Merge remote-tracking branch 'origin/develop' into feature/recordio

.travis.yml

@@ -4,13 +4,17 @@ cache:
     - $HOME/.ccache
     - $HOME/.cache/pip
     - $TRAVIS_BUILD_DIR/build/third_party
+    - $TRAVIS_BUILD_DIR/build_android/third_party
 sudo: required
 dist: trusty
+services:
+  - docker
 os:
   - linux
 env:
   - JOB=build_doc
   - JOB=check_style
+  - JOB=build_android
 addons:
   apt:
     packages:
@@ -41,8 +45,10 @@ before_install:
     function timeout() { perl -e 'alarm shift; exec @ARGV' "$@"; }
 script:
   - |
-    timeout 2580 paddle/scripts/travis/${JOB}.sh # 43min timeout
+    # 43min timeout
-    RESULT=$?; if [ $RESULT -eq 0 ] || [ $RESULT -eq 142 ]; then true ;else exit 1; fi;
+    if [[ "$JOB" == "build_android" ]]; then timeout 2580 docker run -it --rm -v "$TRAVIS_BUILD_DIR:/paddle" paddlepaddle/paddle:latest-dev-android;
+    else timeout 2580 paddle/scripts/travis/${JOB}.sh; fi;
+    RESULT=$?; if [ $RESULT -eq 0 ] || [ $RESULT -eq 142 ]; then true; else exit 1; fi;
   - |
     if [[ "$JOB" != "build_doc" ]]; then exit 0; fi;
     if [[ "$TRAVIS_PULL_REQUEST" != "false" ]]; then exit 0; fi;

AUTHORS.md

@@ -1,5 +1,6 @@
 | Github account | name |
 |---|---|
+| abhinavarora | Abhinav Arora |
 | backyes | Yan-Fei Wang |
 | beckett1124 | Bin Qi |
 | JiayiFeng | Jia-Yi Feng |

CMakeLists.txt

@@ -19,7 +19,7 @@ set(PADDLE_BINARY_DIR ${CMAKE_CURRENT_BINARY_DIR})
 
 include(system)
 
-project(paddle CXX C Go)
+project(paddle CXX C)
 message(STATUS "CXX compiler: ${CMAKE_CXX_COMPILER}, version: "
         "${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION}")
 message(STATUS "C compiler: ${CMAKE_C_COMPILER}, version: "
@@ -166,16 +166,16 @@ include_directories("${CMAKE_CURRENT_BINARY_DIR}/proto")
 include_directories("${CMAKE_CURRENT_BINARY_DIR}/go/pserver/client/c")
 
 set(EXTERNAL_LIBS
-    ${GFLAGS_LIBRARIES}
+    gflags
-    ${GLOG_LIBRARIES}
+    glog
     ${CBLAS_LIBRARIES}
-    ${PROTOBUF_LIBRARY}
+    protobuf
-    ${ZLIB_LIBRARIES}
+    zlib
     ${PYTHON_LIBRARIES}
 )
 
 if(WITH_GPU)
-  include(cuda)
+    include(cuda)
 endif(WITH_GPU)
 
 if(WITH_MKLML)
@@ -202,17 +202,18 @@ endif()
 # "add_subdirectory(paddle)" and "add_subdirectory(python)" should be
 # placed after this block, because they depends on it.
 if(WITH_GOLANG)
+    enable_language(Go)
     add_subdirectory(go)
 endif(WITH_GOLANG)
 
 set(PADDLE_PYTHON_BUILD_DIR "${CMAKE_CURRENT_BINARY_DIR}/python/build")
 
-SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
+set(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
-SET(CMAKE_C_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
+set(CMAKE_C_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
 
 add_subdirectory(paddle)
 if(WITH_PYTHON)
-  add_subdirectory(python)
+    add_subdirectory(python)
 endif()
 
 if(WITH_DOC)

Dockerfile.android

@@ -21,16 +21,6 @@ RUN apt-get update && \
     wget curl tar unzip gcc g++ locales clang-format-3.8 swig cmake && \
     apt-get clean -y
 
-# Install Go and glide
-RUN wget -qO- go.tgz https://storage.googleapis.com/golang/go1.8.1.linux-amd64.tar.gz | \
-    tar -xz -C /usr/local && \
-    mkdir /root/gopath && \
-    mkdir /root/gopath/bin && \
-    mkdir /root/gopath/src
-ENV GOROOT=/usr/local/go GOPATH=/root/gopath
-# should not be in the same line with GOROOT definition, otherwise docker build could not find GOROOT.
-ENV PATH=${PATH}:${GOROOT}/bin:${GOPATH}/bin
-
 # git credential to skip password typing
 RUN git config --global credential.helper store
 

cmake/configure.cmake

@@ -59,7 +59,6 @@ endif(NOT WITH_GOLANG)
 
 if(NOT WITH_GPU)
     add_definitions(-DHPPL_STUB_FUNC)
-    add_definitions("-DCUPTI_LIB_PATH=\"\"")
 
     list(APPEND CMAKE_CXX_SOURCE_FILE_EXTENSIONS cu)
 else()
@@ -77,9 +76,7 @@ else()
     if(CUPTI_FOUND)
         include_directories(${CUPTI_INCLUDE_DIR})
         add_definitions(-DPADDLE_WITH_CUPTI)
-        add_definitions("-DCUPTI_LIB_PATH=\"${CUPTI_LIBRARY_PATH}\"")
     else()
-        add_definitions("-DCUPTI_LIB_PATH=\"\"")
         message(STATUS "Cannot find CUPTI, GPU Profiling is incorrect.")
     endif()
     set(CUDA_NVCC_FLAGS ${CUDA_NVCC_FLAGS} "-Xcompiler ${SIMD_FLAG}")

cmake/external/zlib.cmake

@@ -28,7 +28,7 @@ ENDIF(WIN32)
 INCLUDE_DIRECTORIES(${ZLIB_INCLUDE_DIR})
 
 ExternalProject_Add(
-    zlib
+    extern_zlib
     ${EXTERNAL_PROJECT_LOG_ARGS}
     GIT_REPOSITORY  "https://github.com/madler/zlib.git"
     GIT_TAG         "v1.2.8"
@@ -49,9 +49,11 @@ ExternalProject_Add(
                      -DCMAKE_BUILD_TYPE:STRING=${THIRD_PARTY_BUILD_TYPE}
 )
 
+ADD_LIBRARY(zlib STATIC IMPORTED GLOBAL)
+SET_PROPERTY(TARGET zlib PROPERTY IMPORTED_LOCATION ${ZLIB_LIBRARIES})
+ADD_DEPENDENCIES(zlib extern_zlib)
+
 LIST(APPEND external_project_dependencies zlib)
-ADD_LIBRARY(zlib_target STATIC IMPORTED GLOBAL)
-SET_PROPERTY(TARGET zlib_target PROPERTY IMPORTED_LOCATION ${ZLIB_LIBRARIES})
 
 IF(WITH_C_API)
   INSTALL(DIRECTORY ${ZLIB_INCLUDE_DIR} DESTINATION third_party/zlib)

cmake/generic.cmake

@@ -104,7 +104,9 @@ function(merge_static_libs TARGET_NAME)
   foreach(lib ${libs})
     list(APPEND libs_deps ${${lib}_LIB_DEPENDS})
   endforeach()
-  list(REMOVE_DUPLICATES libs_deps)
+  if(libs_deps)
+    list(REMOVE_DUPLICATES libs_deps)
+  endif()
 
   # To produce a library we need at least one source file.
   # It is created by add_custom_command below and will helps
@@ -191,10 +193,13 @@ function(cc_library TARGET_NAME)
         list(REMOVE_ITEM cc_library_DEPS warpctc)
         add_dependencies(${TARGET_NAME} warpctc)
       endif()
-      # Support linking flags: --whole-archive (Linux) / -force_load (MacOS)
-      target_circle_link_libraries(${TARGET_NAME} ${cc_library_DEPS})
       if("${cc_library_DEPS}" MATCHES "ARCHIVE_START")
+        # Support linking flags: --whole-archive (Linux) / -force_load (MacOS).
+        # WARNING: Please don't use ARCHIVE_START&ARCHIVE_END if TARGET_NAME will be linked by other libraries.
+        target_circle_link_libraries(${TARGET_NAME} ${cc_library_DEPS})
         list(REMOVE_ITEM cc_library_DEPS ARCHIVE_START ARCHIVE_END)
+      else()
+        target_link_libraries(${TARGET_NAME} ${cc_library_DEPS})
       endif()
       add_dependencies(${TARGET_NAME} ${cc_library_DEPS})
     endif()

cmake/inference_lib.cmake

@@ -72,7 +72,7 @@ copy(inference_lib DEPENDS paddle_fluid_shared
 )
 
 set(module "platform")
-copy(platform_lib
+copy(platform_lib DEPS profiler_py_proto
   SRCS ${src_dir}/${module}/*.h ${src_dir}/${module}/dynload/*.h ${src_dir}/${module}/details/*.h
   DSTS ${dst_dir}/${module} ${dst_dir}/${module}/dynload ${dst_dir}/${module}/details
 )

doc/build_and_install/build_from_source_cn.rst

@@ -9,7 +9,7 @@
 为了编译PaddlePaddle，我们需要
 
 1. 一台电脑，可以装的是 Linux, Windows 或者 MacOS 操作系统
-1. Docker
+2. Docker
 
 不需要依赖其他任何软件了。即便是 Python 和 GCC 都不需要，因为我们会把所有编译工具都安装进一个 Docker 镜像里。
 
@@ -189,7 +189,7 @@ PaddlePaddle的编译选项，包括生成CPU/GPU二进制文件、链接何种B
     "WITH_TESTING", "是否开启单元测试", "OFF"
     "WITH_DOC", "是否编译中英文文档", "OFF"
     "WITH_SWIG_PY", "是否编译PYTHON的SWIG接口，该接口可用于预测和定制化训练", "Auto"
-    "WITH_GOLANG", "是否编译go语言的可容错parameter server", "ON"
+    "WITH_GOLANG", "是否编译go语言的可容错parameter server", "OFF"
     "WITH_MKL", "是否使用MKL数学库，如果为否则是用OpenBLAS", "ON"
 
 BLAS

doc/build_and_install/build_from_source_en.rst

@@ -9,7 +9,7 @@ Requirements
 To build PaddlePaddle, you need
 
 1. A computer -- Linux, Windows, MacOS.
-1. Docker.
+2. Docker.
 
 Nothing else.  Not even Python and GCC, because you can install all build tools into a Docker image. 
 We run all the tools by running this image.
@@ -191,7 +191,7 @@ You can add :code:`-D` argument to pass such options, like:
     "WITH_TESTING", "Build unit tests", "OFF"
     "WITH_DOC", "Build documentations", "OFF"
     "WITH_SWIG_PY", "Build Python SWIG interface for V2 API", "Auto"
-    "WITH_GOLANG", "Build fault-tolerant parameter server written in go", "ON"
+    "WITH_GOLANG", "Build fault-tolerant parameter server written in go", "OFF"
     "WITH_MKL", "Use MKL as BLAS library, else use OpenBLAS", "ON"
 
 

doc/build_and_install/pip_install_en.rst

@@ -1,4 +1,4 @@
-Install Using pip
+Install using pip
 ================================
 
 You can use current widely used Python package management
@@ -8,7 +8,7 @@ most of current Linux systems or MacOS.
 
 .. _pip_install:
 
-Install Using pip
+Install using pip
 ------------------------------
 
 Run the following command to install PaddlePaddle on the current

doc/design/parallel_do.md

@@ -39,15 +39,16 @@ In the backward pass
 This implementation allows to write mixed device program like this
 
 ```python
-# get embedding feature on CPU
+W1 = fluid.tensor(size=[100,20], parameter=true)
-feature = some_cpu_only_op(data)
+W2 = fluid.tensor(size=[20,15], parameter=true)
 
-gpu_places = get_place(use_gpu=True)
+data = layers.data()
+
+gpu_places = layers.get_place(use_gpu=True)
 # parallel processing on multiple GPUs
 pd = ParallelDo(gpu_places)
-with pd.do():
+with pd.do(input=data):
-    read_input(feature)
+    prediction = softmax(fc(fc(data, W1), W2))
-    prediction = my_net(feature)
     write_output(prediction)
 prediction = pd()
 loss = cross_entropy(prediction, label)
@@ -66,20 +67,20 @@ start_program
 main_program
 {
 block0 {
-  vars: data, places, w1, w2
+  vars: data, places, w1, w2, w1_grad, w2_grad,
   ops: data, get_place, parallel_do(block1),
        parallel_do_grad(block2),
        sgd(w2, w2_grad),
        sgd(w1, w1_grad)
 }
-block1 {
+block1 { # the forward pass
   parent_block: 0
   vars: data, h1, h2, loss
   ops: fc, fc, softmax
 }
-block2 {
+block2 { # the backward pass
   parent_block: 1
-  vars: data_grad, h1_grad, h2_grad, loss_gard, w1_grad, w2_grad
+  vars: data_grad, h1_grad, h2_grad, loss_gard, local_w1_grad, local_w2_grad
   ops: softmax_grad,
        fc_grad
        fc_grad

doc/dev/contribute_to_paddle_cn.md

@@ -222,6 +222,7 @@ upstream
 ## 提交代码的一些约定
 
 为了使评审人在评审代码时更好地专注于代码本身，请您每次提交代码时，遵守以下约定：
+
 1. 请保证Travis-CI 中单元测试能顺利通过。如果没过，说明提交的代码存在问题，评审人一般不做评审。
 2. 提交PUll Request前：
    - 请注意commit的数量：
@@ -231,6 +232,7 @@ upstream
 3. 如果解决了某个Issue的问题，请在该PUll Request的**第一个**评论框中加上：`fix #issue_number`，这样当该PUll Request被合并后，会自动关闭对应的Issue。关键词包括：close, closes, closed, fix, fixes, fixed, resolve, resolves, resolved，请选择合适的词汇。详细可参考[Closing issues via commit messages](https://help.github.com/articles/closing-issues-via-commit-messages)。
 
 此外，在回复评审人意见时，请您遵守以下约定：
+
 1. 评审人的每个意见都必须回复（这是开源社区的基本礼貌，别人帮了忙，应该说谢谢）：
    - 对评审意见同意且按其修改完的，给个简单的`Done`即可；
    - 对评审意见不同意的，请给出您自己的反驳理由。

doc/dev/index_en.rst

@@ -4,6 +4,5 @@ Development
 ..  toctree::
   :maxdepth: 1
 
-  new_layer_en.rst
   contribute_to_paddle_en.md
   write_docs_en.rst

doc/faq/build_and_install/index_cn.rst

@@ -16,7 +16,7 @@
     $ export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}')
     $ docker run ${CUDA_SO} ${DEVICES} -it paddlepaddle/paddle:latest-gpu
 
-更多关于Docker的安装与使用, 请参考 `PaddlePaddle Docker 文档 <http://www.paddlepaddle.org/doc_cn/build_and_install/install/docker_install.html>`_ 。
+更多关于Docker的安装与使用, 请参考 `PaddlePaddle Docker 文档 <http://www.paddlepaddle.org/docs/0.11.0/documentation/zh/getstarted/build_and_install/docker_install_cn.html>`_ 。
 
 
 2. CMake源码编译, 找到的PythonLibs和PythonInterp版本不一致

doc/faq/build_and_install/index_en.rst

@@ -0,0 +1,5 @@
+############################
+Install, Build and Unit test
+############################
+
+TBD

doc/faq/cluster/index_en.rst

@@ -0,0 +1,5 @@
+###############################
+Cluster Training and Prediction
+###############################
+
+TBD

doc/faq/index_cn.rst

@@ -1,6 +1,8 @@
 FAQ
 ====
 
+本文档对关于PaddlePaddle的一些常见问题提供了解答。如果您的问题未在此处，请您到 `PaddlePaddle社区 <https://github.com/PaddlePaddle/Paddle/issues>`_ 查找答案或直接提 `issue <https://github.com/PaddlePaddle/Paddle/issues/new>`_ ，我们会及时进行回复。
+ 
 ..  toctree::
   :maxdepth: 1
 

doc/faq/index_en.rst

@@ -0,0 +1,12 @@
+FAQ
+====
+
+ 
+..  toctree::
+  :maxdepth: 1
+
+  build_and_install/index_en.rst
+  model/index_en.rst
+  parameter/index_en.rst
+  local/index_en.rst
+  cluster/index_en.rst

doc/faq/local/index_cn.rst

@@ -148,10 +148,10 @@ Paddle二进制在运行时捕获了浮点数异常，只要出现浮点数异
 
 ..  code-block:: python
 
-optimizer = paddle.optimizer.RMSProp(
+    optimizer = paddle.optimizer.RMSProp(
-    learning_rate=1e-3,
+        learning_rate=1e-3,
-    gradient_clipping_threshold=10.0,
+        gradient_clipping_threshold=10.0,
-    regularization=paddle.optimizer.L2Regularization(rate=8e-4))
+        regularization=paddle.optimizer.L2Regularization(rate=8e-4))
 
 具体可以参考  `nmt_without_attention  <https://github.com/PaddlePaddle/models/blob/develop/nmt_without_attention/train.py#L35>`_ 示例。
 
@@ -159,13 +159,13 @@ optimizer = paddle.optimizer.RMSProp(
 
 ..  code-block:: python
 
-decoder_inputs = paddle.layer.fc(
+    decoder_inputs = paddle.layer.fc(
-    act=paddle.activation.Linear(),
+        act=paddle.activation.Linear(),
-    size=decoder_size * 3,
+        size=decoder_size * 3,
-    bias_attr=False,
+        bias_attr=False,
-    input=[context, current_word],
+        input=[context, current_word],
-    layer_attr=paddle.attr.ExtraLayerAttribute(
+        layer_attr=paddle.attr.ExtraLayerAttribute(
-        error_clipping_threshold=100.0))
+            error_clipping_threshold=100.0))
 
 完整代码可以参考示例 `machine translation <https://github.com/PaddlePaddle/book/blob/develop/08.machine_translation/train.py#L66>`_ 。
 

doc/faq/local/index_en.rst

@@ -0,0 +1,5 @@
+#############################
+Local Training and Prediction
+#############################
+
+TBD

doc/faq/model/index_en.rst

@@ -0,0 +1,5 @@
+###################
+Model Configuration
+###################
+
+TBD

doc/faq/parameter/index_cn.rst

@@ -196,6 +196,6 @@ PaddlePaddle保存的模型参数文件内容由16字节头信息和网络参数
         obj="process",
         args={"src_dict_path": src_dict_path})
 
-完整源码可参考 `seqToseq <https://github.com/PaddlePaddle/Paddle/tree/develop/demo/seqToseq>`_ 示例。
+完整源码可参考 `sequence_recurrent <https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/gserver/tests/sequence_recurrent.py>`_ 示例。
 
 

doc/faq/parameter/index_en.rst

@@ -0,0 +1,5 @@
+#################
+Parameter Setting
+#################
+
+TBD

doc/getstarted/concepts/use_concepts_en.rst

@@ -0,0 +1,3 @@
+Basic Concept
+=============
+TBD