Skip to content

P
PaddleDetection

s920243400 / PaddleDetection
与 Fork 源项目一致

Fork自 PaddlePaddle / PaddleDetection

通知 2
Star 0
Fork 0
P
PaddleDetection
提交 a698824a 编写于 作者: W whs 提交者： GitHub
浏览文件
操作

Merge branch 'develop' into voc_dataset

上级 338dd135 6398c15c
展开全部 隐藏空白更改
内联 并排
Showing with 23165 addition and 11686 deletion
.gitignore
浏览文件 @ a698824a
......@@ -16,3 +16,6 @@ third_party/
*~
bazel-*
third_party/
# clion workspace.
cmake-build-*
.pre-commit-config.yaml
浏览文件 @ a698824a
......@@ -3,8 +3,8 @@
hooks:
- id: remove-crlf
files: (?!.*third_party)^.*$ | (?!.*book)^.*$
- repo: https://github.com/reyoung/mirrors-yapf.git
sha: v0.13.2
- repo: https://github.com/PaddlePaddle/mirrors-yapf.git
sha: 0d79c0c469bab64f7229c9aca2b1186ef47f0e37
hooks:
- id: yapf
files: (.*\.(py|bzl)|BUILD|.*\.BUILD|WORKSPACE)$
......@@ -21,3 +21,10 @@
sha: 28c0ea8a67a3e2dbbf4822ef44e85b63a0080a29
hooks:
- id: clang-formater
- repo: https://github.com/dnephin/pre-commit-golang
sha: e4693a4c282b4fc878eda172a929f7a6508e7d16
hooks:
- id: go-fmt
files: (.*\.go)
- id: go-lint
files: (.*\.go)
.travis.yml
浏览文件 @ a698824a
language: cpp
cache:
directories:
- $HOME/third_party
- $HOME/.ccache
- $HOME/.cache/pip
- $TRAVIS_BUILD_DIR/build/third_party
sudo: required
dist: trusty
os:
- linux
env:
- JOB=DOCS
- JOB=BUILD_AND_TEST
- JOB=PRE_COMMIT
- JOB=build_doc
- JOB=check_style
addons:
apt:
packages:
- gcc-4.8
- g++-4.8
- gfortran-4.8
- git
- build-essential
- python
......@@ -34,26 +32,18 @@ addons:
- libtool
- ccache
before_install:
- |
if [ ${JOB} == "BUILD_AND_TEST" ]; then
local change_list=`git diff --name-only $TRAVIS_COMMIT_RANGE`
if [ $? -eq 0 ]; then # if git diff return no zero, then rerun unit test.
if ! echo ${change_list} | grep -qvE '(\.md$)|(\.rst$)|(\.jpg$)|(\.png$)'
then
echo "Only markdown docs were updated, stopping build process."
exit
fi
fi
fi
- if [[ "$JOB" == "PRE_COMMIT" ]]; then sudo ln -s /usr/bin/clang-format-3.8 /usr/bin/clang-format; fi
# Paddle is using protobuf 3.1 currently. Protobuf 3.2 breaks the compatibility. So we specify the python
- if [[ "$JOB" == "check_style" ]]; then sudo ln -s /usr/bin/clang-format-3.8 /usr/bin/clang-format; fi
# Paddle is using protobuf 3.1 currently. Protobuf 3.2 breaks the compatibility. So we specify the python
# protobuf version.
- pip install numpy wheel 'protobuf==3.1' sphinx==1.5.6 recommonmark sphinx-rtd-theme==0.1.9 virtualenv pre-commit requests==2.9.2 LinkChecker
- pip install rarfile
- curl https://glide.sh/get | bash
- eval "$(GIMME_GO_VERSION=1.8.3 gimme)"
- |
function timeout() { perl -e 'alarm shift; exec @ARGV' "$@"; }
script:
- |
timeout 2580 paddle/scripts/travis/main.sh # 43min timeout
- |
timeout 2580 paddle/scripts/travis/${JOB}.sh # 43min timeout
RESULT=$?; if [ $RESULT -eq 0 ] || [ $RESULT -eq 142 ]; then true; else false; fi;
notifications:
email:
......
AUTHORS.md
浏览文件 @ a698824a
| Github account | name |
|---|---|
| reyoung | Yang Yu |
| backyes | Yan-Fei Wang |
| beckett1124 | Bin Qi |
| Canpio | Jia-Yi Feng |
| chengxiaohua1105 | Xiao-Hua Cheng |
| cxwangyi, yiwangbaidu, wangkuiyi | Yi Wang |
| cxysteven | Xing-Yi Cheng |
| dzhwinter | Zhi-Hong Dong |
| emailweixu | Wei Xu |
| gangliao | Gang Liao |
| luotao01 | Tao Luo |
| jacquesqiao | Long-Fei Qiao |
| qingqing01 | Qing-Qing Dang |
| gongweibao | Wei-Bao Gong |
| Guo Sheng | Sheng Guo |
| Haichao-Zhang | Hai-Chao Zhang |
| hedaoyuan | Dao-Yuan He |
| wangyang59 | Yang Wang |
| helinwang | He-Lin Wang |
| jacquesqiao | Long-Fei Qiao |
| kuke | Yi-Bing Liu |
| lcy-seso | Ying Cao |
| lipeng-unisound | Peng Li |
| liuyuan | Yuan Liu |
| livc | Zhao Li |
| llxxxll | Yong-Feng Liu |
| luotao01 | Tao Luo |
| lzhao4ever | Liang Zhao |
| NHZlX | Zhao-Long Xing |
| pakchoi | Chuan-Jiang Song |
| pengli09 | Peng Li |
| pkuyym | Ya-Ming Yang |
| QiJune | Jun Qi |
| qingqing01 | Qing-Qing Dang |
| reyoung | Yang Yu |
| Superjom | Chun-Wei Yan |
| tianbingsz | Tian-Bing Xu |
| cxwangyi, yiwangbaidu, wangkuiyi | Yi Wang |
| typhoonzero | Yi Wu |
| backyes | Yan-Fei Wang |
| pengli09 | Peng Li |
| livc | Zhao Li |
| wanghaoshuang | Hao-Shuang Wang |
| wangyang59 | Yang Wang |
| wangzhen-nlp | Zhen Wang |
| wen-bo-yang | Wen-Bo Yang |
| wwhu | Wei-Wei Hu |
| xinghai-sun | Xing-Hai Sun |
| Xreki | Yi-Qun Liu |
| xujun05 | Jun Xu |
| xushaoyong | Shao-Yong Xu |
| Yancey1989 | Xu Yan |
| emailweixu | Wei Xu |
| wen-bo-yang | Wen-Bo Yang |
| helinwang | He-Lin Wang |
| lcy-seso | Ying Cao |
| Zrachel | Rui-Qing Zhang |
| Haichao-Zhang | Hai-Chao Zhang |
| gongweibao | Wei-Bao Gong |
| lzhao4ever | Liang Zhao |
| zhaopu7 | Pu Zhao |
| zhouxiao-coder | Xiao Zhou |
| lipeng-unisound | Peng Li |
| Zrachel | Rui-Qing Zhang |
CMakeLists.txt
浏览文件 @ a698824a
......@@ -16,10 +16,11 @@ cmake_minimum_required(VERSION 3.0)
set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} "${CMAKE_CURRENT_SOURCE_DIR}/cmake")
set(PROJ_ROOT ${CMAKE_CURRENT_SOURCE_DIR})
set(PROJ_BINARY_ROOT ${CMAKE_CURRENT_BINARY_DIR})
include(system)
project(paddle CXX C)
project(paddle CXX C Go)
find_package(Sphinx)
if(NOT CMAKE_CROSSCOMPILING)
......@@ -27,6 +28,7 @@ if(NOT CMAKE_CROSSCOMPILING)
endif(NOT CMAKE_CROSSCOMPILING)
find_package(Git REQUIRED)
find_package(Threads REQUIRED)
find_package(Boost QUIET)
include(simd)
......@@ -47,6 +49,9 @@ option(WITH_COVERAGE "Compile PaddlePaddle with code coverage" OFF)
option(COVERALLS_UPLOAD "Package code coverage data to coveralls" OFF)
option(ON_TRAVIS "Exclude special unit test on Travis CI" OFF)
option(WITH_C_API "Compile PaddlePaddle with C-API(Prediction)" OFF)
option(WITH_GOLANG "Compile PaddlePaddle with GOLANG" OFF)
option(GLIDE_INSTALL "Download and install go dependencies " ON)
option(USE_NNPACK "Compile PaddlePaddle with NNPACK library" OFF)
# CMAKE_BUILD_TYPE
if(NOT CMAKE_BUILD_TYPE)
......@@ -70,7 +75,7 @@ if(ANDROID)
"Disable RDMA when cross-compiling for Android" FORCE)
endif(ANDROID)
set(THIRD_PARTY_PATH "${PROJ_ROOT}/third_party" CACHE STRING
set(THIRD_PARTY_PATH "${CMAKE_BINARY_DIR}/third_party" CACHE STRING
"A path setting third party libraries download & build directories.")
if (WITH_C_API AND WITH_PYTHON)
......@@ -91,7 +96,10 @@ include(external/openblas) # download, build, install openblas
include(external/swig) # download, build, install swig
include(external/warpctc) # download, build, install warpctc
include(external/any) # download libn::any
include(external/eigen) # download eigen3
include(cudnn) # set cudnn libraries, must before configure
include(configure) # add paddle env configuration
include(generic) # simplify cmake module
include(package) # set paddle packages
include(cpplint) # set paddle c++ style
......@@ -99,14 +107,15 @@ include(ccache) # set ccache for compilation
include(util) # set unittest and link libs
include(rdma) # set rdma libraries
include(flags) # set paddle compile flags
include(cudnn) # set cudnn libraries
include(version) # set PADDLE_VERSION
include(coveralls) # set code coverage
include(configure) # add paddle env configuration
include_directories("${PROJ_ROOT}")
include_directories("${PROJ_ROOT}/paddle/cuda/include")
include_directories("${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}
......@@ -124,10 +133,20 @@ if(WITH_GPU)
endif(NOT WITH_DSO)
endif(WITH_GPU)
if(USE_NNPACK)
list(APPEND EXTERNAL_LIBS ${NNPACK_LIB} ${PTHREADPOOL_LIB} "rt")
endif(USE_NNPACK)
add_subdirectory(proto)
# "add_subdirectory(paddle)" and "add_subdirectory(python)" should be
# placed after this block, because they depends on it.
if(WITH_GOLANG)
add_subdirectory(go)
endif(WITH_GOLANG)
add_subdirectory(paddle)
add_subdirectory(python)
if(WITH_DOC)
add_subdirectory(doc)
endif()
Dockerfile
浏览文件 @ a698824a
......@@ -25,14 +25,28 @@ COPY ./paddle/scripts/docker/root/ /root/
RUN apt-get update && \
apt-get install -y \
git python-pip python-dev openssh-server bison \
wget unzip tar xz-utils bzip2 gzip coreutils \
wget unzip tar xz-utils bzip2 gzip coreutils ntp \
curl sed grep graphviz libjpeg-dev zlib1g-dev \
python-numpy python-matplotlib gcc g++ \
automake locales clang-format-3.8 swig doxygen cmake \
liblapack-dev liblapacke-dev libboost-dev \
clang-3.8 llvm-3.8 libclang-3.8-dev && \
clang-3.8 llvm-3.8 libclang-3.8-dev \
net-tools && \
apt-get clean -y
# Install Go and glide
RUN wget -O go.tgz https://storage.googleapis.com/golang/go1.8.1.linux-amd64.tar.gz && \
tar -C /usr/local -xzf go.tgz && \
mkdir /root/gopath && \
mkdir /root/gopath/bin && \
mkdir /root/gopath/src && \
rm go.tgz
ENV GOROOT=/usr/local/go GOPATH=/root/gopath
# should not be in the same line with GOROOT definition, otherwise docker build could not find GOROOT.
ENV PATH=${PATH}:${GOROOT}/bin:${GOPATH}/bin
# install glide
RUN curl -q https://glide.sh/get | sh
# git credential to skip password typing
RUN git config --global credential.helper store
......@@ -47,7 +61,8 @@ RUN pip install --upgrade pip && \
pip install -U docopt PyYAML sphinx && \
pip install -U sphinx-rtd-theme==0.1.9 recommonmark && \
pip install pre-commit 'requests==2.9.2' 'ipython==5.3.0' && \
pip install 'ipykernel==4.6.0' 'jupyter==1.0.0'
pip install 'ipykernel==4.6.0' 'jupyter==1.0.0' && \
pip install rarfile
# To fix https://github.com/PaddlePaddle/Paddle/issues/1954, we use
# the solution in https://urllib3.readthedocs.io/en/latest/user-guide.html#ssl-py2
......
Dockerfile.android 0 → 100644
浏览文件 @ a698824a
FROM ubuntu:16.04
MAINTAINER PaddlePaddle Authors <paddle-dev@baidu.com>
ARG UBUNTU_MIRROR
RUN /bin/bash -c 'if [[ -n ${UBUNTU_MIRROR} ]]; then sed -i 's#http://archive.ubuntu.com/ubuntu#${UBUNTU_MIRROR}#g' /etc/apt/sources.list; fi'
ENV HOME=/root \
ANDROID_NDK_HOME=/opt/android-ndk-linux \
ANDROID_STANDALONE_TOOLCHAIN=/opt/android-toolchain-gcc
RUN apt-get update && \
apt-get install -y \
git python-dev python-pip python-numpy \
wget curl tar unzip gcc g++ locales clang-format-3.8 swig cmake && \
apt-get clean -y
# git credential to skip password typing
RUN git config --global credential.helper store
# Fix locales to en_US.UTF-8
RUN localedef -i en_US -f UTF-8 en_US.UTF-8
RUN pip install --upgrade pip && \
pip install -U 'protobuf==3.1.0' && \
pip install -U wheel sphinx && \
pip install pre-commit
# Android NDK
RUN mkdir /opt/android-ndk-tmp && \
cd /opt/android-ndk-tmp && \
wget -q https://dl.google.com/android/repository/android-ndk-r14b-linux-x86_64.zip && \
unzip -q android-ndk-r14b-linux-x86_64.zip && \
mv android-ndk-r14b ${ANDROID_NDK_HOME} && \
${ANDROID_NDK_HOME}/build/tools/make-standalone-toolchain.sh --arch=arm --platform=android-21 --install-dir=${ANDROID_STANDALONE_TOOLCHAIN} && \
rm -rf /opt/android-ndk-tmp && \
rm -rf ${ANDROID_NDK_HOME}
CMD ["bash", "/paddle/paddle/scripts/docker/build_android.sh"]
README.md
浏览文件 @ a698824a
......@@ -70,7 +70,7 @@ before looking into the
We provide [English](http://www.paddlepaddle.org/develop/doc/) and
[Chinese](http://www.paddlepaddle.org/doc_cn/) documentation.
- [Deep Learning 101](http://book.paddlepaddle.org/index.en.html)
- [Deep Learning 101](http://book.paddlepaddle.org/index.html)
You might want to start from the this online interactive book that can run in Jupyter Notebook.
......
paddle/go/cclient/cmake/CMakeDetermineGoCompiler.cmake → cmake/CMakeDetermineGoCompiler.cmake
浏览文件 @ a698824a
......@@ -15,7 +15,7 @@ if(NOT CMAKE_Go_COMPILER)
set(Go_BIN_PATH
$ENV{GOPATH}
$ENV{GOROOT}
$ENV{GOROOT}/../bin
$ENV{GOROOT}/bin
$ENV{GO_COMPILER}
/usr/bin
/usr/local/bin
......@@ -28,10 +28,12 @@ if(NOT CMAKE_Go_COMPILER)
NAMES go
PATHS ${Go_BIN_PATH}
)
EXEC_PROGRAM(${CMAKE_Go_COMPILER} ARGS version OUTPUT_VARIABLE GOLANG_VERSION)
STRING(REGEX MATCH "go[0-9]+.[0-9]+.[0-9]+[ /A-Za-z0-9]*" VERSION "${GOLANG_VERSION}")
message("-- The Golang compiler identification is ${VERSION}")
message("-- Check for working Golang compiler: ${CMAKE_Go_COMPILER}")
if(CMAKE_Go_COMPILER)
EXEC_PROGRAM(${CMAKE_Go_COMPILER} ARGS version OUTPUT_VARIABLE GOLANG_VERSION)
STRING(REGEX MATCH "go[0-9]+[.0-9]*[ /A-Za-z0-9]*" VERSION "${GOLANG_VERSION}")
message("-- The Golang compiler identification is ${VERSION}")
message("-- Check for working Golang compiler: ${CMAKE_Go_COMPILER}")
endif()
endif()
endif()
......
cmake/configure.cmake
浏览文件 @ a698824a
# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
#
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
......@@ -40,6 +40,10 @@ if(NOT CMAKE_CROSSCOMPILING)
endif()
endif()
if(NOT WITH_GOLANG)
add_definitions(-DPADDLE_WITHOUT_GOLANG)
endif(NOT WITH_GOLANG)
if(NOT WITH_GPU)
add_definitions(-DPADDLE_ONLY_CPU)
add_definitions(-DHPPL_STUB_FUNC)
......@@ -65,3 +69,45 @@ endif(NOT WITH_GPU)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${SIMD_FLAG}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${SIMD_FLAG}")
if(WITH_GOLANG)
# we need to symlink Paddle directory into GOPATH. If we
# don't do it and we have code that depends on Paddle, go
# get ./... will download a new Paddle repo from Github,
# without the changes in our current Paddle repo that we
# want to build.
set(GOPATH "${CMAKE_CURRENT_BINARY_DIR}/go")
file(MAKE_DIRECTORY ${GOPATH})
set(PADDLE_IN_GOPATH "${GOPATH}/src/github.com/PaddlePaddle/Paddle")
file(MAKE_DIRECTORY "${PADDLE_IN_GOPATH}")
set(PADDLE_GO_PATH "${CMAKE_SOURCE_DIR}/go")
add_custom_target(go_path)
add_custom_command(TARGET go_path
# Symlink Paddle directory into GOPATH
COMMAND mkdir -p ${PADDLE_IN_GOPATH}
COMMAND rm -rf ${PADDLE_IN_GOPATH}
COMMAND ln -sf ${CMAKE_SOURCE_DIR} ${PADDLE_IN_GOPATH}
# Automatically get all dependencies specified in the source code
# We can't run `go get -d ./...` for every target, because
# multiple `go get` can not run concurrently, but make need to be
# able to run with multiple jobs.
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR}
)
if (GLIDE_INSTALL)
if(EXISTS $ENV{GOPATH}/bin/glide)
set(GLIDE "$ENV{GOPATH}/bin/glide")
else()
message(FATAL_ERROR "no glide executeble found: $ENV{GOPATH}/bin/glide")
endif()
add_custom_target(go_vendor)
add_custom_command(TARGET go_vendor
COMMAND env GOPATH=${GOPATH} ${GLIDE} install
WORKING_DIRECTORY "${PADDLE_IN_GOPATH}/go"
)
add_dependencies(go_vendor go_path)
endif()
endif(WITH_GOLANG)
cmake/cpplint.cmake
浏览文件 @ a698824a
......@@ -25,6 +25,7 @@ set(STYLE_FILTER "${STYLE_FILTER}-readability/casting")
set(IGNORE_PATTERN
.*ImportanceSampler.*
.*cblas\\.h.*
.*\\.pb\\.txt
.*LtrDataProvider.*
.*MultiDataProvider.*)
......@@ -59,7 +60,7 @@ macro(add_style_check_target TARGET_NAME)
"--filter=${STYLE_FILTER}"
"--write-success=${CUR_GEN}" ${filename}
DEPENDS ${filename}
WORKING_DIRECTORY ${CMAKE_CURRENT_LIST_DIR})
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
endif()
endforeach()
endif()
......
cmake/cudnn.cmake
浏览文件 @ a698824a
......@@ -11,11 +11,16 @@ find_path(CUDNN_INCLUDE_DIR cudnn.h
get_filename_component(__libpath_hist ${CUDA_CUDART_LIBRARY} PATH)
set(TARGET_ARCH "x86_64")
if(NOT ${CMAKE_SYSTEM_PROCESSOR})
set(TARGET_ARCH ${CMAKE_SYSTEM_PROCESSOR})
endif()
list(APPEND CUDNN_CHECK_LIBRARY_DIRS
${CUDNN_ROOT}
${CUDNN_ROOT}/lib64
${CUDNN_ROOT}/lib
${CUDNN_ROOT}/lib/x86_64-linux-gnu
${CUDNN_ROOT}/lib/${TARGET_ARCH}-linux-gnu
$ENV{CUDNN_ROOT}
$ENV{CUDNN_ROOT}/lib64
$ENV{CUDNN_ROOT}/lib
......
cmake/external/any.cmake
浏览文件 @ a698824a
......@@ -2,10 +2,10 @@ INCLUDE(ExternalProject)
SET(ANY_SOURCE_DIR ${THIRD_PARTY_PATH}/any)
INCLUDE_DIRECTORIES(${ANY_SOURCE_DIR}/src/linb_any)
INCLUDE_DIRECTORIES(${ANY_SOURCE_DIR}/src/extern_lib_any)
ExternalProject_Add(
linb_any
extern_lib_any
${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY "https://github.com/thelink2012/any.git"
GIT_TAG "8fef1e93710a0edf8d7658999e284a1142c4c020"
......@@ -17,4 +17,15 @@ ExternalProject_Add(
TEST_COMMAND ""
)
if (${CMAKE_VERSION} VERSION_LESS "3.3.0")
set(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/lib_any_dummy.c)
file(WRITE ${dummyfile} "const char * dummy_any = \"${dummyfile}\";")
add_library(lib_any STATIC ${dummyfile})
else()
add_library(lib_any INTERFACE)
endif()
add_dependencies(lib_any extern_lib_any)
add_definitions(-DANY_IMPL_ANY_CAST_MOVEABLE)
LIST(APPEND external_project_dependencies lib_any)
cmake/external/eigen.cmake 0 → 100644
浏览文件 @ a698824a
INCLUDE(ExternalProject)
SET(EIGEN_SOURCE_DIR ${THIRD_PARTY_PATH}/eigen3)
INCLUDE_DIRECTORIES(${EIGEN_SOURCE_DIR}/src/extern_eigen3)
ExternalProject_Add(
extern_eigen3
${EXTERNAL_PROJECT_LOG_ARGS}
# for latest version, please get from official website
# URL "https://bitbucket.org/eigen/eigen/get/3.3.4.tar.gz"
# URL_MD5 "1a47e78efe365a97de0c022d127607c3"
# for no-ssl http support, please get from bazel's mirror
# URL "http://mirror.bazel.build/bitbucket.org/eigen/eigen/get/f3a22f35b044.tar.gz"
# URL_MD5 "4645c66075982da6fa0bcf6b20f3e8f7"
# get from github mirror
GIT_REPOSITORY "https://github.com/RLovelett/eigen.git"
GIT_TAG "a46d2e7337c4656f00abe54a8115f6d76153a048"
PREFIX ${EIGEN_SOURCE_DIR}
UPDATE_COMMAND ""
CONFIGURE_COMMAND ""
BUILD_COMMAND ""
INSTALL_COMMAND ""
TEST_COMMAND ""
)
if (${CMAKE_VERSION} VERSION_LESS "3.3.0")
set(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/eigen3_dummy.c)
file(WRITE ${dummyfile} "const char * dummy_eigen3 = \"${dummyfile}\";")
add_library(eigen3 STATIC ${dummyfile})
else()
add_library(eigen3 INTERFACE)
endif()
add_dependencies(eigen3 extern_eigen3)
LIST(APPEND external_project_dependencies eigen3)
cmake/external/gflags.cmake
浏览文件 @ a698824a
# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
#
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
......@@ -26,7 +26,7 @@ ENDIF(WIN32)
INCLUDE_DIRECTORIES(${GFLAGS_INCLUDE_DIR})
ExternalProject_Add(
gflags
extern_gflags
${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY "https://github.com/gflags/gflags.git"
PREFIX ${GFLAGS_SOURCES_DIR}
......@@ -44,4 +44,8 @@ ExternalProject_Add(
-DCMAKE_BUILD_TYPE:STRING=Release
)
ADD_LIBRARY(gflags STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET gflags PROPERTY IMPORTED_LOCATION ${GFLAGS_LIBRARIES})
ADD_DEPENDENCIES(gflags extern_gflags)
LIST(APPEND external_project_dependencies gflags)
cmake/external/glog.cmake
浏览文件 @ a698824a
......@@ -27,7 +27,7 @@ ENDIF(WIN32)
INCLUDE_DIRECTORIES(${GLOG_INCLUDE_DIR})
ExternalProject_Add(
glog
extern_glog
${EXTERNAL_PROJECT_LOG_ARGS}
DEPENDS gflags
GIT_REPOSITORY "https://github.com/google/glog.git"
......@@ -38,14 +38,20 @@ ExternalProject_Add(
CMAKE_ARGS -DCMAKE_CXX_FLAGS=${CMAKE_CXX_FLAGS}
CMAKE_ARGS -DCMAKE_C_FLAGS=${CMAKE_C_FLAGS}
CMAKE_ARGS -DCMAKE_INSTALL_PREFIX=${GLOG_INSTALL_DIR}
CMAKE_ARGS -DCMAKE_INSTALL_LIBDIR=${GLOG_INSTALL_DIR}/lib
CMAKE_ARGS -DCMAKE_POSITION_INDEPENDENT_CODE=ON
CMAKE_ARGS -DWITH_GFLAGS=ON
CMAKE_ARGS -Dgflags_DIR=${GFLAGS_INSTALL_DIR}/lib/cmake/gflags
CMAKE_ARGS -DBUILD_TESTING=OFF
CMAKE_ARGS -DCMAKE_BUILD_TYPE=Release
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${GLOG_INSTALL_DIR}
-DCMAKE_INSTALL_LIBDIR:PATH=${GLOG_INSTALL_DIR}/lib
-DCMAKE_POSITION_INDEPENDENT_CODE:BOOL=ON
-DCMAKE_BUILD_TYPE:STRING=Release
)
ADD_LIBRARY(glog STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET glog PROPERTY IMPORTED_LOCATION ${GLOG_LIBRARIES})
ADD_DEPENDENCIES(glog extern_glog)
LIST(APPEND external_project_dependencies glog)
cmake/external/gtest.cmake
浏览文件 @ a698824a
# 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.
......@@ -35,7 +35,7 @@ IF(WITH_TESTING)
ENDIF(WIN32)
ExternalProject_Add(
gtest
extern_gtest
${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY "https://github.com/google/googletest.git"
GIT_TAG "release-1.8.0"
......@@ -55,5 +55,14 @@ IF(WITH_TESTING)
-DCMAKE_POSITION_INDEPENDENT_CODE:BOOL=ON
-DCMAKE_BUILD_TYPE:STRING=Release
)
LIST(APPEND external_project_dependencies gtest)
ADD_LIBRARY(gtest STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET gtest PROPERTY IMPORTED_LOCATION ${GTEST_LIBRARIES})
ADD_DEPENDENCIES(gtest extern_gtest)
ADD_LIBRARY(gtest_main STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET gtest_main PROPERTY IMPORTED_LOCATION ${GTEST_MAIN_LIBRARIES})
ADD_DEPENDENCIES(gtest_main extern_gtest)
LIST(APPEND external_project_dependencies gtest gtest_main)
ENDIF(WITH_TESTING)
cmake/external/openblas.cmake
浏览文件 @ a698824a
......@@ -21,27 +21,33 @@ IF(NOT ${CBLAS_FOUND})
SET(CBLAS_INSTALL_DIR ${THIRD_PARTY_PATH}/install/openblas)
SET(CBLAS_INC_DIR "${CBLAS_INSTALL_DIR}/include" CACHE PATH "openblas include directory." FORCE)
SET(CBLAS_LIBRARIES "${CBLAS_INSTALL_DIR}/lib/${LIBRARY_PREFIX}openblas${STATIC_LIBRARY_SUFFIX}"
SET(CBLAS_LIBRARIES
"${CBLAS_INSTALL_DIR}/lib/${CMAKE_STATIC_LIBRARY_PREFIX}openblas${CMAKE_STATIC_LIBRARY_SUFFIX}"
CACHE FILEPATH "openblas library." FORCE)
SET(COMMON_ARGS CC=${CMAKE_C_COMPILER} NO_SHARED=1 NO_LAPACK=1)
SET(COMMON_ARGS CC=${CMAKE_C_COMPILER} NO_SHARED=1 NO_LAPACK=1 libs)
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")
SET(OPTIONAL_ARGS HOSTCC=${HOST_C_COMPILER} TARGET=ARMV7 ARM_SOFTFP_ABI=1 USE_THREAD=0 libs)
ELSEIF(RPI)
# use hardfp
SET(OPENBLAS_COMMIT "v0.2.19")
SET(OPTIONAL_ARGS HOSTCC=${HOST_C_COMPILER} TARGET=ARMV7 USE_THREAD=0 libs)
IF(CMAKE_CROSSCOMPILING)
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")
SET(OPTIONAL_ARGS HOSTCC=${HOST_C_COMPILER} TARGET=ARMV7 ARM_SOFTFP_ABI=1 USE_THREAD=0)
ELSEIF(RPI)
# use hardfp
SET(OPENBLAS_COMMIT "v0.2.19")
SET(OPTIONAL_ARGS HOSTCC=${HOST_C_COMPILER} TARGET=ARMV7 USE_THREAD=0)
ENDIF()
ELSE()
SET(OPENBLAS_COMMIT "v0.2.19")
SET(OPENBLAS_ARGS DYNAMIC_ARCH=1 libs)
SET(OPTIONAL_ARGS "")
IF(CMAKE_SYSTEM_PROCESSOR MATCHES "^x86(_64)?$")
SET(OPTIONAL_ARGS DYNAMIC_ARCH=1 NUM_THREADS=64)
ENDIF()
ENDIF()
ExternalProject_Add(
openblas
extern_openblas
${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY https://github.com/xianyi/OpenBLAS.git
GIT_TAG ${OPENBLAS_COMMIT}
......@@ -53,8 +59,14 @@ IF(NOT ${CBLAS_FOUND})
UPDATE_COMMAND ""
CONFIGURE_COMMAND ""
)
LIST(APPEND external_project_dependencies openblas)
ENDIF(NOT ${CBLAS_FOUND})
MESSAGE(STATUS "BLAS library: ${CBLAS_LIBRARIES}")
INCLUDE_DIRECTORIES(${CBLAS_INC_DIR})
ADD_LIBRARY(cblas STATIC IMPORTED)
SET_PROPERTY(TARGET cblas PROPERTY IMPORTED_LOCATION ${CBLAS_LIBRARIES})
IF(NOT ${CBLAS_FOUND})
ADD_DEPENDENCIES(cblas extern_openblas)
LIST(APPEND external_project_dependencies cblas)
ENDIF(NOT ${CBLAS_FOUND})
cmake/external/protobuf.cmake
浏览文件 @ a698824a
......@@ -13,24 +13,153 @@
# limitations under the License.
INCLUDE(ExternalProject)
# Always invoke `FIND_PACKAGE(Protobuf)` for importing function protobuf_generate_cpp
FIND_PACKAGE(Protobuf QUIET)
SET(PROTOBUF_FOUND "OFF")
if(NOT COMMAND protobuf_generate_python) # before cmake 3.4, protobuf_genrerate_python is not defined.
function(protobuf_generate_python SRCS)
# shameless copy from https://github.com/Kitware/CMake/blob/master/Modules/FindProtobuf.cmake
if(NOT ARGN)
message(SEND_ERROR "Error: PROTOBUF_GENERATE_PYTHON() called without any proto files")
return()
endif()
if(PROTOBUF_GENERATE_CPP_APPEND_PATH)
# Create an include path for each file specified
foreach(FIL ${ARGN})
get_filename_component(ABS_FIL ${FIL} ABSOLUTE)
get_filename_component(ABS_PATH ${ABS_FIL} PATH)
list(FIND _protobuf_include_path ${ABS_PATH} _contains_already)
if(${_contains_already} EQUAL -1)
list(APPEND _protobuf_include_path -I ${ABS_PATH})
endif()
endforeach()
else()
set(_protobuf_include_path -I ${CMAKE_CURRENT_SOURCE_DIR})
endif()
if(DEFINED PROTOBUF_IMPORT_DIRS AND NOT DEFINED Protobuf_IMPORT_DIRS)
set(Protobuf_IMPORT_DIRS "${PROTOBUF_IMPORT_DIRS}")
endif()
if(DEFINED Protobuf_IMPORT_DIRS)
foreach(DIR ${Protobuf_IMPORT_DIRS})
get_filename_component(ABS_PATH ${DIR} ABSOLUTE)
list(FIND _protobuf_include_path ${ABS_PATH} _contains_already)
if(${_contains_already} EQUAL -1)
list(APPEND _protobuf_include_path -I ${ABS_PATH})
endif()
endforeach()
endif()
set(${SRCS})
foreach(FIL ${ARGN})
get_filename_component(ABS_FIL ${FIL} ABSOLUTE)
get_filename_component(FIL_WE ${FIL} NAME_WE)
if(NOT PROTOBUF_GENERATE_CPP_APPEND_PATH)
get_filename_component(FIL_DIR ${FIL} DIRECTORY)
if(FIL_DIR)
set(FIL_WE "${FIL_DIR}/${FIL_WE}")
endif()
endif()
list(APPEND ${SRCS} "${CMAKE_CURRENT_BINARY_DIR}/${FIL_WE}_pb2.py")
add_custom_command(
OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/${FIL_WE}_pb2.py"
COMMAND ${Protobuf_PROTOC_EXECUTABLE} --python_out ${CMAKE_CURRENT_BINARY_DIR} ${_protobuf_include_path} ${ABS_FIL}
DEPENDS ${ABS_FIL} ${Protobuf_PROTOC_EXECUTABLE}
COMMENT "Running Python protocol buffer compiler on ${FIL}"
VERBATIM )
endforeach()
set(${SRCS} ${${SRCS}} PARENT_SCOPE)
endfunction()
endif()
# Print and set the protobuf library information,
# finish this cmake process and exit from this file.
macro(PROMPT_PROTOBUF_LIB)
SET(protobuf_DEPS ${ARGN})
MESSAGE(STATUS "Protobuf protoc executable: ${PROTOBUF_PROTOC_EXECUTABLE}")
MESSAGE(STATUS "Protobuf library: ${PROTOBUF_LIBRARY}")
MESSAGE(STATUS "Protobuf version: ${PROTOBUF_VERSION}")
INCLUDE_DIRECTORIES(${PROTOBUF_INCLUDE_DIR})
# Assuming that all the protobuf libraries are of the same type.
IF(${PROTOBUF_LIBRARY} MATCHES "${CMAKE_STATIC_LIBRARY_SUFFIX}$")
SET(protobuf_LIBTYPE STATIC)
ELSEIF(${PROTOBUF_LIBRARY} MATCHES "${CMAKE_SHARED_LIBRARY_SUFFIX}$")
SET(protobuf_LIBTYPE SHARED)
ELSE()
MESSAGE(FATAL_ERROR "Unknown library type: ${PROTOBUF_LIBRARY}")
ENDIF()
ADD_LIBRARY(protobuf ${protobuf_LIBTYPE} IMPORTED GLOBAL)
SET_PROPERTY(TARGET protobuf PROPERTY IMPORTED_LOCATION ${PROTOBUF_LIBRARY})
ADD_LIBRARY(protobuf_lite ${protobuf_LIBTYPE} IMPORTED GLOBAL)
SET_PROPERTY(TARGET protobuf_lite PROPERTY IMPORTED_LOCATION ${PROTOBUF_LITE_LIBRARY})
ADD_LIBRARY(libprotoc ${protobuf_LIBTYPE} IMPORTED GLOBAL)
SET_PROPERTY(TARGET libprotoc PROPERTY IMPORTED_LOCATION ${PROTOC_LIBRARY})
ADD_EXECUTABLE(protoc IMPORTED GLOBAL)
SET_PROPERTY(TARGET protoc PROPERTY IMPORTED_LOCATION ${PROTOBUF_PROTOC_EXECUTABLE})
# FIND_Protobuf.cmake uses `Protobuf_PROTOC_EXECUTABLE`.
# make `protobuf_generate_cpp` happy.
SET(Protobuf_PROTOC_EXECUTABLE ${PROTOBUF_PROTOC_EXECUTABLE})
FOREACH(dep ${protobuf_DEPS})
ADD_DEPENDENCIES(protobuf ${dep})
ADD_DEPENDENCIES(protobuf_lite ${dep})
ADD_DEPENDENCIES(libprotoc ${dep})
ADD_DEPENDENCIES(protoc ${dep})
ENDFOREACH()
LIST(APPEND external_project_dependencies protobuf)
RETURN()
endmacro()
macro(SET_PROTOBUF_VERSION)
EXEC_PROGRAM(${PROTOBUF_PROTOC_EXECUTABLE} ARGS --version OUTPUT_VARIABLE PROTOBUF_VERSION)
STRING(REGEX MATCH "[0-9]+.[0-9]+" PROTOBUF_VERSION "${PROTOBUF_VERSION}")
endmacro()
set(PROTOBUF_ROOT "" CACHE PATH "Folder contains protobuf")
if (NOT "${PROTOBUF_ROOT}" STREQUAL "")
find_path(PROTOBUF_INCLUDE_DIR google/protobuf/message.h PATHS ${PROTOBUF_ROOT}/include)
find_library(PROTOBUF_LIBRARY protobuf PATHS ${PROTOBUF_ROOT}/lib)
find_library(PROTOBUF_LITE_LIBRARY protobuf-lite PATHS ${PROTOBUF_ROOT}/lib)
find_library(PROTOBUF_PROTOC_LIBRARY protoc PATHS ${PROTOBUF_ROOT}/lib)
find_program(PROTOBUF_PROTOC_EXECUTABLE protoc PATHS ${PROTOBUF_ROOT}/bin)
if (PROTOBUF_INCLUDE_DIR AND PROTOBUF_LIBRARY AND PROTOBUF_LITE_LIBRARY AND PROTOBUF_PROTOC_LIBRARY AND PROTOBUF_PROTOC_EXECUTABLE)
message(STATUS "Using custom protobuf library in ${PROTOBUF_ROOT}.")
SET_PROTOBUF_VERSION()
PROMPT_PROTOBUF_LIB()
else()
message(WARNING "Cannot find protobuf library in ${PROTOBUF_ROOT}.")
endif()
endif()
FUNCTION(build_protobuf TARGET_NAME BUILD_FOR_HOST)
SET(PROTOBUF_SOURCES_DIR ${THIRD_PARTY_PATH}/${TARGET_NAME})
SET(PROTOBUF_INSTALL_DIR ${THIRD_PARTY_PATH}/install/${TARGET_NAME})
STRING(REPLACE "extern_" "" TARGET_DIR_NAME "${TARGET_NAME}")
SET(PROTOBUF_SOURCES_DIR ${THIRD_PARTY_PATH}/${TARGET_DIR_NAME})
SET(PROTOBUF_INSTALL_DIR ${THIRD_PARTY_PATH}/install/${TARGET_DIR_NAME})
SET(${TARGET_NAME}_INCLUDE_DIR "${PROTOBUF_INSTALL_DIR}/include" PARENT_SCOPE)
SET(PROTOBUF_INCLUDE_DIR "${PROTOBUF_INSTALL_DIR}/include" PARENT_SCOPE)
SET(${TARGET_NAME}_LITE_LIBRARY
"${PROTOBUF_INSTALL_DIR}/lib/libprotobuf-lite${STATIC_LIBRARY_SUFFIX}"
"${PROTOBUF_INSTALL_DIR}/lib/libprotobuf-lite${CMAKE_STATIC_LIBRARY_SUFFIX}"
PARENT_SCOPE)
SET(${TARGET_NAME}_LIBRARY
"${PROTOBUF_INSTALL_DIR}/lib/libprotobuf${STATIC_LIBRARY_SUFFIX}"
"${PROTOBUF_INSTALL_DIR}/lib/libprotobuf${CMAKE_STATIC_LIBRARY_SUFFIX}"
PARENT_SCOPE)
SET(${TARGET_NAME}_PROTOC_LIBRARY
"${PROTOBUF_INSTALL_DIR}/lib/libprotoc${STATIC_LIBRARY_SUFFIX}"
"${PROTOBUF_INSTALL_DIR}/lib/libprotoc${CMAKE_STATIC_LIBRARY_SUFFIX}"
PARENT_SCOPE)
SET(${TARGET_NAME}_PROTOC_EXECUTABLE
"${PROTOBUF_INSTALL_DIR}/bin/protoc${EXECUTABLE_SUFFIX}"
"${PROTOBUF_INSTALL_DIR}/bin/protoc${CMAKE_EXECUTABLE_SUFFIX}"
PARENT_SCOPE)
SET(OPTIONAL_CACHE_ARGS "")
......@@ -74,17 +203,7 @@ FUNCTION(build_protobuf TARGET_NAME BUILD_FOR_HOST)
ENDFUNCTION()
SET(PROTOBUF_VERSION 3.1)
IF(NOT CMAKE_CROSSCOMPILING)
FIND_PACKAGE(Protobuf ${PROTOBUF_VERSION})
IF(PROTOBUF_FOUND)
EXEC_PROGRAM(${PROTOBUF_PROTOC_EXECUTABLE} ARGS --version OUTPUT_VARIABLE PROTOBUF_VERSION)
STRING(REGEX MATCH "[0-9]+.[0-9]+" PROTOBUF_VERSION "${PROTOBUF_VERSION}")
IF("${PROTOBUF_VERSION}" VERSION_LESS "3.1.0")
SET(PROTOBUF_FOUND OFF)
ENDIF()
ENDIF(PROTOBUF_FOUND)
ELSE()
IF(CMAKE_CROSSCOMPILING)
build_protobuf(protobuf_host TRUE)
LIST(APPEND external_project_dependencies protobuf_host)
......@@ -93,20 +212,22 @@ ELSE()
ENDIF()
IF(NOT PROTOBUF_FOUND)
build_protobuf(protobuf FALSE)
LIST(APPEND external_project_dependencies protobuf)
build_protobuf(extern_protobuf FALSE)
SET(PROTOBUF_INCLUDE_DIR ${protobuf_INCLUDE_DIR}
SET(PROTOBUF_INCLUDE_DIR ${extern_protobuf_INCLUDE_DIR}
CACHE PATH "protobuf include directory." FORCE)
IF(NOT CMAKE_CROSSCOMPILING)
SET(PROTOBUF_PROTOC_EXECUTABLE ${protobuf_PROTOC_EXECUTABLE}
SET(PROTOBUF_LITE_LIBRARY ${extern_protobuf_LITE_LIBRARY}
CACHE FILEPATH "protobuf lite library." FORCE)
SET(PROTOBUF_LIBRARY ${extern_protobuf_LIBRARY}
CACHE FILEPATH "protobuf library." FORCE)
SET(PROTOBUF_PROTOC_LIBRARY ${extern_protobuf_PROTOC_LIBRARY}
CACHE FILEPATH "protoc library." FORCE)
IF(CMAKE_CROSSCOMPILING)
PROMPT_PROTOBUF_LIB(protobuf_host extern_protobuf)
ELSE()
SET(PROTOBUF_PROTOC_EXECUTABLE ${extern_protobuf_PROTOC_EXECUTABLE}
CACHE FILEPATH "protobuf executable." FORCE)
PROMPT_PROTOBUF_LIB(extern_protobuf)
ENDIF()
SET(PROTOBUF_LITE_LIBRARY ${protobuf_LITE_LIBRARY} CACHE FILEPATH "protobuf lite library." FORCE)
SET(PROTOBUF_LIBRARY ${protobuf_LIBRARY} CACHE FILEPATH "protobuf library." FORCE)
SET(PROTOBUF_PROTOC_LIBRARY ${protobuf_PROTOC_LIBRARY} CACHE FILEPATH "protoc library." FORCE)
ENDIF(NOT PROTOBUF_FOUND)
MESSAGE(STATUS "Protobuf protoc executable: ${PROTOBUF_PROTOC_EXECUTABLE}")
MESSAGE(STATUS "Protobuf library: ${PROTOBUF_LIBRARY}")
INCLUDE_DIRECTORIES(${PROTOBUF_INCLUDE_DIR})
cmake/external/python.cmake
浏览文件 @ a698824a
......@@ -32,193 +32,6 @@ IF(PYTHONINTERP_FOUND)
MESSAGE(FATAL_ERROR "Found Python Protobuf ${PY_GOOGLE.PROTOBUF_VERSION} < 3.0.0, "
"please use pip to upgrade protobuf. pip install -U protobuf")
ENDIF()
ELSE(PYTHONINTERP_FOUND)
MESSAGE(FATAL_ERROR "Please install python 2.7 before building PaddlePaddle.")
##################################### PYTHON ########################################
SET(PYTHON_SOURCES_DIR ${THIRD_PARTY_PATH}/python)
SET(PYTHON_INSTALL_DIR ${THIRD_PARTY_PATH}/install/python)
SET(_python_DIR ${PYTHON_INSTALL_DIR})
IF(UNIX)
SET(PYTHON_FOUND ON)
SET(PYTHON_INCLUDE_DIR "${PYTHON_INSTALL_DIR}/include/python2.7" CACHE PATH "Python include dir" FORCE)
SET(PYTHON_LIBRARIES "${PYTHON_INSTALL_DIR}/lib/libpython2.7.a" CACHE FILEPATH "Python library" FORCE)
SET(PYTHON_EXECUTABLE ${PYTHON_INSTALL_DIR}/bin/python CACHE FILEPATH "Python executable" FORCE)
SET(PY_SITE_PACKAGES_PATH "${PYTHON_INSTALL_DIR}/lib/python2.7/site-packages" CACHE PATH "Python site-packages path" FORCE)
ELSEIF(WIN32)
SET(PYTHON_FOUND ON)
SET(PYTHON_INCLUDE_DIR "${PYTHON_INSTALL_DIR}/include" CACHE PATH "Python include dir" FORCE)
SET(PYTHON_LIBRARIES "${PYTHON_INSTALL_DIR}/libs/python27.lib" CACHE FILEPATH "Python library" FORCE)
SET(PYTHON_EXECUTABLE "${PYTHON_INSTALL_DIR}/bin/python.exe" CACHE FILEPATH "Python executable" FORCE)
SET(PY_SITE_PACKAGES_PATH "${PYTHON_INSTALL_DIR}/Lib/site-packages" CACHE PATH "Python site-packages path" FORCE)
ELSE()
MESSAGE(FATAL_ERROR "Unknown system !")
ENDIF()
IF(APPLE)
LIST(APPEND EXTERNAL_PROJECT_OPTIONAL_CMAKE_ARGS
-DCMAKE_BUILD_WITH_INSTALL_RPATH:BOOL=ON
)
ENDIF()
SET(EXTERNAL_PROJECT_OPTIONAL_CMAKE_CACHE_ARGS)
# Force Python build to "Release".
IF(CMAKE_CONFIGURATION_TYPES)
SET(SAVED_CMAKE_CFG_INTDIR ${CMAKE_CFG_INTDIR})
SET(CMAKE_CFG_INTDIR "Release")
ELSE()
LIST(APPEND EXTERNAL_PROJECT_OPTIONAL_CMAKE_CACHE_ARGS
-DCMAKE_BUILD_TYPE:STRING=Release
)
ENDIF()
ExternalProject_Add(python
${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY "https://github.com/python-cmake-buildsystem/python-cmake-buildsystem.git"
PREFIX ${PYTHON_SOURCES_DIR}
UPDATE_COMMAND ""
CMAKE_ARGS -DPYTHON_VERSION=2.7.12
CMAKE_ARGS -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
CMAKE_ARGS -DCMAKE_C_COMPILER=${CMAKE_C_COMPILER}
CMAKE_CACHE_ARGS
-DCMAKE_INSTALL_PREFIX:PATH=${PYTHON_INSTALL_DIR}
-DBUILD_LIBPYTHON_SHARED:BOOL=OFF
-DUSE_SYSTEM_LIBRARIES:BOOL=OFF
-DZLIB_ROOT:FILEPATH=${ZLIB_ROOT}
-DZLIB_INCLUDE_DIR:PATH=${ZLIB_INCLUDE_DIR}
-DZLIB_LIBRARY:FILEPATH=${ZLIB_LIBRARIES}
-DDOWNLOAD_SOURCES:BOOL=ON
-DINSTALL_WINDOWS_TRADITIONAL:BOOL=OFF
${EXTERNAL_PROJECT_OPTIONAL_CMAKE_CACHE_ARGS}
${EXTERNAL_PROJECT_OPTIONAL_CMAKE_ARGS}
DEPENDS zlib
)
SET(py_env
PATH=${PYTHON_INSTALL_DIR}/bin
PYTHONHOME=${PYTHON_INSTALL_DIR}
PYTHONPATH=${PYTHON_INSTALL_DIR}/lib:${PYTHON_INSTALL_DIR}/lib/python2.7:${PY_SITE_PACKAGES_PATH})
####################################################################################
##################################### SETUPTOOLS ###################################
SET(SETUPTOOLS_SOURCES_DIR ${PYTHON_SOURCES_DIR}/setuptools)
ExternalProject_Add(setuptools
${EXTERNAL_PROJECT_LOG_ARGS}
PREFIX ${SETUPTOOLS_SOURCES_DIR}
URL "https://pypi.python.org/packages/source/s/setuptools/setuptools-18.3.2.tar.gz"
BUILD_IN_SOURCE 1
PATCH_COMMAND ""
UPDATE_COMMAND ""
CONFIGURE_COMMAND ""
INSTALL_COMMAND ""
BUILD_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py install
DEPENDS python zlib
)
#####################################################################################
##################################### SIX ###########################################
SET(SIX_SOURCES_DIR ${PYTHON_SOURCES_DIR}/six)
ExternalProject_Add(six
${EXTERNAL_PROJECT_LOG_ARGS}
PREFIX ${SIX_SOURCES_DIR}
URL https://pypi.python.org/packages/source/s/six/six-1.10.0.tar.gz
BUILD_IN_SOURCE 1
PATCH_COMMAND ""
UPDATE_COMMAND ""
CONFIGURE_COMMAND ""
INSTALL_COMMAND ""
BUILD_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py install
DEPENDS python setuptools
)
#####################################################################################
##################################### CYTHON ########################################
SET(CYTHON_SOURCES_DIR ${PYTHON_SOURCES_DIR}/cython)
ExternalProject_Add(cython
${EXTERNAL_PROJECT_LOG_ARGS}
PREFIX ${CYTHON_SOURCES_DIR}
URL https://github.com/cython/cython/archive/0.25.2.tar.gz
GIT_TAG 0.25.2
BUILD_IN_SOURCE 1
CONFIGURE_COMMAND ""
PATCH_COMMAND ""
UPDATE_COMMAND ""
INSTALL_COMMAND ""
BUILD_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py install
DEPENDS python
)
####################################################################################
##################################### NUMPY ########################################
SET(NUMPY_SOURCES_DIR ${PYTHON_SOURCES_DIR}/numpy)
SET(NUMPY_TAG_VERSION "v1.11.3")
SET(NUMPY_VERSION "1.11.3")
SET(EGG_NAME "")
SET(PYTHON_NUMPY_INCLUDE_DIR "")
IF(WIN32)
SET(EGG_NAME "numpy-${NUMPY_VERSION}-py2.7-${HOST_SYSTEM}.egg")
ELSE(WIN32)
IF(APPLE)
SET(EGG_NAME "numpy-${NUMPY_VERSION}-py2.7-${HOST_SYSTEM}-${MACOS_VERSION}")
ELSE(APPLE)
SET(EGG_NAME "numpy-${NUMPY_VERSION}-py2.7-linux")
SET(EGG_NAME "numpy-${NUMPY_VERSION}-py2.7-linux")
ENDIF(APPLE)
FOREACH(suffix x86_64 intel fat64 fat32 universal)
LIST(APPEND PYTHON_NUMPY_INCLUDE_DIR ${PY_SITE_PACKAGES_PATH}/${EGG_NAME}-${suffix}.egg/numpy/core/include)
ENDFOREACH()
ENDIF(WIN32)
ExternalProject_Add(numpy
${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY https://github.com/numpy/numpy.git
GIT_TAG ${NUMPY_TAG_VERSION}
CONFIGURE_COMMAND ""
UPDATE_COMMAND ""
PREFIX ${NUMPY_SOURCES_DIR}
BUILD_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py build
INSTALL_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py install
BUILD_IN_SOURCE 1
DEPENDS python setuptools cython
)
####################################################################################
##################################### WHEEL ########################################
SET(WHEEL_SOURCES_DIR ${PYTHON_SOURCES_DIR}/wheel)
ExternalProject_Add(wheel
${EXTERNAL_PROJECT_LOG_ARGS}
URL https://pypi.python.org/packages/source/w/wheel/wheel-0.29.0.tar.gz
PREFIX ${WHEEL_SOURCES_DIR}
CONFIGURE_COMMAND ""
UPDATE_COMMAND ""
BUILD_COMMAND ""
INSTALL_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py install
BUILD_IN_SOURCE 1
DEPENDS python setuptools
)
####################################################################################
################################### PROTOBUF #######################################
SET(PY_PROTOBUF_SOURCES_DIR ${PYTHON_SOURCES_DIR}/protobuf)
ExternalProject_Add(python-protobuf
${EXTERNAL_PROJECT_LOG_ARGS}
URL https://pypi.python.org/packages/e0/b0/0a1b364fe8a7d177b4b7d4dca5b798500dc57a7273b93cca73931b305a6a/protobuf-3.1.0.post1.tar.gz
URL_MD5 38b5fb160c768d2f8444d0c6d637ff91
PREFIX ${PY_PROTOBUF_SOURCES_DIR}
BUILD_IN_SOURCE 1
PATCH_COMMAND ""
CONFIGURE_COMMAND ""
BUILD_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py build
INSTALL_COMMAND env ${py_env} ${PYTHON_EXECUTABLE} setup.py install
DEPENDS python setuptools six
)
####################################################################################
LIST(APPEND external_project_dependencies python setuptools six cython wheel python-protobuf numpy)
ENDIF(PYTHONINTERP_FOUND)
IF(WITH_PYTHON)
......
cmake/external/warpctc.cmake
浏览文件 @ a698824a
# 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.
......@@ -43,7 +43,7 @@ ELSE()
ENDIF()
ExternalProject_Add(
warpctc
extern_warpctc
${EXTERNAL_PROJECT_LOG_ARGS}
GIT_REPOSITORY "https://github.com/gangliao/warp-ctc.git"
PREFIX ${WARPCTC_SOURCES_DIR}
......@@ -65,4 +65,8 @@ ExternalProject_Add(
-DCMAKE_INSTALL_PREFIX:PATH=${WARPCTC_INSTALL_DIR}
)
ADD_LIBRARY(warpctc STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET warpctc PROPERTY IMPORTED_LOCATION ${WARPCTC_LIBRARIES})
ADD_DEPENDENCIES(warpctc extern_warpctc)
LIST(APPEND external_project_dependencies warpctc)
cmake/generic.cmake
浏览文件 @ a698824a
# 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.
#
# To simplify the build process of PaddlePaddle, we defined couple of
# fundamental abstractions, e.g., how to build library, binary and
# test in C++, CUDA and Go.
#
# generic.cmake defines CMakes functions that look like Bazel's
# building rules (https://bazel.build/).
#
#
# -------------------------------------------
# C++ CUDA C++ Go
# C++ CUDA C++ Go
# -------------------------------------------
# cc_library nv_library go_library
# cc_binary nv_binary go_binary
# cc_test nv_test go_test
# cc_library nv_library go_library
# cc_binary nv_binary go_binary
# cc_test nv_test go_test
# -------------------------------------------
#
# cmake_parse_arguments can help us to achieve this goal.
# https://cmake.org/cmake/help/v3.0/module/CMakeParseArguments.html
# To build a static library example.a from example.cc using the system
# compiler (like GCC):
#
# cc_library(example SRCS example.cc)
#
# To build a static library example.a from multiple source files
# example{1,2,3}.cc:
#
# cc_library(example SRCS example1.cc example2.cc example3.cc)
#
# To build a shared library example.so from example.cc:
#
# cc_library(example SHARED SRCS example.cc)
#
# To build a library using Nvidia's NVCC from .cu file(s), use the nv_
# prefixed version:
#
# nv_library(example SRCS example.cu)
#
# To specify that a library new_example.a depends on other libraies:
#
# cc_library(new_example SRCS new_example.cc DEPS example)
#
# Static libraries can be composed of other static libraries:
#
# cc_library(composed DEPS dependent1 dependent2 dependent3)
#
# To build an executable binary file from some source files and
# dependent libraries:
#
# cc_binary(example SRCS main.cc something.cc DEPS example1 example2)
#
# To build an executable binary file using NVCC, use the nv_ prefixed
# version:
#
# nv_binary(example SRCS main.cc something.cu DEPS example1 example2)
#
# To build a unit test binary, which is an executable binary with
# GoogleTest linked:
#
# cc_test(example_test SRCS example_test.cc DEPS example)
#
# To build a unit test binary using NVCC, use the nv_ prefixed version:
#
# nv_test(example_test SRCS example_test.cu DEPS example)
#
# It is pretty often that executable and test binaries depend on
# pre-defined external libaries like glog and gflags defined in
# /cmake/external/*.cmake:
#
# cc_test(example_test SRCS example_test.cc DEPS example glog gflags)
#
# To build a go static library using Golang, use the go_ prefixed version:
#
# go_library(example STATIC)
#
# To build a go shared library using Golang, use the go_ prefixed version:
#
# go_library(example SHARED)
#
# including binary directory for generated headers.
include_directories(${CMAKE_CURRENT_BINARY_DIR})
if(NOT APPLE)
find_package(Threads REQUIRED)
link_libraries(${CMAKE_THREAD_LIBS_INIT})
endif(NOT APPLE)
function(merge_static_libs TARGET_NAME)
set(libs ${ARGN})
list(REMOVE_DUPLICATES libs)
# Get all propagation dependencies from the merged libraries
foreach(lib ${libs})
list(APPEND libs_deps ${${lib}_LIB_DEPENDS})
endforeach()
if(APPLE) # Use OSX's libtool to merge archives
# To produce a library we need at least one source file.
# It is created by add_custom_command below and will helps
# also help to track dependencies.
set(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/${TARGET_NAME}_dummy.c)
# Make the generated dummy source file depended on all static input
# libs. If input lib changes,the source file is touched
# which causes the desired effect (relink).
add_custom_command(OUTPUT ${dummyfile}
COMMAND ${CMAKE_COMMAND} -E touch ${dummyfile}
DEPENDS ${libs})
# Generate dummy staic lib
file(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
add_library(${TARGET_NAME} STATIC ${dummyfile})
target_link_libraries(${TARGET_NAME} ${libs_deps})
foreach(lib ${libs})
# Get the file names of the libraries to be merged
set(libfiles ${libfiles} $<TARGET_FILE:${lib}>)
endforeach()
add_custom_command(TARGET ${TARGET_NAME} POST_BUILD
COMMAND rm "${CMAKE_CURRENT_BINARY_DIR}/lib${TARGET_NAME}.a"
COMMAND /usr/bin/libtool -static -o "${CMAKE_CURRENT_BINARY_DIR}/lib${TARGET_NAME}.a" ${libfiles})
else() # general UNIX: use "ar" to extract objects and re-add to a common lib
foreach(lib ${libs})
set(objlistfile ${lib}.objlist) # list of objects in the input library
set(objdir ${lib}.objdir)
add_custom_command(OUTPUT ${objdir}
COMMAND ${CMAKE_COMMAND} -E make_directory ${objdir}
DEPENDS ${lib})
add_custom_command(OUTPUT ${objlistfile}
COMMAND ${CMAKE_AR} -x "$<TARGET_FILE:${lib}>"
COMMAND ${CMAKE_AR} -t "$<TARGET_FILE:${lib}>" > ../${objlistfile}
DEPENDS ${lib} ${objdir}
WORKING_DIRECTORY ${objdir})
# Empty dummy source file that goes into merged library
set(mergebase ${lib}.mergebase.c)
add_custom_command(OUTPUT ${mergebase}
COMMAND ${CMAKE_COMMAND} -E touch ${mergebase}
DEPENDS ${objlistfile})
list(APPEND mergebases "${mergebase}")
endforeach()
add_library(${TARGET_NAME} STATIC ${mergebases})
target_link_libraries(${TARGET_NAME} ${libs_deps})
# Get the file name of the generated library
set(outlibfile "$<TARGET_FILE:${TARGET_NAME}>")
foreach(lib ${libs})
add_custom_command(TARGET ${TARGET_NAME} POST_BUILD
COMMAND ${CMAKE_AR} cr ${outlibfile} *.o
COMMAND ${CMAKE_RANLIB} ${outlibfile}
WORKING_DIRECTORY ${lib}.objdir)
endforeach()
endif()
endfunction(merge_static_libs)
# cc_library parses tensor.cc and figures out that target also depend on tensor.h.
# cc_library(tensor
# SRCS
# tensor.cc
# DEPS
# variant)
function(cc_library TARGET_NAME)
set(options OPTIONAL)
set(options STATIC static SHARED shared)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(cc_library "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
if (${cc_library_OPTIONAL} STREQUAL "SHARED")
add_library(${TARGET_NAME} SHARED ${cc_library_SRCS})
else()
add_library(${TARGET_NAME} STATIC ${cc_library_SRCS})
endif()
add_dependencies(${TARGET_NAME} ${cc_library_DEPS} ${external_project_dependencies})
if (cc_library_SRCS)
if (cc_library_SHARED OR cc_library_shared) # build *.so
add_library(${TARGET_NAME} SHARED ${cc_library_SRCS})
else()
add_library(${TARGET_NAME} STATIC ${cc_library_SRCS})
endif()
if (cc_library_DEPS)
add_dependencies(${TARGET_NAME} ${cc_library_DEPS})
target_link_libraries(${TARGET_NAME} ${cc_library_DEPS})
endif()
else(cc_library_SRCS)
if (cc_library_DEPS)
merge_static_libs(${TARGET_NAME} ${cc_library_DEPS})
else()
message(FATAL "Please specify source file or library in cc_library.")
endif()
endif(cc_library_SRCS)
endfunction(cc_library)
# cc_binary parses tensor.cc and figures out that target also depend on tensor.h.
# cc_binary(tensor
# SRCS
# tensor.cc)
function(cc_binary TARGET_NAME)
set(options OPTIONAL)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(cc_binary "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
add_executable(${TARGET_NAME} ${cc_binary_SRCS})
link_paddle_exe(${TARGET_NAME})
if(cc_binary_DEPS)
if(cc_binary_DEPS)
target_link_libraries(${TARGET_NAME} ${cc_binary_DEPS})
add_dependencies(${TARGET_NAME} ${cc_binary_DEPS})
endif()
endfunction(cc_binary)
# The dependency to target tensor implies that if any of
# tensor{.h,.cc,_test.cc} is changed, tensor_test need to be re-built.
# cc_test(tensor_test
# SRCS
# tensor_test.cc
# DEPS
# tensor)
function(cc_test TARGET_NAME)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(cc_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
add_executable(${TARGET_NAME} ${cc_test_SRCS})
link_paddle_test(${TARGET_NAME})
if(cc_test_DEPS)
target_link_libraries(${TARGET_NAME} ${cc_test_DEPS})
add_dependencies(${TARGET_NAME} ${cc_test_DEPS})
if(WITH_TESTING)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(cc_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
add_executable(${TARGET_NAME} ${cc_test_SRCS})
target_link_libraries(${TARGET_NAME} ${cc_test_DEPS} gtest gtest_main)
add_dependencies(${TARGET_NAME} ${cc_test_DEPS} gtest gtest_main)
add_test(NAME ${TARGET_NAME} COMMAND ${TARGET_NAME} WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
endif()
add_test(${TARGET_NAME} ${TARGET_NAME})
endfunction(cc_test)
# Suppose that ops.cu includes global functions that take Tensor as
# their parameters, so ops depend on tensor. This implies that if
# any of tensor.{h.cc}, ops.{h,cu} is changed, ops need to be re-built.
# nv_library(ops
# SRCS
# ops.cu
# DEPS
# tensor)
function(nv_library TARGET_NAME)
set(options OPTIONAL)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(nv_library "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
if (${nv_library_OPTIONAL} STREQUAL "SHARED")
cuda_add_library(${TARGET_NAME} SHARED ${nv_library_SRCS})
else()
cuda_add_library(${TARGET_NAME} STATIC ${nv_library_SRCS})
if (WITH_GPU)
set(options STATIC static SHARED shared)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(nv_library "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
if(nv_library_SRCS)
if (nv_library_SHARED OR nv_library_shared) # build *.so
cuda_add_library(${TARGET_NAME} SHARED ${nv_library_SRCS})
else()
cuda_add_library(${TARGET_NAME} STATIC ${nv_library_SRCS})
endif()
if (nv_library_DEPS)
add_dependencies(${TARGET_NAME} ${nv_library_DEPS})
target_link_libraries(${TARGET_NAME} ${nv_library_DEPS})
endif()
else(nv_library_SRCS)
if (nv_library_DEPS)
merge_static_libs(${TARGET_NAME} ${nv_library_DEPS})
else()
message(FATAL "Please specify source file or library in nv_library.")
endif()
endif(nv_library_SRCS)
endif()
add_dependencies(${TARGET_NAME} ${nv_library_DEPS} ${external_project_dependencies})
endfunction(nv_library)
function(nv_binary TARGET_NAME)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(nv_binary "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
cuda_add_executable(${TARGET_NAME} ${nv_binary_SRCS})
link_paddle_exe(${TARGET_NAME})
if(nv_binary_DEPS)
target_link_libraries(${TARGET_NAME} ${nv_binary_DEPS})
add_dependencies(${TARGET_NAME} ${nv_binary_DEPS})
if (WITH_GPU)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(nv_binary "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
cuda_add_executable(${TARGET_NAME} ${nv_binary_SRCS})
if(nv_binary_DEPS)
target_link_libraries(${TARGET_NAME} ${nv_binary_DEPS})
add_dependencies(${TARGET_NAME} ${nv_binary_DEPS})
endif()
endif()
endfunction(nv_binary)
# The dependency to target tensor implies that if any of
# ops{.h,.cu,_test.cu} is changed, ops_test need to be re-built.
# nv_test(ops_test
# SRCS
# ops_test.cu
# DEPS
# ops)
function(nv_test TARGET_NAME)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(nv_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
cuda_add_executable(${TARGET_NAME} ${nv_test_SRCS})
link_paddle_test(${TARGET_NAME})
if(nv_test_DEPS)
target_link_libraries(${TARGET_NAME} ${nv_test_DEPS})
add_dependencies(${TARGET_NAME} ${nv_test_DEPS})
if (WITH_GPU AND WITH_TESTING)
set(options "")
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(nv_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
cuda_add_executable(${TARGET_NAME} ${nv_test_SRCS})
target_link_libraries(${TARGET_NAME} ${nv_test_DEPS} gtest gtest_main)
add_dependencies(${TARGET_NAME} ${nv_test_DEPS} gtest gtest_main)
add_test(${TARGET_NAME} ${TARGET_NAME})
endif()
add_test(${TARGET_NAME} ${TARGET_NAME})
endfunction(nv_test)
function(go_library TARGET_NAME)
set(options STATIC static SHARED shared)
set(oneValueArgs "")
set(multiValueArgs DEPS)
cmake_parse_arguments(go_library "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
if (go_library_SHARED OR go_library_shared)
set(BUILD_MODE "-buildmode=c-shared")
set(${TARGET_NAME}_LIB_NAME "${CMAKE_SHARED_LIBRARY_PREFIX}${TARGET_NAME}${CMAKE_SHARED_LIBRARY_SUFFIX}" CACHE STRING "output library name for target ${TARGET_NAME}")
else()
set(BUILD_MODE "-buildmode=c-archive")
set(${TARGET_NAME}_LIB_NAME "${CMAKE_STATIC_LIBRARY_PREFIX}${TARGET_NAME}${CMAKE_STATIC_LIBRARY_SUFFIX}" CACHE STRING "output library name for target ${TARGET_NAME}")
endif()
# Add dummy code to support `make target_name` under Terminal Command
set(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/${TARGET_NAME}_dummy.c)
file(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
if (go_library_SHARED OR go_library_shared)
add_library(${TARGET_NAME} SHARED ${dummyfile})
else()
add_library(${TARGET_NAME} STATIC ${dummyfile})
endif()
if(go_library_DEPS)
add_dependencies(${TARGET_NAME} ${go_library_DEPS})
endif(go_library_DEPS)
set(${TARGET_NAME}_LIB_PATH "${CMAKE_CURRENT_BINARY_DIR}/${${TARGET_NAME}_LIB_NAME}" CACHE STRING "output library path for target ${TARGET_NAME}")
file(GLOB GO_SOURCE RELATIVE "${CMAKE_CURRENT_SOURCE_DIR}" "*.go")
string(REPLACE "${PADDLE_GO_PATH}/" "" CMAKE_CURRENT_SOURCE_REL_DIR ${CMAKE_CURRENT_SOURCE_DIR})
# FIXME: link path
add_custom_command(TARGET ${TARGET_NAME} POST_BUILD
COMMAND rm "${${TARGET_NAME}_LIB_PATH}"
# Golang build source code
COMMAND GOPATH=${GOPATH} ${CMAKE_Go_COMPILER} build ${BUILD_MODE}
-o "${${TARGET_NAME}_LIB_PATH}"
"./${CMAKE_CURRENT_SOURCE_REL_DIR}/${GO_SOURCE}"
# must run under GOPATH
WORKING_DIRECTORY "${PADDLE_IN_GOPATH}/go")
add_dependencies(${TARGET_NAME} go_vendor)
endfunction(go_library)
function(go_binary TARGET_NAME)
set(options OPTIONAL)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(go_binary "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
string(REPLACE "${PADDLE_GO_PATH}/" "" CMAKE_CURRENT_SOURCE_REL_DIR ${CMAKE_CURRENT_SOURCE_DIR})
# FIXME: link path
add_custom_command(OUTPUT ${TARGET_NAME}_timestamp
COMMAND env LIBRARY_PATH=${CMAKE_BINARY_DIR}/go/pserver/client/c/:$ENV{LIBRARY_PATH}
GOPATH=${GOPATH} ${CMAKE_Go_COMPILER} build
-o "${CMAKE_CURRENT_BINARY_DIR}/${TARGET_NAME}"
"./${CMAKE_CURRENT_SOURCE_REL_DIR}/${go_binary_SRCS}"
WORKING_DIRECTORY "${PADDLE_IN_GOPATH}/go")
# TODO: don't know what ${TARGET_NAME}_link does
add_custom_target(${TARGET_NAME} ALL DEPENDS go_vendor ${TARGET_NAME}_timestamp ${go_binary_DEPS})
install(PROGRAMS ${CMAKE_CURRENT_BINARY_DIR}/${TARGET_NAME} DESTINATION bin)
endfunction(go_binary)
function(go_test TARGET_NAME)
set(options OPTIONAL)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(go_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
add_custom_command(OUTPUT ${TARGET_NAME}_timestamp
COMMAND env GOPATH=${GOPATH} ${CMAKE_Go_COMPILER} test
-c -o "${CMAKE_CURRENT_BINARY_DIR}/${TARGET_NAME}"
${go_test_SRCS}
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
add_custom_target(${TARGET_NAME} ALL DEPENDS ${TARGET_NAME}_timestamp ${go_test_DEPS})
add_test(${TARGET_NAME} ${CMAKE_CURRENT_BINARY_DIR}/${TARGET_NAME})
endfunction(go_test)
function(proto_library TARGET_NAME)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(proto_library "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
set(proto_srcs)
set(proto_hdrs)
protobuf_generate_cpp(proto_srcs proto_hdrs ${proto_library_SRCS})
cc_library(${TARGET_NAME} SRCS ${proto_srcs} DEPS ${proto_library_DEPS} protobuf)
endfunction()
function(py_proto_compile TARGET_NAME)
set(oneValueArgs "")
set(multiValueArgs SRCS)
cmake_parse_arguments(py_proto_compile "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
set(py_srcs)
protobuf_generate_python(py_srcs ${py_proto_compile_SRCS})
add_custom_target(${TARGET_NAME} ALL DEPENDS ${py_srcs})
endfunction()
cmake/rdma.cmake
浏览文件 @ a698824a
......@@ -10,7 +10,7 @@ if(WITH_RDMA)
function(generate_rdma_links)
#redirect to current DIR to isolate the pollution from system runtime environment
#it can benifits unified control for different gcc environment.
#it can benifits unified control for different gcc environment.
#e.g, by default gcc48 did not refer /usr/lib64 which could contain low version
#runtime libraries that will crash process while loading it. That redirect trick
#can fix it.
......@@ -19,7 +19,9 @@ if(WITH_RDMA)
COMMAND ln -s -f /usr/lib64/libibverbs.so.1.0.0 librdma/libibverbs.so.1
COMMAND ln -s -f /usr/lib64/libibverbs.so.1.0.0 librdma/libibverbs.so
COMMAND ln -s -f /usr/lib64/librdmacm.so.1.0.0 librdma/librdmacm.so.1
COMMAND ln -s -f /usr/lib64/librdmacm.so.1.0.0 librdma/librdmacm.so
COMMAND ln -s -f /usr/lib64/librdmacm.so.1.0.0 librdma/librdmacm.so
COMMAND ln -s -f /lib64/libnl.so.1.1.4 librdma/libnl.so.1
COMMAND ln -s -f /lib64/libnl.so.1.1.4 librdma/libnl.so
WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR}
)
endfunction(generate_rdma_links)
......@@ -44,7 +46,7 @@ if(WITH_RDMA)
RDMA_INC_XIO AND
RDMA_INC_EVENT AND
RDMA_INC_NUMA AND
RDMA_LIB_SXISOCK AND
RDMA_LIB_SXISOCK AND
RDMA_LIB_XIO AND
RDMA_LIB_EVENT AND
RDMA_LIB_EVENT_CORE AND
......@@ -53,19 +55,19 @@ if(WITH_RDMA)
RDMA_LIB_NUMA
)
set(RDMA_INC_DIR
${RDMA_INC_SXISOCK}
set(RDMA_INC_DIR
${RDMA_INC_SXISOCK}
${RDMA_INC_XIO}
${RDMA_INC_EVENT}
${RDMA_INC_NUMA})
set(RDMA_LIBS
${RDMA_LIB_SXISOCK}
${RDMA_LIB_XIO}
${RDMA_LIB_EVENT}
${RDMA_LIB_EVENT_CORE}
${RDMA_LIB_EVENT_EXTRA}
${RDMA_LIB_EVENT_PTHREADS}
${RDMA_LIB_NUMA}
set(RDMA_LIBS
${RDMA_LIB_SXISOCK}
${RDMA_LIB_XIO}
${RDMA_LIB_EVENT}
${RDMA_LIB_EVENT_CORE}
${RDMA_LIB_EVENT_EXTRA}
${RDMA_LIB_EVENT_PTHREADS}
${RDMA_LIB_NUMA}
)
set(RDMA_LD_FLAGS "-L./librdma -libverbs -lrdmacm -Xlinker -rpath ./librdma")
include_directories("${RDMA_INC_DIR}")
......
cmake/system.cmake
浏览文件 @ a698824a
......@@ -33,6 +33,7 @@ ELSE(WIN32)
SET(CMAKE_OSX_DEPLOYMENT_TARGET ${MACOS_VERSION} CACHE STRING
"Minimum OS X version to target for deployment (at runtime); newer APIs weak linked. Set to empty string for default value.")
ENDIF()
set(CMAKE_EXE_LINKER_FLAGS "-framework CoreFoundation -framework Security")
ELSE(APPLE)
IF(EXISTS "/etc/issue")
......@@ -84,24 +85,6 @@ IF(DEFINED CMAKE_SYSTEM_NAME)
ENDIF()
ENDIF()
# prefix and suffix on different os
IF(WIN32)
SET(LIBRARY_PREFIX "")
SET(SHARED_LIBRARY_SUFFIX ".dll")
SET(STATIC_LIBRARY_SUFFIX ".lib")
SET(EXECUTABLE_SUFFIX ".exe")
ELSE(WIN32)
SET(LIBRARY_PREFIX "lib")
IF(APPLE)
SET(SHARED_LIBRARY_SUFFIX ".dylib")
ELSE(APPLE)
SET(SHARED_LIBRARY_SUFFIX ".so")
ENDIF(APPLE)
SET(STATIC_LIBRARY_SUFFIX ".a")
SET(EXECUTABLE_SUFFIX "")
ENDIF(WIN32)
# external dependencies log output
SET(EXTERNAL_PROJECT_LOG_ARGS
LOG_DOWNLOAD 0 # Wrap download in script to log output
......
cmake/util.cmake
浏览文件 @ a698824a
......@@ -84,6 +84,7 @@ function(link_paddle_exe TARGET_NAME)
paddle_parameter
paddle_proto
paddle_cuda
paddle_optimizer
${EXTERNAL_LIBS}
${CMAKE_THREAD_LIBS_INIT}
${CMAKE_DL_LIBS}
......@@ -149,8 +150,9 @@ endfunction()
# Create a python unittest using run_python_tests.sh,
# which takes care of making correct running environment
function(add_python_test TEST_NAME)
add_test(NAME ${TEST_NAME}
COMMAND bash ${PROJ_ROOT}/paddle/scripts/run_python_tests.sh
add_test(NAME ${TEST_NAME}
COMMAND env PADDLE_PACKAGE_DIR=${PADDLE_PYTHON_PACKAGE_DIR}
bash ${PROJ_ROOT}/paddle/scripts/run_python_tests.sh
${USE_VIRTUALENV_FOR_TEST} ${PYTHON_EXECUTABLE} ${ARGN}
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
endfunction()
demo/image_classification/.gitignore 已删除 100644 → 0
浏览文件 @ 338dd135
data/cifar-10-batches-py
data/cifar-out
cifar_vgg_model/*
plot.png
train.log
image_provider_copy_1.py
*pyc
train.list
test.list
demo/image_classification/api_v2_resnet.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 paddle.v2 as paddle
__all__ = ['resnet_cifar10']
def conv_bn_layer(input,
ch_out,
filter_size,
stride,
padding,
active_type=paddle.activation.Relu(),
ch_in=None):
tmp = paddle.layer.img_conv(
input=input,
filter_size=filter_size,
num_channels=ch_in,
num_filters=ch_out,
stride=stride,
padding=padding,
act=paddle.activation.Linear(),
bias_attr=False)
return paddle.layer.batch_norm(input=tmp, act=active_type)
def shortcut(ipt, n_in, n_out, stride):
if n_in != n_out:
return conv_bn_layer(ipt, n_out, 1, stride, 0,
paddle.activation.Linear())
else:
return ipt
def basicblock(ipt, ch_out, stride):
ch_in = ch_out * 2
tmp = conv_bn_layer(ipt, ch_out, 3, stride, 1)
tmp = conv_bn_layer(tmp, ch_out, 3, 1, 1, paddle.activation.Linear())
short = shortcut(ipt, ch_in, ch_out, stride)
return paddle.layer.addto(input=[tmp, short], act=paddle.activation.Relu())
def layer_warp(block_func, ipt, features, count, stride):
tmp = block_func(ipt, features, stride)
for i in range(1, count):
tmp = block_func(tmp, features, 1)
return tmp
def resnet_cifar10(ipt, depth=32):
# depth should be one of 20, 32, 44, 56, 110, 1202
assert (depth - 2) % 6 == 0
n = (depth - 2) / 6
nStages = {16, 64, 128}
conv1 = conv_bn_layer(
ipt, ch_in=3, ch_out=16, filter_size=3, stride=1, padding=1)
res1 = layer_warp(basicblock, conv1, 16, n, 1)
res2 = layer_warp(basicblock, res1, 32, n, 2)
res3 = layer_warp(basicblock, res2, 64, n, 2)
pool = paddle.layer.img_pool(
input=res3, pool_size=8, stride=1, pool_type=paddle.pooling.Avg())
return pool
demo/image_classification/api_v2_train.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 paddle.v2 as paddle
from api_v2_vgg import vgg_bn_drop
def main():
datadim = 3 * 32 * 32
classdim = 10
# PaddlePaddle init
paddle.init(use_gpu=False, trainer_count=1)
image = paddle.layer.data(
name="image", type=paddle.data_type.dense_vector(datadim))
# Add neural network config
# option 1. resnet
# net = resnet_cifar10(image, depth=32)
# option 2. vgg
net = vgg_bn_drop(image)
out = paddle.layer.fc(input=net,
size=classdim,
act=paddle.activation.Softmax())
lbl = paddle.layer.data(
name="label", type=paddle.data_type.integer_value(classdim))
cost = paddle.layer.classification_cost(input=out, label=lbl)
# Create parameters
parameters = paddle.parameters.create(cost)
# Create optimizer
momentum_optimizer = paddle.optimizer.Momentum(
momentum=0.9,
regularization=paddle.optimizer.L2Regularization(rate=0.0002 * 128),
learning_rate=0.1 / 128.0,
learning_rate_decay_a=0.1,
learning_rate_decay_b=50000 * 100,
learning_rate_schedule='discexp',
batch_size=128)
# End batch and end pass event handler
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
print "\nPass %d, Batch %d, Cost %f, %s" % (
event.pass_id, event.batch_id, event.cost, event.metrics)
else:
sys.stdout.write('.')
sys.stdout.flush()
if isinstance(event, paddle.event.EndPass):
result = trainer.test(
reader=paddle.batch(
paddle.dataset.cifar.test10(), batch_size=128),
feeding={'image': 0,
'label': 1})
print "\nTest with Pass %d, %s" % (event.pass_id, result.metrics)
# Create trainer
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=momentum_optimizer)
trainer.train(
reader=paddle.batch(
paddle.reader.shuffle(
paddle.dataset.cifar.train10(), buf_size=50000),
batch_size=128),
num_passes=5,
event_handler=event_handler,
feeding={'image': 0,
'label': 1})
if __name__ == '__main__':
main()
demo/image_classification/api_v2_vgg.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 paddle.v2 as paddle
__all__ = ['vgg_bn_drop']
def vgg_bn_drop(input):
def conv_block(ipt, num_filter, groups, dropouts, num_channels=None):
return paddle.networks.img_conv_group(
input=ipt,
num_channels=num_channels,
pool_size=2,
pool_stride=2,
conv_num_filter=[num_filter] * groups,
conv_filter_size=3,
conv_act=paddle.activation.Relu(),
conv_with_batchnorm=True,
conv_batchnorm_drop_rate=dropouts,
pool_type=paddle.pooling.Max())
conv1 = conv_block(input, 64, 2, [0.3, 0], 3)
conv2 = conv_block(conv1, 128, 2, [0.4, 0])
conv3 = conv_block(conv2, 256, 3, [0.4, 0.4, 0])
conv4 = conv_block(conv3, 512, 3, [0.4, 0.4, 0])
conv5 = conv_block(conv4, 512, 3, [0.4, 0.4, 0])
drop = paddle.layer.dropout(input=conv5, dropout_rate=0.5)
fc1 = paddle.layer.fc(input=drop, size=512, act=paddle.activation.Linear())
bn = paddle.layer.batch_norm(
input=fc1,
act=paddle.activation.Relu(),
layer_attr=paddle.attr.Extra(drop_rate=0.5))
fc2 = paddle.layer.fc(input=bn, size=512, act=paddle.activation.Linear())
return fc2
demo/image_classification/data/download_cifar.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
wget https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz
tar zxf cifar-10-python.tar.gz
rm cifar-10-python.tar.gz
rm -rf cifar-out/*
echo Converting CIFAR data to images.....
python process_cifar.py ./cifar-10-batches-py ./cifar-out
demo/image_classification/data/process_cifar.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 numpy as np
import sys
import os
import PIL.Image as Image
"""
Usage: python process_cifar input_dir output_dir
"""
def mkdir_not_exist(path):
"""
Make dir if the path does not exist.
path: the path to be created.
"""
if not os.path.exists(path):
os.mkdir(path)
def create_dir_structure(output_dir):
"""
Create the directory structure for the directory.
output_dir: the direcotry structure path.
"""
mkdir_not_exist(os.path.join(output_dir))
mkdir_not_exist(os.path.join(output_dir, "train"))
mkdir_not_exist(os.path.join(output_dir, "test"))
def convert_batch(batch_path, label_set, label_map, output_dir, data_split):
"""
Convert CIFAR batch to the structure of Paddle format.
batch_path: the batch to be converted.
label_set: the set of labels.
output_dir: the output path.
data_split: whether it is training or testing data.
"""
data = np.load(batch_path)
for data, label, filename in zip(data['data'], data['labels'],
data['filenames']):
data = data.reshape((3, 32, 32))
data = np.transpose(data, (1, 2, 0))
label = label_map[label]
output_dir_this = os.path.join(output_dir, data_split, str(label))
output_filename = os.path.join(output_dir_this, filename)
if not label in label_set:
label_set[label] = True
mkdir_not_exist(output_dir_this)
Image.fromarray(data).save(output_filename)
if __name__ == '__main__':
input_dir = sys.argv[1]
output_dir = sys.argv[2]
num_batch = 5
create_dir_structure(output_dir)
label_map = {
0: "airplane",
1: "automobile",
2: "bird",
3: "cat",
4: "deer",
5: "dog",
6: "frog",
7: "horse",
8: "ship",
9: "truck"
}
labels = {}
for i in range(1, num_batch + 1):
convert_batch(
os.path.join(input_dir, "data_batch_%d" % i), labels, label_map,
output_dir, "train")
convert_batch(
os.path.join(input_dir, "test_batch"), {}, label_map, output_dir,
"test")
demo/image_classification/image_provider.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 io
import random
import paddle.utils.image_util as image_util
from paddle.trainer.PyDataProvider2 import *
#
# {'img_size': 32,
# 'settings': a global object,
# 'color': True,
# 'mean_img_size': 32,
# 'meta': './data/cifar-out/batches/batches.meta',
# 'num_classes': 10,
# 'file_list': ('./data/cifar-out/batches/train_batch_000',),
# 'use_jpeg': True}
def hook(settings, img_size, mean_img_size, num_classes, color, meta, use_jpeg,
is_train, **kwargs):
settings.mean_img_size = mean_img_size
settings.img_size = img_size
settings.num_classes = num_classes
settings.color = color
settings.is_train = is_train
if settings.color:
settings.img_raw_size = settings.img_size * settings.img_size * 3
else:
settings.img_raw_size = settings.img_size * settings.img_size
settings.meta_path = meta
settings.use_jpeg = use_jpeg
settings.img_mean = image_util.load_meta(settings.meta_path,
settings.mean_img_size,
settings.img_size, settings.color)
settings.logger.info('Image size: %s', settings.img_size)
settings.logger.info('Meta path: %s', settings.meta_path)
settings.input_types = {
'image': dense_vector(settings.img_raw_size),
'label': integer_value(settings.num_classes)
}
settings.logger.info('DataProvider Initialization finished')
@provider(init_hook=hook, min_pool_size=0)
def processData(settings, file_list):
"""
The main function for loading data.
Load the batch, iterate all the images and labels in this batch.
file_list: the batch file list.
"""
with open(file_list, 'r') as fdata:
lines = [line.strip() for line in fdata]
random.shuffle(lines)
for file_name in lines:
with io.open(file_name.strip(), 'rb') as file:
data = cPickle.load(file)
indexes = list(range(len(data['images'])))
if settings.is_train:
random.shuffle(indexes)
for i in indexes:
if settings.use_jpeg == 1:
img = image_util.decode_jpeg(data['images'][i])
else:
img = data['images'][i]
img_feat = image_util.preprocess_img(
img, settings.img_mean, settings.img_size,
settings.is_train, settings.color)
label = data['labels'][i]
yield {
'image': img_feat.astype('float32'),
'label': int(label)
}
demo/image_classification/image_util.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 numpy as np
from PIL import Image
from cStringIO import StringIO
def resize_image(img, target_size):
"""
Resize an image so that the shorter edge has length target_size.
img: the input image to be resized.
target_size: the target resized image size.
"""
percent = (target_size / float(min(img.size[0], img.size[1])))
resized_size = int(round(img.size[0] * percent)), int(
round(img.size[1] * percent))
img = img.resize(resized_size, Image.ANTIALIAS)
return img
def flip(im):
"""
Return the flipped image.
Flip an image along the horizontal direction.
im: input image, (H x W x K) ndarrays
"""
if len(im.shape) == 3:
return im[:, :, ::-1]
else:
return im[:, ::-1]
def crop_img(im, inner_size, color=True, test=True):
"""
Return cropped image.
The size of the cropped image is inner_size * inner_size.
im: (K x H x W) ndarrays
inner_size: the cropped image size.
color: whether it is color image.
test: whether in test mode.
If False, does random cropping and flipping.
If True, crop the center of images.
"""
if color:
height, width = max(inner_size, im.shape[1]), max(inner_size,
im.shape[2])
padded_im = np.zeros((3, height, width))
startY = (height - im.shape[1]) / 2
startX = (width - im.shape[2]) / 2
endY, endX = startY + im.shape[1], startX + im.shape[2]
padded_im[:, startY:endY, startX:endX] = im
else:
im = im.astype('float32')
height, width = max(inner_size, im.shape[0]), max(inner_size,
im.shape[1])
padded_im = np.zeros((height, width))
startY = (height - im.shape[0]) / 2
startX = (width - im.shape[1]) / 2
endY, endX = startY + im.shape[0], startX + im.shape[1]
padded_im[startY:endY, startX:endX] = im
if test:
startY = (height - inner_size) / 2
startX = (width - inner_size) / 2
else:
startY = np.random.randint(0, height - inner_size + 1)
startX = np.random.randint(0, width - inner_size + 1)
endY, endX = startY + inner_size, startX + inner_size
if color:
pic = padded_im[:, startY:endY, startX:endX]
else:
pic = padded_im[startY:endY, startX:endX]
if (not test) and (np.random.randint(2) == 0):
pic = flip(pic)
return pic
def decode_jpeg(jpeg_string):
np_array = np.array(Image.open(StringIO(jpeg_string)))
if len(np_array.shape) == 3:
np_array = np.transpose(np_array, (2, 0, 1))
return np_array
def preprocess_img(im, img_mean, crop_size, is_train, color=True):
"""
Does data augmentation for images.
If is_train is false, cropping the center region from the image.
If is_train is true, randomly crop a region from the image,
and randomy does flipping.
im: (K x H x W) ndarrays
"""
im = im.astype('float32')
test = not is_train
pic = crop_img(im, crop_size, color, test)
pic -= img_mean
return pic.flatten()
def load_meta(meta_path, mean_img_size, crop_size, color=True):
"""
Return the loaded meta file.
Load the meta image, which is the mean of the images in the dataset.
The mean image is subtracted from every input image so that the expected mean
of each input image is zero.
"""
mean = np.load(meta_path)['data_mean']
border = (mean_img_size - crop_size) / 2
if color:
assert (mean_img_size * mean_img_size * 3 == mean.shape[0])
mean = mean.reshape(3, mean_img_size, mean_img_size)
mean = mean[:, border:border + crop_size, border:border +
crop_size].astype('float32')
else:
assert (mean_img_size * mean_img_size == mean.shape[0])
mean = mean.reshape(mean_img_size, mean_img_size)
mean = mean[border:border + crop_size, border:border +
crop_size].astype('float32')
return mean
def load_image(img_path, is_color=True):
"""
Load image and return.
img_path: image path.
is_color: is color image or not.
"""
img = Image.open(img_path)
img.load()
return img
def oversample(img, crop_dims):
"""
image : iterable of (H x W x K) ndarrays
crop_dims: (height, width) tuple for the crops.
Returned data contains ten crops of input image, namely,
four corner patches and the center patch as well as their
horizontal reflections.
"""
# Dimensions and center.
im_shape = np.array(img[0].shape)
crop_dims = np.array(crop_dims)
im_center = im_shape[:2] / 2.0
# Make crop coordinates
h_indices = (0, im_shape[0] - crop_dims[0])
w_indices = (0, im_shape[1] - crop_dims[1])
crops_ix = np.empty((5, 4), dtype=int)
curr = 0
for i in h_indices:
for j in w_indices:
crops_ix[curr] = (i, j, i + crop_dims[0], j + crop_dims[1])
curr += 1
crops_ix[4] = np.tile(im_center, (1, 2)) + np.concatenate(
[-crop_dims / 2.0, crop_dims / 2.0])
crops_ix = np.tile(crops_ix, (2, 1))
# Extract crops
crops = np.empty(
(10 * len(img), crop_dims[0], crop_dims[1], im_shape[-1]),
dtype=np.float32)
ix = 0
for im in img:
for crop in crops_ix:
crops[ix] = im[crop[0]:crop[2], crop[1]:crop[3], :]
ix += 1
crops[ix - 5:ix] = crops[ix - 5:ix, :, ::-1, :] # flip for mirrors
return crops
class ImageTransformer:
def __init__(self,
transpose=None,
channel_swap=None,
mean=None,
is_color=True):
self.transpose = transpose
self.channel_swap = None
self.mean = None
self.is_color = is_color
def set_transpose(self, order):
if self.is_color:
assert 3 == len(order)
self.transpose = order
def set_channel_swap(self, order):
if self.is_color:
assert 3 == len(order)
self.channel_swap = order
def set_mean(self, mean):
# mean value, may be one value per channel
if mean.ndim == 1:
mean = mean[:, np.newaxis, np.newaxis]
else:
# elementwise mean
if self.is_color:
assert len(mean.shape) == 3
self.mean = mean
def transformer(self, data):
if self.transpose is not None:
data = data.transpose(self.transpose)
if self.channel_swap is not None:
data = data[self.channel_swap, :, :]
if self.mean is not None:
data -= self.mean
return data
demo/image_classification/prediction.py 已删除 100755 → 0
浏览文件 @ 338dd135
# 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 os, sys
import numpy as np
import logging
from PIL import Image
from optparse import OptionParser
import paddle.utils.image_util as image_util
from py_paddle import swig_paddle, DataProviderConverter
from paddle.trainer.PyDataProvider2 import dense_vector
from paddle.trainer.config_parser import parse_config
logging.basicConfig(
format='[%(levelname)s %(asctime)s %(filename)s:%(lineno)s] %(message)s')
logging.getLogger().setLevel(logging.INFO)
class ImageClassifier():
def __init__(self,
train_conf,
use_gpu=True,
model_dir=None,
resize_dim=None,
crop_dim=None,
mean_file=None,
oversample=False,
is_color=True):
"""
train_conf: network configure.
model_dir: string, directory of model.
resize_dim: int, resized image size.
crop_dim: int, crop size.
mean_file: string, image mean file.
oversample: bool, oversample means multiple crops, namely five
patches (the four corner patches and the center
patch) as well as their horizontal reflections,
ten crops in all.
"""
self.train_conf = train_conf
self.model_dir = model_dir
if model_dir is None:
self.model_dir = os.path.dirname(train_conf)
self.resize_dim = resize_dim
self.crop_dims = [crop_dim, crop_dim]
self.oversample = oversample
self.is_color = is_color
self.transformer = image_util.ImageTransformer(is_color=is_color)
self.transformer.set_transpose((2, 0, 1))
self.mean_file = mean_file
mean = np.load(self.mean_file)['data_mean']
mean = mean.reshape(3, self.crop_dims[0], self.crop_dims[1])
self.transformer.set_mean(mean) # mean pixel
gpu = 1 if use_gpu else 0
conf_args = "is_test=1,use_gpu=%d,is_predict=1" % (gpu)
conf = parse_config(train_conf, conf_args)
swig_paddle.initPaddle("--use_gpu=%d" % (gpu))
self.network = swig_paddle.GradientMachine.createFromConfigProto(
conf.model_config)
assert isinstance(self.network, swig_paddle.GradientMachine)
self.network.loadParameters(self.model_dir)
data_size = 3 * self.crop_dims[0] * self.crop_dims[1]
slots = [dense_vector(data_size)]
self.converter = DataProviderConverter(slots)
def get_data(self, img_path):
"""
1. load image from img_path.
2. resize or oversampling.
3. transformer data: transpose, sub mean.
return K x H x W ndarray.
img_path: image path.
"""
image = image_util.load_image(img_path, self.is_color)
if self.oversample:
# image_util.resize_image: short side is self.resize_dim
image = image_util.resize_image(image, self.resize_dim)
image = np.array(image)
input = np.zeros(
(1, image.shape[0], image.shape[1], 3), dtype=np.float32)
input[0] = image.astype(np.float32)
input = image_util.oversample(input, self.crop_dims)
else:
image = image.resize(self.crop_dims, Image.ANTIALIAS)
input = np.zeros(
(1, self.crop_dims[0], self.crop_dims[1], 3), dtype=np.float32)
input[0] = np.array(image).astype(np.float32)
data_in = []
for img in input:
img = self.transformer.transformer(img).flatten()
data_in.append([img.tolist()])
return data_in
def forward(self, input_data):
in_arg = self.converter(input_data)
return self.network.forwardTest(in_arg)
def forward(self, data, output_layer):
"""
input_data: py_paddle input data.
output_layer: specify the name of probability, namely the layer with
softmax activation.
return: the predicting probability of each label.
"""
input = self.converter(data)
self.network.forwardTest(input)
output = self.network.getLayerOutputs(output_layer)
# For oversampling, average predictions across crops.
# If not, the shape of output[name]: (1, class_number),
# the mean is also applicable.
return output[output_layer]['value'].mean(0)
def predict(self, image=None, output_layer=None):
assert isinstance(image, basestring)
assert isinstance(output_layer, basestring)
data = self.get_data(image)
prob = self.forward(data, output_layer)
lab = np.argsort(-prob)
logging.info("Label of %s is: %d", image, lab[0])
if __name__ == '__main__':
image_size = 32
crop_size = 32
multi_crop = True
config = "vgg_16_cifar.py"
output_layer = "__fc_layer_1__"
mean_path = "data/cifar-out/batches/batches.meta"
model_path = sys.argv[1]
image = sys.argv[2]
use_gpu = bool(int(sys.argv[3]))
obj = ImageClassifier(
train_conf=config,
model_dir=model_path,
resize_dim=image_size,
crop_dim=crop_size,
mean_file=mean_path,
use_gpu=use_gpu,
oversample=multi_crop)
obj.predict(image, output_layer)
demo/image_classification/preprocess.py 已删除 100755 → 0
浏览文件 @ 338dd135
# 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 paddle.utils.preprocess_img import ImageClassificationDatasetCreater
from optparse import OptionParser
def option_parser():
parser = OptionParser(usage="usage: python preprcoess.py "\
"-i data_dir [options]")
parser.add_option(
"-i",
"--input",
action="store",
dest="input",
help="Input data directory.")
parser.add_option(
"-s",
"--size",
action="store",
dest="size",
help="Processed image size.")
parser.add_option(
"-c",
"--color",
action="store",
dest="color",
help="whether to use color images.")
return parser.parse_args()
if __name__ == '__main__':
options, args = option_parser()
data_dir = options.input
processed_image_size = int(options.size)
color = options.color == "1"
data_creator = ImageClassificationDatasetCreater(
data_dir, processed_image_size, color)
data_creator.train_list_name = "train.txt"
data_creator.test_list_name = "test.txt"
data_creator.num_per_batch = 1000
data_creator.overwrite = True
data_creator.create_batches()
demo/image_classification/preprocess.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
data_dir=./data/cifar-out
python preprocess.py -i $data_dir -s 32 -c 1
echo "data/cifar-out/batches/train.txt" > train.list
echo "data/cifar-out/batches/test.txt" > test.list
demo/image_classification/train.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
set -e
config=vgg_16_cifar.py
output=./cifar_vgg_model
log=train.log
paddle train \
--config=$config \
--dot_period=10 \
--log_period=100 \
--test_all_data_in_one_period=1 \
--use_gpu=1 \
--trainer_count=1 \
--num_passes=300 \
--save_dir=$output \
2>&1 | tee $log
paddle usage -l $log -e $? -n "image_classification_train" >/dev/null 2>&1
python -m paddle.utils.plotcurve -i $log > plot.png
demo/image_classification/vgg_16_cifar.py 已删除 100755 → 0
浏览文件 @ 338dd135
# 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 paddle.trainer_config_helpers import *
is_predict = get_config_arg("is_predict", bool, False)
####################Data Configuration ##################
if not is_predict:
data_dir = 'data/cifar-out/batches/'
meta_path = data_dir + 'batches.meta'
args = {
'meta': meta_path,
'mean_img_size': 32,
'img_size': 32,
'num_classes': 10,
'use_jpeg': 1,
'color': "color"
}
define_py_data_sources2(
train_list="train.list",
test_list="train.list",
module='image_provider',
obj='processData',
args=args)
######################Algorithm Configuration #############
settings(
batch_size=128,
learning_rate=0.1 / 128.0,
learning_method=MomentumOptimizer(0.9),
regularization=L2Regularization(0.0005 * 128))
#######################Network Configuration #############
data_size = 3 * 32 * 32
label_size = 10
img = data_layer(name='image', size=data_size)
# small_vgg is predefined in trainer_config_helpers.networks
predict = small_vgg(input_image=img, num_channels=3, num_classes=label_size)
if not is_predict:
lbl = data_layer(name="label", size=label_size)
outputs(classification_cost(input=predict, label=lbl))
else:
outputs(predict)
demo/introduction/.gitignore 已删除 100644 → 0
浏览文件 @ 338dd135
dataprovider.pyc
empty.list
train.log
output
train.list
demo/introduction/README.md 已删除 100644 → 0
浏览文件 @ 338dd135
This folder contains scripts used in PaddlePaddle introduction.
- use `bash train.sh` to train a simple linear regression model
- use `python evaluate_model.py` to read model parameters. You can see that `w` and `b` are very close to [2, 0.3].
demo/introduction/dataprovider.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 paddle.trainer.PyDataProvider2 import *
import random
# define data types of input: 2 real numbers
@provider(
input_types={'x': dense_vector(1),
'y': dense_vector(1)}, use_seq=False)
def process(settings, input_file):
for i in xrange(2000):
x = random.random()
yield {'x': [x], 'y': [2 * x + 0.3]}
demo/introduction/train.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
paddle train \
--config=trainer_config.py \
--save_dir=./output \
--num_passes=30 \
2>&1 |tee 'train.log'
paddle usage -l "train.log" -e $? -n "introduction" >/dev/null 2>&1
demo/introduction/trainer_config.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 paddle.trainer_config_helpers import *
# 1. read data. Suppose you saved above python code as dataprovider.py
define_py_data_sources2(
train_list=['no_matter.txt'],
test_list=None,
module='dataprovider',
obj='process',
args={})
# 2. learning algorithm
settings(batch_size=12, learning_rate=1e-3, learning_method=MomentumOptimizer())
# 3. Network configuration
x = data_layer(name='x', size=1)
y = data_layer(name='y', size=1)
y_predict = fc_layer(
input=x,
param_attr=ParamAttr(name='w'),
size=1,
act=LinearActivation(),
bias_attr=ParamAttr(name='b'))
cost = mse_cost(input=y_predict, label=y)
outputs(cost)
demo/recommendation/.gitignore 已删除 100644 → 0
浏览文件 @ 338dd135
log.txt
data/meta.bin
data/ml-1m
data/ratings.dat.train
data/ratings.dat.test
data/train.list
data/test.list
dataprovider_copy_1.py
*.pyc
output
demo/recommendation/api_train_v2.py 已删除 100644 → 0
浏览文件 @ 338dd135
import paddle.v2 as paddle
import cPickle
import copy
def main():
paddle.init(use_gpu=False)
movie_title_dict = paddle.dataset.movielens.get_movie_title_dict()
uid = paddle.layer.data(
name='user_id',
type=paddle.data_type.integer_value(
paddle.dataset.movielens.max_user_id() + 1))
usr_emb = paddle.layer.embedding(input=uid, size=32)
usr_gender_id = paddle.layer.data(
name='gender_id', type=paddle.data_type.integer_value(2))
usr_gender_emb = paddle.layer.embedding(input=usr_gender_id, size=16)
usr_age_id = paddle.layer.data(
name='age_id',
type=paddle.data_type.integer_value(
len(paddle.dataset.movielens.age_table)))
usr_age_emb = paddle.layer.embedding(input=usr_age_id, size=16)
usr_job_id = paddle.layer.data(
name='job_id',
type=paddle.data_type.integer_value(paddle.dataset.movielens.max_job_id(
) + 1))
usr_job_emb = paddle.layer.embedding(input=usr_job_id, size=16)
usr_combined_features = paddle.layer.fc(
input=[usr_emb, usr_gender_emb, usr_age_emb, usr_job_emb],
size=200,
act=paddle.activation.Tanh())
mov_id = paddle.layer.data(
name='movie_id',
type=paddle.data_type.integer_value(
paddle.dataset.movielens.max_movie_id() + 1))
mov_emb = paddle.layer.embedding(input=mov_id, size=32)
mov_categories = paddle.layer.data(
name='category_id',
type=paddle.data_type.sparse_binary_vector(
len(paddle.dataset.movielens.movie_categories())))
mov_categories_hidden = paddle.layer.fc(input=mov_categories, size=32)
mov_title_id = paddle.layer.data(
name='movie_title',
type=paddle.data_type.integer_value_sequence(len(movie_title_dict)))
mov_title_emb = paddle.layer.embedding(input=mov_title_id, size=32)
mov_title_conv = paddle.networks.sequence_conv_pool(
input=mov_title_emb, hidden_size=32, context_len=3)
mov_combined_features = paddle.layer.fc(
input=[mov_emb, mov_categories_hidden, mov_title_conv],
size=200,
act=paddle.activation.Tanh())
inference = paddle.layer.cos_sim(
a=usr_combined_features, b=mov_combined_features, size=1, scale=5)
cost = paddle.layer.mse_cost(
input=inference,
label=paddle.layer.data(
name='score', type=paddle.data_type.dense_vector(1)))
parameters = paddle.parameters.create(cost)
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=paddle.optimizer.Adam(
learning_rate=1e-4))
feeding = {
'user_id': 0,
'gender_id': 1,
'age_id': 2,
'job_id': 3,
'movie_id': 4,
'category_id': 5,
'movie_title': 6,
'score': 7
}
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
print "Pass %d Batch %d Cost %.2f" % (
event.pass_id, event.batch_id, event.cost)
trainer.train(
reader=paddle.batch(
paddle.reader.shuffle(
paddle.dataset.movielens.train(), buf_size=8192),
batch_size=256),
event_handler=event_handler,
feeding=feeding,
num_passes=1)
user_id = 234
movie_id = 345
user = paddle.dataset.movielens.user_info()[user_id]
movie = paddle.dataset.movielens.movie_info()[movie_id]
feature = user.value() + movie.value()
def reader():
yield feature
infer_dict = copy.copy(feeding)
del infer_dict['score']
prediction = paddle.infer(
output=inference,
parameters=parameters,
reader=paddle.batch(
reader, batch_size=32),
feeding=infer_dict)
print(prediction + 5) / 2
if __name__ == '__main__':
main()
demo/recommendation/common_utils.py 已删除 100755 → 0
浏览文件 @ 338dd135
# 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 paddle.trainer.PyDataProvider2 import *
def meta_to_header(meta, name):
metas = meta[name]['__meta__']['raw_meta']
for each_meta in metas:
slot_name = each_meta.get('name', '%s_id' % name)
if each_meta['type'] == 'id':
yield slot_name, integer_value(each_meta['max'])
elif each_meta['type'] == 'embedding':
is_seq = each_meta['seq'] == 'sequence'
yield slot_name, integer_value(
len(each_meta['dict']),
seq_type=SequenceType.SEQUENCE
if is_seq else SequenceType.NO_SEQUENCE)
elif each_meta['type'] == 'one_hot_dense':
yield slot_name, dense_vector(len(each_meta['dict']))
demo/recommendation/data/config.json 已删除 100644 → 0
浏览文件 @ 338dd135
{
"user": {
"file": {
"name": "users.dat",
"delimiter": "::"
},
"fields": ["id", "gender", "age", "occupation"]
},
"movie": {
"file": {
"name": "movies.dat",
"delimiter": "::"
},
"fields": ["id", "title", "genres"]
}
}
demo/recommendation/data/config_generator.py 已删除 100644 → 0
浏览文件 @ 338dd135
#!/bin/env python2
# 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.
"""
config_generator.py
Usage:
./config_generator.py <config_file> [--output_format=<output_format>]
./config_generator.py -h | --help
Options:
-h --help Show this screen.
--output_format=<output_format> Output Config format(json or yaml) [default: json].
"""
import json
import docopt
import copy
DEFAULT_FILE = {"type": "split", "delimiter": ","}
DEFAULT_FIELD = {
"id": {
"type": "id"
},
"gender": {
"name": "gender",
"type": "embedding",
"dict": {
"type": "char_based"
}
},
"age": {
"name": "age",
"type": "embedding",
"dict": {
"type": "whole_content",
"sort": True
}
},
"occupation": {
"name": "occupation",
"type": "embedding",
"dict": {
"type": "whole_content",
"sort": "true"
}
},
"title": {
"regex": {
"pattern": r"^(.*)\((\d+)\)$",
"group_id": 1,
"strip": True
},
"name": "title",
"type": {
"name": "embedding",
"seq_type": "sequence",
},
"dict": {
"type": "char_based"
}
},
"genres": {
"type": "one_hot_dense",
"dict": {
"type": "split",
"delimiter": "|"
},
"name": "genres"
}
}
def merge_dict(master_dict, slave_dict):
return dict(((k, master_dict.get(k) or slave_dict.get(k))
for k in set(slave_dict) | set(master_dict)))
def main(filename, fmt):
with open(filename, 'r') as f:
conf = json.load(f)
obj = dict()
for k in conf:
val = conf[k]
file_dict = val['file']
file_dict = merge_dict(file_dict, DEFAULT_FILE)
fields = []
for pos, field_key in enumerate(val['fields']):
assert isinstance(field_key, basestring)
field = copy.deepcopy(DEFAULT_FIELD[field_key])
field['pos'] = pos
fields.append(field)
obj[k] = {"file": file_dict, "fields": fields}
meta = {"meta": obj}
# print meta
if fmt == 'json':
def formatter(x):
import json
return json.dumps(x, indent=2)
elif fmt == 'yaml':
def formatter(x):
import yaml
return yaml.safe_dump(x, default_flow_style=False)
else:
raise NotImplementedError("Dump format %s is not implemented" % fmt)
print formatter(meta)
if __name__ == '__main__':
args = docopt.docopt(__doc__, version="0.1.0")
main(args["<config_file>"], args["--output_format"])
demo/recommendation/data/meta_config.json 已删除 100644 → 0
浏览文件 @ 338dd135
{
"meta": {
"movie": {
"fields": [
{
"type": "id",
"pos": 0
},
{
"regex": {
"pattern": "^(.*)\\((\\d+)\\)$",
"group_id": 1,
"strip": true
},
"type": {
"seq_type": "sequence",
"name": "embedding"
},
"dict": {
"type": "char_based"
},
"name": "title",
"pos": 1
},
{
"type": "one_hot_dense",
"dict": {
"delimiter": "|",
"type": "split"
},
"name": "genres",
"pos": 2
}
],
"file": {
"delimiter": "::",
"type": "split",
"name": "movies.dat"
}
},
"user": {
"fields": [
{
"type": "id",
"pos": 0
},
{
"type": "embedding",
"dict": {
"type": "char_based"
},
"name": "gender",
"pos": 1
},
{
"type": "embedding",
"dict": {
"sort": true,
"type": "whole_content"
},
"name": "age",
"pos": 2
},
{
"type": "embedding",
"dict": {
"sort": "true",
"type": "whole_content"
},
"name": "occupation",
"pos": 3
}
],
"file": {
"delimiter": "::",
"type": "split",
"name": "users.dat"
}
}
}
}
demo/recommendation/data/meta_generator.py 已删除 100644 → 0
浏览文件 @ 338dd135
#!/bin/env python2
# 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.
"""
Preprocess Movielens dataset, to get movie/user object.
Usage:
./preprocess.py <dataset_dir> <binary_filename> [--config=<config_file>]
./preprocess.py -h | --help
Options:
-h --help Show this screen.
--version Show version.
--config=<config_file> Get MetaData config file [default: config.json].
"""
import docopt
import os
import sys
import re
import collections
try:
import cPickle as pickle
except ImportError:
import pickle
class UniqueIDGenerator(object):
def __init__(self):
self.pool = collections.defaultdict(self.__next_id__)
self.next_id = 0
def __next_id__(self):
tmp = self.next_id
self.next_id += 1
return tmp
def __call__(self, k):
return self.pool[k]
def to_list(self):
ret_val = [None] * len(self.pool)
for k in self.pool.keys():
ret_val[self.pool[k]] = k
return ret_val
class SortedIDGenerator(object):
def __init__(self):
self.__key_set__ = set()
self.dict = None
def scan(self, key):
self.__key_set__.add(key)
def finish_scan(self, compare=None, key=None, reverse=False):
self.__key_set__ = sorted(
list(self.__key_set__), cmp=compare, key=key, reverse=reverse)
self.dict = dict()
for idx, each_key in enumerate(self.__key_set__):
self.dict[each_key] = idx
def __call__(self, key):
return self.dict[key]
def to_list(self):
return self.__key_set__
class SplitFileReader(object):
def __init__(self, work_dir, config):
assert isinstance(config, dict)
self.filename = config['name']
self.delimiter = config.get('delimiter', ',')
self.work_dir = work_dir
def read(self):
with open(os.path.join(self.work_dir, self.filename), 'r') as f:
for line in f:
line = line.strip()
if isinstance(self.delimiter, unicode):
self.delimiter = str(self.delimiter)
yield line.split(self.delimiter)
@staticmethod
def create(work_dir, config):
assert isinstance(config, dict)
if config['type'] == 'split':
return SplitFileReader(work_dir, config)
class IFileReader(object):
READERS = [SplitFileReader]
def read(self):
raise NotImplementedError()
@staticmethod
def create(work_dir, config):
for reader_cls in IFileReader.READERS:
val = reader_cls.create(work_dir, config)
if val is not None:
return val
class IDFieldParser(object):
TYPE = 'id'
def __init__(self, config):
self.__max_id__ = -sys.maxint - 1
self.__min_id__ = sys.maxint
self.__id_count__ = 0
def scan(self, line):
idx = int(line)
self.__max_id__ = max(self.__max_id__, idx)
self.__min_id__ = min(self.__min_id__, idx)
self.__id_count__ += 1
def parse(self, line):
return int(line)
def meta_field(self):
return {
"is_key": True,
'max': self.__max_id__,
'min': self.__min_id__,
'count': self.__id_count__,
'type': 'id'
}
class SplitEmbeddingDict(object):
def __init__(self, delimiter):
self.__id__ = UniqueIDGenerator()
self.delimiter = delimiter
def scan(self, multi):
for val in multi.split(self.delimiter):
self.__id__(val)
def parse(self, multi):
return map(self.__id__, multi.split(self.delimiter))
def meta_field(self):
return self.__id__.to_list()
class EmbeddingFieldParser(object):
TYPE = 'embedding'
NO_SEQUENCE = "no_sequence"
SEQUENCE = "sequence"
class CharBasedEmbeddingDict(object):
def __init__(self, is_seq=True):
self.__id__ = UniqueIDGenerator()
self.is_seq = is_seq
def scan(self, s):
for ch in s:
self.__id__(ch)
def parse(self, s):
return map(self.__id__, s) if self.is_seq else self.__id__(s[0])
def meta_field(self):
return self.__id__.to_list()
class WholeContentDict(object):
def __init__(self, need_sort=True):
assert need_sort
self.__id__ = SortedIDGenerator()
self.__has_finished__ = False
def scan(self, txt):
self.__id__.scan(txt)
def meta_field(self):
if not self.__has_finished__:
self.__id__.finish_scan()
self.__has_finished__ = True
return self.__id__.to_list()
def parse(self, txt):
return self.__id__(txt)
def __init__(self, config):
try:
self.seq_type = config['type']['seq_type']
except TypeError:
self.seq_type = EmbeddingFieldParser.NO_SEQUENCE
if config['dict']['type'] == 'char_based':
self.dict = EmbeddingFieldParser.CharBasedEmbeddingDict(
self.seq_type == EmbeddingFieldParser.SEQUENCE)
elif config['dict']['type'] == 'split':
self.dict = SplitEmbeddingDict(config['dict'].get('delimiter', ','))
elif config['dict']['type'] == 'whole_content':
self.dict = EmbeddingFieldParser.WholeContentDict(config['dict'][
'sort'])
else:
print config
assert False
self.name = config['name']
def scan(self, s):
self.dict.scan(s)
def meta_field(self):
return {
'name': self.name,
'dict': self.dict.meta_field(),
'type': 'embedding',
'seq': self.seq_type
}
def parse(self, s):
return self.dict.parse(s)
class OneHotDenseFieldParser(object):
TYPE = 'one_hot_dense'
def __init__(self, config):
if config['dict']['type'] == 'split':
self.dict = SplitEmbeddingDict(config['dict']['delimiter'])
self.name = config['name']
def scan(self, s):
self.dict.scan(s)
def meta_field(self):
# print self.dict.meta_field()
return {
'dict': self.dict.meta_field(),
'name': self.name,
'type': 'one_hot_dense'
}
def parse(self, s):
ids = self.dict.parse(s)
retv = [0.0] * len(self.dict.meta_field())
for idx in ids:
retv[idx] = 1.0
# print retv
return retv
class FieldParserFactory(object):
PARSERS = [IDFieldParser, EmbeddingFieldParser, OneHotDenseFieldParser]
@staticmethod
def create(config):
if isinstance(config['type'], basestring):
config_type = config['type']
elif isinstance(config['type'], dict):
config_type = config['type']['name']
assert config_type is not None
for each_parser_cls in FieldParserFactory.PARSERS:
if config_type == each_parser_cls.TYPE:
return each_parser_cls(config)
print config
class CompositeFieldParser(object):
def __init__(self, parser, extractor):
self.extractor = extractor
self.parser = parser
def scan(self, *args, **kwargs):
self.parser.scan(self.extractor.extract(*args, **kwargs))
def parse(self, *args, **kwargs):
return self.parser.parse(self.extractor.extract(*args, **kwargs))
def meta_field(self):
return self.parser.meta_field()
class PositionContentExtractor(object):
def __init__(self, pos):
self.pos = pos
def extract(self, line):
assert isinstance(line, list)
return line[self.pos]
class RegexPositionContentExtractor(PositionContentExtractor):
def __init__(self, pos, pattern, group_id, strip=True):
PositionContentExtractor.__init__(self, pos)
pattern = pattern.strip()
self.pattern = re.compile(pattern)
self.group_id = group_id
self.strip = strip
def extract(self, line):
line = PositionContentExtractor.extract(self, line)
match = self.pattern.match(line)
# print line, self.pattern.pattern, match
assert match is not None
txt = match.group(self.group_id)
if self.strip:
txt.strip()
return txt
class ContentExtractorFactory(object):
def extract(self, line):
pass
@staticmethod
def create(config):
if 'pos' in config:
if 'regex' not in config:
return PositionContentExtractor(config['pos'])
else:
extra_args = config['regex']
return RegexPositionContentExtractor(
pos=config['pos'], **extra_args)
class MetaFile(object):
def __init__(self, work_dir):
self.work_dir = work_dir
self.obj = dict()
def parse(self, config):
config = config['meta']
ret_obj = dict()
for key in config.keys():
val = config[key]
assert 'file' in val
reader = IFileReader.create(self.work_dir, val['file'])
assert reader is not None
assert 'fields' in val and isinstance(val['fields'], list)
fields_config = val['fields']
field_parsers = map(MetaFile.__field_config_mapper__, fields_config)
for each_parser in field_parsers:
assert each_parser is not None
for each_block in reader.read():
for each_parser in field_parsers:
each_parser.scan(each_block)
metas = map(lambda x: x.meta_field(), field_parsers)
# print metas
key_index = filter(
lambda x: x is not None,
map(lambda (idx, meta): idx if 'is_key' in meta and meta['is_key'] else None,
enumerate(metas)))[0]
key_map = []
for i in range(min(key_index, len(metas))):
key_map.append(i)
for i in range(key_index + 1, len(metas)):
key_map.append(i)
obj = {'__meta__': {'raw_meta': metas, 'feature_map': key_map}}
for each_block in reader.read():
idx = field_parsers[key_index].parse(each_block)
val = []
for i, each_parser in enumerate(field_parsers):
if i != key_index:
val.append(each_parser.parse(each_block))
obj[idx] = val
ret_obj[key] = obj
self.obj = ret_obj
return ret_obj
@staticmethod
def __field_config_mapper__(conf):
assert isinstance(conf, dict)
extrator = ContentExtractorFactory.create(conf)
field_parser = FieldParserFactory.create(conf)
assert extrator is not None
assert field_parser is not None
return CompositeFieldParser(field_parser, extrator)
def dump(self, fp):
pickle.dump(self.obj, fp, pickle.HIGHEST_PROTOCOL)
def preprocess(binary_filename, dataset_dir, config, **kwargs):
assert isinstance(config, str)
with open(config, 'r') as config_file:
file_loader = None
if config.lower().endswith('.yaml'):
import yaml
file_loader = yaml
elif config.lower().endswith('.json'):
import json
file_loader = json
config = file_loader.load(config_file)
meta = MetaFile(dataset_dir)
meta.parse(config)
with open(binary_filename, 'wb') as outf:
meta.dump(outf)
if __name__ == '__main__':
args = docopt.docopt(__doc__, version='0.1.0')
kwargs = dict()
for key in args.keys():
if key != '--help':
param_name = key
assert isinstance(param_name, str)
param_name = param_name.replace('<', '')
param_name = param_name.replace('>', '')
param_name = param_name.replace('--', '')
kwargs[param_name] = args[key]
preprocess(**kwargs)
demo/recommendation/data/ml_data.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -ex
cd "$(dirname "$0")"
# download the dataset
wget http://files.grouplens.org/datasets/movielens/ml-1m.zip
# unzip the dataset
unzip ml-1m.zip
# remove the unused zip file
rm ml-1m.zip
demo/recommendation/data/split.py 已删除 100644 → 0
浏览文件 @ 338dd135
#!/bin/env python2
# 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.
"""
Separate movielens 1m dataset to train/test file.
Usage:
./separate.py <input_file> [--test_ratio=<test_ratio>] [--delimiter=<delimiter>]
./separate.py -h | --help
Options:
-h --help Show this screen.
--version Show version.
--test_ratio=<test_ratio> Test ratio for separate [default: 0.1].
--delimiter=<delimiter> File delimiter [default: ,].
"""
import docopt
import collections
import random
def process(test_ratio, input_file, delimiter, **kwargs):
test_ratio = float(test_ratio)
rating_dict = collections.defaultdict(list)
with open(input_file, 'r') as f:
for line in f:
user_id = int(line.split(delimiter)[0])
rating_dict[user_id].append(line.strip())
with open(input_file + ".train", 'w') as train_file:
with open(input_file + ".test", 'w') as test_file:
for k in rating_dict.keys():
lines = rating_dict[k]
assert isinstance(lines, list)
random.shuffle(lines)
test_len = int(len(lines) * test_ratio)
for line in lines[:test_len]:
print >> test_file, line
for line in lines[test_len:]:
print >> train_file, line
if __name__ == '__main__':
args = docopt.docopt(__doc__, version='0.1.0')
kwargs = dict()
for key in args.keys():
if key != '--help':
param_name = key
assert isinstance(param_name, str)
param_name = param_name.replace('<', '')
param_name = param_name.replace('>', '')
param_name = param_name.replace('--', '')
kwargs[param_name] = args[key]
process(**kwargs)
demo/recommendation/dataprovider.py 已删除 100755 → 0
浏览文件 @ 338dd135
# 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 paddle.trainer.PyDataProvider2 import *
import common_utils # parse
def __list_to_map__(lst):
ret_val = dict()
for each in lst:
k, v = each
ret_val[k] = v
return ret_val
def hook(settings, meta, **kwargs):
"""
Init hook is invoked before process data. It will set obj.slots and store
data meta.
:param obj: global object. It will passed to process routine.
:type obj: object
:param meta: the meta file object, which passed from trainer_config. Meta
file record movie/user features.
:param kwargs: unused other arguments.
"""
del kwargs # unused kwargs
# Header define slots that used for paddle.
# first part is movie features.
# second part is user features.
# final part is rating score.
# header is a list of [USE_SEQ_OR_NOT?, SlotType]
movie_headers = list(common_utils.meta_to_header(meta, 'movie'))
settings.movie_names = [h[0] for h in movie_headers]
headers = movie_headers
user_headers = list(common_utils.meta_to_header(meta, 'user'))
settings.user_names = [h[0] for h in user_headers]
headers.extend(user_headers)
headers.append(("rating", dense_vector(1))) # Score
# slot types.
settings.input_types = __list_to_map__(headers)
settings.meta = meta
@provider(init_hook=hook, cache=CacheType.CACHE_PASS_IN_MEM)
def process(settings, filename):
with open(filename, 'r') as f:
for line in f:
# Get a rating from file.
user_id, movie_id, score = map(int, line.split('::')[:-1])
# Scale score to [-5, +5]
score = float(score) * 2 - 5.0
# Get movie/user features by movie_id, user_id
movie_meta = settings.meta['movie'][movie_id]
user_meta = settings.meta['user'][user_id]
outputs = [('movie_id', movie_id - 1)]
# Then add movie features
for i, each_meta in enumerate(movie_meta):
outputs.append((settings.movie_names[i + 1], each_meta))
# Then add user id.
outputs.append(('user_id', user_id - 1))
# Then add user features.
for i, each_meta in enumerate(user_meta):
outputs.append((settings.user_names[i + 1], each_meta))
# Finally, add score
outputs.append(('rating', [score]))
# Return data to paddle
yield __list_to_map__(outputs)
demo/recommendation/evaluate.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
function get_best_pass() {
cat $1 | grep -Pzo 'Test .*\n.*pass-.*' | sed -r 'N;s/Test.* cost=([0-9]+\.[0-9]+).*\n.*pass-([0-9]+)/\1 \2/g' | sort | head -n 1
}
LOG=`get_best_pass log.txt`
LOG=(${LOG})
echo 'Best pass is '${LOG[1]}, ' error is '${LOG[0]}, 'which means predict get error as '`echo ${LOG[0]} | python -c 'import math; print math.sqrt(float(raw_input()))/2'`
evaluate_pass="output/pass-${LOG[1]}"
echo 'evaluating from pass '$evaluate_pass
demo/recommendation/prediction.py 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/env python2
# 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 common_utils import *
from paddle.trainer.config_parser import parse_config
try:
import cPickle as pickle
except ImportError:
import pickle
import sys
if __name__ == '__main__':
model_path = sys.argv[1]
swig_paddle.initPaddle('--use_gpu=0')
conf = parse_config("trainer_config.py", "is_predict=1")
network = swig_paddle.GradientMachine.createFromConfigProto(
conf.model_config)
assert isinstance(network, swig_paddle.GradientMachine)
network.loadParameters(model_path)
with open('./data/meta.bin', 'rb') as f:
meta = pickle.load(f)
headers = [h[1] for h in meta_to_header(meta, 'movie')]
headers.extend([h[1] for h in meta_to_header(meta, 'user')])
cvt = DataProviderConverter(headers)
while True:
movie_id = int(raw_input("Input movie_id: "))
user_id = int(raw_input("Input user_id: "))
movie_meta = meta['movie'][movie_id] # Query Data From Meta.
user_meta = meta['user'][user_id]
data = [movie_id - 1]
data.extend(movie_meta)
data.append(user_id - 1)
data.extend(user_meta)
print "Prediction Score is %.2f" % (
(network.forwardTest(cvt.convert([data]))[0]['value'][0][0] + 5)
/ 2)
demo/recommendation/preprocess.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
UNAME_STR=`uname`
if [[ ${UNAME_STR} == 'Linux' ]]; then
SHUF_PROG='shuf'
else
SHUF_PROG='gshuf'
fi
cd "$(dirname "$0")"
delimiter='::'
dir=ml-1m
cd data
echo 'generate meta config file'
python config_generator.py config.json > meta_config.json
echo 'generate meta file'
python meta_generator.py $dir meta.bin --config=meta_config.json
echo 'split train/test file'
python split.py $dir/ratings.dat --delimiter=${delimiter} --test_ratio=0.1
echo 'shuffle train file'
${SHUF_PROG} $dir/ratings.dat.train > ratings.dat.train
cp $dir/ratings.dat.test .
echo "./data/ratings.dat.train" > train.list
echo "./data/ratings.dat.test" > test.list
demo/recommendation/run.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
paddle train \
--config=trainer_config.py \
--save_dir=./output \
--use_gpu=false \
--trainer_count=4\
--test_all_data_in_one_period=true \
--log_period=100 \
--dot_period=1 \
--num_passes=50 2>&1 | tee 'log.txt'
paddle usage -l log.txt -e $? -n "recommendation" >/dev/null 2>&1
demo/recommendation/trainer_config.py 已删除 100755 → 0
浏览文件 @ 338dd135
# 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 paddle.trainer_config_helpers import *
try:
import cPickle as pickle
except ImportError:
import pickle
is_predict = get_config_arg('is_predict', bool, False)
META_FILE = 'data/meta.bin'
with open(META_FILE, 'rb') as f:
# load meta file
meta = pickle.load(f)
settings(
batch_size=1600, learning_rate=1e-3, learning_method=RMSPropOptimizer())
def construct_feature(name):
"""
Construct movie/user features.
This method read from meta data. Then convert feature to neural network due
to feature type. The map relation as follow.
* id: embedding => fc
* embedding:
is_sequence: embedding => context_projection => fc => pool
not sequence: embedding => fc
* one_hot_dense: fc => fc
Then gather all features vector, and use a fc layer to combined them as
return.
:param name: 'movie' or 'user'
:type name: basestring
:return: combined feature output
:rtype: LayerOutput
"""
__meta__ = meta[name]['__meta__']['raw_meta']
fusion = []
for each_meta in __meta__:
type_name = each_meta['type']
slot_name = each_meta.get('name', '%s_id' % name)
if type_name == 'id':
slot_dim = each_meta['max']
embedding = embedding_layer(
input=data_layer(
slot_name, size=slot_dim), size=256)
fusion.append(fc_layer(input=embedding, size=256))
elif type_name == 'embedding':
is_seq = each_meta['seq'] == 'sequence'
slot_dim = len(each_meta['dict'])
din = data_layer(slot_name, slot_dim)
embedding = embedding_layer(input=din, size=256)
if is_seq:
fusion.append(
text_conv_pool(
input=embedding, context_len=5, hidden_size=256))
else:
fusion.append(fc_layer(input=embedding, size=256))
elif type_name == 'one_hot_dense':
slot_dim = len(each_meta['dict'])
hidden = fc_layer(input=data_layer(slot_name, slot_dim), size=256)
fusion.append(fc_layer(input=hidden, size=256))
return fc_layer(name="%s_fusion" % name, input=fusion, size=256)
movie_feature = construct_feature("movie")
user_feature = construct_feature("user")
similarity = cos_sim(a=movie_feature, b=user_feature)
if not is_predict:
outputs(mse_cost(input=similarity, label=data_layer('rating', size=1)))
define_py_data_sources2(
'data/train.list',
'data/test.list',
module='dataprovider',
obj='process',
args={'meta': meta})
else:
outputs(similarity)
demo/semantic_role_labeling/.gitignore 已删除 100644 → 0
浏览文件 @ 338dd135
*.pyc
train.log
data/feature
data/conll05st-release/
data/src.dict
data/test.wsj.props
data/test.wsj.seq_pair
data/test.wsj.words
data/tgt.dict
output
data/emb
data/targetDict.txt
data/verbDict.txt
data/wordDict.txt
demo/semantic_role_labeling/api_train_v2.py 已删除 100644 → 0
浏览文件 @ 338dd135
import math
import numpy as np
import gzip
import logging
import paddle.v2.dataset.conll05 as conll05
import paddle.v2.evaluator as evaluator
import paddle.v2 as paddle
logger = logging.getLogger('paddle')
word_dict, verb_dict, label_dict = conll05.get_dict()
word_dict_len = len(word_dict)
label_dict_len = len(label_dict)
pred_len = len(verb_dict)
mark_dict_len = 2
word_dim = 32
mark_dim = 5
hidden_dim = 512
depth = 8
default_std = 1 / math.sqrt(hidden_dim) / 3.0
mix_hidden_lr = 1e-3
def d_type(size):
return paddle.data_type.integer_value_sequence(size)
def db_lstm():
#8 features
word = paddle.layer.data(name='word_data', type=d_type(word_dict_len))
predicate = paddle.layer.data(name='verb_data', type=d_type(pred_len))
ctx_n2 = paddle.layer.data(name='ctx_n2_data', type=d_type(word_dict_len))
ctx_n1 = paddle.layer.data(name='ctx_n1_data', type=d_type(word_dict_len))
ctx_0 = paddle.layer.data(name='ctx_0_data', type=d_type(word_dict_len))
ctx_p1 = paddle.layer.data(name='ctx_p1_data', type=d_type(word_dict_len))
ctx_p2 = paddle.layer.data(name='ctx_p2_data', type=d_type(word_dict_len))
mark = paddle.layer.data(name='mark_data', type=d_type(mark_dict_len))
emb_para = paddle.attr.Param(name='emb', initial_std=0., is_static=True)
std_0 = paddle.attr.Param(initial_std=0.)
std_default = paddle.attr.Param(initial_std=default_std)
predicate_embedding = paddle.layer.embedding(
size=word_dim,
input=predicate,
param_attr=paddle.attr.Param(
name='vemb', initial_std=default_std))
mark_embedding = paddle.layer.embedding(
size=mark_dim, input=mark, param_attr=std_0)
word_input = [word, ctx_n2, ctx_n1, ctx_0, ctx_p1, ctx_p2]
emb_layers = [
paddle.layer.embedding(
size=word_dim, input=x, param_attr=emb_para) for x in word_input
]
emb_layers.append(predicate_embedding)
emb_layers.append(mark_embedding)
hidden_0 = paddle.layer.mixed(
size=hidden_dim,
bias_attr=std_default,
input=[
paddle.layer.full_matrix_projection(
input=emb, param_attr=std_default) for emb in emb_layers
])
lstm_para_attr = paddle.attr.Param(initial_std=0.0, learning_rate=1.0)
hidden_para_attr = paddle.attr.Param(
initial_std=default_std, learning_rate=mix_hidden_lr)
lstm_0 = paddle.layer.lstmemory(
input=hidden_0,
act=paddle.activation.Relu(),
gate_act=paddle.activation.Sigmoid(),
state_act=paddle.activation.Sigmoid(),
bias_attr=std_0,
param_attr=lstm_para_attr)
#stack L-LSTM and R-LSTM with direct edges
input_tmp = [hidden_0, lstm_0]
for i in range(1, depth):
mix_hidden = paddle.layer.mixed(
size=hidden_dim,
bias_attr=std_default,
input=[
paddle.layer.full_matrix_projection(
input=input_tmp[0], param_attr=hidden_para_attr),
paddle.layer.full_matrix_projection(
input=input_tmp[1], param_attr=lstm_para_attr)
])
lstm = paddle.layer.lstmemory(
input=mix_hidden,
act=paddle.activation.Relu(),
gate_act=paddle.activation.Sigmoid(),
state_act=paddle.activation.Sigmoid(),
reverse=((i % 2) == 1),
bias_attr=std_0,
param_attr=lstm_para_attr)
input_tmp = [mix_hidden, lstm]
feature_out = paddle.layer.mixed(
size=label_dict_len,
bias_attr=std_default,
input=[
paddle.layer.full_matrix_projection(
input=input_tmp[0], param_attr=hidden_para_attr),
paddle.layer.full_matrix_projection(
input=input_tmp[1], param_attr=lstm_para_attr)
], )
return feature_out
def load_parameter(file_name, h, w):
with open(file_name, 'rb') as f:
f.read(16) # skip header.
return np.fromfile(f, dtype=np.float32).reshape(h, w)
def train():
paddle.init(use_gpu=False, trainer_count=1)
# define network topology
feature_out = db_lstm()
target = paddle.layer.data(name='target', type=d_type(label_dict_len))
crf_cost = paddle.layer.crf(size=label_dict_len,
input=feature_out,
label=target,
param_attr=paddle.attr.Param(
name='crfw',
initial_std=default_std,
learning_rate=mix_hidden_lr))
crf_dec = paddle.layer.crf_decoding(
size=label_dict_len,
input=feature_out,
label=target,
param_attr=paddle.attr.Param(name='crfw'))
evaluator.sum(input=crf_dec)
# create parameters
parameters = paddle.parameters.create(crf_cost)
parameters.set('emb', load_parameter(conll05.get_embedding(), 44068, 32))
# create optimizer
optimizer = paddle.optimizer.Momentum(
momentum=0,
learning_rate=2e-2,
regularization=paddle.optimizer.L2Regularization(rate=8e-4),
model_average=paddle.optimizer.ModelAverage(
average_window=0.5, max_average_window=10000), )
trainer = paddle.trainer.SGD(cost=crf_cost,
parameters=parameters,
update_equation=optimizer,
extra_layers=crf_dec)
reader = paddle.batch(
paddle.reader.shuffle(
conll05.test(), buf_size=8192), batch_size=10)
feeding = {
'word_data': 0,
'ctx_n2_data': 1,
'ctx_n1_data': 2,
'ctx_0_data': 3,
'ctx_p1_data': 4,
'ctx_p2_data': 5,
'verb_data': 6,
'mark_data': 7,
'target': 8
}
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
logger.info("Pass %d, Batch %d, Cost %f, %s" % (
event.pass_id, event.batch_id, event.cost, event.metrics))
if event.batch_id and event.batch_id % 1000 == 0:
result = trainer.test(reader=reader, feeding=feeding)
logger.info("\nTest with Pass %d, Batch %d, %s" %
(event.pass_id, event.batch_id, result.metrics))
if isinstance(event, paddle.event.EndPass):
# save parameters
with gzip.open('params_pass_%d.tar.gz' % event.pass_id, 'w') as f:
parameters.to_tar(f)
result = trainer.test(reader=reader, feeding=feeding)
logger.info("\nTest with Pass %d, %s" %
(event.pass_id, result.metrics))
trainer.train(
reader=reader,
event_handler=event_handler,
num_passes=10,
feeding=feeding)
def infer_a_batch(inferer, test_data, word_dict, pred_dict, label_dict):
probs = inferer.infer(input=test_data, field='id')
assert len(probs) == sum(len(x[0]) for x in test_data)
for idx, test_sample in enumerate(test_data):
start_id = 0
pred_str = "%s\t" % (pred_dict[test_sample[6][0]])
for w, tag in zip(test_sample[0],
probs[start_id:start_id + len(test_sample[0])]):
pred_str += "%s[%s] " % (word_dict[w], label_dict[tag])
print(pred_str.strip())
start_id += len(test_sample[0])
def infer():
label_dict_reverse = dict((value, key)
for key, value in label_dict.iteritems())
word_dict_reverse = dict((value, key)
for key, value in word_dict.iteritems())
pred_dict_reverse = dict((value, key)
for key, value in verb_dict.iteritems())
test_creator = paddle.dataset.conll05.test()
paddle.init(use_gpu=False, trainer_count=1)
# define network topology
feature_out = db_lstm()
predict = paddle.layer.crf_decoding(
size=label_dict_len,
input=feature_out,
param_attr=paddle.attr.Param(name='crfw'))
test_pass = 0
with gzip.open('params_pass_%d.tar.gz' % (test_pass)) as f:
parameters = paddle.parameters.Parameters.from_tar(f)
inferer = paddle.inference.Inference(
output_layer=predict, parameters=parameters)
# prepare test data
test_data = []
test_batch_size = 50
for idx, item in enumerate(test_creator()):
test_data.append(item[0:8])
if idx and (not idx % test_batch_size):
infer_a_batch(
inferer,
test_data,
word_dict_reverse,
pred_dict_reverse,
label_dict_reverse, )
test_data = []
infer_a_batch(
inferer,
test_data,
word_dict_reverse,
pred_dict_reverse,
label_dict_reverse, )
test_data = []
def main(is_inferring=False):
if is_inferring:
infer()
else:
train()
if __name__ == '__main__':
main(is_inferring=False)
demo/semantic_role_labeling/data/extract_dict_feature.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 os
from optparse import OptionParser
def extract_dict_features(pair_file, feature_file):
with open(pair_file) as fin, open(feature_file, 'w') as feature_out:
for line in fin:
sentence, predicate, labels = line.strip().split('\t')
sentence_list = sentence.split()
labels_list = labels.split()
verb_index = labels_list.index('B-V')
mark = [0] * len(labels_list)
if verb_index > 0:
mark[verb_index - 1] = 1
ctx_n1 = sentence_list[verb_index - 1]
else:
ctx_n1 = 'bos'
if verb_index > 1:
mark[verb_index - 2] = 1
ctx_n2 = sentence_list[verb_index - 2]
else:
ctx_n2 = 'bos'
mark[verb_index] = 1
ctx_0 = sentence_list[verb_index]
if verb_index < len(labels_list) - 1:
mark[verb_index + 1] = 1
ctx_p1 = sentence_list[verb_index + 1]
else:
ctx_p1 = 'eos'
if verb_index < len(labels_list) - 2:
mark[verb_index + 2] = 1
ctx_p2 = sentence_list[verb_index + 2]
else:
ctx_p2 = 'eos'
feature_str = sentence + '\t' \
+ predicate + '\t' \
+ ctx_n2 + '\t' \
+ ctx_n1 + '\t' \
+ ctx_0 + '\t' \
+ ctx_p1 + '\t' \
+ ctx_p2 + '\t' \
+ ' '.join([str(i) for i in mark]) + '\t' \
+ labels
feature_out.write(feature_str + '\n')
if __name__ == '__main__':
usage = '-p pair_file -f feature_file'
parser = OptionParser(usage)
parser.add_option('-p', dest='pair_file', help='the pair file')
parser.add_option('-f', dest='feature_file', help='the feature file')
(options, args) = parser.parse_args()
extract_dict_features(options.pair_file, options.feature_file)
demo/semantic_role_labeling/data/extract_pairs.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 os
from optparse import OptionParser
def read_labels(props_file):
'''
a sentence maybe has more than one verb, each verb has its label sequence
label[], is a 3-dimension list.
the first dim is to store all sentence's label seqs, len is the sentence number
the second dim is to store all label sequences for one sentences
the third dim is to store each label for one word
'''
labels = []
with open(props_file) as fin:
label_seqs_for_one_sentences = []
one_seg_in_file = []
for line in fin:
line = line.strip()
if line == '':
for i in xrange(len(one_seg_in_file[0])):
a_kind_lable = [x[i] for x in one_seg_in_file]
label_seqs_for_one_sentences.append(a_kind_lable)
labels.append(label_seqs_for_one_sentences)
one_seg_in_file = []
label_seqs_for_one_sentences = []
else:
part = line.split()
one_seg_in_file.append(part)
return labels
def read_sentences(words_file):
sentences = []
with open(words_file) as fin:
s = ''
for line in fin:
line = line.strip()
if line == '':
sentences.append(s)
s = ''
else:
s += line + ' '
return sentences
def transform_labels(sentences, labels):
sen_lab_pair = []
for i in xrange(len(sentences)):
if len(labels[i]) == 1:
continue
else:
verb_list = []
for x in labels[i][0]:
if x != '-':
verb_list.append(x)
for j in xrange(1, len(labels[i])):
label_list = labels[i][j]
current_tag = 'O'
is_in_bracket = False
label_seq = []
verb_word = ''
for ll in label_list:
if ll == '*' and is_in_bracket == False:
label_seq.append('O')
elif ll == '*' and is_in_bracket == True:
label_seq.append('I-' + current_tag)
elif ll == '*)':
label_seq.append('I-' + current_tag)
is_in_bracket = False
elif ll.find('(') != -1 and ll.find(')') != -1:
current_tag = ll[1:ll.find('*')]
label_seq.append('B-' + current_tag)
is_in_bracket = False
elif ll.find('(') != -1 and ll.find(')') == -1:
current_tag = ll[1:ll.find('*')]
label_seq.append('B-' + current_tag)
is_in_bracket = True
else:
print 'error:', ll
sen_lab_pair.append((sentences[i], verb_list[j - 1], label_seq))
return sen_lab_pair
def write_file(sen_lab_pair, output_file):
with open(output_file, 'w') as fout:
for x in sen_lab_pair:
sentence = x[0]
label_seq = ' '.join(x[2])
assert len(sentence.split()) == len(x[2])
fout.write(sentence + '\t' + x[1] + '\t' + label_seq + '\n')
if __name__ == '__main__':
usage = '-w words_file -p props_file -o output_file'
parser = OptionParser(usage)
parser.add_option('-w', dest='words_file', help='the words file')
parser.add_option('-p', dest='props_file', help='the props file')
parser.add_option('-o', dest='output_file', help='the output_file')
(options, args) = parser.parse_args()
sentences = read_sentences(options.words_file)
labels = read_labels(options.props_file)
sen_lab_pair = transform_labels(sentences, labels)
write_file(sen_lab_pair, options.output_file)
demo/semantic_role_labeling/data/get_data.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
wget http://www.cs.upc.edu/~srlconll/conll05st-tests.tar.gz
wget http://paddlepaddle.bj.bcebos.com/demo/srl_dict_and_embedding/verbDict.txt
wget http://paddlepaddle.bj.bcebos.com/demo/srl_dict_and_embedding/targetDict.txt
wget http://paddlepaddle.bj.bcebos.com/demo/srl_dict_and_embedding/wordDict.txt
wget http://paddlepaddle.bj.bcebos.com/demo/srl_dict_and_embedding/emb
tar -xzvf conll05st-tests.tar.gz
rm conll05st-tests.tar.gz
cp ./conll05st-release/test.wsj/words/test.wsj.words.gz .
cp ./conll05st-release/test.wsj/props/test.wsj.props.gz .
gunzip test.wsj.words.gz
gunzip test.wsj.props.gz
python extract_pairs.py -w test.wsj.words -p test.wsj.props -o test.wsj.seq_pair
python extract_dict_feature.py -p test.wsj.seq_pair -f feature
demo/semantic_role_labeling/dataprovider.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 paddle.trainer.PyDataProvider2 import *
UNK_IDX = 0
def hook(settings, word_dict, label_dict, predicate_dict, **kwargs):
settings.word_dict = word_dict
settings.label_dict = label_dict
settings.predicate_dict = predicate_dict
#all inputs are integral and sequential type
settings.slots = [
integer_value_sequence(len(word_dict)),
integer_value_sequence(len(word_dict)),
integer_value_sequence(len(word_dict)),
integer_value_sequence(len(word_dict)),
integer_value_sequence(len(word_dict)),
integer_value_sequence(len(word_dict)),
integer_value_sequence(len(predicate_dict)), integer_value_sequence(2),
integer_value_sequence(len(label_dict))
]
def get_batch_size(yeild_data):
return len(yeild_data[0])
@provider(
init_hook=hook,
should_shuffle=True,
calc_batch_size=get_batch_size,
can_over_batch_size=True,
cache=CacheType.CACHE_PASS_IN_MEM)
def process(settings, file_name):
with open(file_name, 'r') as fdata:
for line in fdata:
sentence, predicate, ctx_n2, ctx_n1, ctx_0, ctx_p1, ctx_p2, mark, label = \
line.strip().split('\t')
words = sentence.split()
sen_len = len(words)
word_slot = [settings.word_dict.get(w, UNK_IDX) for w in words]
predicate_slot = [settings.predicate_dict.get(predicate)] * sen_len
ctx_n2_slot = [settings.word_dict.get(ctx_n2, UNK_IDX)] * sen_len
ctx_n1_slot = [settings.word_dict.get(ctx_n1, UNK_IDX)] * sen_len
ctx_0_slot = [settings.word_dict.get(ctx_0, UNK_IDX)] * sen_len
ctx_p1_slot = [settings.word_dict.get(ctx_p1, UNK_IDX)] * sen_len
ctx_p2_slot = [settings.word_dict.get(ctx_p2, UNK_IDX)] * sen_len
marks = mark.split()
mark_slot = [int(w) for w in marks]
label_list = label.split()
label_slot = [settings.label_dict.get(w) for w in label_list]
yield word_slot, ctx_n2_slot, ctx_n1_slot, \
ctx_0_slot, ctx_p1_slot, ctx_p2_slot, predicate_slot, mark_slot, label_slot
demo/semantic_role_labeling/db_lstm.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 math
import os
import sys
from paddle.trainer_config_helpers import *
#file paths
word_dict_file = './data/wordDict.txt'
label_dict_file = './data/targetDict.txt'
predicate_file = './data/verbDict.txt'
train_list_file = './data/train.list'
test_list_file = './data/test.list'
is_test = get_config_arg('is_test', bool, False)
is_predict = get_config_arg('is_predict', bool, False)
if not is_predict:
#load dictionaries
word_dict = dict()
label_dict = dict()
predicate_dict = dict()
with open(word_dict_file, 'r') as f_word, \
open(label_dict_file, 'r') as f_label, \
open(predicate_file, 'r') as f_pre:
for i, line in enumerate(f_word):
w = line.strip()
word_dict[w] = i
for i, line in enumerate(f_label):
w = line.strip()
label_dict[w] = i
for i, line in enumerate(f_pre):
w = line.strip()
predicate_dict[w] = i
if is_test:
train_list_file = None
#define data provider
define_py_data_sources2(
train_list=train_list_file,
test_list=test_list_file,
module='dataprovider',
obj='process',
args={
'word_dict': word_dict,
'label_dict': label_dict,
'predicate_dict': predicate_dict
})
word_dict_len = len(word_dict)
label_dict_len = len(label_dict)
pred_len = len(predicate_dict)
else:
word_dict_len = get_config_arg('dict_len', int)
label_dict_len = get_config_arg('label_len', int)
pred_len = get_config_arg('pred_len', int)
############################## Hyper-parameters ##################################
mark_dict_len = 2
word_dim = 32
mark_dim = 5
hidden_dim = 512
depth = 8
########################### Optimizer #######################################
settings(
batch_size=150,
learning_method=MomentumOptimizer(momentum=0),
learning_rate=2e-2,
regularization=L2Regularization(8e-4),
is_async=False,
model_average=ModelAverage(
average_window=0.5, max_average_window=10000), )
####################################### network ##############################
#8 features and 1 target
word = data_layer(name='word_data', size=word_dict_len)
predicate = data_layer(name='verb_data', size=pred_len)
ctx_n2 = data_layer(name='ctx_n2_data', size=word_dict_len)
ctx_n1 = data_layer(name='ctx_n1_data', size=word_dict_len)
ctx_0 = data_layer(name='ctx_0_data', size=word_dict_len)
ctx_p1 = data_layer(name='ctx_p1_data', size=word_dict_len)
ctx_p2 = data_layer(name='ctx_p2_data', size=word_dict_len)
mark = data_layer(name='mark_data', size=mark_dict_len)
if not is_predict:
target = data_layer(name='target', size=label_dict_len)
default_std = 1 / math.sqrt(hidden_dim) / 3.0
emb_para = ParameterAttribute(name='emb', initial_std=0., learning_rate=0.)
std_0 = ParameterAttribute(initial_std=0.)
std_default = ParameterAttribute(initial_std=default_std)
predicate_embedding = embedding_layer(
size=word_dim,
input=predicate,
param_attr=ParameterAttribute(
name='vemb', initial_std=default_std))
mark_embedding = embedding_layer(
name='word_ctx-in_embedding', size=mark_dim, input=mark, param_attr=std_0)
word_input = [word, ctx_n2, ctx_n1, ctx_0, ctx_p1, ctx_p2]
emb_layers = [
embedding_layer(
size=word_dim, input=x, param_attr=emb_para) for x in word_input
]
emb_layers.append(predicate_embedding)
emb_layers.append(mark_embedding)
hidden_0 = mixed_layer(
name='hidden0',
size=hidden_dim,
bias_attr=std_default,
input=[
full_matrix_projection(
input=emb, param_attr=std_default) for emb in emb_layers
])
mix_hidden_lr = 1e-3
lstm_para_attr = ParameterAttribute(initial_std=0.0, learning_rate=1.0)
hidden_para_attr = ParameterAttribute(
initial_std=default_std, learning_rate=mix_hidden_lr)
lstm_0 = lstmemory(
name='lstm0',
input=hidden_0,
act=ReluActivation(),
gate_act=SigmoidActivation(),
state_act=SigmoidActivation(),
bias_attr=std_0,
param_attr=lstm_para_attr)
#stack L-LSTM and R-LSTM with direct edges
input_tmp = [hidden_0, lstm_0]
for i in range(1, depth):
mix_hidden = mixed_layer(
name='hidden' + str(i),
size=hidden_dim,
bias_attr=std_default,
input=[
full_matrix_projection(
input=input_tmp[0], param_attr=hidden_para_attr),
full_matrix_projection(
input=input_tmp[1], param_attr=lstm_para_attr)
])
lstm = lstmemory(
name='lstm' + str(i),
input=mix_hidden,
act=ReluActivation(),
gate_act=SigmoidActivation(),
state_act=SigmoidActivation(),
reverse=((i % 2) == 1),
bias_attr=std_0,
param_attr=lstm_para_attr)
input_tmp = [mix_hidden, lstm]
feature_out = mixed_layer(
name='output',
size=label_dict_len,
bias_attr=std_default,
input=[
full_matrix_projection(
input=input_tmp[0], param_attr=hidden_para_attr),
full_matrix_projection(
input=input_tmp[1], param_attr=lstm_para_attr)
], )
if not is_predict:
crf_l = crf_layer(
name='crf',
size=label_dict_len,
input=feature_out,
label=target,
param_attr=ParameterAttribute(
name='crfw', initial_std=default_std, learning_rate=mix_hidden_lr))
crf_dec_l = crf_decoding_layer(
name='crf_dec_l',
size=label_dict_len,
input=feature_out,
label=target,
param_attr=ParameterAttribute(name='crfw'))
eval = sum_evaluator(input=crf_dec_l)
outputs(crf_l)
else:
crf_dec_l = crf_decoding_layer(
name='crf_dec_l',
size=label_dict_len,
input=feature_out,
param_attr=ParameterAttribute(name='crfw'))
outputs(crf_dec_l)
demo/semantic_role_labeling/predict.py 已删除 100644 → 0
浏览文件 @ 338dd135
# 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 os
import numpy as np
from optparse import OptionParser
from py_paddle import swig_paddle, DataProviderConverter
from paddle.trainer.PyDataProvider2 import integer_value_sequence
from paddle.trainer.config_parser import parse_config
"""
Usage: run following command to show help message.
python predict.py -h
"""
UNK_IDX = 0
class Prediction():
def __init__(self, train_conf, dict_file, model_dir, label_file,
predicate_dict_file):
"""
train_conf: trainer configure.
dict_file: word dictionary file name.
model_dir: directory of model.
"""
self.dict = {}
self.labels = {}
self.predicate_dict = {}
self.labels_reverse = {}
self.load_dict_label(dict_file, label_file, predicate_dict_file)
len_dict = len(self.dict)
len_label = len(self.labels)
len_pred = len(self.predicate_dict)
conf = parse_config(
train_conf, 'dict_len=' + str(len_dict) + ',label_len=' +
str(len_label) + ',pred_len=' + str(len_pred) + ',is_predict=True')
self.network = swig_paddle.GradientMachine.createFromConfigProto(
conf.model_config)
self.network.loadParameters(model_dir)
slots = [
integer_value_sequence(len_dict), integer_value_sequence(len_dict),
integer_value_sequence(len_dict), integer_value_sequence(len_dict),
integer_value_sequence(len_dict), integer_value_sequence(len_dict),
integer_value_sequence(len_pred), integer_value_sequence(2)
]
self.converter = DataProviderConverter(slots)
def load_dict_label(self, dict_file, label_file, predicate_dict_file):
"""
Load dictionary from self.dict_file.
"""
for line_count, line in enumerate(open(dict_file, 'r')):
self.dict[line.strip()] = line_count
for line_count, line in enumerate(open(label_file, 'r')):
self.labels[line.strip()] = line_count
self.labels_reverse[line_count] = line.strip()
for line_count, line in enumerate(open(predicate_dict_file, 'r')):
self.predicate_dict[line.strip()] = line_count
def get_data(self, data_file):
"""
Get input data of paddle format.
"""
with open(data_file, 'r') as fdata:
for line in fdata:
sentence, predicate, ctx_n2, ctx_n1, ctx_0, ctx_p1, ctx_p2, mark, label = line.strip(
).split('\t')
words = sentence.split()
sen_len = len(words)
word_slot = [self.dict.get(w, UNK_IDX) for w in words]
predicate_slot = [self.predicate_dict.get(predicate, UNK_IDX)
] * sen_len
ctx_n2_slot = [self.dict.get(ctx_n2, UNK_IDX)] * sen_len
ctx_n1_slot = [self.dict.get(ctx_n1, UNK_IDX)] * sen_len
ctx_0_slot = [self.dict.get(ctx_0, UNK_IDX)] * sen_len
ctx_p1_slot = [self.dict.get(ctx_p1, UNK_IDX)] * sen_len
ctx_p2_slot = [self.dict.get(ctx_p2, UNK_IDX)] * sen_len
marks = mark.split()
mark_slot = [int(w) for w in marks]
yield word_slot, ctx_n2_slot, ctx_n1_slot, \
ctx_0_slot, ctx_p1_slot, ctx_p2_slot, predicate_slot, mark_slot
def predict(self, data_file, output_file):
"""
data_file: file name of input data.
"""
input = self.converter(self.get_data(data_file))
output = self.network.forwardTest(input)
lab = output[0]["id"].tolist()
with open(data_file, 'r') as fin, open(output_file, 'w') as fout:
index = 0
for line in fin:
sen = line.split('\t')[0]
len_sen = len(sen.split())
line_labels = lab[index:index + len_sen]
index += len_sen
fout.write(sen + '\t' + ' '.join(
[self.labels_reverse[i] for i in line_labels]) + '\n')
def option_parser():
usage = (
"python predict.py -c config -w model_dir "
"-d word dictionary -l label_file -i input_file -p pred_dict_file")
parser = OptionParser(usage="usage: %s [options]" % usage)
parser.add_option(
"-c",
"--tconf",
action="store",
dest="train_conf",
help="network config")
parser.add_option(
"-d",
"--dict",
action="store",
dest="dict_file",
help="dictionary file")
parser.add_option(
"-l",
"--label",
action="store",
dest="label_file",
default=None,
help="label file")
parser.add_option(
"-p",
"--predict_dict_file",
action="store",
dest="predict_dict_file",
default=None,
help="predict_dict_file")
parser.add_option(
"-i",
"--data",
action="store",
dest="data_file",
help="data file to predict")
parser.add_option(
"-w",
"--model",
action="store",
dest="model_path",
default=None,
help="model path")
parser.add_option(
"-o",
"--output_file",
action="store",
dest="output_file",
default=None,
help="output file")
return parser.parse_args()
def main():
options, args = option_parser()
train_conf = options.train_conf
data_file = options.data_file
dict_file = options.dict_file
model_path = options.model_path
label_file = options.label_file
predict_dict_file = options.predict_dict_file
output_file = options.output_file
swig_paddle.initPaddle("--use_gpu=0")
predict = Prediction(train_conf, dict_file, model_path, label_file,
predict_dict_file)
predict.predict(data_file, output_file)
if __name__ == '__main__':
main()
demo/semantic_role_labeling/predict.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/semantic_role_labeling/test.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/semantic_role_labeling/train.sh 已删除 100755 → 0
浏览文件 @ 338dd135
#!/bin/bash
# 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.
set -e
paddle train \
--config=./db_lstm.py \
--use_gpu=0 \
--log_period=5000 \
--trainer_count=1 \
--show_parameter_stats_period=5000 \
--save_dir=./output \
--num_passes=10000 \
--average_test_period=10000000 \
--init_model_path=./data \
--load_missing_parameter_strategy=rand \
--test_all_data_in_one_period=1 \
2>&1 | tee 'train.log'
paddle usage -l train.log -e $? -n "semantic_role_labeling_train" >/dev/null 2>&1
demo/sentiment/.gitignore 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/dataprovider.py 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/predict.py 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/predict.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/preprocess.py 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/sentiment_net.py 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/test.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/train.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/train_v2.py 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/sentiment/trainer_config.py 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/.gitignore 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/api_train_v2.py 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/data/paraphrase_data.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/data/paraphrase_model.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/data/wmt14_data.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/data/wmt14_model.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/dataprovider.py 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/paraphrase/train.conf 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/paraphrase/train.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/preprocess.py 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/seqToseq_net.py 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/translation/eval_bleu.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/translation/gen.conf 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/translation/gen.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/translation/moses_bleu.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/translation/train.conf 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/seqToseq/translation/train.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
demo/word2vec/api_train_v2.py 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
doc/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
doc/api/v2/config/evaluators.rst
浏览文件 @ a698824a
此差异已折叠。
doc/api/v2/config/layer.rst
浏览文件 @ a698824a
此差异已折叠。
doc/api/v2/config/networks.rst
浏览文件 @ a698824a
此差异已折叠。
doc/design/build_system/README.md
浏览文件 @ a698824a
此差异已折叠。
doc/design/cluster_train/remote_parameter_updater.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
doc/design/cluster_train/save_model.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
doc/design/parameters_in_cpp.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
doc/design/scope.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
doc/design/simple_op_design.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
doc/design/speech/README.MD 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
doc/design/speech/image/ds2_network.png 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
doc/getstarted/index_cn.rst
浏览文件 @ a698824a
此差异已折叠。
doc/getstarted/index_en.rst
浏览文件 @ a698824a
此差异已折叠。
doc_theme/templates/layout.html
浏览文件 @ a698824a
此差异已折叠。
go/.gitignore 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
demo/sentiment/preprocess.sh → go/CMakeLists.txt 100755 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/cmd/master/master.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/cmd/pserver/.gitignore 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/cmd/pserver/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/cmd/pserver/pserver.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/connection/conn.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/glide.lock 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/glide.yaml 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/c/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/c/client.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/client.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/client_internal_test.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/client_test.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/etcd_client.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/inmem_store.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/service.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/master/service_internal_test.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/client/c/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/client/c/test/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/client/c/test/test_cclient.c 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/client/c/test/testdata/optimizer.pb 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/client/client.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/client/client_test.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/client/etcd_client.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/etcd_client.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/optimizer.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/optimizer_test.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/service.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/pserver/service_test.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/utils/networkhelper/helper.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
go/utils/networkhelper/helper_test.go 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/api/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/api/Paddle.i
浏览文件 @ a698824a
此差异已折叠。
paddle/api/PaddleAPI.h
浏览文件 @ a698824a
此差异已折叠。
paddle/api/ParameterUpdater.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/capi/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/cuda/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/cuda/include/hl_cpu_scalar.cuh 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/cuda/include/hl_cpu_simd_neon.cuh 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/cuda/include/hl_cpu_simd_sse.cuh 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/cuda/include/hl_matrix_base_sse.cuh 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/cuda/include/hl_neon_matrix_kernel.cuh 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/cuda/include/hl_tensor_ops.h
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/.clang-format 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/attr_checker.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/attr_type.proto 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/ddim.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/ddim.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/ddim_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/dim.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/dim_test.cu 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/enforce.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/enforce_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/net_design.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/op_desc.proto 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/op_desc_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/op_proto.proto 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/op_proto_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/op_registry.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/op_registry_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/scope.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/scope_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/tensor.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/tensor_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/variable.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/variable.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/framework/variable_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/function/ConvOp.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/ConvOpTest.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/CosSimOpTest.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/function/FunctionTest.h
浏览文件 @ a698824a
此差异已折叠。
paddle/function/GemmConvOp.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/GemmConvOp.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/GemmConvOpGpu.cu 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/GemmFunctor.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/MulOpTest.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/function/NaiveConvOp.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/PadOpTest.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/function/RowConvOp.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/RowConvOp.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/RowConvOpGpu.cu 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/RowConvOpTest.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/nnpack/NNPACKConvOp.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/nnpack/NNPACKConvOpTest.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/function/nnpack/nnpack.cmake 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/go/cclient/CMakeLists.txt 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/go/cclient/cmake/flags.cmake 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/go/cclient/cmake/golang.cmake 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/go/cclient/test/CMakeLists.txt 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/go/cclient/test/main.c 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/go/pserver/client.go 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/gserver/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/evaluators/DetectionMAPEvaluator.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/AgentLayer.h
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/CostLayer.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/DetectionOutputLayer.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/DetectionOutputLayer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/DetectionUtil.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/DetectionUtil.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/Layer.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/MaxLayer.h
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/MultiBoxLossLayer.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/MultiBoxLossLayer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/NormLayer.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/layers/RowConvLayer.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/tests/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/tests/sequence_rnn_matched_inputs.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/tests/sequence_rnn_mixed_inputs.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/gserver/tests/test_DetectionOutput.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/majel/.gitignore 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/majel/CMakeLists.txt 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/majel/place.cc 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/majel/place.h 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/majel/test/CMakeLists.txt 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/majel/test/place_test.cc 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/math/BaseMatrix.h
浏览文件 @ a698824a
此差异已折叠。
paddle/math/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/math/MathFunctions.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/math/Matrix.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/math/Matrix.h
浏览文件 @ a698824a
此差异已折叠。
paddle/math/SparseRowMatrix.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/math/SparseRowMatrix.h
浏览文件 @ a698824a
此差异已折叠。
paddle/math/TensorEvaluate.h
浏览文件 @ a698824a
此差异已折叠。
paddle/math/Vector.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/math/Vector.h
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/.clang-format 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/README.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/detail/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/detail/buddy_allocator.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/detail/buddy_allocator.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/detail/system_allocator.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/detail/system_allocator.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/detail/system_allocator_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/memory/memory.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/adadelta_optimizer.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/adadelta_optimizer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/adagrad_optimizer.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/adagrad_optimizer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/adam_optimizer.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/adam_optimizer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/lr_policy.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/optimizer.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/optimizer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/parameter_optimizer.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/parameter_optimizer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/parameter_optimizer_test.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/serialization.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/serialization_test.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/sgd_optimizer.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/sgd_optimizer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/optimizer/tensor.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/parameter/Argument.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/parameter/Argument.h
浏览文件 @ a698824a
此差异已折叠。
paddle/parameter/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/parameter/Parameter.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/parameter/Parameter.h
浏览文件 @ a698824a
此差异已折叠。
paddle/parameter/ThreadLocalBuffer.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/parameter/ThreadLocalBuffer.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/.clang-format 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/assert.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/cuda.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/dynload/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/dynload/cublas.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/dynload/cudnn.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/dynload/curand.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/dynload/dynamic_loader.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/dynload/dynamic_loader.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/hostdevice.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/place.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/place.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/platform/place_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/pserver/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/pserver/LightNetwork.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/pserver/ParameterClient2.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/pserver/ParameterServer2.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/pserver/ParameterServer2.h
浏览文件 @ a698824a
此差异已折叠。
paddle/pserver/SocketChannel.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/pserver/test/SocketTest.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/scripts/docker/build.sh
浏览文件 @ a698824a
此差异已折叠。
paddle/scripts/docker/build_android.sh 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/scripts/run_python_tests.sh
浏览文件 @ a698824a
此差异已折叠。
paddle/scripts/travis/build_and_test.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/scripts/travis/common.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/scripts/travis/main.sh 已删除 100755 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/string/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/string/piece.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/string/piece.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/string/piece_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/string/printf.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/string/printf_test.cc 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/string/tinyformat/tinyformat.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/testing/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/trainer/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/trainer/NewRemoteParameterUpdater.cpp 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/trainer/NewRemoteParameterUpdater.h 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
paddle/trainer/Tester.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/BarrierStat.cpp 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/utils/BarrierStat.h 已删除 100644 → 0
浏览文件 @ 338dd135
此差异已折叠。
paddle/utils/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/CpuId.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/CustomStackTrace.h
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/Error.h
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/Stat.cpp
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/Stat.h
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/ThreadLocal.h
浏览文件 @ a698824a
此差异已折叠。
paddle/utils/arch/linux/Locks.cpp
浏览文件 @ a698824a
此差异已折叠。
proto/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
proto/ModelConfig.proto
浏览文件 @ a698824a
此差异已折叠。
proto/OptimizerConfig.proto 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
proto/ParameterConfig.proto
浏览文件 @ a698824a
此差异已折叠。
proto/TrainerConfig.proto
浏览文件 @ a698824a
此差异已折叠。
python/CMakeLists.txt
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/__init__.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/attr.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/config_base.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/data_type.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/cifar.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/common.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/conll05.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/flowers.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/imdb.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/mnist.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/mq2007.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/dataset/wmt14.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/evaluator.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/framework/__init__.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/framework/tests/CMakeLists.txt 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/framework/tests/test_protobuf.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/image.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/inference.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/layer.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/master/.gitignore 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/master/__init__.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/master/client.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/networks.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/op.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/optimizer.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/parameters.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/reader/creator.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/reader/tests/test_recordio_creator.dat 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/topology.py
浏览文件 @ a698824a
此差异已折叠。
python/paddle/v2/trainer.py
浏览文件 @ a698824a
此差异已折叠。
python/setup.py.in
浏览文件 @ a698824a
此差异已折叠。
v1_api_demo/README.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
v1_api_demo/vae/README.md 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
v1_api_demo/vae/data/get_mnist_data.sh 0 → 100755
浏览文件 @ a698824a
此差异已折叠。
v1_api_demo/vae/vae_conf.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
v1_api_demo/vae/vae_train.py 0 → 100644
浏览文件 @ a698824a
此差异已折叠。
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或