponding `op`'s outputs, and calculate gradients of the `op`'s inputs. During the building of a model's backward part, the framework creates each forward `op`'s `grad_op`, and then string them together in reverse order of forwarding part. In this way, gradients spread from the end to the beginning of the model, in another word, from the loss to parameters.
+
+## Challenges
+
+The motivation of backward building is apparent. However, implementation it correctly is not so easy. In the **Fluid** design, a deep learning model is described by `Program`, `Block`, `Op` and `Variable`. The `Block` itself can be nested. It means that the `op`s and `variable`s are scattered across different blocks rather than all be gathered in a single graph. Our backward building algorithm shall visit blocks in recursive order and be able to insert `grad_op`s and new created `variable`s into the right place. 
+
+## Usage
+
+Although the whole algorithm is comprised of many functions, only one is exposed as API:
+
+```python
+def append_backward(loss, parameter_list=None, no_grad_set=None):
+    """
+    Append backward part to main_program
+
+    Args:
+        loss(Variable): The variable generated by the cost function.
+        parameter_list(list): Parameters that need to be updated by optimizers.
+            If None, it means all parameters need to be updated.
+
+        no_grad_set(set): Variables that have no gradients in Block 0. 
+            If None, the set will be generated inside the function and 
+            contains all variables with `step_gradient=True` from all blocks.
+        
+    Return:
+        (list[Variable]): list of (parameters, gradients) pair.
+    """
+```
+
+By invoking this API, the framework appends backward part of the program where the `loss` is. It takes three arguments. `loss` means the final loss value. It must be a scalar and is usually the output of the loss layer. It is also where the gradient generated and backpropagation starts. `parameter_list` marks all parameters needs updating. If it's `None`, all parameter will be updated by optimizers. `no_grad_set` marks variables without gradient. if all outputs of some `grad_op` are in `no_grad_set`, the `grad_op` will not be run.
+
+This API will be invoked automatically before optimizer building. 
+As a result, in most cases, users do not need to invoke the API by themselves to append backward part.
+
+## Implementation
+
+The implementation of backward building algorithm is in `backward.py` file. The whole algorithm can be divided into two independent parts: creating `grad_op`s and creating new variables. 
+
+### Creating `grad_op`s
+
+The creating of `grad_op`s is implemented by:
+
+```python
+def _append_backward_ops_(target,
+                          block,
+                          target_block,
+                          no_grad_dict,
+                          grad_to_var):
+    """
+    Create all grad ops, and insert them into given block
+
+    Args:
+        target(Variable): the target variable of forward pass
+        block(Block): the block where forward ops are
+        target_block(Block): the block which is going to hold new generated grad ops
+        no_grad_dict(dict): 
+            key(int)  block index
+            val(set) a set of varibale names. These varibales have no gradient
+        grad_to_var(dict)(output argument):
+            key(str): grad variable name
+            val(str): corresponding forward variable name
+    """
+```
+
+Given a `block`, the function will traverses all `op`s in this block in reverse order, gets corresponding `grad_op` from the C++ core via `core.get_grad_op_desc()`, then append it to `target_block`. 
+
+However, some specific `op`(e.g. `while_op`, `if_else_op`) can hold its own sub-block. For these sub-blocks contains `op`s as well, the `grad_op` creating should be recursive.
+
+During the reverse traversal, we check each `op` whether it has an attribute named `sub_block`. If so, it means there is a sub-block and we need to deal with it first. After creating a new block whose father is the one in `op`'s attribute, we invoke `_append_backward_ops_()` recursively, assigning the new block to parameter `target_block` and the one in `op`'s attribute to `block`. The *pseudo-code* shows this process:
+
+```
+******* pseudo-code ********
+for op in reversed(block.ops):
+    if op has an attribute named 'sub_block':
+        Get the sub-block(`s_block`) from op's attribute.
+        Create a new block(`grad_s_block`), whose father is `s_block`.
+        Invoke _append_backward_ops_(), with `block=s_block` and `target_block=grad_s_block`
+    
+    Invoke `core.get_grad_op_desc()` to get op's grad_op.
+    Insert name correspondings between variables and their gradients of the grad_op to grad_to_var
+    Assign grad_s_block to grad_op as it's 'sub_block' attribute.
+    Append grad_op to current target_block.
+```
+
+The first invoking of `_append_backward_ops_()` is initiated by `append_backward()`, in which parameters `block` and `target_block` are all assigned with root block(the block with index 0).
+
+### Corner Cases of `grad_op` Creating
+
+In the previous section, we show the regular process of `grad_op` creating. However, in some corner cases, the conventional algorithm is not enough to get the correct result and appending handling is required. These additional processes run after the algorithm mentioned above and do some special adjusts on its output `grad_op`s.
+
+#### Shared Variables
+
+If a variable is read by more than one `op` in the forward pass, its gradient is likely to be written by more than one `grad_op`s in the next backward pass. To make the gradient result being the sum of all `grad_op`s' outputs instead of the last running one, we assign each output with a temporary variable and then add a `sum_op` to add them up. 
+
+For the debug convenience, if the final gradient name is `w@GRAD`, it's corresponding temporary variables will be named as `w@GRAD@RENAME@0`, `w@GRAD@RENAME@1`...
+
+See function `_addup_repetitive_outputs_` in `backward.py` for implementation details.
+
+#### No Gradient Variables
+
+In our framework, variables can be marked as *no_gradient*, it means that the gradient of this variable is unnecessary and can be considered as zero in model training. Apparently, when all the outputs of some `grad_op` are marked as *no_gradient*, the `grad_op` itself can be skipped in backward pass. 
+
+Another situation is all the gradient inputs of some `grad_op` are marked as *no_gradient*, which means all of them can be considered as zeros. For `grad_op`s are in essence the propagation of gradients, all the outputs are definitely zeros when all gradient inputs are zeros. Therefore the `grad_op` can also be skipped.
+
+It should be noted that all these zero gradients still need to be creating and initialized by something, otherwise following `grad_op`s who take these gradients as inputs take the risk of using uninitialized memory. In our code, we employ `fill_zeros_like_op` to initialize them as all zeros. 
+
+This features are implemented in function `_remove_no_grad_branch_`. It checks new created `grad_op`s one-by-one, removes who can be skipped and inserts `fill_zeros_like_op` when its necessary. We can get the `no_grad_set` from the `_append_backward_ops_` argument `no_grad_dict` or generate it on the fly by scanning all variables' `no_gradient` attribute(True or False). 
+
+### Creating Backward Variables
+
+Up to now, we have completed all creating and adjusting jobs of `grad_op`s. However, backward variables have not been created. Now they are only represented by `grad_op`'s input and output arguments. The backward variable creating job will be done by:
+
+```python
+def _append_backward_vars_(block, 
+                           start_op_idx, 
+                           grad_to_var, 
+                           grad_info_map):
+    """
+    Create new variables required by backward pass.
+
+    Args:
+        block(Block): the block where new variables will be created
+        start_op_idx(int): Only variables required by ops in block.ops[start_op_idx : ] will be created
+        grad_to_var(dict):
+            key(str): grad variable name
+            val(str): corresponding forward variable name
+            In most cases, this dict is generated by _append_backward_ops_()
+        grad_info_map(dict)(output argument):
+            key(str): forward variable name
+            val(tuple): a tuple of (str, int), str is the corresponding grad name, int is the block index
+    """
+```
+
+Given a `block`, this function traverses all the `grad_op`s in it(The argument `start_op_idx` indicates where the grad_op sequence starts.) and creates all the uncreated outputs. The *pseudo-code* shows this process:
+
+```
+for op in block.ops[start_op_idx : ]:
+
+    if op has an attribute named 'sub_block':
+        Get the sub-block(`s_block`) from op's attribute.
+        Invoke _append_backward_vars_(), with `block=s_block`
+        
+    for var_name in op.all_output_names():
+        if block.has_var_recursive(var_name) or var_name is the name of empty variable:
+            continue
+        create a new variable named 'var_name' in block
+        if grad_to_var.has_key(var_name):
+            set grad_info_map[grad_to_var[var_name]] as a tuple of (var_name. block)
+            
+    do op's var type inference
+    do op's shape inference
+```
--- a/doc/design/block.md
+++ b/doc/design/block.md
@@ -202,8 +202,8 @@ This `OpDesc` value is in the `ops` field of the `BlockDesc` value representing

 During the generation of the Protobuf message, the Block should store VarDesc (the Protobuf message which describes Variable) and OpDesc (the Protobuf message which describes Operator).

-VarDesc in a block should have its name scope to avoid local variables affect parent block's name scope.
-Child block's name scopes should inherit the parent's so that OpDesc in child block can reference a VarDesc that stored in parent block. For example:
+VarDesc in a block should have its name scope to avoid local variables affecting parent block's name scope.
+Child block's name scopes should inherit the parent's so that OpDesc in child block can reference a VarDesc that is stored in the parent block. For example:

 ```python
 a = pd.Variable(shape=[20, 20])
@@ -291,10 +291,10 @@ public:
  }

  void Run(const framework::Scope& scope,
-           const platform::DeviceContext& dev_ctx) const override {
+           const platform::Place& place) const override {
    PADDLE_ENFORCE(symbols_ready_, "operators and variables should be created first.");
    for (auto& op : runtime_table_.ops()) {
-      op->Run(scope, dev_ctx);
+      op->Run(scope, place);
    }
  }


--- a/doc/design/ci_build_whl.png
+++ b/doc/design/ci_build_whl.png
--- a/doc/design/concurrent_programming.md
+++ b/doc/design/concurrent_programming.md
+# Design Doc: Concurrent Programming with Fluid
+
+With PaddlePaddle Fluid, users describe a program other than a model.  The program is a [`ProgramDesc`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/framework.proto) protobuf message. TensorFlow/MxNet/Caffe2 applications generate protobuf messages too, but their protobuf messages represent the model, a graph of operators, but not the program that trains/uses the model.   
+
+Many know that when we program TensorFlow, we can specify the device on which each operator runs.  This allows us to create a concurrent/parallel AI application.   An interesting questions is **how does a `ProgramDesc` represents a concurrent program?**  
+
+The answer relies on the fact that a `ProgramDesc` is similar to an abstract syntax tree (AST) that describes a program.  So users just program a concurrent program that they do with any concurrent programming language, e.g., [Go](https://golang.org).
+
+## An Analogy
+
+The following table compares concepts in Fluid and Go
+
+| Go | Fluid |
+|----|-------|
+|user-defined functions | [layers](https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/v2/fluid) |
+| control-flow and built-in functions | [intrinsics/operators](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/operators) |
+| goroutines, channels | [class ThreadPool](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/framework/thread_pool.h) |
+| runtime | [class Executor](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.h) |
+
+## An Example Concurrent Program
+
+To review all above concepts in an example, let us take a simple program and writes its distributed version.
+
+Suppose that we want to parallelize a naive Fluid program (written in Go and calling Fluid's Go binding) that multiplies two tensors.
+
+```go
+import "fluid"
+
+func paddlepaddle() {
+  X = fluid.read(...)
+  W = fluid.Tensor(...)
+  Y = fluid.mult(X, W)
+}
+```
+
+Please be aware that the Fluid's Go binding provides the default `main` function, which calls the `paddlepaddle` function, which, in this case, is defined in above program and creates the following `ProgramDesc` message.
+
+```protobuf
+message ProgramDesc {
+  block[0] = Block {
+    vars = [X, W, Y],
+    ops = [
+      read(output = X)
+      assign(input = ..., output = W)
+      mult(input = {X, W}, output = Y)
+    ],
+  }
+}
+```
+
+Then, the default `main` function calls `fluid.run()`, which creates an instance of the [`class Executor`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.h) and calls `Executor.Run(block[0])`, where `block[0]` is the first and only block defined in above `ProgramDesc` message.
+
+The default `main` function is defined as follows:
+
+```go
+func main() {
+  paddlepaddle()
+  fluid.run()
+}
+```
+
+## The Concurrent Version
+
+By parallelizing the above program, we could support very big tensor X by splitting into small pieces {x_1, x_2, ...} and sent each piece to worker process/node for parallel multiplication.
+
+In this case, we can write a transpiler that takes a `ProgramDesc` message that represents the above example program and outputs two `ProgramDesc` messages, one for running on the master process/node, and the other one for worker processes/nodes.
+
+### The Master Program
+
+The master program could look like the following:
+
+```protobuf
+message ProgramDesc {
+  block[0] = Block {
+    vars = [X, L, Y],
+    ops = [
+      read(output = X)
+      kube_get_workers_addrs(output = L)
+      Y = tensor_array(len(L))
+      parallel_for(input = X, output = Y, 
+                   attrs = {L, block_id(1)}) # referring to block 1
+    ]
+  }
+  
+  block[1] = Block {
+    parent = 0,
+    vars = [x, y, index],
+    ops = [
+      slice(input = [X, index], output = x) # index is initialized by parallel_for
+      send(input = x, attrs = L[index])
+      recv(outputs = y, attrs = L[index])
+      assign(input = y, output = Y[index])
+    ]
+  }
+}
+```
+
+The equivalent Fluid program (calling the Go binding) is:
+
+```go
+func main() {  //// block 0
+  X = fluid.read(...)
+  L = fluid.k8s.get_worker_addrs()
+  Y = fluid.tensor_array(len(L))
+  fluid.parallel_for(X, L, 
+                     func(index int) {  //// block 1
+                       x = X[index]
+                       fluid.send(L[index], x)
+                       y = fluid.recv(L[index])
+                       Y[index] = y
+                     })
+}
+```
+
+An explanation of the above program:
+
+- `fluid.k8s` is a package that provides access to Kubernetes API.  
+- `fluid.k8s.get_worker_addrs` returns the list of IP and ports of all pods of the current job except for the current one (the master pod).  
+- `fluid.tensor_array` creates a [tensor array](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/lod_tensor_array.h).  `fluid.parallel_for` creates a `ParallelFor` intrinsic, which, when executed, 
+
+  1. creates `len(L)` scopes, each for the concurrent running of the sub-block (block 1 in this case), and initializes a variable named "index" in the scope to an integer value in the range `[0, len(L)-1]`, and
+  2. creates `len(L)` threads by calling into the `ThreadPool` singleton, each thread  
+     1. creates an Executor instance, and
+     2. calls `Executor.Run(block)`, where `block` is block 1 as explained above.
+1. Please be aware that block 1 is a sub-block of block 0, so ops in block 1 could refer to variables defined in block 0.
+
+### The Worker Program
+
+The worker program looks like
+
+```go
+func main() {
+  W = Tensor(...)
+  x = fluid.listen_and_do(
+        fluid.k8s.self_addr(),
+        func(input Tensor) {
+          output = fluid.mult(input, W)
+        })
+}
+```
+
+where
+
+- `fluid.listen_and_do` creates a `ListenAndDo` intrinsic, which, when executed,
+  1. listens on the current pod's IP address, as returned by `fliud.k8s.self_addr()`,
+  2. once a connection is established,
+     1. creates a scope of two parameters, "input" and "output",
+     2. reads a [Fluid variable](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/variable.h) and saves it into "input",
+     3. creates an Executor instance and calls `Executor.Run(block)`, where the block is generated by running the lambda specified as the second parameter of `fluid.listen_and_do`.
+
+## Summarization
+
+From the above example, we see that:
+
+1. Fluid enables the imperative programming paradigm by:
+   1. letting users describe a program, but not a model (a sequence of layers, or a graph of operators), and
+   2. call the `fluid.run` function that runs the program implicitly.
+1. The program is described as a `ProgramDesc` protobuf message.
+2. Function `Executor.Run` takes a block, instead of a `ProgramDesc`, as its parameter.
+3. `fluid.run` calls `Executor.Run` to run the first block in the `ProgramDesc` message.
+4. `Executor.Run`'s implementation is extremely simple -- it doesn't plan the execution nor create threads; instead, it runs on the current thread and execute intrinsics/operators' `Run` method sequentially as they appear in the `Block.ops` array.
+5. Intrinsics/operators' `Run` method might create threads.  For example, the `ListenAndDo` operator creates a thread to handle each incoming request.
+6. Threads are not necessarily OS thread; instead, they could be [green threads](https://en.wikipedia.org/wiki/Green_threads) managed by ThreadPool.  Multiple green threads might run on the same OS thread.  An example green threads is Go's [goroutines](https://tour.golang.org/concurrency/1).
--- a/doc/design/csp.md
+++ b/doc/design/csp.md
+# Design Doc: CSP in PaddlePaddle Fluid
+
+## Motivation
+
+Concurrent programming is important for deep learning.  Few example applications are:
+
+1.  The main thread keeps reading the next mini-batch while another thread uses the GPU for computing.
+2.  The main thread performs the computation while another thread uploads the local gradients from each trainer to the parameter server.
+
+Most DL systems, including TensorFlow, Caffe2, and MxNet, can asynchronously execute operators in a graph. However, Fluid doesn't have the concept of a graph at all, as the design goal of Fluid is that of a programming language.
+
+## Concurrent Programming Models
+
+There were many concurrent programming models, implemented in various forms:
+
+| concurrent programming model | implementation |
+|-----|-----|
+| mutex | types and functions in standard libraries |
+| semaphore | types and functions in standard libraries |
+| communicating sequential processes (CSP) | Go programming language |
+| actor model | Erlang programming language |
+| message passing | MPI |
+| bulk synchronous parallel (BSP) | Pregel distributed programming framework |
+
+Since Fluid was designed to be a programming language, we would like to implement CSP in Fluid.
+
+### CSP v.s. Actor Model
+
+A well-known implementation of Actor Model is the Erlang programming language.  In Actor Model, *processes* could send messages to another process and receive messages from another process given the process IDs.  We can find the three ingredients, process with ID, send, and recv, in MPI too.  Indeed, we can rewrite Erlang programs in Python + MPI with possibly fewer lines of code.  Our concern with Actor Model is that it doesn't seem reasonable to implement process management in a programming language's runtime library; instead, it should be the operating systems' responsibility to manage processes and libraries like MPI for send/recv.
+
+## CSP in Fluid
+
+Fluid has two fundamental control-flows: *if-else* and *while*.  If we are to implement CSP, we need the following:
+
+1. a new data type: *channel* and operators *send* and *recv*,
+1. *goroutine* or thread, and
+1. a new control-flow: select.
+
+We also need Python wrappers for the above components.
+
+The type *channel* is conceptually the blocking queue.  In Go, its implemented is a [blocking circular queue](https://github.com/golang/go/blob/68ce117cf17b8debf5754bfd476345779b5b6616/src/runtime/chan.go#L31-L50), which supports send and recv.
+
+The `select` operation has been in OS kernels long before Go language.  All Unix kernels implement system calls *poll* and *select*.  They monitor multiple file descriptors to see if I/O is possible on any of them.  This takes O(N) time.  Since Linux 2.6, a new system call, *epoll*, can do the same in O(1) time.  In BSD systems, there is a similar system call *kqueue*.  Go's Linux implementation uses epoll.
+
+It might be a good idea to implement Fluid's select using epoll too.  In this design doc, we start from the O(N) way, so we could focus on Python binding and the syntax.
+
+### Type Channel
+
+Fluid supports many data types:
+
+1. Tensor,
+1. Row-sparse Tensor
+1. LoD Tensor,
+1. Tensor array, etc
+
+Each data type is registered in the [`framework.proto`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/framework.proto#L117-L127) as an enum value.  To add a new type channel, we need to add a new type enum.
+
+To expose a C++ type to Python, we need to edit the [`pybind.cc`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/pybind/pybind.cc) file.  [Here](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/pybind/pybind.cc#L120-L164) is an example how we expose C++ class LoDTensor.
+
+## Syntax Design
+
+### Create Channel
+
+In Go, we create a channel by specifying the element type and buffer size:
+
+```go
+ch  := make(chan int)       // a channel without buffer
+ch1 := make(chan int, 100)  // a channel that can buffer 100 ints.
+```
+
+In Fluid, we should be able to do the same:
+
+```python
+ch  = fluid.make_chan(dtype=INT)
+ch1 = fluid.make_chan(dtype=INT, 100)
+```
+
+In addition to that, we want channels that can hold more complex element types, e.g., Tensors of float16:
+
+```python
+ch = fluid.make_chan(dtype=Tensor, etype=float16)
+```
+
+or Tensors of Tensors of float16 etc.
+
+The point here is that we need a consistent way to compose types, like in C++ we can have `Tensor<Tensor<...<float16>...> >`.
+
+### Send and Recv
+
+### Select
+
+## Example Programs
+
+### 1. RPC between Trainers and Parameter Servers
+
+### 2. Concurrent Minibatch Loading
--- a/doc/design/refactor/distributed_architecture.md
+++ b/doc/design/refactor/distributed_architecture.md
@@ -52,8 +52,9 @@ The IR for PaddlePaddle after refactoring is called a `Block`, it specifies the

 The user can not directly specify the parameter update rule for the parameter server in the Python module, since the parameter server does not use the same computation definition as the trainer. Instead, the update rule is baked inside the parameter server. The user can not specify the update rule explicitly.

-This could be fixed by making the parameter server run the same computation definition as the trainer (the user's Python module). For a detailed explanation, refer to this document -
-[Design Doc: Operation Graph Based Parameter Server](./parameter_server.md)
+This could be fixed by making the parameter server also run an IR, which can be different to the trainer side
+For a detailed explanation, refer to this document -
+[Design Doc: Parameter Server](./parameter_server.md)

 ## Distributed Training Architecture

@@ -61,68 +62,111 @@ The revamped distributed training architecture can address the above discussed l

 <img src="src/distributed_architecture.png"/>

-The major components in the architecture are: *PaddlePaddle Python*, *PaddlePaddle converter* and *PaddlePaddle runtime*.
+The major components are: *Python API*, *Distribute Transpiler* and *Remote Executor*.

-### PaddlePaddle Python
+### Python API

-PaddlePaddle Python is the Python library that user's Python code invokes, to read the data. build the neural network topology, start training, etc.
+Python API is the Python library that user's Python code invokes, to read the data, build the neural network topology, and start training, etc.

 ```Python
-paddle.init()
-input = paddle.op.recordIO("/home/data/mnist.recordio") # file stored on the cluster
-img, label = input[0], input[1]
-hidden = paddle.layer.fc(input=img, size=200, act=paddle.activation.Tanh())
-prediction = paddle.layer.fc(input=img, size=10, act=paddle.activation.Softmax())
-cost = paddle.layer.classification_cost(input=prediction, label=label)
-optimizer = paddle.optimizer.SGD(cost, learning_rate=0.01)
-session = paddle.session.NewRemote(num_trainer=3, num_ps=2, GPU_per_trainer=1)
-for i in range(1000):
-	_, cost_val = session.eval(targets=[cost, optimizer])
-	print cost_val
+images = fluid.layers.data(name='pixel', shape=[1, 28, 28], dtype='float32')
+label = fluid.layers.data(name='label', shape=[1], dtype='int64')
+...
+predict = fluid.layers.fc(input=conv_pool_2, size=10, act="softmax")
+cost = fluid.layers.cross_entropy(input=predict, label=label)
+avg_cost = fluid.layers.mean(x=cost)
+optimizer = fluid.optimizer.Adam(learning_rate=0.01)
+optimizer.minimize(avg_cost)
+
+train_reader = paddle.batch(
+    paddle.reader.shuffle(
+        paddle.dataset.mnist.train(), buf_size=500),
+    batch_size=BATCH_SIZE)
+
+place = fluid.CPUPlace()
+exe = fluid.Executor(place)
+
+for pass_id in range(10):
+    for data in train_reader():
+        loss, acc = exe.run(trainer_prog,
+                            feed=feeder.feed(data),
+                            fetch_list=[avg_cost])
 ```

-The above code is what a typical Python trainer code is, the neural network topology is built using the helper functions such as `paddle.layer.fc`. Training is done by calling `session.eval` iteratively.
-
-#### session.eval
-
-As shown in the graph, `session.eval` sends the IR and the evaluation inputs or targets to the PaddlePaddle cluster for evaluation.
-The targets can be any variable in the computation graph. When the target is say, the `optimizer` variable, the neural network will be optimized once. When the target is the `cost` variable, `session.eval` returns the cost value. Based on what the target is, an appropriate action is taken.
-
-The Python `session` is a wrapper of the C++ `Session` class. For more information about `Session`, refer to this document - [Design Doc: Session](./session.md).
-
-### PaddlePaddle Converter
-
-The PaddlePaddle converter automatically converts the IR in the request (IR and evaluation inputs/targets) from PaddlePaddle Python to partitioned IRs and dispatches the new IRs and evaluation inputs/targets to different PaddlePaddle runtimes. Below are the steps that are followed :
-
-1. Add a `feed` OP that feeds the eval inputs, and a `fetch` OP that fetches the eval targets to the IR.
-
-2. Extract a new computation (sub)graph with the `feed` and `fetch` OPs as the boundary. The runtime does not need to run the OP that is not dependent on the `fetch` OP.
-
-3. Optimize the computation graph.
-
-4. Place the OPs in the graph onto different devices on different PaddlePaddle runtime according to a placement algorithm and the device constraints specified by the user.
-
-5. Partition the graph according to runtime boundaries and add `send` / `recv` OP pair on the runtime boundaries.
+The code above is a typical local training program, the "Training Program" is built using helper functions such as
+`fluid.layer.fc`. The training is done by calling `Executor.run`
+iteratively.
+
+For more details, the implementation of IR is [Program](../program.md), and `ProgramDesc` is the protobuf type.
+
+[Executor](../executor.md) simply runs the `ProgramDesc`. For local training you generally use
+`Executor` to run the program locally. For any kind of distributed training, you can use
+`RemoteExecutor` to specify desired distributed training method with some optional arguments.
+
+### Distributed Transpiler
+
+The Distributed Transpiler automatically converts the IR (in protobuf format) to partitioned IRs. Then
+the Remote Executor dispatches the new IRs to Remote Executors across the cluster.
+Below are the steps that are followed :
+
+1. User only need to change `Executor` to `RemoteExecutor` to change local program to distributed program.
+1. `RemoteExecutor` calls `Distributed Transpiler` to "transpile" user's program to several IRs representing a
+   distributed training program:
+   1. Parse configurations from `RemoteExecutor`.
+   1. Determine the type of distributed program, can be DataParallelism, ModelParallelism or Streaming.
+   1. Partition the `ProgramDesc` according to type and add `send` / `recv` OP pair on the boundaries. Take
+      DataParallelism type for example, it removes the optimization operators and add a `send` OP to the
+      "trainer" role, then add the optimization operators to the parameter server role within the `recv` OP.
+1. Dispatch the partitioned graph to different `RemoteExecutor` in the cluster.
+1. `RemoteExecutor` on each node run the received `ProgramDesc` utill the end.
+
+
+### RemoteExecutor
+
+As shown in the graph, `RemoteExecutor.run` sends the IR to the cluster for Execution.
+You can also use parameter `fetch_list` to interactively fetch variable back to local for
+log printing.
+
+The Python `RemoteExecutor` is derived from `Executor` class.
+
+```python
+exe = RemoteExecutor(
+    feed=feeder.feed(data),
+    fetch_list=[avg_cost],
+    job_desc=JobDesc(
+      jobname,
+      num_trainer,
+      num_pserver,
+      cpu_per_trainer,
+      gpu_per_trainer,
+      mem_per_trainer,
+      cpu_per_pserver,
+      mem_per_pserver
+    ))
+for data in train_reader():
+    loss, acc = exe.run(trainer_prog,
+                        feed=feeder.feed(data),
+                        fetch_list=[avg_cost])
+```

-6. Dispatch the partitioned graph to different PaddlePaddle runtimes.
+`JobDesc` object describe the distributed job resource specification to run on
+Cluster environment.

-7. PaddlePaddle runtimes with the `fetch` OP reports evaluation results back to the converter, the converter reports the evaluation results back to the PaddlePaddle Python.
+<img src="src/remote_executor.png" width="500" align="center" />

-The output IRs will be cached to optimize the conversion latency.
+`RemoteExecutor.run` sends the `ProgramDesc` and
+[TrainingJob](https://github.com/PaddlePaddle/cloud/blob/develop/doc/autoscale/README.md#training-job-resource)
+to a server in the cluster which executes `RemoteExecutor.listen`. This server is responsible
+to start the final Kubernetes Jobs to run the different role of `ProgramDesc` from `ConfigMap`.


-#### Placement Algorithm
+### Placement Algorithm

 Our first implementation will only support "trainer-parameter server" placement: the parameters, initializers, and optimizers are all placed on the PaddlePaddle runtimes with the parameter server role. Everything else will be placed on the PaddlePaddle runtimes with the trainer role. This has the same functionality as the "trainer-parameter server" architecture of PaddlePaddle v0.10.0, but is more generic and flexible.

 In the future, a more general placement algorithm should be implemented, which makes placements according to the input IR, and a model of device computation time and device communication time. Model parallelism requires the generic placement algorithm.


-### PaddlePaddle Runtime
-
-The PaddlePaddle runtime owns multiple devices (e.g., CPUs, GPUs) and runs the IR. The runtime does not need to do OP placement since it is already done by the converter.
-
-
 ### Local Training Architecture

 The local training architecture will be the same as the distributed training architecture, the difference is that everything runs locally, and there is just one PaddlePaddle runtime:
@@ -132,9 +176,18 @@ The local training architecture will be the same as the distributed training arc

 ### Training Data

-In PaddlePaddle v0.10.0, training data is typically read with a [data reader](../reader/README.md) from Python. This approach is no longer efficient when training in a distributed fashion since the Python process no longer runs on the same node with the trainer processes. The Python reader will need to read from the distributed filesystem (assuming it has the required access) and send to the trainers, doubling the network traffic.
-
-When doing distributed training, the user can still use Python data reader: the training data are sent with `session.eval`. However this should be used for debugging purpose only. The users are encouraged to use the read data OPs.
+In PaddlePaddle v0.10.0, training data is typically read
+with [data reader](../reader/README.md) from Python. This approach is
+no longer efficient when training distributedly since the Python
+process no longer runs on the same node with the trainer processes,
+the Python reader will need to read from the distributed filesystem
+(assuming it has the access) and send to the trainers, doubling the
+network traffic.
+
+When doing distributed training, the user can still use Python data
+reader: the training data are sent with `Executor.run`. However, should
+be used for debugging purpose only. The users are encouraged to use
+the read data OPs.


 ## References:

--- a/doc/design/dist_refactor/multi_cpu.md
+++ b/doc/design/dist_refactor/multi_cpu.md
+# Design Doc: Execute the Program with Multi CPU
+
+## Abstract
+
+This Design Doc propose an approach to make the user-defined Op graph
+running with multi-CPU, we will use an auto transpiler to convert the user-defined
+Op graph to a multi-CPU Op graph, and run `ParallelDo` Op to run the graph.
+
+## Transpiler
+
+<img src="src/multi-threads/single-thread@3x.png" width="300">
+
+After converted:
+
+<img src="src/multi-threads/multi-threads@3x.png" width="1000">
+
+## Implement
+
+- `Multi-CPU Transpiler` will convert the graph to a multi-CPU graph
+  which would be executed with multi-threads.
+- `BlockingCounter` will `Init/Decrement` an atomic counter, and Blocking `Wait`
+  for the atomic counter become `0`:
+  ```cpp
+  BlockingCounter bc(thread_count);
+  for (int i = 0; i < thread_count; ++i) {
+    thread_pool->Start([&bc] {bc.DecrementCount(); })
+  }
+  bc.Wait();
+  ```
+- `ParallelDo` Operator
+  - Initialize a thread pool which is a Singleton.
+  - Use a block id as the input, and create run the specify Block on independent scope
+    with multi-threads.
+  - Initialize a `BlockingCounter` instance and wait until all threads are done.
+- `Split` Operator will split the Input Tensor into a TensorArray.
+- `Merge` merge all the gradients which calculated in different threads
+  with `mean/sum/max/min...` method, and then run the Optimizer Op to optimize `W`.
+
+## TODO
+
+- Improve the optimizer stage with multi-threads, since we could
+  assign the parameters to the different threads and execute
+  optimizer with multi-threads.
--- a/doc/design/refactor/parameter_server.md
+++ b/doc/design/refactor/parameter_server.md
-# Design Doc: Operation Graph Based Parameter Server
+# Design Doc: Parameter Server

 ## Abstract

@@ -9,32 +9,31 @@ different purposes.

 ## Background

-The previous implementations of the parameter server does not run a
-subgraph. parameter initialization, optimizer computation, network
+The previous implementations of the parameter server do not run a
+fluid sub-program. Parameter initialization, optimizer computation, network
 communication and checkpointing are implemented twice on both the
-trainer and the parameter server.
+trainer as well as the parameter server.

-It would be great if we can write code once and use them on both the
-trainer and the parameter server: reduces code duplication and
-improves extensibility. Given that after the current refactor, we are
-representing everything as a computing graph on the
-trainer. Representing everything as a computing graph on the parameter
+It would be great if we can write code once and use them on both: the
+trainer and the parameter server, since this reduces code duplication and
+improves extensibility. Given that after the current refactoring, we are
+representing everything as a computation graph on the
+trainer. Representing everything as a computation graph on the parameter
 server becomes a natural extension.

 ## Design

-### Graph Converter
+### Distributed Transpiler

-The *graph converter* converts the user-defined operation (OP) graph
-into subgraphs to be scheduled on different nodes with the following
+The *Distributed Transpiler* converts the user-defined fluid program
+into sub-programs to be scheduled on different nodes with the following
 steps:

 1. OP placement: the OPs will be placed on different nodes according
-   to heuristic that minimizes estimated total computation
+   to a heuristic that minimizes the estimated total computation
   time. Currently we will use a simple heuristic that puts parameter
-   varable on parameter server workers and everything else on trainer
+   variable on parameter server workers and everything else on trainer
   workers.
-
 1. Add communication OPs to enable the communication between nodes.

 We will need these OPs: *Send*, *Recv*, *Enqueue*, *Dequeue*.
@@ -48,56 +47,47 @@ After converting:

 <img src="src/dist-graph.png" width="700"/>

-1. The parameter variable W and it's optimizer subgraph are placed on the parameter server.
-1. Operators are added to the subgraphs.
+1. The parameter variable W and its optimizer program are placed on the parameter server.
+1. Operators are added to the program.
   - *Send* sends data to the connected *Recv* operator.  The
 	 scheduler on the receive node will only schedule *Recv* operator
 	 to run when the *Send* operator has ran (the *Send* OP will mark
 	 the *Recv* OP runnable automatically).
-   - *Enueue* enqueues the input variable, it can block until space
+   - *Enqueue* enqueues the input variable, it can block until space
     become available in the queue.
   - *Dequeue* outputs configurable numbers of tensors from the
-     queue. It will block until the queue have the required number of
+     queue. It will block until the queue has the required number of
     tensors.


 ### Benefits

- Model parallelism become easier to implement: it's an extension to
-  the trainer - parameter server approach. we already have the
-  communication OPs, but need to extend the graph converter's
-  placement functionality.
-
+- Model parallelism becomes easier to implement: it is an extension to
+  the trainer - parameter server approach. We can have several "Transpilers"
+  to achieve different goals.
 - User-defined optimizer is easier to add - user can now express it as
-  a subgraph.
-
+  a sub-program.
 - No more duplication logic inside the trainer and the parameter
  server mentioned in the background section.

 ### Challenges

- It might be hard for the graph converter to cut a general graph
-  (without any hint for which subgraph is the optimizer). We may need
-  to label which subgraph inside the OP graph is the optimizer.
-
- It's important to balance the parameter shards of on multiple
-  parameter server. If a single parameter is very big (some
+- It is important to balance the parameter shards on multiple
+  parameter servers. If a single parameter is very big (for example: some
  word-embedding, fully connected, softmax layer), we need to
  automatically partition the single parameter onto different
  parameter servers when possible (only element-wise optimizer depends
  on the parameter variable).
+- In the "Async SGD" figure, the "W" variable on the parameter server
+  could be read and written concurrently. See
+  [here](https://github.com/PaddlePaddle/Paddle/pull/6394) for more
+  details about concurrent program in Fluid.

 ### Discussion

- In the "Aync SGD" figure, the "W" variable on the parameter server
-  could be read and wrote concurrently, what is our locking strategy?
-  E.g., each variable have a lock cpp method to be invoked by every
-  OP, or, have a lock OP.
-
 - Can the Enqueue OP be implemented under our current tensor design
-  (puts the input tensor into the queue tensor)?
-
- *Dequeue* OP will have variable numbers of output (depends on the
+  (put the input tensor into the queue tensor)?
+- *Dequeue* OP will have variable numbers of output (depending on the
  `min_count` attribute), does our current design support it? (similar
  question for the *Add* OP)


--- a/doc/design/refactor/src/compiler.graffle
+++ b/doc/design/refactor/src/compiler.graffle
--- a/doc/design/refactor/src/compiler.png
+++ b/doc/design/refactor/src/compiler.png
--- a/doc/design/refactor/src/dist-graph.graffle
+++ b/doc/design/refactor/src/dist-graph.graffle
--- a/doc/design/refactor/src/dist-graph.png
+++ b/doc/design/refactor/src/dist-graph.png
--- a/doc/design/dist_refactor/src/distributed_architecture.graffle
+++ b/doc/design/dist_refactor/src/distributed_architecture.graffle
--- a/doc/design/dist_refactor/src/distributed_architecture.png
+++ b/doc/design/dist_refactor/src/distributed_architecture.png
--- a/doc/design/refactor/src/local-graph.graffle
+++ b/doc/design/refactor/src/local-graph.graffle
--- a/doc/design/refactor/src/local-graph.png
+++ b/doc/design/refactor/src/local-graph.png
--- a/doc/design/dist_refactor/src/local_architecture.graffle
+++ b/doc/design/dist_refactor/src/local_architecture.graffle
--- a/doc/design/dist_refactor/src/local_architecture.png
+++ b/doc/design/dist_refactor/src/local_architecture.png
--- a/doc/design/dist_refactor/src/multi-threads.graffle
+++ b/doc/design/dist_refactor/src/multi-threads.graffle
--- a/doc/design/dist_refactor/src/multi-threads/multi-threads@3x.png
+++ b/doc/design/dist_refactor/src/multi-threads/multi-threads@3x.png
--- a/doc/design/dist_refactor/src/multi-threads/single-thread@3x.png
+++ b/doc/design/dist_refactor/src/multi-threads/single-thread@3x.png
--- a/doc/design/refactor/src/paddle-compile.graffle
+++ b/doc/design/refactor/src/paddle-compile.graffle
--- a/doc/design/refactor/src/paddle-compile.png
+++ b/doc/design/refactor/src/paddle-compile.png
--- a/doc/design/dist_refactor/src/remote_executor.graffle
+++ b/doc/design/dist_refactor/src/remote_executor.graffle
--- a/doc/design/dist_refactor/src/remote_executor.png
+++ b/doc/design/dist_refactor/src/remote_executor.png
--- a/doc/design/error_clip.md
+++ b/doc/design/error_clip.md
+# Error Clip
+
+## Overview
+
+Error clip is widely used in model training to prevent gradient exploding. It takes some specific rules to adjust variables' gradients and prevent them from being too large. With it, values of a gradient will be checked before they are taken by the next `grad_op` and be shrunk if necessary.
+## Usage
+
+Users are allowed to assign different error clip methods or attributes to different `Variable`s. Users can specify it as a parameter of `Variable`'s constructor:
+
+```python
+var = framework.Variable(..., error_clip=myErrorClip, ...)
+```
+
+The default value of `error_clip` is `None`, which means no error clip is employed. When it's not `None`, it should take an object of `BaseErrorClipAttr`'s derived class. So far, `BaseErrorClipAttr` has only one derived class: `ErrorClipByValue`, whose constructor is:
+
+```python
+ErrorClipByValue(max, min=None)
+```
+
+`max` and `min` represent the maximal and minimal clip threshold respectively. In backward pass, all values of `var`'s gradient greater than `max` or less than `min` will be clipped to `max` and `min` respectively. When the `min` is None, the minimal threshold will be assigned with `-max` automatically.
+
+So we can enable the error clip with threshold `[-5.0, 5.0]` for variable `var` by:
+
+```python
+var = framework.Variable(..., error_clip=ErrorClipByValue(max=5.0), ...)
+```
+
+## Implementation
+
+The `BaseErrorClipAttr` and its derived class `ErrorClipByValue` are defined in *clip.py*.
+
+```python
+class BaseErrorClipAttr(object):
+    def append_clip_op(self, block, grad_name):
+        raise NotImplementedError()
+
+
+class ErrorClipByValue(BaseErrorClipAttr):
+    def __init__(self, max, min=None):
+        max = float(max)
+        if min is None:
+            min = -max
+        else:
+            min = float(min)
+        self.max = max
+        self.min = min
+
+    def append_clip_op(self, block, grad_name):
+        clip_op_desc = block.desc.append_op()
+        clip_op_desc.set_type("clip")
+        clip_op_desc.set_input("X", [grad_name])
+        clip_op_desc.set_output("Out", [grad_name])
+        clip_op_desc.set_attr("min", self.min)
+        clip_op_desc.set_attr("max", self.max)
+```
+
+The `BaseErrorClipAttr` have one main member functions: `append_clip_op(self, block, grad_name)`.
+
+This function is used to create a `clip_op` and append it to the end of given `block`. For different error clip algorithm require different `clip_op`, the function is defined as virtual in the base class. All derived classes must implement their own versions of this function.
+
+These `clip_op`s should be inserted after `grad_op`s whose output gradients need to be clipped. It is equivalent to appending some `clip_op`s to the end of the target block every time a new `grad_op` is added.
+
+```python
+for op_desc in grad_op_descs:
+        new_op_desc = target_block.desc.append_op()
+        new_op_desc.copy_from(op_desc)
+        callback(block=target_block, context=grad_to_var)
+```
+
+Here we employ a callback function to complete this kind of jobs. In `_append_backward_ops_` function, each time after a `grad_op` is added to the `target_block`, a callback function is invoked. The logic of `clip_op` appending can be implemented inside the callback function.
+
+The callback function for `clip_op` appending is defined in *clip.py*:
+
+```python
+def error_clip_callback(block, context):
+    # the context is a grad_to_var map
+    grad_to_var = context
+    op_desc = block.desc.op(block.desc.op_size() - 1)
+    for grad_n in filter(lambda n: grad_to_var.has_key(n),
+                         op_desc.output_arg_names()):
+        fwd_var = block.var_recursive(grad_to_var[grad_n])
+        error_clip = getattr(fwd_var, "error_clip", None)
+        if not (error_clip is None or isinstance(error_clip,
+                                                 BaseErrorClipAttr)):
+            raise TypeError(
+                "Variable's error_clip should be an instance of BaseErrorClipAttr or None."
+            )
+        if error_clip is not None:
+            error_clip.append_clip_op(block, grad_n)
+```
+
+This function takes a `block` and a `context`(which is actually a grad\_to\_var map) as inputs. It checks each output of the last `OpDesc` in the `block`. Notice that the last `OpDesc` of the `block` must be a `grad_op` and its outputs must be some forward variables' gradients. If an output gradient's corresponding forward variable has an attribute of `error_clip`, `error_clip_callback` will call the `error_clip`'s `append_clip_op` function to append the required `clip_op` into the `block`.
--- a/doc/design/executor.md
+++ b/doc/design/executor.md
 # Executor Design Doc

 ## Motivation
+In [fluid](https://github.com/PaddlePaddle/Paddle/blob/develop/doc/design/fluid.md), we encourage the user to use deep learning programming paradigms to describe the training process. When the user-written Python program is executed, it will first create a protobuf message
+[`ProgramDesc`](https://github.com/PaddlePaddle/Paddle/blob/a91efdde6910ce92a78e3aa7157412c4c88d9ee8/paddle/framework/framework.proto#L145) that describes the process and is conceptually like an [abstract syntax tree](https://en.wikipedia.org/wiki/Abstract_syntax_tree).

-We use executor to do the runtime evaluation of a `ProgramDesc`.
+The executor runs the `ProgramDesc` like an interpreter. `ProgramDesc` contains the intrinsics (operators in this case) and variables which will be used, executor explicitly executes the stored precompiled code.

 ## Overview

-An executor takes a `ProgramDesc`, a `block_id` and a `Scope`.  The `ProgramDesc` is a list of blocks and each block contains the protobuf definition of all the parameters and operators. The `block_id` specifies the entrance block. And the `Scope` is the container of all the variable instance, which is persistent throughout different runs.
+An executor takes a `ProgramDesc`, a `block_id` and a `Scope`.  The `ProgramDesc` is a list of blocks and each block contains the protobuf definition of all the parameters and operators in the block. The `block_id` specifies the entrance block. And the `Scope` is the container of all the variable instances, which is persistent throughout different runs.

-### What does executor do?
+## Executor

-It evaluates all the operators in the `block_id`th block of a `ProgramDesc`.
+The `Executor` explicitly executes all the intrinsics (operators here) in the `block_id`th block of a `ProgramDesc`. Essentially, it instantiates Variables and Operators, then runs all the operators in sequence one-by-one.
+It is very similar to how a push stack frame works when entering a block, following which it cleans up all the temporary variables when a mini-batch is finished. It does not however, have the stack frame pop process.

-### What does executor NOT do?
+### The interface
+```c++
+  Executor(places);
+```
+A executor does not own any computing resources, a user can only construct an executor using the specified places.

-It does not do runtime optimization, meaning intelligently parse the dependency of each op a choose which one to be run and in which order they should be run.
+### Running an Executor

-It does not do graph partitioning, meaning dividing the `ProgramDesc` into several small pieces and executing them on different devices.
-
-## Implementation
-
-`Executor` evaluates a `ProgramDesc`. Essentially, it instantiates Variables and Operators, then run all the operators in sequence. [[code]](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.cc)
+```
+  void Run(ProgramDesc, Scope, block_id, create_local_scope);
+```
+An `Executor` only provides a unified way to execute `ProgramDesc`. `ProgramDesc` is the target that will be executed, the `Scope` specifies the variable container, the `block_id` indicates the entrance block and `create_local_scope` is a boolean that states whether it will destroy the temporary variables after the execution is finished.
--- a/doc/design/fluid-compiler.graffle
+++ b/doc/design/fluid-compiler.graffle
--- a/doc/design/fluid-compiler.png
+++ b/doc/design/fluid-compiler.png
--- a/doc/design/fluid.md
+++ b/doc/design/fluid.md
--- a/doc/design/fluid_compiler.md
+++ b/doc/design/fluid_compiler.md
--- a/doc/design/images/control_flow_graph.png
+++ b/doc/design/images/control_flow_graph.png
--- a/doc/design/images/dataflow_equations.png
+++ b/doc/design/images/dataflow_equations.png
--- a/doc/design/images/deep_learning.png
+++ b/doc/design/images/deep_learning.png
--- a/paddle/framework/images/duplicate_op.graffle
+++ b/paddle/framework/images/duplicate_op.graffle
--- a/paddle/framework/images/duplicate_op.png
+++ b/paddle/framework/images/duplicate_op.png
--- a/paddle/framework/images/duplicate_op2.graffle
+++ b/paddle/framework/images/duplicate_op2.graffle
--- a/paddle/framework/images/duplicate_op2.png
+++ b/paddle/framework/images/duplicate_op2.png
--- a/doc/design/images/multigpu_allreduce.graffle
+++ b/doc/design/images/multigpu_allreduce.graffle
--- a/doc/design/images/multigpu_allreduce.png
+++ b/doc/design/images/multigpu_allreduce.png
--- a/doc/design/images/multigpu_before_convert.graffle
+++ b/doc/design/images/multigpu_before_convert.graffle
--- a/doc/design/images/multigpu_before_convert.png
+++ b/doc/design/images/multigpu_before_convert.png
--- a/doc/design/images/profiler.png
+++ b/doc/design/images/profiler.png
--- a/doc/design/kernel_hint_design.md
+++ b/doc/design/kernel_hint_design.md
--- a/doc/design/memory_optimization.md
+++ b/doc/design/memory_optimization.md
--- a/doc/design/mkldnn/image/engine.png
+++ b/doc/design/mkldnn/image/engine.png
--- a/doc/design/mkldnn/image/gradients.png
+++ b/doc/design/mkldnn/image/gradients.png
--- a/doc/design/mkldnn/image/layers.png
+++ b/doc/design/mkldnn/image/layers.png
--- a/doc/design/mkldnn/image/matrix.png
+++ b/doc/design/mkldnn/image/matrix.png
--- a/doc/design/mkldnn/image/overview.png
+++ b/doc/design/mkldnn/image/overview.png
--- a/doc/design/mkl/mkl_packed.md
+++ b/doc/design/mkl/mkl_packed.md
--- a/doc/design/mkldnn/README.MD
+++ b/doc/design/mkldnn/README.MD
--- a/doc/design/mkl/mkldnn_fluid.md
+++ b/doc/design/mkl/mkldnn_fluid.md
--- a/doc/design/operator_kernel_type.md
+++ b/doc/design/operator_kernel_type.md
--- a/doc/design/ops/sequence_decoder.md
+++ b/doc/design/ops/sequence_decoder.md
--- a/doc/design/optimizer.md
+++ b/doc/design/optimizer.md
--- a/doc/design/paddle_nccl.md
+++ b/doc/design/paddle_nccl.md
--- a/doc/design/profiler.md
+++ b/doc/design/profiler.md
--- a/doc/design/python_api.md
+++ b/doc/design/python_api.md
--- a/doc/design/refactor/session.md
+++ b/doc/design/refactor/session.md
--- a/doc/design/refactor/src/distributed_architecture.graffle
+++ b/doc/design/refactor/src/distributed_architecture.graffle
--- a/doc/design/refactor/src/distributed_architecture.png
+++ b/doc/design/refactor/src/distributed_architecture.png
--- a/doc/design/refactor/src/local_architecture.graffle
+++ b/doc/design/refactor/src/local_architecture.graffle
--- a/doc/design/refactor/src/local_architecture.png
+++ b/doc/design/refactor/src/local_architecture.png
--- a/doc/design/releasing_process.md
+++ b/doc/design/releasing_process.md
--- a/doc/design/support_new_device.md
+++ b/doc/design/support_new_device.md
--- a/doc/design/switch_kernel.md
+++ b/doc/design/switch_kernel.md
--- a/doc/design/var_desc.md
+++ b/doc/design/var_desc.md
--- a/doc/faq/build_and_install/index_cn.rst
+++ b/doc/faq/build_and_install/index_cn.rst
--- a/doc/faq/local/index_cn.rst
+++ b/doc/faq/local/index_cn.rst
--- a/doc/faq/local/src/reduce_min_pool_size.py
+++ b/doc/faq/local/src/reduce_min_pool_size.py
--- a/doc/faq/local/src/word2vec_config.py
+++ b/doc/faq/local/src/word2vec_config.py
--- a/doc/faq/local/src/word2vec_dataprovider.py
+++ b/doc/faq/local/src/word2vec_dataprovider.py
--- a/doc/faq/model/index_cn.rst
+++ b/doc/faq/model/index_cn.rst
--- a/doc/getstarted/build_and_install/build_from_source_cn.rst
+++ b/doc/getstarted/build_and_install/build_from_source_cn.rst
--- a/doc/getstarted/build_and_install/build_from_source_en.rst
+++ b/doc/getstarted/build_and_install/build_from_source_en.rst
--- a/doc/getstarted/build_and_install/docker_install_cn.rst
+++ b/doc/getstarted/build_and_install/docker_install_cn.rst
--- a/doc/getstarted/build_and_install/docker_install_en.rst
+++ b/doc/getstarted/build_and_install/docker_install_en.rst
--- a/doc/getstarted/build_and_install/pip_install_cn.rst
+++ b/doc/getstarted/build_and_install/pip_install_cn.rst
--- a/doc/getstarted/build_and_install/pip_install_en.rst
+++ b/doc/getstarted/build_and_install/pip_install_en.rst
--- a/v1_api_demo/model_zoo/resnet/net_diagram.sh
+++ b/v1_api_demo/model_zoo/resnet/net_diagram.sh
--- a/doc/getstarted/concepts/src/train.py
+++ b/doc/getstarted/concepts/src/train.py
--- a/doc/getstarted/concepts/use_concepts_cn.rst
+++ b/doc/getstarted/concepts/use_concepts_cn.rst
--- a/doc/getstarted/index_cn.rst
+++ b/doc/getstarted/index_cn.rst
--- a/doc/getstarted/index_en.rst
+++ b/doc/getstarted/index_en.rst
--- a/doc/howto/dev/new_op_cn.md
+++ b/doc/howto/dev/new_op_cn.md
--- a/doc/howto/dev/new_op_en.md
+++ b/doc/howto/dev/new_op_en.md
--- a/doc/howto/dev/new_op_kernel_en.md
+++ b/doc/howto/dev/new_op_kernel_en.md
--- a/doc/howto/index_cn.rst
+++ b/doc/howto/index_cn.rst
--- a/doc/howto/index_en.rst
+++ b/doc/howto/index_en.rst
--- a/doc/howto/optimization/cpu_profiling.md
+++ b/doc/howto/optimization/cpu_profiling.md
--- a/doc/howto/read_source.md
+++ b/doc/howto/read_source.md
--- a/doc/howto/usage/capi/compile_paddle_lib_cn.md
+++ b/doc/howto/usage/capi/compile_paddle_lib_cn.md
--- a/doc/howto/usage/capi/images/csr.png
+++ b/doc/howto/usage/capi/images/csr.png
--- a/doc/howto/usage/capi/images/sequence_data.png
+++ b/doc/howto/usage/capi/images/sequence_data.png
--- a/doc/howto/usage/capi/images/workflow_of_CAPI.png
+++ b/doc/howto/usage/capi/images/workflow_of_CAPI.png
--- a/doc/howto/usage/capi/index_cn.rst
+++ b/doc/howto/usage/capi/index_cn.rst
--- a/doc/howto/usage/capi/organization_of_the_inputs_cn.md
+++ b/doc/howto/usage/capi/organization_of_the_inputs_cn.md
--- a/doc/howto/usage/capi/workflow_of_capi_cn.md
+++ b/doc/howto/usage/capi/workflow_of_capi_cn.md
--- a/doc/howto/usage/cluster/cluster_train_cn.md
+++ b/doc/howto/usage/cluster/cluster_train_cn.md
--- a/doc/howto/usage/cluster/cluster_train_en.md
+++ b/doc/howto/usage/cluster/cluster_train_en.md
--- a/doc/howto/usage/cluster/fabric_cn.md
+++ b/doc/howto/usage/cluster/fabric_cn.md
--- a/doc/howto/usage/cluster/fabric_en.md
+++ b/doc/howto/usage/cluster/fabric_en.md
--- a/doc/howto/usage/cluster/fluid_cluster_train_en.md
+++ b/doc/howto/usage/cluster/fluid_cluster_train_en.md
--- a/doc/howto/usage/cluster/k8s_aws_cn.md
+++ b/doc/howto/usage/cluster/k8s_aws_cn.md
--- a/doc/howto/usage/k8s/k8s_aws_en.md
+++ b/doc/howto/usage/k8s/k8s_aws_en.md
--- a/doc/howto/usage/k8s/k8s_cn.md
+++ b/doc/howto/usage/k8s/k8s_cn.md
--- a/doc/howto/usage/k8s/k8s_distributed_cn.md
+++ b/doc/howto/usage/k8s/k8s_distributed_cn.md
--- a/doc/howto/usage/k8s/k8s_en.md
+++ b/doc/howto/usage/k8s/k8s_en.md
--- a/doc/howto/usage/cluster/openmpi_cn.md
+++ b/doc/howto/usage/cluster/openmpi_cn.md
--- a/doc/howto/usage/cluster/openmpi_en.md
+++ b/doc/howto/usage/cluster/openmpi_en.md
--- a/doc/howto/usage/k8s/src/Dockerfile
+++ b/doc/howto/usage/k8s/src/Dockerfile
--- a/doc/howto/usage/k8s/src/add_security_group.png
+++ b/doc/howto/usage/k8s/src/add_security_group.png
--- a/doc/howto/usage/k8s/src/create_efs.png
+++ b/doc/howto/usage/k8s/src/create_efs.png
--- a/doc/howto/usage/k8s/src/efs_mount.png
+++ b/doc/howto/usage/k8s/src/efs_mount.png
--- a/doc/howto/usage/cluster/src/k8s-paddle-arch.png
+++ b/doc/howto/usage/cluster/src/k8s-paddle-arch.png
--- a/doc/howto/usage/k8s/src/k8s_data/Dockerfile
+++ b/doc/howto/usage/k8s/src/k8s_data/Dockerfile
--- a/doc/howto/usage/k8s/src/k8s_data/README.md
+++ b/doc/howto/usage/k8s/src/k8s_data/README.md
--- a/doc/howto/usage/k8s/src/k8s_data/get_data.sh
+++ b/doc/howto/usage/k8s/src/k8s_data/get_data.sh
--- a/doc/howto/usage/k8s/src/k8s_train/Dockerfile
+++ b/doc/howto/usage/k8s/src/k8s_train/Dockerfile
--- a/doc/howto/usage/k8s/src/k8s_train/README.md
+++ b/doc/howto/usage/k8s/src/k8s_train/README.md
--- a/doc/howto/usage/k8s/src/k8s_train/start.sh
+++ b/doc/howto/usage/k8s/src/k8s_train/start.sh
--- a/doc/howto/usage/k8s/src/k8s_train/start_paddle.py
+++ b/doc/howto/usage/k8s/src/k8s_train/start_paddle.py
--- a/doc/howto/usage/k8s/src/managed_policy.png
+++ b/doc/howto/usage/k8s/src/managed_policy.png
--- a/doc/howto/usage/k8s/src/pserver_and_trainer.png
+++ b/doc/howto/usage/k8s/src/pserver_and_trainer.png
--- a/doc/howto/usage/k8s/src/route53_create_recordset.png
+++ b/doc/howto/usage/k8s/src/route53_create_recordset.png
--- a/doc/howto/usage/k8s/src/route53_create_zone.png
+++ b/doc/howto/usage/k8s/src/route53_create_zone.png
--- a/doc/howto/usage/cluster/src/word2vec/api_train_v2.py
+++ b/doc/howto/usage/cluster/src/word2vec/api_train_v2.py
--- a/doc/howto/usage/cluster/src/word2vec/api_train_v2_cluster.py
+++ b/doc/howto/usage/cluster/src/word2vec/api_train_v2_cluster.py
--- a/doc/howto/usage/cluster/src/word2vec/prepare.py
+++ b/doc/howto/usage/cluster/src/word2vec/prepare.py
--- a/doc/howto/usage/k8s/src/worker_security_group.png
+++ b/doc/howto/usage/k8s/src/worker_security_group.png
--- a/doc/howto/usage/k8s/k8s_basis_cn.md
+++ b/doc/howto/usage/k8s/k8s_basis_cn.md
--- a/doc/howto/usage/k8s/src/k8s-paddle-arch.png
+++ b/doc/howto/usage/k8s/src/k8s-paddle-arch.png
--- a/doc/mobile/cross_compiling_for_android_cn.md
+++ b/doc/mobile/cross_compiling_for_android_cn.md
--- a/doc/mobile/cross_compiling_for_android_en.md
+++ b/doc/mobile/cross_compiling_for_android_en.md
--- a/doc/mobile/cross_compiling_for_ios_cn.md
+++ b/doc/mobile/cross_compiling_for_ios_cn.md
--- a/doc/mobile/cross_compiling_for_ios_en.md
+++ b/doc/mobile/cross_compiling_for_ios_en.md
--- a/doc/mobile/index_en.rst
+++ b/doc/mobile/index_en.rst
--- a/doc/v1_api_tutorials/README.md
+++ b/doc/v1_api_tutorials/README.md
--- a/doc/v1_api_tutorials/embedding_model/index_cn.md
+++ b/doc/v1_api_tutorials/embedding_model/index_cn.md
--- a/doc/v1_api_tutorials/embedding_model/index_en.md
+++ b/doc/v1_api_tutorials/embedding_model/index_en.md
--- a/doc/v1_api_tutorials/embedding_model/neural-n-gram-model.png
+++ b/doc/v1_api_tutorials/embedding_model/neural-n-gram-model.png
--- a/doc/v1_api_tutorials/gan/gan.png
+++ b/doc/v1_api_tutorials/gan/gan.png
--- a/doc/v1_api_tutorials/gan/index_en.md
+++ b/doc/v1_api_tutorials/gan/index_en.md
--- a/doc/v1_api_tutorials/gan/mnist_sample.png
+++ b/doc/v1_api_tutorials/gan/mnist_sample.png
--- a/doc/v1_api_tutorials/gan/uniform_sample.png
+++ b/doc/v1_api_tutorials/gan/uniform_sample.png
--- a/doc/v1_api_tutorials/imagenet_model/resnet_block.jpg
+++ b/doc/v1_api_tutorials/imagenet_model/resnet_block.jpg
--- a/doc/v1_api_tutorials/imagenet_model/resnet_model_cn.md
+++ b/doc/v1_api_tutorials/imagenet_model/resnet_model_cn.md
--- a/doc/v1_api_tutorials/imagenet_model/resnet_model_en.md
+++ b/doc/v1_api_tutorials/imagenet_model/resnet_model_en.md
--- a/doc/v1_api_tutorials/quick_start/index_cn.rst
+++ b/doc/v1_api_tutorials/quick_start/index_cn.rst
--- a/doc/v1_api_tutorials/quick_start/index_en.md
+++ b/doc/v1_api_tutorials/quick_start/index_en.md
--- a/doc/v1_api_tutorials/quick_start/src/NetContinuous_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/NetContinuous_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/NetContinuous_en.png
+++ b/doc/v1_api_tutorials/quick_start/src/NetContinuous_en.png
--- a/doc/v1_api_tutorials/quick_start/src/NetConv_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/NetConv_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/NetConv_en.png
+++ b/doc/v1_api_tutorials/quick_start/src/NetConv_en.png
--- a/doc/v1_api_tutorials/quick_start/src/NetLR_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/NetLR_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/NetLR_en.png
+++ b/doc/v1_api_tutorials/quick_start/src/NetLR_en.png
--- a/doc/v1_api_tutorials/quick_start/src/NetRNN_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/NetRNN_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/NetRNN_en.png
+++ b/doc/v1_api_tutorials/quick_start/src/NetRNN_en.png
--- a/doc/v1_api_tutorials/quick_start/src/PipelineNetwork_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/PipelineNetwork_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/PipelineNetwork_en.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/PipelineNetwork_en.jpg
--- a/doc/v1_api_tutorials/quick_start/src/PipelineTest_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/PipelineTest_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/PipelineTest_en.png
+++ b/doc/v1_api_tutorials/quick_start/src/PipelineTest_en.png
--- a/doc/v1_api_tutorials/quick_start/src/PipelineTrain_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/PipelineTrain_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/PipelineTrain_en.png
+++ b/doc/v1_api_tutorials/quick_start/src/PipelineTrain_en.png
--- a/doc/v1_api_tutorials/quick_start/src/Pipeline_cn.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/Pipeline_cn.jpg
--- a/doc/v1_api_tutorials/quick_start/src/Pipeline_en.jpg
+++ b/doc/v1_api_tutorials/quick_start/src/Pipeline_en.jpg
--- a/go/pserver/client/c/test/test_cclient.c
+++ b/go/pserver/client/c/test/test_cclient.c
--- a/go/pserver/client/c/test/test_mnist.py
+++ b/go/pserver/client/c/test/test_mnist.py
--- a/go/pserver/client/c/test/test_train.py
+++ b/go/pserver/client/c/test/test_train.py
--- a/paddle/CMakeLists.txt
+++ b/paddle/CMakeLists.txt
--- a/paddle/api/test/testTrainConfig.py
+++ b/paddle/api/test/testTrainConfig.py
--- a/paddle/capi/error.cpp
+++ b/paddle/capi/error.cpp
--- a/paddle/capi/error.h
+++ b/paddle/capi/error.h
--- a/paddle/capi/examples/model_inference/common/common.h
+++ b/paddle/capi/examples/model_inference/common/common.h
--- a/paddle/capi/examples/model_inference/dense/main.c
+++ b/paddle/capi/examples/model_inference/dense/main.c
--- a/v1_api_demo/mnist/data/generate_list.py
+++ b/v1_api_demo/mnist/data/generate_list.py
--- a/paddle/capi/examples/model_inference/dense/mnist_v2.py
+++ b/paddle/capi/examples/model_inference/dense/mnist_v2.py
--- a/paddle/capi/examples/model_inference/dense/trainer_config.py
+++ b/paddle/capi/examples/model_inference/dense/trainer_config.py
--- a/paddle/capi/examples/model_inference/multi_thread/main.c
+++ b/paddle/capi/examples/model_inference/multi_thread/main.c
--- a/paddle/capi/examples/model_inference/multi_thread/main_gpu.c
+++ b/paddle/capi/examples/model_inference/multi_thread/main_gpu.c
--- a/paddle/capi/examples/model_inference/sequence/main.c
+++ b/paddle/capi/examples/model_inference/sequence/main.c
--- a/paddle/capi/examples/model_inference/sequence/trainer_config.py
+++ b/paddle/capi/examples/model_inference/sequence/trainer_config.py
--- a/paddle/capi/examples/model_inference/sparse_binary/main.c
+++ b/paddle/capi/examples/model_inference/sparse_binary/main.c
--- a/paddle/capi/gradient_machine.cpp
+++ b/paddle/capi/gradient_machine.cpp
--- a/paddle/capi/gradient_machine.h
+++ b/paddle/capi/gradient_machine.h
--- a/paddle/capi/tests/test_predict_network.py
+++ b/paddle/capi/tests/test_predict_network.py
--- a/paddle/cuda/src/avx_mathfun.h
+++ b/paddle/cuda/src/avx_mathfun.h
--- a/paddle/framework/CMakeLists.txt
+++ b/paddle/framework/CMakeLists.txt
--- a/paddle/framework/attribute.cc
+++ b/paddle/framework/attribute.cc
--- a/paddle/framework/attribute.h
+++ b/paddle/framework/attribute.h
--- a/paddle/framework/backward.cc
+++ b/paddle/framework/backward.cc
--- a/paddle/framework/backward.h
+++ b/paddle/framework/backward.h
--- a/paddle/framework/backward.md
+++ b/paddle/framework/backward.md
--- a/paddle/framework/backward_test.cc
+++ b/paddle/framework/backward_test.cc
--- a/paddle/framework/block_desc.cc
+++ b/paddle/framework/block_desc.cc
--- a/paddle/framework/block_desc.h
+++ b/paddle/framework/block_desc.h
--- a/paddle/framework/channel.h
+++ b/paddle/framework/channel.h
--- a/paddle/framework/channel_test.cc
+++ b/paddle/framework/channel_test.cc
--- a/paddle/framework/data_device_transform.cc
+++ b/paddle/framework/data_device_transform.cc
--- a/paddle/framework/data_device_transform.h
+++ b/paddle/framework/data_device_transform.h
--- a/paddle/framework/data_device_transform_test.cu
+++ b/paddle/framework/data_device_transform_test.cu
--- a/paddle/framework/data_layout.h
+++ b/paddle/framework/data_layout.h
--- a/paddle/framework/data_layout_transform.cc
+++ b/paddle/framework/data_layout_transform.cc
--- a/paddle/framework/data_layout_transform.h
+++ b/paddle/framework/data_layout_transform.h
--- a/paddle/framework/data_layout_transform_test.cc
+++ b/paddle/framework/data_layout_transform_test.cc
--- a/paddle/framework/data_transform.cc
+++ b/paddle/framework/data_transform.cc
--- a/paddle/framework/data_transform.h
+++ b/paddle/framework/data_transform.h
--- a/paddle/framework/data_type.h
+++ b/paddle/framework/data_type.h
--- a/paddle/framework/data_type_transform.cc
+++ b/paddle/framework/data_type_transform.cc
--- a/paddle/framework/data_type_transform.h
+++ b/paddle/framework/data_type_transform.h
--- a/paddle/framework/data_type_transform_test.cc
+++ b/paddle/framework/data_type_transform_test.cc
--- a/paddle/framework/ddim_test.cc
+++ b/paddle/framework/ddim_test.cc
--- a/paddle/framework/details/buffered_channel.h
+++ b/paddle/framework/details/buffered_channel.h
--- a/paddle/framework/details/cow_ptr.h
+++ b/paddle/framework/details/cow_ptr.h
--- a/paddle/operators/detail/recv_impl.cc
+++ b/paddle/operators/detail/recv_impl.cc
--- a/paddle/framework/details/op_registry.h
+++ b/paddle/framework/details/op_registry.h
--- a/paddle/framework/details/unbuffered_channel.h
+++ b/paddle/framework/details/unbuffered_channel.h
--- a/paddle/framework/dim.h
+++ b/paddle/framework/dim.h
--- a/paddle/framework/dim_test.cu
+++ b/paddle/framework/dim_test.cu
--- a/paddle/framework/eigen_test.cc
+++ b/paddle/framework/eigen_test.cc
--- a/paddle/framework/executor.cc
+++ b/paddle/framework/executor.cc
--- a/paddle/framework/executor.h
+++ b/paddle/framework/executor.h
--- a/paddle/framework/feed_fetch_method.cc
+++ b/paddle/framework/feed_fetch_method.cc
--- a/paddle/framework/feed_fetch_method.h
+++ b/paddle/framework/feed_fetch_method.h
--- a/paddle/framework/feed_fetch_type.h
+++ b/paddle/framework/feed_fetch_type.h
--- a/paddle/framework/framework.proto
+++ b/paddle/framework/framework.proto
--- a/paddle/framework/grad_op_desc_maker.h
+++ b/paddle/framework/grad_op_desc_maker.h
--- a/paddle/framework/init.cc
+++ b/paddle/framework/init.cc
--- a/paddle/framework/init.h
+++ b/paddle/framework/init.h
--- a/paddle/framework/init_test.cc
+++ b/paddle/framework/init_test.cc
--- a/paddle/framework/library_type.h
+++ b/paddle/framework/library_type.h
--- a/paddle/framework/lod_rank_table.cc
+++ b/paddle/framework/lod_rank_table.cc
--- a/paddle/framework/lod_rank_table.h
+++ b/paddle/framework/lod_rank_table.h
--- a/paddle/framework/lod_tensor.cc
+++ b/paddle/framework/lod_tensor.cc
--- a/paddle/framework/lod_tensor.h
+++ b/paddle/framework/lod_tensor.h
--- a/paddle/framework/lod_tensor_array.h
+++ b/paddle/framework/lod_tensor_array.h
--- a/paddle/framework/lod_tensor_test.cc
+++ b/paddle/framework/lod_tensor_test.cc
--- a/paddle/framework/lod_tensor_test.cu
+++ b/paddle/framework/lod_tensor_test.cu
--- a/paddle/framework/op_desc.cc
+++ b/paddle/framework/op_desc.cc
--- a/paddle/framework/op_desc.h
+++ b/paddle/framework/op_desc.h
--- a/paddle/framework/op_info.cc
+++ b/paddle/framework/op_info.cc
--- a/paddle/framework/op_info.h
+++ b/paddle/framework/op_info.h
--- a/paddle/framework/op_kernel_type.h
+++ b/paddle/framework/op_kernel_type.h
--- a/paddle/framework/op_kernel_type_test.cc
+++ b/paddle/framework/op_kernel_type_test.cc
--- a/paddle/framework/op_proto_maker.h
+++ b/paddle/framework/op_proto_maker.h
--- a/paddle/framework/op_proto_maker_test.cc
+++ b/paddle/framework/op_proto_maker_test.cc
--- a/paddle/framework/op_registry.cc
+++ b/paddle/framework/op_registry.cc
--- a/paddle/framework/op_registry.h
+++ b/paddle/framework/op_registry.h
--- a/paddle/framework/op_registry_test.cc
+++ b/paddle/framework/op_registry_test.cc
--- a/paddle/framework/operator.cc
+++ b/paddle/framework/operator.cc
--- a/paddle/framework/operator.h
+++ b/paddle/framework/operator.h
--- a/paddle/framework/operator_test.cc
+++ b/paddle/framework/operator_test.cc
--- a/paddle/framework/program_desc.cc
+++ b/paddle/framework/program_desc.cc
--- a/paddle/framework/program_desc.h
+++ b/paddle/framework/program_desc.h
--- a/paddle/framework/program_desc_test.cc
+++ b/paddle/framework/program_desc_test.cc
--- a/paddle/framework/prune.cc
+++ b/paddle/framework/prune.cc
--- a/paddle/framework/prune.h
+++ b/paddle/framework/prune.h
--- a/paddle/framework/prune_test.cc
+++ b/paddle/framework/prune_test.cc
--- a/paddle/framework/scope.cc
+++ b/paddle/framework/scope.cc
--- a/paddle/framework/scope.h
+++ b/paddle/framework/scope.h
--- a/paddle/framework/scope_test.cc
+++ b/paddle/framework/scope_test.cc
--- a/paddle/framework/selected_rows.cc
+++ b/paddle/framework/selected_rows.cc
--- a/paddle/framework/selected_rows.h
+++ b/paddle/framework/selected_rows.h
--- a/paddle/framework/selected_rows_test.cc
+++ b/paddle/framework/selected_rows_test.cc
--- a/paddle/framework/shape_inference.cc
+++ b/paddle/framework/shape_inference.cc
--- a/paddle/framework/shape_inference.h
+++ b/paddle/framework/shape_inference.h
--- a/paddle/framework/tensor.cc
+++ b/paddle/framework/tensor.cc
--- a/paddle/framework/tensor.h
+++ b/paddle/framework/tensor.h
--- a/paddle/framework/tensor.md
+++ b/paddle/framework/tensor.md
--- a/paddle/framework/tensor_impl.h
+++ b/paddle/framework/tensor_impl.h
--- a/paddle/framework/tensor_test.cc
+++ b/paddle/framework/tensor_test.cc
--- a/paddle/framework/tensor_util.cc
+++ b/paddle/framework/tensor_util.cc
--- a/paddle/framework/tensor_util.cu
+++ b/paddle/framework/tensor_util.cu
--- a/paddle/framework/tensor_util.h
+++ b/paddle/framework/tensor_util.h
--- a/paddle/framework/tensor_util_test.cc
+++ b/paddle/framework/tensor_util_test.cc
--- a/paddle/framework/tensor_util_test.cu
+++ b/paddle/framework/tensor_util_test.cu
--- a/paddle/framework/threadpool.cc
+++ b/paddle/framework/threadpool.cc
--- a/paddle/framework/threadpool.h
+++ b/paddle/framework/threadpool.h
--- a/paddle/framework/threadpool_test.cc
+++ b/paddle/framework/threadpool_test.cc
--- a/paddle/framework/type_defs.h
+++ b/paddle/framework/type_defs.h
--- a/paddle/framework/var_desc.cc
+++ b/paddle/framework/var_desc.cc
--- a/paddle/framework/var_desc.h
+++ b/paddle/framework/var_desc.h
--- a/paddle/framework/var_type.h
+++ b/paddle/framework/var_type.h
--- a/paddle/framework/var_type_inference.h
+++ b/paddle/framework/var_type_inference.h
--- a/paddle/framework/var_type_inference_test.cc
+++ b/paddle/framework/var_type_inference_test.cc
--- a/paddle/framework/variable.h
+++ b/paddle/framework/variable.h
--- a/paddle/framework/variable_test.cc
+++ b/paddle/framework/variable_test.cc
--- a/paddle/function/GemmConvOp.cpp
+++ b/paddle/function/GemmConvOp.cpp
--- a/paddle/function/Im2Col.h
+++ b/paddle/function/Im2Col.h
--- a/paddle/function/Im2ColTest.cpp
+++ b/paddle/function/Im2ColTest.cpp
--- a/paddle/gserver/CMakeLists.txt
+++ b/paddle/gserver/CMakeLists.txt
--- a/paddle/gserver/gradientmachines/GradientMachine.h
+++ b/paddle/gserver/gradientmachines/GradientMachine.h
--- a/paddle/gserver/gradientmachines/NeuralNetwork.cpp
+++ b/paddle/gserver/gradientmachines/NeuralNetwork.cpp
--- a/paddle/gserver/gradientmachines/NeuralNetwork.h
+++ b/paddle/gserver/gradientmachines/NeuralNetwork.h
--- a/paddle/gserver/layers/MKLDNNConcatLayer.cpp
+++ b/paddle/gserver/layers/MKLDNNConcatLayer.cpp
--- a/paddle/gserver/layers/MKLDNNLRNLayer.cpp
+++ b/paddle/gserver/layers/MKLDNNLRNLayer.cpp
--- a/paddle/gserver/layers/MKLDNNLayer.cpp
+++ b/paddle/gserver/layers/MKLDNNLayer.cpp
--- a/paddle/gserver/layers/MKLDNNLayer.h
+++ b/paddle/gserver/layers/MKLDNNLayer.h
--- a/paddle/gserver/layers/MKLPackedRecurrentLayer.cpp
+++ b/paddle/gserver/layers/MKLPackedRecurrentLayer.cpp
--- a/paddle/gserver/layers/MKLPackedRecurrentLayer.h
+++ b/paddle/gserver/layers/MKLPackedRecurrentLayer.h
--- a/paddle/gserver/layers/MKLPackedWeight.h
+++ b/paddle/gserver/layers/MKLPackedWeight.h
--- a/paddle/gserver/layers/PriorBox.cpp
+++ b/paddle/gserver/layers/PriorBox.cpp
--- a/paddle/gserver/layers/ROIPoolLayer.cpp
+++ b/paddle/gserver/layers/ROIPoolLayer.cpp
--- a/paddle/gserver/layers/RecurrentLayer.cpp
+++ b/paddle/gserver/layers/RecurrentLayer.cpp
--- a/paddle/gserver/layers/RecurrentLayer.h
+++ b/paddle/gserver/layers/RecurrentLayer.h
--- a/paddle/gserver/layers/SequenceToBatch.cpp
+++ b/paddle/gserver/layers/SequenceToBatch.cpp
--- a/paddle/gserver/tests/img_conv_cudnn.py
+++ b/paddle/gserver/tests/img_conv_cudnn.py
--- a/paddle/gserver/tests/img_conv_exconv.py
+++ b/paddle/gserver/tests/img_conv_exconv.py
--- a/paddle/gserver/tests/pyDataProvider.py
+++ b/paddle/gserver/tests/pyDataProvider.py
--- a/paddle/gserver/tests/rnn_data_provider.py
+++ b/paddle/gserver/tests/rnn_data_provider.py
--- a/paddle/gserver/tests/sequenceGen.py
+++ b/paddle/gserver/tests/sequenceGen.py
--- a/paddle/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py
+++ b/paddle/gserver/tests/sequence_nest_rnn_multi_unequalength_inputs.py
--- a/paddle/gserver/tests/sequence_recurrent.py
+++ b/paddle/gserver/tests/sequence_recurrent.py
--- a/paddle/gserver/tests/sequence_recurrent_group.py
+++ b/paddle/gserver/tests/sequence_recurrent_group.py
--- a/paddle/gserver/tests/sequence_rnn_matched_inputs.py
+++ b/paddle/gserver/tests/sequence_rnn_matched_inputs.py
--- a/paddle/gserver/tests/sequence_rnn_mixed_inputs.py
+++ b/paddle/gserver/tests/sequence_rnn_mixed_inputs.py
--- a/paddle/gserver/tests/sequence_rnn_multi_unequalength_inputs.py
+++ b/paddle/gserver/tests/sequence_rnn_multi_unequalength_inputs.py
--- a/paddle/gserver/tests/test_LayerGrad.cpp
+++ b/paddle/gserver/tests/test_LayerGrad.cpp
--- a/paddle/gserver/tests/test_PyDataProvider2.py
+++ b/paddle/gserver/tests/test_PyDataProvider2.py
--- a/paddle/gserver/tests/test_RecurrentLayer.cpp
+++ b/paddle/gserver/tests/test_RecurrentLayer.cpp
--- a/paddle/inference/CMakeLists.txt
+++ b/paddle/inference/CMakeLists.txt
--- a/paddle/inference/example.cc
+++ b/paddle/inference/example.cc
--- a/paddle/inference/io.cc
+++ b/paddle/inference/io.cc
--- a/paddle/inference/io.h
+++ b/paddle/inference/io.h
--- a/paddle/math/float16.h
+++ b/paddle/math/float16.h
--- a/paddle/memory/CMakeLists.txt
+++ b/paddle/memory/CMakeLists.txt
--- a/paddle/memory/README.md
+++ b/paddle/memory/README.md
--- a/paddle/memory/detail/buddy_allocator.cc
+++ b/paddle/memory/detail/buddy_allocator.cc
--- a/paddle/memory/detail/buddy_allocator.h
+++ b/paddle/memory/detail/buddy_allocator.h
--- a/paddle/memory/detail/memory_block.cc
+++ b/paddle/memory/detail/memory_block.cc
--- a/paddle/memory/detail/memory_block.h
+++ b/paddle/memory/detail/memory_block.h
--- a/paddle/memory/detail/meta_cache.cc
+++ b/paddle/memory/detail/meta_cache.cc
--- a/paddle/memory/detail/meta_cache.h
+++ b/paddle/memory/detail/meta_cache.h
--- a/paddle/memory/detail/meta_data.cc
+++ b/paddle/memory/detail/meta_data.cc
--- a/paddle/memory/detail/meta_data.h
+++ b/paddle/memory/detail/meta_data.h
--- a/paddle/memory/detail/system_allocator.cc
+++ b/paddle/memory/detail/system_allocator.cc
--- a/paddle/memory/memcpy.cc
+++ b/paddle/memory/memcpy.cc
--- a/paddle/memory/memory.cc
+++ b/paddle/memory/memory.cc
--- a/paddle/memory/memory.h
+++ b/paddle/memory/memory.h
--- a/paddle/memory/memory_test.cc
+++ b/paddle/memory/memory_test.cc
--- a/paddle/operators/CMakeLists.txt
+++ b/paddle/operators/CMakeLists.txt
--- a/paddle/operators/accuracy_op.cc
+++ b/paddle/operators/accuracy_op.cc
--- a/paddle/operators/accuracy_op.cu
+++ b/paddle/operators/accuracy_op.cu
--- a/paddle/operators/activation_op.cc
+++ b/paddle/operators/activation_op.cc
--- a/paddle/operators/activation_op.cu
+++ b/paddle/operators/activation_op.cu
--- a/paddle/operators/activation_op.h
+++ b/paddle/operators/activation_op.h
--- a/paddle/operators/adadelta_op.cc
+++ b/paddle/operators/adadelta_op.cc
--- a/paddle/operators/adadelta_op.cu
+++ b/paddle/operators/adadelta_op.cu
--- a/paddle/operators/adagrad_op.cc
+++ b/paddle/operators/adagrad_op.cc
--- a/paddle/operators/adagrad_op.cu
+++ b/paddle/operators/adagrad_op.cu
--- a/paddle/operators/adagrad_op.h
+++ b/paddle/operators/adagrad_op.h
--- a/paddle/operators/adam_op.cc
+++ b/paddle/operators/adam_op.cc
--- a/paddle/operators/adam_op.cu
+++ b/paddle/operators/adam_op.cu
--- a/paddle/operators/adam_op.h
+++ b/paddle/operators/adam_op.h
--- a/paddle/operators/adamax_op.cc
+++ b/paddle/operators/adamax_op.cc
--- a/paddle/operators/adamax_op.cu
+++ b/paddle/operators/adamax_op.cu
--- a/paddle/operators/array_operator.h
+++ b/paddle/operators/array_operator.h
--- a/paddle/operators/array_to_lod_tensor_op.cc
+++ b/paddle/operators/array_to_lod_tensor_op.cc
--- a/paddle/operators/assign_op.cc
+++ b/paddle/operators/assign_op.cc
--- a/paddle/operators/assign_value_op.cc
+++ b/paddle/operators/assign_value_op.cc
--- a/paddle/operators/assign_value_op.cu.cc
+++ b/paddle/operators/assign_value_op.cu.cc
--- a/paddle/operators/assign_value_op.h
+++ b/paddle/operators/assign_value_op.h
--- a/paddle/operators/auc_op.cc
+++ b/paddle/operators/auc_op.cc
--- a/paddle/operators/batch_norm_op.cc
+++ b/paddle/operators/batch_norm_op.cc
--- a/paddle/operators/batch_norm_op.cu.cc
+++ b/paddle/operators/batch_norm_op.cu.cc
--- a/paddle/operators/batch_norm_op.h
+++ b/paddle/operators/batch_norm_op.h
--- a/paddle/operators/beam_search_decode_op.cc
+++ b/paddle/operators/beam_search_decode_op.cc
--- a/paddle/operators/beam_search_op.cc
+++ b/paddle/operators/beam_search_op.cc
--- a/paddle/operators/beam_search_op.h
+++ b/paddle/operators/beam_search_op.h
--- a/paddle/operators/beam_search_op_test.cc
+++ b/paddle/operators/beam_search_op_test.cc
--- a/paddle/operators/bilinear_tensor_product_op.cc
+++ b/paddle/operators/bilinear_tensor_product_op.cc
--- a/paddle/operators/bipartite_match_op.cc
+++ b/paddle/operators/bipartite_match_op.cc
--- a/paddle/operators/cast_op.cc
+++ b/paddle/operators/cast_op.cc
--- a/paddle/operators/cast_op.cu
+++ b/paddle/operators/cast_op.cu
--- a/paddle/operators/cast_op.h
+++ b/paddle/operators/cast_op.h
--- a/paddle/operators/chunk_eval_op.cc
+++ b/paddle/operators/chunk_eval_op.cc
--- a/paddle/operators/chunk_eval_op.h
+++ b/paddle/operators/chunk_eval_op.h
--- a/paddle/operators/clip_by_norm_op.cc
+++ b/paddle/operators/clip_by_norm_op.cc
--- a/paddle/operators/clip_by_norm_op.cu
+++ b/paddle/operators/clip_by_norm_op.cu
--- a/paddle/operators/clip_by_norm_op.h
+++ b/paddle/operators/clip_by_norm_op.h
--- a/paddle/operators/clip_op.cc
+++ b/paddle/operators/clip_op.cc
--- a/paddle/operators/clip_op.cu
+++ b/paddle/operators/clip_op.cu
--- a/paddle/operators/clip_op.h
+++ b/paddle/operators/clip_op.h
--- a/paddle/operators/compare_op.cc
+++ b/paddle/operators/compare_op.cc
--- a/paddle/operators/compare_op.cu
+++ b/paddle/operators/compare_op.cu
--- a/paddle/operators/compare_op.h
+++ b/paddle/operators/compare_op.h
--- a/paddle/operators/concat_op.cc
+++ b/paddle/operators/concat_op.cc
--- a/paddle/operators/cond_op.cc
+++ b/paddle/operators/cond_op.cc
--- a/paddle/operators/cond_op.h
+++ b/paddle/operators/cond_op.h
--- a/paddle/operators/conditional_block_op.cc
+++ b/paddle/operators/conditional_block_op.cc
--- a/paddle/operators/conv_cudnn_op.cc
+++ b/paddle/operators/conv_cudnn_op.cc
--- a/paddle/operators/conv_cudnn_op.cu.cc
+++ b/paddle/operators/conv_cudnn_op.cu.cc
--- a/paddle/operators/conv_op.cc
+++ b/paddle/operators/conv_op.cc
--- a/paddle/operators/conv_op.cu.cc
+++ b/paddle/operators/conv_op.cu.cc
--- a/paddle/operators/conv_op.h
+++ b/paddle/operators/conv_op.h
--- a/paddle/operators/conv_shift_op.cc
+++ b/paddle/operators/conv_shift_op.cc
--- a/paddle/operators/conv_shift_op.cu
+++ b/paddle/operators/conv_shift_op.cu
--- a/paddle/operators/conv_shift_op.h
+++ b/paddle/operators/conv_shift_op.h
--- a/paddle/operators/conv_transpose_cudnn_op.cc
+++ b/paddle/operators/conv_transpose_cudnn_op.cc
--- a/paddle/operators/conv_transpose_cudnn_op.cu.cc
+++ b/paddle/operators/conv_transpose_cudnn_op.cu.cc
--- a/paddle/operators/conv_transpose_op.cc
+++ b/paddle/operators/conv_transpose_op.cc
--- a/paddle/operators/conv_transpose_op.cu.cc
+++ b/paddle/operators/conv_transpose_op.cu.cc
--- a/paddle/operators/conv_transpose_op.h
+++ b/paddle/operators/conv_transpose_op.h
--- a/paddle/operators/cos_sim_op.cc
+++ b/paddle/operators/cos_sim_op.cc
--- a/paddle/operators/cos_sim_op.cu
+++ b/paddle/operators/cos_sim_op.cu
--- a/paddle/operators/cos_sim_op.h
+++ b/paddle/operators/cos_sim_op.h
--- a/paddle/operators/crf_decoding_op.cc
+++ b/paddle/operators/crf_decoding_op.cc
--- a/paddle/operators/crf_decoding_op.h
+++ b/paddle/operators/crf_decoding_op.h
--- a/paddle/operators/crop_op.cc
+++ b/paddle/operators/crop_op.cc
--- a/paddle/operators/crop_op.cu
+++ b/paddle/operators/crop_op.cu
--- a/paddle/operators/crop_op.h
+++ b/paddle/operators/crop_op.h
--- a/paddle/operators/cross_entropy_op.cc
+++ b/paddle/operators/cross_entropy_op.cc
--- a/paddle/operators/cross_entropy_op.cu
+++ b/paddle/operators/cross_entropy_op.cu
--- a/paddle/operators/ctc_align_op.cc
+++ b/paddle/operators/ctc_align_op.cc
--- a/paddle/operators/ctc_align_op.cu
+++ b/paddle/operators/ctc_align_op.cu
--- a/paddle/operators/ctc_align_op.h
+++ b/paddle/operators/ctc_align_op.h
--- a/paddle/operators/decayed_adagrad_op.cc
+++ b/paddle/operators/decayed_adagrad_op.cc
--- a/paddle/operators/decayed_adagrad_op.cu
+++ b/paddle/operators/decayed_adagrad_op.cu
--- a/paddle/operators/detail/CMakeLists.txt
+++ b/paddle/operators/detail/CMakeLists.txt
--- a/paddle/operators/detail/grpc_client.cc
+++ b/paddle/operators/detail/grpc_client.cc
--- a/paddle/operators/detail/grpc_client.h
+++ b/paddle/operators/detail/grpc_client.h
--- a/paddle/operators/detail/grpc_server.cc
+++ b/paddle/operators/detail/grpc_server.cc
--- a/paddle/operators/detail/grpc_server.h
+++ b/paddle/operators/detail/grpc_server.h
--- a/paddle/operators/detail/safe_ref.h
+++ b/paddle/operators/detail/safe_ref.h
--- a/paddle/operators/detail/send_impl.cc
+++ b/paddle/operators/detail/send_impl.cc
--- a/paddle/operators/detail/send_recv.proto
+++ b/paddle/operators/detail/send_recv.proto
--- a/paddle/operators/detail/send_recv_impl.h
+++ b/paddle/operators/detail/send_recv_impl.h
--- a/paddle/operators/detail/sendrecvop_utils.cc
+++ b/paddle/operators/detail/sendrecvop_utils.cc
--- a/paddle/operators/detail/sendrecvop_utils.h
+++ b/paddle/operators/detail/sendrecvop_utils.h
--- a/paddle/operators/detail/simple_block_queue.h
+++ b/paddle/operators/detail/simple_block_queue.h
--- a/paddle/operators/detail/strided_memcpy.h
+++ b/paddle/operators/detail/strided_memcpy.h
--- a/paddle/operators/detection_map_op.cc
+++ b/paddle/operators/detection_map_op.cc
--- a/paddle/operators/detection_output_op.cc
+++ b/paddle/operators/detection_output_op.cc
--- a/paddle/operators/detection_output_op.cu.cc
+++ b/paddle/operators/detection_output_op.cu.cc
--- a/paddle/operators/detection_output_op.h
+++ b/paddle/operators/detection_output_op.h
--- a/paddle/operators/dropout_op.cc
+++ b/paddle/operators/dropout_op.cc
--- a/paddle/operators/dropout_op.cu
+++ b/paddle/operators/dropout_op.cu
--- a/paddle/operators/dropout_op.h
+++ b/paddle/operators/dropout_op.h
--- a/paddle/operators/edit_distance_op.cc
+++ b/paddle/operators/edit_distance_op.cc
--- a/paddle/operators/edit_distance_op.cu
+++ b/paddle/operators/edit_distance_op.cu
--- a/paddle/operators/edit_distance_op.h
+++ b/paddle/operators/edit_distance_op.h
--- a/paddle/operators/elementwise_add_op.cc
+++ b/paddle/operators/elementwise_add_op.cc
--- a/paddle/operators/elementwise_add_op.cu
+++ b/paddle/operators/elementwise_add_op.cu
--- a/paddle/operators/elementwise_add_op.h
+++ b/paddle/operators/elementwise_add_op.h
--- a/paddle/operators/elementwise_div_op.cc
+++ b/paddle/operators/elementwise_div_op.cc
--- a/paddle/operators/elementwise_div_op.cu
+++ b/paddle/operators/elementwise_div_op.cu
--- a/paddle/operators/elementwise_div_op.h
+++ b/paddle/operators/elementwise_div_op.h
--- a/paddle/operators/elementwise_max_op.cc
+++ b/paddle/operators/elementwise_max_op.cc
--- a/paddle/operators/pool_cudnn_op.cc
+++ b/paddle/operators/pool_cudnn_op.cc
--- a/paddle/operators/elementwise_max_op.h
+++ b/paddle/operators/elementwise_max_op.h
--- a/paddle/operators/elementwise_min_op.cc
+++ b/paddle/operators/elementwise_min_op.cc
--- a/paddle/operators/elementwise_min_op.cu
+++ b/paddle/operators/elementwise_min_op.cu
--- a/paddle/operators/elementwise_min_op.h
+++ b/paddle/operators/elementwise_min_op.h
--- a/paddle/operators/elementwise_mul_op.cc
+++ b/paddle/operators/elementwise_mul_op.cc
--- a/paddle/operators/elementwise_mul_op.cu
+++ b/paddle/operators/elementwise_mul_op.cu
--- a/paddle/operators/elementwise_mul_op.h
+++ b/paddle/operators/elementwise_mul_op.h
--- a/paddle/operators/elementwise_op.h
+++ b/paddle/operators/elementwise_op.h
--- a/paddle/operators/elementwise_op_function.h
+++ b/paddle/operators/elementwise_op_function.h
--- a/paddle/operators/elementwise_sub_op.cc
+++ b/paddle/operators/elementwise_sub_op.cc
--- a/paddle/operators/elementwise_sub_op.cu
+++ b/paddle/operators/elementwise_sub_op.cu
--- a/paddle/operators/elementwise_sub_op.h
+++ b/paddle/operators/elementwise_sub_op.h
--- a/paddle/operators/expand_op.cc
+++ b/paddle/operators/expand_op.cc
--- a/paddle/operators/expand_op.cu
+++ b/paddle/operators/expand_op.cu
--- a/paddle/operators/expand_op.h
+++ b/paddle/operators/expand_op.h
--- a/paddle/operators/feed_op.cc
+++ b/paddle/operators/feed_op.cc
--- a/paddle/operators/fetch_op.cc
+++ b/paddle/operators/fetch_op.cc
--- a/paddle/operators/fill_constant_batch_size_like_op.cc
+++ b/paddle/operators/fill_constant_batch_size_like_op.cc
--- a/paddle/operators/fill_constant_batch_size_like_op.cu.cc
+++ b/paddle/operators/fill_constant_batch_size_like_op.cu.cc
--- a/paddle/operators/fill_constant_op.cc
+++ b/paddle/operators/fill_constant_op.cc
--- a/paddle/operators/fill_op.cc
+++ b/paddle/operators/fill_op.cc
--- a/paddle/operators/fill_zeros_like_op.cc
+++ b/paddle/operators/fill_zeros_like_op.cc
--- a/paddle/operators/fill_zeros_like_op.cu.cc
+++ b/paddle/operators/fill_zeros_like_op.cu.cc
--- a/paddle/operators/fill_zeros_like_op.h
+++ b/paddle/operators/fill_zeros_like_op.h
--- a/paddle/operators/ftrl_op.cc
+++ b/paddle/operators/ftrl_op.cc
--- a/paddle/operators/gather.cu.h
+++ b/paddle/operators/gather.cu.h
--- a/paddle/operators/gather_op.cc
+++ b/paddle/operators/gather_op.cc
--- a/paddle/operators/gather_op.cu
+++ b/paddle/operators/gather_op.cu
--- a/paddle/operators/gaussian_random_op.cc
+++ b/paddle/operators/gaussian_random_op.cc
--- a/paddle/operators/gaussian_random_op.cu
+++ b/paddle/operators/gaussian_random_op.cu
--- a/paddle/operators/get_places_op.cc
+++ b/paddle/operators/get_places_op.cc
--- a/paddle/operators/gru_op.cc
+++ b/paddle/operators/gru_op.cc
--- a/paddle/operators/gru_op.cu.cc
+++ b/paddle/operators/gru_op.cu.cc
--- a/paddle/operators/gru_op.h
+++ b/paddle/operators/gru_op.h
--- a/paddle/operators/gru_unit_op.cc
+++ b/paddle/operators/gru_unit_op.cc
--- a/paddle/operators/gru_unit_op.cu
+++ b/paddle/operators/gru_unit_op.cu
--- a/paddle/operators/gru_unit_op.h
+++ b/paddle/operators/gru_unit_op.h
--- a/paddle/operators/hinge_loss_op.cc
+++ b/paddle/operators/hinge_loss_op.cc
--- a/paddle/operators/hinge_loss_op.cu
+++ b/paddle/operators/hinge_loss_op.cu
--- a/paddle/operators/huber_loss_op.cc
+++ b/paddle/operators/huber_loss_op.cc
--- a/paddle/operators/huber_loss_op.cu
+++ b/paddle/operators/huber_loss_op.cu
--- a/paddle/operators/im2sequence_op.cc
+++ b/paddle/operators/im2sequence_op.cc
--- a/paddle/operators/seq_expand_op.cu
+++ b/paddle/operators/seq_expand_op.cu
--- a/paddle/operators/im2sequence_op.h
+++ b/paddle/operators/im2sequence_op.h
--- a/paddle/operators/increment_op.cc
+++ b/paddle/operators/increment_op.cc
--- a/paddle/operators/iou_similarity_op.cc
+++ b/paddle/operators/iou_similarity_op.cc
--- a/paddle/operators/iou_similarity_op.cu
+++ b/paddle/operators/iou_similarity_op.cu
--- a/paddle/operators/iou_similarity_op.h
+++ b/paddle/operators/iou_similarity_op.h
--- a/paddle/operators/is_empty_op.cc
+++ b/paddle/operators/is_empty_op.cc
--- a/paddle/operators/l1_norm_op.cc
+++ b/paddle/operators/l1_norm_op.cc
--- a/paddle/operators/l1_norm_op.cu
+++ b/paddle/operators/l1_norm_op.cu
--- a/paddle/operators/l1_norm_op.h
+++ b/paddle/operators/l1_norm_op.h
--- a/paddle/operators/linear_chain_crf_op.cc
+++ b/paddle/operators/linear_chain_crf_op.cc
--- a/paddle/operators/linear_chain_crf_op.cu
+++ b/paddle/operators/linear_chain_crf_op.cu
--- a/paddle/operators/linear_chain_crf_op.h
+++ b/paddle/operators/linear_chain_crf_op.h
--- a/paddle/operators/load_op.cc
+++ b/paddle/operators/load_op.cc
--- a/paddle/operators/lod_array_length_op.cc
+++ b/paddle/operators/lod_array_length_op.cc
--- a/paddle/operators/lod_rank_table_op.cc
+++ b/paddle/operators/lod_rank_table_op.cc
--- a/paddle/operators/lod_reset_op.cc
+++ b/paddle/operators/lod_reset_op.cc
--- a/paddle/operators/lod_reset_op.cu
+++ b/paddle/operators/lod_reset_op.cu
--- a/paddle/operators/lod_reset_op.h
+++ b/paddle/operators/lod_reset_op.h
--- a/paddle/operators/lod_tensor_to_array_op.cc
+++ b/paddle/operators/lod_tensor_to_array_op.cc
--- a/paddle/operators/log_loss_op.cc
+++ b/paddle/operators/log_loss_op.cc
--- a/paddle/operators/log_loss_op.cu
+++ b/paddle/operators/log_loss_op.cu
--- a/paddle/operators/logical_op.cc
+++ b/paddle/operators/logical_op.cc
--- a/paddle/operators/logical_op.cu
+++ b/paddle/operators/logical_op.cu
--- a/paddle/operators/logical_op.h
+++ b/paddle/operators/logical_op.h
--- a/paddle/operators/lookup_table_op.cc
+++ b/paddle/operators/lookup_table_op.cc
--- a/paddle/operators/lookup_table_op.cu
+++ b/paddle/operators/lookup_table_op.cu
--- a/paddle/operators/lookup_table_op.h
+++ b/paddle/operators/lookup_table_op.h
--- a/paddle/operators/lrn_op.cc
+++ b/paddle/operators/lrn_op.cc
--- a/paddle/operators/lrn_op.cu
+++ b/paddle/operators/lrn_op.cu
--- a/paddle/operators/lrn_op.h
+++ b/paddle/operators/lrn_op.h
--- a/paddle/operators/lstm_op.cc
+++ b/paddle/operators/lstm_op.cc
--- a/paddle/operators/lstm_op.cu.cc
+++ b/paddle/operators/lstm_op.cu.cc
--- a/paddle/operators/lstm_op.h
+++ b/paddle/operators/lstm_op.h
--- a/paddle/operators/lstm_unit_op.cc
+++ b/paddle/operators/lstm_unit_op.cc
--- a/paddle/operators/lstm_unit_op.cu
+++ b/paddle/operators/lstm_unit_op.cu
--- a/paddle/operators/lstm_unit_op.h
+++ b/paddle/operators/lstm_unit_op.h
--- a/paddle/operators/lstmp_op.cc
+++ b/paddle/operators/lstmp_op.cc
--- a/paddle/operators/lstmp_op.cu
+++ b/paddle/operators/lstmp_op.cu
--- a/paddle/operators/lstmp_op.h
+++ b/paddle/operators/lstmp_op.h
--- a/paddle/operators/margin_rank_loss_op.cc
+++ b/paddle/operators/margin_rank_loss_op.cc
--- a/paddle/operators/margin_rank_loss_op.cu
+++ b/paddle/operators/margin_rank_loss_op.cu
--- a/paddle/operators/margin_rank_loss_op.h
+++ b/paddle/operators/margin_rank_loss_op.h
--- a/paddle/operators/math/CMakeLists.txt
+++ b/paddle/operators/math/CMakeLists.txt
--- a/paddle/operators/math/context_project.h
+++ b/paddle/operators/math/context_project.h
--- a/paddle/operators/math/cos_sim_functor.cc
+++ b/paddle/operators/math/cos_sim_functor.cc
--- a/paddle/operators/math/cos_sim_functor.cu
+++ b/paddle/operators/math/cos_sim_functor.cu
--- a/paddle/operators/math/cos_sim_functor.h
+++ b/paddle/operators/math/cos_sim_functor.h
--- a/paddle/operators/math/cross_entropy.cc
+++ b/paddle/operators/math/cross_entropy.cc
--- a/paddle/operators/math/cross_entropy.cu
+++ b/paddle/operators/math/cross_entropy.cu
--- a/paddle/operators/math/cross_entropy.h
+++ b/paddle/operators/math/cross_entropy.h
--- a/paddle/operators/math/detail/activation_functions.h
+++ b/paddle/operators/math/detail/activation_functions.h
--- a/paddle/operators/math/detail/gru_cpu_kernel.h
+++ b/paddle/operators/math/detail/gru_cpu_kernel.h
--- a/paddle/operators/math/detail/gru_gpu_kernel.h
+++ b/paddle/operators/math/detail/gru_gpu_kernel.h
--- a/paddle/operators/math/detail/gru_kernel.h
+++ b/paddle/operators/math/detail/gru_kernel.h
--- a/paddle/operators/math/detail/lstm_cpu_kernel.h
+++ b/paddle/operators/math/detail/lstm_cpu_kernel.h
--- a/paddle/operators/math/detail/lstm_gpu_kernel.h
+++ b/paddle/operators/math/detail/lstm_gpu_kernel.h
--- a/paddle/operators/math/detail/lstm_kernel.h
+++ b/paddle/operators/math/detail/lstm_kernel.h
--- a/paddle/operators/math/detection_util.h
+++ b/paddle/operators/math/detection_util.h
--- a/paddle/operators/math/gru_compute.cc
+++ b/paddle/operators/math/gru_compute.cc
--- a/paddle/operators/math/gru_compute.cu
+++ b/paddle/operators/math/gru_compute.cu
--- a/paddle/operators/math/gru_compute.h
+++ b/paddle/operators/math/gru_compute.h
--- a/paddle/operators/math/im2col.cc
+++ b/paddle/operators/math/im2col.cc
--- a/paddle/operators/math/im2col_test.cc
+++ b/paddle/operators/math/im2col_test.cc
--- a/paddle/operators/math/lstm_compute.cc
+++ b/paddle/operators/math/lstm_compute.cc
--- a/paddle/operators/math/lstm_compute.cu
+++ b/paddle/operators/math/lstm_compute.cu
--- a/paddle/operators/math/lstm_compute.h
+++ b/paddle/operators/math/lstm_compute.h
--- a/paddle/operators/math/math_function.cc
+++ b/paddle/operators/math/math_function.cc
--- a/paddle/operators/math/math_function.cu
+++ b/paddle/operators/math/math_function.cu
--- a/paddle/operators/math/math_function_impl.h
+++ b/paddle/operators/math/math_function_impl.h
--- a/paddle/operators/math/math_function_test.cc
+++ b/paddle/operators/math/math_function_test.cc
--- a/paddle/operators/math/math_function_test.cu
+++ b/paddle/operators/math/math_function_test.cu
--- a/paddle/operators/math/matmul.h
+++ b/paddle/operators/math/matmul.h
--- a/paddle/operators/math/sampler.cc
+++ b/paddle/operators/math/sampler.cc
--- a/paddle/operators/math/sampler.h
+++ b/paddle/operators/math/sampler.h
--- a/paddle/operators/math/selected_rows_functor.cc
+++ b/paddle/operators/math/selected_rows_functor.cc
--- a/paddle/operators/math/selected_rows_functor.cu
+++ b/paddle/operators/math/selected_rows_functor.cu
--- a/paddle/operators/math/selected_rows_functor.h
+++ b/paddle/operators/math/selected_rows_functor.h
--- a/paddle/operators/math/selected_rows_functor_test.cu
+++ b/paddle/operators/math/selected_rows_functor_test.cu
--- a/paddle/operators/math/sequence2batch.cc
+++ b/paddle/operators/math/sequence2batch.cc
--- a/paddle/operators/math/sequence_padding.cc
+++ b/paddle/operators/math/sequence_padding.cc
--- a/paddle/operators/math/sequence_padding.cu
+++ b/paddle/operators/math/sequence_padding.cu
--- a/paddle/operators/math/sequence_padding.h
+++ b/paddle/operators/math/sequence_padding.h
--- a/paddle/operators/math/sequence_padding_test.cc
+++ b/paddle/operators/math/sequence_padding_test.cc
--- a/paddle/operators/math/sequence_scale.cc
+++ b/paddle/operators/math/sequence_scale.cc
--- a/paddle/operators/math/sequence_scale.cu
+++ b/paddle/operators/math/sequence_scale.cu
--- a/paddle/operators/math/sequence_scale.h
+++ b/paddle/operators/math/sequence_scale.h
--- a/paddle/operators/math/vol2col_test.cc
+++ b/paddle/operators/math/vol2col_test.cc
--- a/paddle/operators/matmul_op.cc
+++ b/paddle/operators/matmul_op.cc
--- a/paddle/operators/matmul_op.cu.cc
+++ b/paddle/operators/matmul_op.cu.cc
--- a/paddle/operators/matmul_op.h
+++ b/paddle/operators/matmul_op.h
--- a/paddle/operators/max_sequence_len_op.cc
+++ b/paddle/operators/max_sequence_len_op.cc
--- a/paddle/operators/maxout_op.cc
+++ b/paddle/operators/maxout_op.cc
--- a/paddle/operators/maxout_op.cu.cc
+++ b/paddle/operators/maxout_op.cu.cc
--- a/paddle/operators/mean_op.cc
+++ b/paddle/operators/mean_op.cc
--- a/paddle/operators/mean_op.cu
+++ b/paddle/operators/mean_op.cu
--- a/paddle/operators/merge_lod_tensor_op.cc
+++ b/paddle/operators/merge_lod_tensor_op.cc
--- a/paddle/operators/minus_op.cc
+++ b/paddle/operators/minus_op.cc
--- a/paddle/operators/minus_op.cu
+++ b/paddle/operators/minus_op.cu
--- a/paddle/operators/minus_op.h
+++ b/paddle/operators/minus_op.h
--- a/paddle/operators/modified_huber_loss_op.cc
+++ b/paddle/operators/modified_huber_loss_op.cc
--- a/paddle/operators/modified_huber_loss_op.cu
+++ b/paddle/operators/modified_huber_loss_op.cu
--- a/paddle/operators/modified_huber_loss_op.h
+++ b/paddle/operators/modified_huber_loss_op.h
--- a/paddle/operators/momentum_op.cc
+++ b/paddle/operators/momentum_op.cc
--- a/paddle/operators/momentum_op.cu
+++ b/paddle/operators/momentum_op.cu
--- a/paddle/operators/mul_op.cc
+++ b/paddle/operators/mul_op.cc
--- a/paddle/operators/mul_op.cu.cc
+++ b/paddle/operators/mul_op.cu.cc
--- a/paddle/operators/mul_op.h
+++ b/paddle/operators/mul_op.h
--- a/paddle/operators/multiplex_op.cc
+++ b/paddle/operators/multiplex_op.cc
--- a/paddle/operators/multiplex_op.cu
+++ b/paddle/operators/multiplex_op.cu
--- a/paddle/operators/multiplex_op.h
+++ b/paddle/operators/multiplex_op.h
--- a/paddle/operators/nccl/nccl_gpu_common.cc
+++ b/paddle/operators/nccl/nccl_gpu_common.cc
--- a/paddle/operators/nccl/nccl_gpu_common.h
+++ b/paddle/operators/nccl/nccl_gpu_common.h
--- a/paddle/operators/nccl_op.cc
+++ b/paddle/operators/nccl_op.cc
--- a/paddle/operators/nccl_op.cu.cc
+++ b/paddle/operators/nccl_op.cu.cc
--- a/paddle/operators/nccl_op_test.cu.cc
+++ b/paddle/operators/nccl_op_test.cu.cc
--- a/paddle/operators/nce_op.cc
+++ b/paddle/operators/nce_op.cc
--- a/paddle/operators/nce_op.h
+++ b/paddle/operators/nce_op.h
--- a/paddle/operators/net_op.cc
+++ b/paddle/operators/net_op.cc
--- a/paddle/operators/net_op.h
+++ b/paddle/operators/net_op.h
--- a/paddle/operators/net_op_test.cc
+++ b/paddle/operators/net_op_test.cc
--- a/paddle/operators/norm_op.cc
+++ b/paddle/operators/norm_op.cc
--- a/paddle/operators/norm_op.cu
+++ b/paddle/operators/norm_op.cu
--- a/paddle/operators/norm_op.h
+++ b/paddle/operators/norm_op.h
--- a/paddle/operators/one_hot_op.cc
+++ b/paddle/operators/one_hot_op.cc
--- a/paddle/operators/one_hot_op.cu
+++ b/paddle/operators/one_hot_op.cu
--- a/paddle/operators/one_hot_op.h
+++ b/paddle/operators/one_hot_op.h
--- a/paddle/operators/batch_norm_op.md
+++ b/paddle/operators/batch_norm_op.md
--- a/paddle/operators/name_convention.md
+++ b/paddle/operators/name_convention.md
--- a/paddle/operators/net_op_design.md
+++ b/paddle/operators/net_op_design.md
--- a/paddle/operators/op_documentation/op_markdown_format.md
+++ b/paddle/operators/op_documentation/op_markdown_format.md
--- a/paddle/operators/rnn_design.md
+++ b/paddle/operators/rnn_design.md
--- a/paddle/operators/pad_op.cc
+++ b/paddle/operators/pad_op.cc
--- a/paddle/operators/pad_op.cu
+++ b/paddle/operators/pad_op.cu
--- a/paddle/operators/pad_op.h
+++ b/paddle/operators/pad_op.h
--- a/paddle/operators/parallel_do_op.cc
+++ b/paddle/operators/parallel_do_op.cc
--- a/paddle/operators/pool_cudnn_op.cu.cc
+++ b/paddle/operators/pool_cudnn_op.cu.cc
--- a/paddle/operators/pool_op.cc
+++ b/paddle/operators/pool_op.cc
--- a/paddle/operators/pool_op.h
+++ b/paddle/operators/pool_op.h
--- a/paddle/operators/pool_with_index_op.cc
+++ b/paddle/operators/pool_with_index_op.cc
--- a/paddle/operators/positive_negative_pair_op.cc
+++ b/paddle/operators/positive_negative_pair_op.cc
--- a/paddle/operators/precision_recall_op.cc
+++ b/paddle/operators/precision_recall_op.cc
--- a/paddle/operators/prelu_op.cc
+++ b/paddle/operators/prelu_op.cc
--- a/paddle/operators/prelu_op.cu
+++ b/paddle/operators/prelu_op.cu
--- a/paddle/operators/print_op.cc
+++ b/paddle/operators/print_op.cc
--- a/paddle/operators/prior_box_op.cc
+++ b/paddle/operators/prior_box_op.cc
--- a/paddle/operators/prior_box_op.h
+++ b/paddle/operators/prior_box_op.h
--- a/paddle/operators/proximal_adagrad_op.cc
+++ b/paddle/operators/proximal_adagrad_op.cc
--- a/paddle/operators/proximal_gd_op.cc
+++ b/paddle/operators/proximal_gd_op.cc
--- a/paddle/operators/rank_loss_op.cc
+++ b/paddle/operators/rank_loss_op.cc
--- a/paddle/operators/rank_loss_op.cu
+++ b/paddle/operators/rank_loss_op.cu
--- a/paddle/operators/rank_loss_op.h
+++ b/paddle/operators/rank_loss_op.h
--- a/paddle/operators/recurrent_op.cc
+++ b/paddle/operators/recurrent_op.cc
--- a/paddle/operators/recv_op.cc
+++ b/paddle/operators/recv_op.cc
--- a/paddle/operators/reduce_op.cc
+++ b/paddle/operators/reduce_op.cc
--- a/paddle/operators/reduce_op.cu
+++ b/paddle/operators/reduce_op.cu
--- a/paddle/operators/reduce_op.h
+++ b/paddle/operators/reduce_op.h
--- a/paddle/operators/reorder_lod_tensor_by_rank_op.cc
+++ b/paddle/operators/reorder_lod_tensor_by_rank_op.cc
--- a/paddle/operators/reshape_op.cc
+++ b/paddle/operators/reshape_op.cc
--- a/paddle/operators/reshape_op.cu
+++ b/paddle/operators/reshape_op.cu
--- a/paddle/operators/reshape_op.h
+++ b/paddle/operators/reshape_op.h
--- a/paddle/operators/rmsprop_op.cc
+++ b/paddle/operators/rmsprop_op.cc
--- a/paddle/operators/rmsprop_op.cu
+++ b/paddle/operators/rmsprop_op.cu
--- a/paddle/operators/rnn_memory_helper_op.cc
+++ b/paddle/operators/rnn_memory_helper_op.cc
--- a/paddle/operators/roi_pool_op.cc
+++ b/paddle/operators/roi_pool_op.cc
--- a/paddle/operators/row_conv_op.cc
+++ b/paddle/operators/row_conv_op.cc
--- a/paddle/operators/row_conv_op.cu
+++ b/paddle/operators/row_conv_op.cu
--- a/paddle/operators/row_conv_op.h
+++ b/paddle/operators/row_conv_op.h
--- a/paddle/operators/save_load_op_test.cc
+++ b/paddle/operators/save_load_op_test.cc
--- a/paddle/operators/save_op.cc
+++ b/paddle/operators/save_op.cc
--- a/paddle/operators/scale_op.cc
+++ b/paddle/operators/scale_op.cc
--- a/paddle/operators/scale_op.cu
+++ b/paddle/operators/scale_op.cu
--- a/paddle/operators/scale_op.h
+++ b/paddle/operators/scale_op.h
--- a/paddle/operators/scatter.cu.h
+++ b/paddle/operators/scatter.cu.h
--- a/paddle/operators/scatter_op.cc
+++ b/paddle/operators/scatter_op.cc
--- a/paddle/operators/scatter_op.cu
+++ b/paddle/operators/scatter_op.cu
--- a/paddle/operators/send_op.cc
+++ b/paddle/operators/send_op.cc
--- a/paddle/operators/send_recv_op_test.cc
+++ b/paddle/operators/send_recv_op_test.cc
--- a/paddle/operators/sequence_concat_op.cc
+++ b/paddle/operators/sequence_concat_op.cc
--- a/paddle/operators/sequence_conv_op.cc
+++ b/paddle/operators/sequence_conv_op.cc
--- a/paddle/operators/sequence_conv_op.cu.cc
+++ b/paddle/operators/sequence_conv_op.cu.cc
--- a/paddle/operators/sequence_erase_op.cc
+++ b/paddle/operators/sequence_erase_op.cc
--- a/paddle/operators/sequence_erase_op.cu
+++ b/paddle/operators/sequence_erase_op.cu
--- a/paddle/operators/sequence_erase_op.h
+++ b/paddle/operators/sequence_erase_op.h
--- a/paddle/operators/seq_expand_op.cc
+++ b/paddle/operators/seq_expand_op.cc
--- a/paddle/operators/sequence_expand_op.cu
+++ b/paddle/operators/sequence_expand_op.cu
--- a/paddle/operators/seq_expand_op.h
+++ b/paddle/operators/seq_expand_op.h
--- a/paddle/operators/sequence_pool_op.cc
+++ b/paddle/operators/sequence_pool_op.cc
--- a/paddle/operators/sequence_pool_op.cu
+++ b/paddle/operators/sequence_pool_op.cu
--- a/paddle/operators/sequence_reshape_op.cc
+++ b/paddle/operators/sequence_reshape_op.cc
--- a/paddle/operators/sequence_reshape_op.cu
+++ b/paddle/operators/sequence_reshape_op.cu
--- a/paddle/operators/sequence_reshape_op.h
+++ b/paddle/operators/sequence_reshape_op.h
--- a/paddle/operators/sequence_slice_op.cc
+++ b/paddle/operators/sequence_slice_op.cc
--- a/paddle/operators/sequence_slice_op.h
+++ b/paddle/operators/sequence_slice_op.h
--- a/paddle/operators/sequence_softmax_op.cc
+++ b/paddle/operators/sequence_softmax_op.cc
--- a/paddle/operators/sgd_op.cc
+++ b/paddle/operators/sgd_op.cc
--- a/paddle/operators/sgd_op.cu
+++ b/paddle/operators/sgd_op.cu
--- a/paddle/operators/sgd_op.h
+++ b/paddle/operators/sgd_op.h
--- a/paddle/operators/shrink_rnn_memory_op.cc
+++ b/paddle/operators/shrink_rnn_memory_op.cc
--- a/paddle/operators/sigmoid_cross_entropy_with_logits_op.cc
+++ b/paddle/operators/sigmoid_cross_entropy_with_logits_op.cc
--- a/paddle/operators/sigmoid_cross_entropy_with_logits_op.cu
+++ b/paddle/operators/sigmoid_cross_entropy_with_logits_op.cu
--- a/paddle/operators/sigmoid_cross_entropy_with_logits_op.h
+++ b/paddle/operators/sigmoid_cross_entropy_with_logits_op.h
--- a/paddle/operators/sign_op.cc
+++ b/paddle/operators/sign_op.cc
--- a/paddle/operators/sign_op.cu
+++ b/paddle/operators/sign_op.cu
--- a/paddle/operators/sign_op.h
+++ b/paddle/operators/sign_op.h
--- a/paddle/operators/smooth_l1_loss_op.cc
+++ b/paddle/operators/smooth_l1_loss_op.cc
--- a/paddle/operators/smooth_l1_loss_op.cu
+++ b/paddle/operators/smooth_l1_loss_op.cu
--- a/paddle/operators/smooth_l1_loss_op.h
+++ b/paddle/operators/smooth_l1_loss_op.h
--- a/paddle/operators/softmax_op.cc
+++ b/paddle/operators/softmax_op.cc
--- a/paddle/operators/softmax_op.cu.cc
+++ b/paddle/operators/softmax_op.cu.cc
--- a/paddle/operators/softmax_op.h
+++ b/paddle/operators/softmax_op.h
--- a/paddle/operators/softmax_with_cross_entropy_op.cc
+++ b/paddle/operators/softmax_with_cross_entropy_op.cc
--- a/paddle/operators/softmax_with_cross_entropy_op.cu
+++ b/paddle/operators/softmax_with_cross_entropy_op.cu
--- a/paddle/operators/softmax_with_cross_entropy_op.h
+++ b/paddle/operators/softmax_with_cross_entropy_op.h
--- a/paddle/operators/split_lod_tensor_op.cc
+++ b/paddle/operators/split_lod_tensor_op.cc
--- a/paddle/operators/split_op.cc
+++ b/paddle/operators/split_op.cc
--- a/paddle/operators/split_selected_rows_op.cc
+++ b/paddle/operators/split_selected_rows_op.cc
--- a/paddle/operators/split_selected_rows_op.cu
+++ b/paddle/operators/split_selected_rows_op.cu
--- a/paddle/operators/split_selected_rows_op.h
+++ b/paddle/operators/split_selected_rows_op.h
--- a/paddle/operators/spp_op.cc
+++ b/paddle/operators/spp_op.cc
--- a/paddle/operators/spp_op.cu.cc
+++ b/paddle/operators/spp_op.cu.cc
--- a/paddle/operators/spp_op.h
+++ b/paddle/operators/spp_op.h
--- a/paddle/operators/squared_l2_distance_op.cc
+++ b/paddle/operators/squared_l2_distance_op.cc
--- a/paddle/operators/squared_l2_distance_op.cu
+++ b/paddle/operators/squared_l2_distance_op.cu
--- a/paddle/operators/squared_l2_norm_op.cc
+++ b/paddle/operators/squared_l2_norm_op.cc
--- a/paddle/operators/squared_l2_norm_op.cu
+++ b/paddle/operators/squared_l2_norm_op.cu
--- a/paddle/operators/squared_l2_norm_op.h
+++ b/paddle/operators/squared_l2_norm_op.h
--- a/paddle/operators/strided_memcpy.h
+++ b/paddle/operators/strided_memcpy.h
--- a/paddle/operators/strided_memcpy_test.cc
+++ b/paddle/operators/strided_memcpy_test.cc
--- a/paddle/operators/sum_op.cc
+++ b/paddle/operators/sum_op.cc
--- a/paddle/operators/sum_op.h
+++ b/paddle/operators/sum_op.h
--- a/paddle/operators/tensor.save
+++ b/paddle/operators/tensor.save
--- a/paddle/operators/tensor_array_read_write_op.cc
+++ b/paddle/operators/tensor_array_read_write_op.cc
--- a/paddle/operators/top_k_op.cc
+++ b/paddle/operators/top_k_op.cc
--- a/paddle/operators/top_k_op.cu
+++ b/paddle/operators/top_k_op.cu
--- a/paddle/operators/top_k_op.h
+++ b/paddle/operators/top_k_op.h
--- a/paddle/operators/transpose_op.cc
+++ b/paddle/operators/transpose_op.cc
--- a/paddle/operators/transpose_op.cu.cc
+++ b/paddle/operators/transpose_op.cu.cc
--- a/paddle/operators/transpose_op.h
+++ b/paddle/operators/transpose_op.h
--- a/paddle/operators/uniform_random_op.cc
+++ b/paddle/operators/uniform_random_op.cc
--- a/paddle/operators/uniform_random_op.cu
+++ b/paddle/operators/uniform_random_op.cu
--- a/paddle/operators/unpool_op.cc
+++ b/paddle/operators/unpool_op.cc
--- a/paddle/operators/warpctc_op.cc
+++ b/paddle/operators/warpctc_op.cc
--- a/paddle/operators/warpctc_op.cu.cc
+++ b/paddle/operators/warpctc_op.cu.cc
--- a/paddle/operators/warpctc_op.h
+++ b/paddle/operators/warpctc_op.h
--- a/paddle/operators/while_op.cc
+++ b/paddle/operators/while_op.cc
--- a/paddle/optimizer/adadelta_optimizer.cc
+++ b/paddle/optimizer/adadelta_optimizer.cc
--- a/paddle/optimizer/adadelta_optimizer.h
+++ b/paddle/optimizer/adadelta_optimizer.h
--- a/paddle/optimizer/adagrad_optimizer.cc
+++ b/paddle/optimizer/adagrad_optimizer.cc
--- a/paddle/optimizer/adagrad_optimizer.h
+++ b/paddle/optimizer/adagrad_optimizer.h
--- a/paddle/optimizer/adam_optimizer.cc
+++ b/paddle/optimizer/adam_optimizer.cc
--- a/paddle/optimizer/adam_optimizer.h
+++ b/paddle/optimizer/adam_optimizer.h
--- a/paddle/optimizer/lr_policy.h
+++ b/paddle/optimizer/lr_policy.h
--- a/paddle/optimizer/optimizer.cc
+++ b/paddle/optimizer/optimizer.cc
--- a/paddle/optimizer/optimizer.h
+++ b/paddle/optimizer/optimizer.h
--- a/paddle/optimizer/parameter_optimizer.cc
+++ b/paddle/optimizer/parameter_optimizer.cc
--- a/paddle/optimizer/parameter_optimizer.h
+++ b/paddle/optimizer/parameter_optimizer.h
--- a/paddle/optimizer/parameter_optimizer_test.cc
+++ b/paddle/optimizer/parameter_optimizer_test.cc
--- a/paddle/optimizer/serialization.h
+++ b/paddle/optimizer/serialization.h
--- a/paddle/optimizer/serialization_test.cc
+++ b/paddle/optimizer/serialization_test.cc
--- a/paddle/optimizer/sgd_optimizer.cc
+++ b/paddle/optimizer/sgd_optimizer.cc
--- a/paddle/optimizer/sgd_optimizer.h
+++ b/paddle/optimizer/sgd_optimizer.h
--- a/paddle/optimizer/tensor.h
+++ b/paddle/optimizer/tensor.h
--- a/paddle/platform/CMakeLists.txt
+++ b/paddle/platform/CMakeLists.txt
--- a/paddle/platform/assert.h
+++ b/paddle/platform/assert.h
--- a/paddle/platform/call_once.h
+++ b/paddle/platform/call_once.h
--- a/paddle/platform/cpu_info_test.cc
+++ b/paddle/platform/cpu_info_test.cc
--- a/paddle/platform/cuda_profiler.h
+++ b/paddle/platform/cuda_profiler.h
--- a/paddle/platform/device_context.cc
+++ b/paddle/platform/device_context.cc
--- a/paddle/platform/device_context.h
+++ b/paddle/platform/device_context.h
--- a/paddle/platform/device_context_test.cc
+++ b/paddle/platform/device_context_test.cc
--- a/paddle/platform/dynload/CMakeLists.txt
+++ b/paddle/platform/dynload/CMakeLists.txt
--- a/paddle/platform/dynload/cublas.cc
+++ b/paddle/platform/dynload/cublas.cc
--- a/paddle/platform/dynload/cudnn.cc
+++ b/paddle/platform/dynload/cudnn.cc
--- a/paddle/platform/dynload/cudnn.h
+++ b/paddle/platform/dynload/cudnn.h
--- a/paddle/platform/dynload/dynamic_loader.cc
+++ b/paddle/platform/dynload/dynamic_loader.cc
--- a/paddle/platform/dynload/dynamic_loader.h
+++ b/paddle/platform/dynload/dynamic_loader.h
--- a/paddle/platform/dynload/nccl.cc
+++ b/paddle/platform/dynload/nccl.cc
--- a/paddle/platform/dynload/nccl.h
+++ b/paddle/platform/dynload/nccl.h
--- a/paddle/platform/dynload/warpctc.cc
+++ b/paddle/platform/dynload/warpctc.cc
--- a/paddle/platform/dynload/warpctc.h
+++ b/paddle/platform/dynload/warpctc.h
--- a/paddle/platform/enforce.h
+++ b/paddle/platform/enforce.h
--- a/paddle/platform/for_range.h
+++ b/paddle/platform/for_range.h
--- a/paddle/platform/gpu_info.cc
+++ b/paddle/platform/gpu_info.cc
--- a/paddle/platform/gpu_info.h
+++ b/paddle/platform/gpu_info.h
--- a/paddle/platform/hostdevice.h
+++ b/paddle/platform/hostdevice.h
--- a/paddle/platform/mkldnn_helper.h
+++ b/paddle/platform/mkldnn_helper.h
--- a/paddle/platform/nccl_test.cu
+++ b/paddle/platform/nccl_test.cu
--- a/paddle/platform/place.cc
+++ b/paddle/platform/place.cc
--- a/paddle/platform/place.h
+++ b/paddle/platform/place.h
--- a/paddle/platform/place_test.cc
+++ b/paddle/platform/place_test.cc
--- a/paddle/platform/profiler.cc
+++ b/paddle/platform/profiler.cc
--- a/paddle/platform/profiler.h
+++ b/paddle/platform/profiler.h
--- a/paddle/platform/profiler_test.cc
+++ b/paddle/platform/profiler_test.cc
--- a/paddle/platform/transform.h
+++ b/paddle/platform/transform.h
--- a/paddle/platform/transform_test.cu
+++ b/paddle/platform/transform_test.cu
--- a/paddle/platform/variant.h
+++ b/paddle/platform/variant.h
--- a/paddle/pybind/CMakeLists.txt
+++ b/paddle/pybind/CMakeLists.txt
--- a/paddle/pybind/const_value.cc
+++ b/paddle/pybind/const_value.cc
--- a/paddle/operators/pool_cudnn_op.h
+++ b/paddle/operators/pool_cudnn_op.h
--- a/paddle/pybind/exception.cc
+++ b/paddle/pybind/exception.cc
--- a/paddle/pybind/exception.h
+++ b/paddle/pybind/exception.h
--- a/paddle/pybind/print_operators_doc.cc
+++ b/paddle/pybind/print_operators_doc.cc
--- a/paddle/pybind/protobuf.cc
+++ b/paddle/pybind/protobuf.cc
--- a/paddle/pybind/protobuf.h
+++ b/paddle/pybind/protobuf.h
--- a/paddle/pybind/pybind.cc
+++ b/paddle/pybind/pybind.cc
--- a/paddle/pybind/tensor_py.h
+++ b/paddle/pybind/tensor_py.h
--- a/paddle/scripts/CMakeLists.txt
+++ b/paddle/scripts/CMakeLists.txt
--- a/paddle/scripts/cluster_train/paddle.py
+++ b/paddle/scripts/cluster_train/paddle.py
--- a/paddle/scripts/cluster_train_v2/openmpi/docker_cluster/Dockerfile
+++ b/paddle/scripts/cluster_train_v2/openmpi/docker_cluster/Dockerfile
--- a/paddle/scripts/docker/README.md
+++ b/paddle/scripts/docker/README.md
--- a/paddle/scripts/docker/build.sh
+++ b/paddle/scripts/docker/build.sh
--- a/paddle/scripts/submit_local.sh.in
+++ b/paddle/scripts/submit_local.sh.in
--- a/paddle/scripts/tools/build_docs/build_docs.sh
+++ b/paddle/scripts/tools/build_docs/build_docs.sh
--- a/paddle/scripts/tools/usage_stat/usage.sh
+++ b/paddle/scripts/tools/usage_stat/usage.sh
--- a/paddle/scripts/travis/build_doc.sh
+++ b/paddle/scripts/travis/build_doc.sh
--- a/paddle/string/CMakeLists.txt
+++ b/paddle/string/CMakeLists.txt
--- a/paddle/string/piece.cc
+++ b/paddle/string/piece.cc
--- a/paddle/string/piece.h
+++ b/paddle/string/piece.h
--- a/paddle/string/piece_test.cc
+++ b/paddle/string/piece_test.cc
--- a/paddle/string/printf.h
+++ b/paddle/string/printf.h
--- a/paddle/string/printf_test.cc
+++ b/paddle/string/printf_test.cc
--- a/paddle/string/tinyformat/tinyformat.h
+++ b/paddle/string/tinyformat/tinyformat.h
--- a/paddle/string/to_string.h
+++ b/paddle/string/to_string.h
--- a/paddle/string/to_string_test.cc
+++ b/paddle/string/to_string_test.cc
--- a/paddle/testing/CMakeLists.txt
+++ b/paddle/testing/CMakeLists.txt
--- a/paddle/testing/paddle_gtest_main.cc
+++ b/paddle/testing/paddle_gtest_main.cc
--- a/paddle/trainer/TrainerConfigHelper.cpp
+++ b/paddle/trainer/TrainerConfigHelper.cpp
--- a/paddle/trainer/tests/picojson.h
+++ b/paddle/trainer/tests/picojson.h
--- a/paddle/trainer/tests/simple_sparse_neural_network.py
+++ b/paddle/trainer/tests/simple_sparse_neural_network.py
--- a/paddle/trainer/tests/simple_sparse_neural_network_dp.py
+++ b/paddle/trainer/tests/simple_sparse_neural_network_dp.py
--- a/paddle/utils/Flags.cpp
+++ b/paddle/utils/Flags.cpp
--- a/paddle/utils/Flags.h
+++ b/paddle/utils/Flags.h
--- a/paddle/utils/enable_virtualenv.py
+++ b/paddle/utils/enable_virtualenv.py
--- a/proto/OptimizerConfig.proto
+++ b/proto/OptimizerConfig.proto
--- a/python/.gitignore
+++ b/python/.gitignore
--- a/python/CMakeLists.txt
+++ b/python/CMakeLists.txt
--- a/python/paddle/trainer/config_parser.py
+++ b/python/paddle/trainer/config_parser.py
--- a/python/paddle/trainer_config_helpers/attrs.py
+++ b/python/paddle/trainer_config_helpers/attrs.py
--- a/python/paddle/trainer_config_helpers/evaluators.py
+++ b/python/paddle/trainer_config_helpers/evaluators.py
--- a/python/paddle/trainer_config_helpers/layers.py
+++ b/python/paddle/trainer_config_helpers/layers.py
--- a/python/paddle/trainer_config_helpers/networks.py
+++ b/python/paddle/trainer_config_helpers/networks.py
--- a/python/paddle/trainer_config_helpers/tests/configs/img_layers.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/img_layers.py
--- a/python/paddle/trainer_config_helpers/tests/configs/img_trans_layers.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/img_trans_layers.py
--- a/python/paddle/trainer_config_helpers/tests/configs/last_first_seq.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/last_first_seq.py
--- a/python/paddle/trainer_config_helpers/tests/configs/layer_activations.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/layer_activations.py
--- a/python/paddle/trainer_config_helpers/tests/configs/math_ops.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/math_ops.py
--- a/python/paddle/trainer_config_helpers/tests/configs/projections.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/projections.py
--- a/python/paddle/trainer_config_helpers/tests/configs/shared_fc.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/shared_fc.py
--- a/python/paddle/trainer_config_helpers/tests/configs/shared_gru.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/shared_gru.py
--- a/python/paddle/trainer_config_helpers/tests/configs/shared_lstm.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/shared_lstm.py
--- a/python/paddle/trainer_config_helpers/tests/configs/simple_rnn_layers.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/simple_rnn_layers.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_BatchNorm3D.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_BatchNorm3D.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_bi_grumemory.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_bi_grumemory.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_bilinear_interp.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_bilinear_interp.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_clip_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_clip_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_conv3d_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_conv3d_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_cost_layers.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_cost_layers.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_cost_layers_with_weight.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_cost_layers_with_weight.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_crop.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_crop.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_deconv3d_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_deconv3d_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_detection_output_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_detection_output_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_dot_prod_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_dot_prod_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_expand_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_expand_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_factorization_machine.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_factorization_machine.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_fc.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_fc.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_gated_unit_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_gated_unit_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_grumemory_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_grumemory_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_hsigmoid.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_hsigmoid.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_l2_distance_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_l2_distance_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_lstmemory_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_lstmemory_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_maxout.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_maxout.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_multibox_loss_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_multibox_loss_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_multiplex_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_multiplex_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_ntm_layers.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_ntm_layers.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_pad.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_pad.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_pooling3D_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_pooling3D_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_prelu_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_prelu_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_print_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_print_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_recursive_topology.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_recursive_topology.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_repeat_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_repeat_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_resize_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_resize_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_rnn_group.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_rnn_group.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_roi_pool_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_roi_pool_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_row_conv.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_row_conv.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_row_l2_norm_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_row_l2_norm_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_scale_shift_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_scale_shift_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_scale_sub_region_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_scale_sub_region_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_seq_concat_reshape.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_seq_concat_reshape.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_sequence_pooling.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_sequence_pooling.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_smooth_l1.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_smooth_l1.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_split_datasource.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_split_datasource.py
--- a/python/paddle/trainer_config_helpers/tests/configs/test_spp_layer.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/test_spp_layer.py
--- a/python/paddle/trainer_config_helpers/tests/configs/unused_layers.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/unused_layers.py
--- a/python/paddle/trainer_config_helpers/tests/configs/util_layers.py
+++ b/python/paddle/trainer_config_helpers/tests/configs/util_layers.py
--- a/python/paddle/trainer_config_helpers/tests/test_reset_hook.py
+++ b/python/paddle/trainer_config_helpers/tests/test_reset_hook.py
--- a/python/paddle/utils/dump_v2_config.py
+++ b/python/paddle/utils/dump_v2_config.py
--- a/python/paddle/utils/image_multiproc.py
+++ b/python/paddle/utils/image_multiproc.py
--- a/python/paddle/utils/merge_model.py
+++ b/python/paddle/utils/merge_model.py
--- a/python/paddle/v2/__init__.py
+++ b/python/paddle/v2/__init__.py
--- a/python/paddle/v2/dataset/__init__.py
+++ b/python/paddle/v2/dataset/__init__.py
--- a/python/paddle/v2/dataset/common.py
+++ b/python/paddle/v2/dataset/common.py
--- a/python/paddle/v2/dataset/flowers.py
+++ b/python/paddle/v2/dataset/flowers.py
--- a/python/paddle/v2/dataset/imdb.py
+++ b/python/paddle/v2/dataset/imdb.py
--- a/python/paddle/v2/dataset/mq2007.py
+++ b/python/paddle/v2/dataset/mq2007.py
--- a/python/paddle/v2/dataset/tests/imikolov_test.py
+++ b/python/paddle/v2/dataset/tests/imikolov_test.py
--- a/python/paddle/v2/dataset/tests/wmt16_test.py
+++ b/python/paddle/v2/dataset/tests/wmt16_test.py
--- a/python/paddle/v2/dataset/wmt14.py
+++ b/python/paddle/v2/dataset/wmt14.py
--- a/python/paddle/v2/dataset/wmt16.py
+++ b/python/paddle/v2/dataset/wmt16.py
--- a/python/paddle/v2/event.py
+++ b/python/paddle/v2/event.py
--- a/python/paddle/v2/fluid/__init__.py
+++ b/python/paddle/v2/fluid/__init__.py
--- a/python/paddle/v2/fluid/backward.py
+++ b/python/paddle/v2/fluid/backward.py
--- a/python/paddle/v2/fluid/clip.py
+++ b/python/paddle/v2/fluid/clip.py
--- a/python/paddle/v2/fluid/data_feeder.py
+++ b/python/paddle/v2/fluid/data_feeder.py
--- a/python/paddle/v2/fluid/default_scope_funcs.py
+++ b/python/paddle/v2/fluid/default_scope_funcs.py
--- a/python/paddle/v2/fluid/distribute_transpiler.py
+++ b/python/paddle/v2/fluid/distribute_transpiler.py
--- a/python/paddle/v2/fluid/distribute_transpiler_simple.py
+++ b/python/paddle/v2/fluid/distribute_transpiler_simple.py
--- a/python/paddle/v2/fluid/distributed_spliter.py
+++ b/python/paddle/v2/fluid/distributed_spliter.py
--- a/python/paddle/v2/fluid/evaluator.py
+++ b/python/paddle/v2/fluid/evaluator.py
--- a/python/paddle/v2/fluid/executor.py
+++ b/python/paddle/v2/fluid/executor.py
--- a/python/paddle/v2/fluid/framework.py
+++ b/python/paddle/v2/fluid/framework.py
--- a/python/paddle/v2/fluid/initializer.py
+++ b/python/paddle/v2/fluid/initializer.py
--- a/python/paddle/v2/fluid/io.py
+++ b/python/paddle/v2/fluid/io.py
--- a/python/paddle/v2/fluid/layer_helper.py
+++ b/python/paddle/v2/fluid/layer_helper.py
--- a/v1_api_demo/quick_start/data/get_data.sh
+++ b/v1_api_demo/quick_start/data/get_data.sh
--- a/python/paddle/v2/fluid/layers.py
+++ b/python/paddle/v2/fluid/layers.py
--- a/python/paddle/v2/fluid/layers/device.py
+++ b/python/paddle/v2/fluid/layers/device.py
--- a/python/paddle/v2/fluid/layers/io.py
+++ b/python/paddle/v2/fluid/layers/io.py
--- a/python/paddle/v2/fluid/registry.py
+++ b/python/paddle/v2/fluid/registry.py
--- a/python/paddle/v2/fluid/layers/math_op_patch.py
+++ b/python/paddle/v2/fluid/layers/math_op_patch.py
--- a/python/paddle/v2/fluid/layers/nn.py
+++ b/python/paddle/v2/fluid/layers/nn.py
--- a/python/paddle/v2/fluid/layers/ops.py
+++ b/python/paddle/v2/fluid/layers/ops.py
--- a/python/paddle/v2/fluid/layers/tensor.py
+++ b/python/paddle/v2/fluid/layers/tensor.py
--- a/python/paddle/v2/fluid/memory_optimization_transpiler.py
+++ b/python/paddle/v2/fluid/memory_optimization_transpiler.py
--- a/python/paddle/v2/fluid/net_drawer.py
+++ b/python/paddle/v2/fluid/net_drawer.py
--- a/python/paddle/v2/fluid/nets.py
+++ b/python/paddle/v2/fluid/nets.py
--- a/python/paddle/v2/fluid/op.py
+++ b/python/paddle/v2/fluid/op.py