Skip to content

P
Paddle

机器未来 / Paddle
与 Fork 源项目一致

Fork自 PaddlePaddle / Paddle

通知 1
Star 1
Fork 0
P
Paddle
未验证 提交 048cc30e 编写于 作者: J Jeff Wang 提交者： GitHub
浏览文件
操作

Merge branch 'develop' into updateWriteDocsCN

上级 2488b81f 35572355
展开全部 隐藏空白更改
内联 并排
Showing with 13079 addition and 16833 deletion
.clang-format
浏览文件 @ 048cc30e
......@@ -25,4 +25,3 @@ AllowAllParametersOfDeclarationOnNextLine: true
BinPackParameters: false
BinPackArguments: false
...
.travis.yml
浏览文件 @ 048cc30e
......@@ -42,7 +42,7 @@ before_install:
script:
- |
timeout 2580 paddle/scripts/travis/${JOB}.sh # 43min timeout
RESULT=$?; if [ $RESULT -eq 0 ] || [ $RESULT -eq 142 ]; then true; else false; fi;
RESULT=$?; if [ $RESULT -eq 0 ] || [ $RESULT -eq 142 ]; then true ;else exit 1; fi;
- |
if [[ "$JOB" != "build_doc" ]]; then exit 0; fi;
if [[ "$TRAVIS_PULL_REQUEST" != "false" ]]; then exit 0; fi;
......
CMakeLists.txt
浏览文件 @ 048cc30e
......@@ -36,8 +36,7 @@ include(simd)
################################ Configurations #######################################
option(WITH_GPU "Compile PaddlePaddle with NVIDIA GPU" ${CUDA_FOUND})
option(WITH_AVX "Compile PaddlePaddle with AVX intrinsics" ${AVX_FOUND})
option(WITH_MKLDNN "Compile PaddlePaddle with mkl-dnn support." ${AVX_FOUND})
option(WITH_MKLML "Compile PaddlePaddle with mklml package." ${AVX_FOUND})
option(WITH_MKL "Compile PaddlePaddle with MKL support." ${AVX_FOUND})
option(WITH_DSO "Compile PaddlePaddle with dynamic linked CUDA" ON)
option(WITH_TESTING "Compile PaddlePaddle with unit testing" ON)
option(WITH_SWIG_PY "Compile PaddlePaddle with inference api" ON)
......@@ -82,10 +81,8 @@ if(ANDROID OR IOS)
"Disable PYTHON when cross-compiling for Android and iOS" FORCE)
set(WITH_RDMA OFF CACHE STRING
"Disable RDMA when cross-compiling for Android and iOS" FORCE)
set(WITH_MKLDNN OFF CACHE STRING
"Disable MKLDNN when cross-compiling for Android and iOS" FORCE)
set(WITH_MKLML OFF CACHE STRING
"Disable MKLML package when cross-compiling for Android and iOS" FORCE)
set(WITH_MKL OFF CACHE STRING
"Disable MKL when cross-compiling for Android and iOS" FORCE)
# Compile PaddlePaddle mobile inference library
if (NOT WITH_C_API)
......@@ -111,6 +108,14 @@ else()
set(THIRD_PARTY_BUILD_TYPE Release)
endif()
set(WITH_MKLML ${WITH_MKL})
if (WITH_MKL AND AVX2_FOUND)
set(WITH_MKLDNN ON)
else()
message(STATUS "Do not have AVX2 intrinsics and disabled MKL-DNN")
set(WITH_MKLDNN OFF)
endif()
########################################################################################
include(external/mklml) # download mklml package
......@@ -128,6 +133,8 @@ include(external/any) # download libn::any
include(external/eigen) # download eigen3
include(external/pybind11) # download pybind11
include(external/nccl)
include(external/cares)
include(external/grpc)
include(cudnn) # set cudnn libraries, must before configure
include(configure) # add paddle env configuration
......@@ -158,14 +165,15 @@ set(EXTERNAL_LIBS
)
if(WITH_GPU)
list(APPEND EXTERNAL_LIBS ${CUDA_LIBRARIES} ${CUDA_rt_LIBRARY})
if(NOT WITH_DSO)
list(APPEND EXTERNAL_LIBS ${CUDNN_LIBRARY} ${CUDA_CUBLAS_LIBRARIES} ${CUDA_curand_LIBRARY} ${NCCL_LIBRARY})
endif(NOT WITH_DSO)
include(cuda)
endif(WITH_GPU)
if(WITH_MKLML)
list(APPEND EXTERNAL_LIBS ${MKLML_IOMP_LIB})
endif()
if(WITH_MKLDNN)
list(APPEND EXTERNAL_LIBS ${MKLDNN_LIB} ${MKLDNN_IOMP_LIB})
list(APPEND EXTERNAL_LIBS ${MKLDNN_LIB})
endif()
if(USE_NNPACK)
......
Dockerfile
浏览文件 @ 048cc30e
......@@ -29,7 +29,7 @@ RUN apt-get update && \
automake locales clang-format swig doxygen cmake \
liblapack-dev liblapacke-dev libboost-dev \
clang-3.8 llvm-3.8 libclang-3.8-dev \
net-tools && \
net-tools libtool && \
apt-get clean -y
# Install Go and glide
......
benchmark/IntelOptimizedPaddle.md
浏览文件 @ 048cc30e
......@@ -12,11 +12,11 @@ Machine:
System: CentOS release 6.3 (Final), Docker 1.12.1.
PaddlePaddle: paddlepaddle/paddle:latest (TODO: will rerun after 0.11.0)
- MKL-DNN tag v0.10
- MKLML 2018.0.20170720
PaddlePaddle: paddlepaddle/paddle:latest (for MKLML and MKL-DNN), paddlepaddle/paddle:latest-openblas (for OpenBLAS)
- MKL-DNN tag v0.11
- MKLML 2018.0.1.20171007
- OpenBLAS v0.2.20
(TODO: will rerun after 0.11.0)
On each machine, we will test and compare the performance of training on single node using MKL-DNN / MKLML / OpenBLAS respectively.
......@@ -31,17 +31,37 @@ Input image size - 3 * 224 * 224, Time: images/second
| BatchSize | 64 | 128 | 256 |
|--------------|-------| -----| --------|
| OpenBLAS | 7.82 | 8.62 | 10.34 |
| MKLML | 11.02 | 12.86 | 15.33 |
| MKL-DNN | 27.69 | 28.8 | 29.27 |
| OpenBLAS | 7.80 | 9.00 | 10.80 |
| MKLML | 12.12 | 13.70 | 16.18 |
| MKL-DNN | 28.46 | 29.83 | 30.44 |
chart on batch size 128
TBD
- ResNet-50
| BatchSize | 64 | 128 | 256 |
|--------------|-------| ------| -------|
| OpenBLAS | 25.22 | 25.68 | 27.12 |
| MKLML | 32.52 | 31.89 | 33.12 |
| MKL-DNN | 81.69 | 82.35 | 84.08 |
chart on batch size 128
TBD
- ResNet
- GoogLeNet
| BatchSize | 64 | 128 | 256 |
|--------------|-------| ------| -------|
| OpenBLAS | 89.52 | 96.97 | 108.25 |
| MKLML | 128.46| 137.89| 158.63 |
| MKL-DNN     | 250.46| 264.83| 269.50 |
chart on batch size 128
TBD
### Laptop
TBD
### Desktop
......
benchmark/paddle/image/googlenet.py
浏览文件 @ 048cc30e
......@@ -5,6 +5,7 @@ height = 224
width = 224
num_class = 1000
batch_size = get_config_arg('batch_size', int, 128)
use_gpu = get_config_arg('use_gpu', bool, True)
args = {'height': height, 'width': width, 'color': True, 'num_class': num_class}
define_py_data_sources2(
......@@ -16,6 +17,8 @@ settings(
learning_method=MomentumOptimizer(0.9),
regularization=L2Regularization(0.0005 * batch_size))
conv_projection = conv_projection if use_gpu else img_conv_layer
def inception2(name, input, channels, \
filter1,
filter3R, filter3,
......@@ -138,7 +141,7 @@ def inception(name, input, channels, \
cat = concat_layer(
name=name,
input=[cov1, cov3, cov5, covprj],
bias_attr=True,
bias_attr=True if use_gpu else False,
act=ReluActivation())
return cat
......
benchmark/paddle/image/run_mkldnn.sh
浏览文件 @ 048cc30e
......@@ -40,6 +40,7 @@ fi
for use_mkldnn in True False; do
for batchsize in 64 128 256; do
train vgg 19 $batchsize $use_mkldnn
train resnet 50 $batchsize $use_mkldnn
train resnet 50 $batchsize $use_mkldnn
train googlenet v1 $batchsize $use_mkldnn
done
done
cmake/configure.cmake
浏览文件 @ 048cc30e
......@@ -76,27 +76,14 @@ else()
include_directories(${CUDA_TOOLKIT_INCLUDE})
endif(NOT WITH_GPU)
if(WITH_MKLDNN)
add_definitions(-DPADDLE_USE_MKLDNN)
if (WITH_MKLML AND MKLDNN_IOMP_DIR)
message(STATUS "Enable Intel OpenMP at ${MKLDNN_IOMP_DIR}")
set(OPENMP_FLAGS "-fopenmp")
set(CMAKE_C_CREATE_SHARED_LIBRARY_FORBIDDEN_FLAGS ${OPENMP_FLAGS})
set(CMAKE_CXX_CREATE_SHARED_LIBRARY_FORBIDDEN_FLAGS ${OPENMP_FLAGS})
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${OPENMP_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${OPENMP_FLAGS}")
else()
find_package(OpenMP)
if(OPENMP_FOUND)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${OpenMP_C_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${OpenMP_CXX_FLAGS}")
else()
message(WARNING "Can not find OpenMP."
"Some performance features in MKLDNN may not be available")
endif()
endif()
endif(WITH_MKLDNN)
if (WITH_MKLML AND MKLML_IOMP_LIB)
message(STATUS "Enable Intel OpenMP with ${MKLML_IOMP_LIB}")
set(OPENMP_FLAGS "-fopenmp")
set(CMAKE_C_CREATE_SHARED_LIBRARY_FORBIDDEN_FLAGS ${OPENMP_FLAGS})
set(CMAKE_CXX_CREATE_SHARED_LIBRARY_FORBIDDEN_FLAGS ${OPENMP_FLAGS})
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${OPENMP_FLAGS}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${OPENMP_FLAGS}")
endif()
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${SIMD_FLAG}")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${SIMD_FLAG}")
......
cmake/cross_compiling/ios.cmake
浏览文件 @ 048cc30e
......@@ -76,11 +76,9 @@ set(IOS_PLATFORM ${IOS_PLATFORM} CACHE STRING "Type of iOS Platform")
# Set the architecture for iOS
if(NOT DEFINED IOS_ARCH)
if(IOS_PLATFORM STREQUAL "OS")
# FIXME(liuyiqun): support "armv7;armv7s;arm64" future
set(IOS_ARCH "arm64")
set(IOS_ARCH "armv7;armv7s;arm64")
elseif(IOS_PLATFORM STREQUAL "SIMULATOR")
# FIXME(liuyiqun): support "i386;x86_64" future
set(IOS_ARCH "x86_64")
set(IOS_ARCH "i386;x86_64")
endif()
endif()
set(CMAKE_OSX_ARCHITECTURES ${IOS_ARCH} CACHE string "Build architecture for iOS")
......@@ -248,7 +246,7 @@ set(IOS_COMPILER_FLAGS "${XCODE_IOS_PLATFORM_VERSION_FLAGS} ${XCODE_IOS_BITCODE_
# Hidden visibilty is required for cxx on iOS
set(CMAKE_C_FLAGS "${IOS_COMPILER_FLAGS} ${CMAKE_C_FLAGS}" CACHE STRING "C flags")
set(CMAKE_CXX_FLAGS "${IOS_COMPILER_FLAGS} -fvisibility-inlines-hidden ${CMAKE_CXX_FLAGS}" CACHE STRING "CXX flags")
set(CMAKE_CXX_FLAGS "${IOS_COMPILER_FLAGS} -fvisibility=hidden -fvisibility-inlines-hidden ${CMAKE_CXX_FLAGS}" CACHE STRING "CXX flags")
set(IOS_LINK_FLAGS "${XCODE_IOS_PLATFORM_VERSION_FLAGS} -Wl,-search_paths_first")
......
cmake/cuda.cmake 0 → 100644
浏览文件 @ 048cc30e
if(NOT WITH_GPU)
return()
endif()
set(paddle_known_gpu_archs "30 35 50 52 60 61 70")
set(paddle_known_gpu_archs7 "30 35 50 52")
set(paddle_known_gpu_archs8 "30 35 50 52 60 61")
######################################################################################
# A function for automatic detection of GPUs installed (if autodetection is enabled)
# Usage:
# detect_installed_gpus(out_variable)
function(detect_installed_gpus out_variable)
if(NOT CUDA_gpu_detect_output)
set(cufile ${PROJECT_BINARY_DIR}/detect_cuda_archs.cu)
file(WRITE ${cufile} ""
"#include <cstdio>\n"
"int main() {\n"
" int count = 0;\n"
" if (cudaSuccess != cudaGetDeviceCount(&count)) return -1;\n"
" if (count == 0) return -1;\n"
" for (int device = 0; device < count; ++device) {\n"
" cudaDeviceProp prop;\n"
" if (cudaSuccess == cudaGetDeviceProperties(&prop, device))\n"
" std::printf(\"%d.%d \", prop.major, prop.minor);\n"
" }\n"
" return 0;\n"
"}\n")
execute_process(COMMAND "${CUDA_NVCC_EXECUTABLE}" "-ccbin=${CUDA_HOST_COMPILER}"
"--run" "${cufile}"
WORKING_DIRECTORY "${PROJECT_BINARY_DIR}/CMakeFiles/"
RESULT_VARIABLE nvcc_res OUTPUT_VARIABLE nvcc_out
ERROR_QUIET OUTPUT_STRIP_TRAILING_WHITESPACE)
if(nvcc_res EQUAL 0)
# only keep the last line of nvcc_out
STRING(REGEX REPLACE ";" "\\\\;" nvcc_out "${nvcc_out}")
STRING(REGEX REPLACE "\n" ";" nvcc_out "${nvcc_out}")
list(GET nvcc_out -1 nvcc_out)
string(REPLACE "2.1" "2.1(2.0)" nvcc_out "${nvcc_out}")
set(CUDA_gpu_detect_output ${nvcc_out} CACHE INTERNAL "Returned GPU architetures from detect_installed_gpus tool" FORCE)
endif()
endif()
if(NOT CUDA_gpu_detect_output)
message(STATUS "Automatic GPU detection failed. Building for all known architectures.")
set(${out_variable} ${paddle_known_gpu_archs} PARENT_SCOPE)
else()
set(${out_variable} ${CUDA_gpu_detect_output} PARENT_SCOPE)
endif()
endfunction()
########################################################################
# Function for selecting GPU arch flags for nvcc based on CUDA_ARCH_NAME
# Usage:
# select_nvcc_arch_flags(out_variable)
function(select_nvcc_arch_flags out_variable)
# List of arch names
set(archs_names "Kepler" "Maxwell" "Pascal" "All" "Manual")
set(archs_name_default "All")
if(NOT CMAKE_CROSSCOMPILING)
list(APPEND archs_names "Auto")
endif()
# set CUDA_ARCH_NAME strings (so it will be seen as dropbox in CMake-Gui)
set(CUDA_ARCH_NAME ${archs_name_default} CACHE STRING "Select target NVIDIA GPU achitecture.")
set_property( CACHE CUDA_ARCH_NAME PROPERTY STRINGS "" ${archs_names} )
mark_as_advanced(CUDA_ARCH_NAME)
# verify CUDA_ARCH_NAME value
if(NOT ";${archs_names};" MATCHES ";${CUDA_ARCH_NAME};")
string(REPLACE ";" ", " archs_names "${archs_names}")
message(FATAL_ERROR "Only ${archs_names} architeture names are supported.")
endif()
if(${CUDA_ARCH_NAME} STREQUAL "Manual")
set(CUDA_ARCH_BIN ${paddle_known_gpu_archs} CACHE STRING "Specify 'real' GPU architectures to build binaries for, BIN(PTX) format is supported")
set(CUDA_ARCH_PTX "50" CACHE STRING "Specify 'virtual' PTX architectures to build PTX intermediate code for")
mark_as_advanced(CUDA_ARCH_BIN CUDA_ARCH_PTX)
else()
unset(CUDA_ARCH_BIN CACHE)
unset(CUDA_ARCH_PTX CACHE)
endif()
if(${CUDA_ARCH_NAME} STREQUAL "Kepler")
set(cuda_arch_bin "30 35")
elseif(${CUDA_ARCH_NAME} STREQUAL "Maxwell")
set(cuda_arch_bin "50")
elseif(${CUDA_ARCH_NAME} STREQUAL "Pascal")
set(cuda_arch_bin "60 61")
elseif(${CUDA_ARCH_NAME} STREQUAL "Volta")
set(cuda_arch_bin "70")
elseif(${CUDA_ARCH_NAME} STREQUAL "All")
set(cuda_arch_bin ${paddle_known_gpu_archs})
elseif(${CUDA_ARCH_NAME} STREQUAL "Auto")
detect_installed_gpus(cuda_arch_bin)
else() # (${CUDA_ARCH_NAME} STREQUAL "Manual")
set(cuda_arch_bin ${CUDA_ARCH_BIN})
endif()
# remove dots and convert to lists
string(REGEX REPLACE "\\." "" cuda_arch_bin "${cuda_arch_bin}")
string(REGEX REPLACE "\\." "" cuda_arch_ptx "${CUDA_ARCH_PTX}")
string(REGEX MATCHALL "[0-9()]+" cuda_arch_bin "${cuda_arch_bin}")
string(REGEX MATCHALL "[0-9]+" cuda_arch_ptx "${cuda_arch_ptx}")
list(REMOVE_DUPLICATES cuda_arch_bin)
list(REMOVE_DUPLICATES cuda_arch_ptx)
set(nvcc_flags "")
set(nvcc_archs_readable "")
# Tell NVCC to add binaries for the specified GPUs
foreach(arch ${cuda_arch_bin})
if(arch MATCHES "([0-9]+)\\(([0-9]+)\\)")
# User explicitly specified PTX for the concrete BIN
list(APPEND nvcc_flags -gencode arch=compute_${CMAKE_MATCH_2},code=sm_${CMAKE_MATCH_1})
list(APPEND nvcc_archs_readable sm_${CMAKE_MATCH_1})
else()
# User didn't explicitly specify PTX for the concrete BIN, we assume PTX=BIN
list(APPEND nvcc_flags -gencode arch=compute_${arch},code=sm_${arch})
list(APPEND nvcc_archs_readable sm_${arch})
endif()
endforeach()
# Tell NVCC to add PTX intermediate code for the specified architectures
foreach(arch ${cuda_arch_ptx})
list(APPEND nvcc_flags -gencode arch=compute_${arch},code=compute_${arch})
list(APPEND nvcc_archs_readable compute_${arch})
endforeach()
string(REPLACE ";" " " nvcc_archs_readable "${nvcc_archs_readable}")
set(${out_variable} ${nvcc_flags} PARENT_SCOPE)
set(${out_variable}_readable ${nvcc_archs_readable} PARENT_SCOPE)
endfunction()
message(STATUS "CUDA detected: " ${CUDA_VERSION})
if (${CUDA_VERSION} LESS 7.0)
set(paddle_known_gpu_archs ${paddle_known_gpu_archs})
elseif (${CUDA_VERSION} LESS 8.0) # CUDA 7.x
set(paddle_known_gpu_archs ${paddle_known_gpu_archs7})
list(APPEND CUDA_NVCC_FLAGS "-D_MWAITXINTRIN_H_INCLUDED")
list(APPEND CUDA_NVCC_FLAGS "-D__STRICT_ANSI__")
elseif (${CUDA_VERSION} LESS 9.0) # CUDA 8.x
set(paddle_known_gpu_archs ${paddle_known_gpu_archs8})
list(APPEND CUDA_NVCC_FLAGS "-D_MWAITXINTRIN_H_INCLUDED")
list(APPEND CUDA_NVCC_FLAGS "-D__STRICT_ANSI__")
# CUDA 8 may complain that sm_20 is no longer supported. Suppress the
# warning for now.
list(APPEND CUDA_NVCC_FLAGS "-Wno-deprecated-gpu-targets")
endif()
include_directories(${CUDA_INCLUDE_DIRS})
list(APPEND EXTERNAL_LIBS ${CUDA_LIBRARIES} ${CUDA_rt_LIBRARY})
if(NOT WITH_DSO)
list(APPEND EXTERNAL_LIBS ${CUDNN_LIBRARY} ${CUDA_CUBLAS_LIBRARIES} ${CUDA_curand_LIBRARY} ${NCCL_LIBRARY})
endif(NOT WITH_DSO)
# setting nvcc arch flags
select_nvcc_arch_flags(NVCC_FLAGS_EXTRA)
list(APPEND CUDA_NVCC_FLAGS ${NVCC_FLAGS_EXTRA})
message(STATUS "Added CUDA NVCC flags for: ${NVCC_FLAGS_EXTRA_readable}")
# Set C++11 support
set(CUDA_PROPAGATE_HOST_FLAGS OFF)
# Release/Debug flags set by cmake. Such as -O3 -g -DNDEBUG etc.
# So, don't set these flags here.
list(APPEND CUDA_NVCC_FLAGS "-std=c++11")
list(APPEND CUDA_NVCC_FLAGS "--use_fast_math")
list(APPEND CUDA_NVCC_FLAGS "-Xcompiler -fPIC")
# Set :expt-relaxed-constexpr to suppress Eigen warnings
list(APPEND CUDA_NVCC_FLAGS "--expt-relaxed-constexpr")
if(CMAKE_BUILD_TYPE STREQUAL "Debug")
list(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_DEBUG})
elseif(CMAKE_BUILD_TYPE STREQUAL "Release")
list(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_RELEASE})
elseif(CMAKE_BUILD_TYPE STREQUAL "RelWithDebInfo")
list(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_RELWITHDEBINFO})
elseif(CMAKE_BUILD_TYPE STREQUAL "MinSizeRel")
list(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_MINSIZEREL})
endif()
mark_as_advanced(CUDA_BUILD_CUBIN CUDA_BUILD_EMULATION CUDA_VERBOSE_BUILD)
mark_as_advanced(CUDA_SDK_ROOT_DIR CUDA_SEPARABLE_COMPILATION)
cmake/external/cares.cmake 0 → 100644
浏览文件 @ 048cc30e
# 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.
#
IF(MOBILE_INFERENCE)
return()
ENDIF()
include (ExternalProject)
# NOTE: c-ares is needed when linking with grpc.
SET(CARES_SOURCES_DIR ${THIRD_PARTY_PATH}/cares)
SET(CARES_INSTALL_DIR ${THIRD_PARTY_PATH}/install/cares)
SET(CARES_INCLUDE_DIR "${CARES_INSTALL_DIR}/include/" CACHE PATH "cares include directory." FORCE)
ExternalProject_Add(
extern_cares
GIT_REPOSITORY "https://github.com/c-ares/c-ares.git"
GIT_TAG "cares-1_13_0"
PREFIX ${CARES_SOURCES_DIR}
UPDATE_COMMAND ""
CONFIGURE_COMMAND ./buildconf && ./configure --disable-shared --prefix=${CARES_INSTALL_DIR}
BUILD_IN_SOURCE 1
BUILD_COMMAND make
INSTALL_COMMAND make install
)
ADD_LIBRARY(cares STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET cares PROPERTY IMPORTED_LOCATION
"${CARES_INSTALL_DIR}/lib/libcares.a")
include_directories(${CARES_INCLUDE_DIR})
ADD_DEPENDENCIES(cares extern_cares)
cmake/external/gflags.cmake
浏览文件 @ 048cc30e
......@@ -28,15 +28,8 @@ INCLUDE_DIRECTORIES(${GFLAGS_INCLUDE_DIR})
ExternalProject_Add(
extern_gflags
${EXTERNAL_PROJECT_LOG_ARGS}
# TODO(yiwang): The annoying warnings mentioned in
# https://github.com/PaddlePaddle/Paddle/issues/3277 are caused by
# gflags. I fired a PR https://github.com/gflags/gflags/pull/230
# to fix it. Before it gets accepted by the gflags team, we use
# my personal fork, which contains above fix, temporarily. Let's
# change this back to the official Github repo once my PR is
# merged.
GIT_REPOSITORY "https://github.com/wangkuiyi/gflags.git"
GIT_TAG 986964c07427ecb9cdb5bd73f73ebbd40e54dadb
GIT_REPOSITORY "https://github.com/gflags/gflags.git"
GIT_TAG 77592648e3f3be87d6c7123eb81cbad75f9aef5a
PREFIX ${GFLAGS_SOURCES_DIR}
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
......
cmake/external/grpc.cmake 0 → 100644
浏览文件 @ 048cc30e
# 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.
#
IF(MOBILE_INFERENCE)
return()
ENDIF()
include (ExternalProject)
SET(GRPC_SOURCES_DIR ${THIRD_PARTY_PATH}/grpc)
SET(GRPC_INSTALL_DIR ${THIRD_PARTY_PATH}/install/grpc)
SET(GRPC_INCLUDE_DIR "${GRPC_INSTALL_DIR}/include/" CACHE PATH "grpc include directory." FORCE)
SET(GRPC_CPP_PLUGIN "${GRPC_INSTALL_DIR}/bin/grpc_cpp_plugin" CACHE FILEPATH "GRPC_CPP_PLUGIN" FORCE)
IF(APPLE)
SET(BUILD_CMD make -n | sed "s/-Werror//g" | sh)
ELSE()
SET(BUILD_CMD make)
ENDIF()
ExternalProject_Add(
extern_grpc
DEPENDS protobuf zlib
GIT_REPOSITORY "https://github.com/grpc/grpc.git"
GIT_TAG "v1.7.x"
PREFIX ${GRPC_SOURCES_DIR}
UPDATE_COMMAND ""
CONFIGURE_COMMAND ""
BUILD_IN_SOURCE 1
# NOTE(yuyang18):
# Disable -Werror, otherwise the compile will fail in MacOS.
# It seems that we cannot configure that by make command.
# Just dry run make command and remove `-Werror`, then use a shell to run make commands
BUILD_COMMAND ${BUILD_CMD} HAS_SYSTEM_PROTOBUF=false -s -j8 static grpc_cpp_plugin
INSTALL_COMMAND make prefix=${GRPC_INSTALL_DIR} install
)
# FIXME(typhoonzero): hack to get static lib path, try a better way like merge them.
ADD_LIBRARY(grpc++_unsecure STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET grpc++_unsecure PROPERTY IMPORTED_LOCATION
"${GRPC_INSTALL_DIR}/lib/libgrpc++_unsecure.a")
ADD_LIBRARY(grpc++ STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET grpc++ PROPERTY IMPORTED_LOCATION
"${GRPC_INSTALL_DIR}/lib/libgrpc++.a")
ADD_LIBRARY(gpr STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET gpr PROPERTY IMPORTED_LOCATION
"${GRPC_INSTALL_DIR}/lib/libgpr.a")
ADD_LIBRARY(grpc_unsecure STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET grpc_unsecure PROPERTY IMPORTED_LOCATION
"${GRPC_INSTALL_DIR}/lib/libgrpc_unsecure.a")
include_directories(${GRPC_INCLUDE_DIR})
ADD_DEPENDENCIES(grpc++_unsecure extern_grpc)
cmake/external/mkldnn.cmake
浏览文件 @ 048cc30e
......@@ -40,10 +40,9 @@ INCLUDE_DIRECTORIES(${MKLDNN_INC_DIR})
IF(${CBLAS_PROVIDER} STREQUAL "MKLML")
SET(MKLDNN_DEPENDS ${MKLML_PROJECT})
SET(MKLDNN_MKLROOT ${MKLML_ROOT})
SET(MKLDNN_IOMP_LIB ${MKLML_IOMP_LIB})
SET(MKLDNN_IOMP_DIR ${MKLML_LIB_DIR})
MESSAGE(STATUS "Build MKLDNN with ${MKLDNN_MKLROOT}")
MESSAGE(STATUS "Build MKLDNN with MKLML ${MKLML_ROOT}")
ELSE()
MESSAGE(FATAL_ERROR "Should enable MKLML when build MKLDNN")
ENDIF()
SET(MKLDNN_CFLAG "${CMAKE_C_FLAGS} -Wno-error=strict-overflow")
......@@ -57,15 +56,16 @@ ExternalProject_Add(
PREFIX ${MKLDNN_SOURCES_DIR}
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_INSTALL_PREFIX=${MKLDNN_INSTALL_DIR}
CMAKE_ARGS -DMKLROOT=${MKLDNN_MKLROOT}
CMAKE_ARGS -DMKLROOT=${MKLML_ROOT}
CMAKE_ARGS -DCMAKE_C_FLAGS=${MKLDNN_CFLAG}
CMAKE_ARGS -DCMAKE_CXX_FLAGS=${MKLDNN_CXXFLAG}
CMAKE_CACHE_ARGS -DCMAKE_INSTALL_PREFIX:PATH=${MKLDNN_INSTALL_DIR}
-DMKLROOT:PATH=${MKLDNN_MKLROOT}
-DMKLROOT:PATH=${MKLML_ROOT}
)
ADD_LIBRARY(mkldnn SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET mkldnn PROPERTY IMPORTED_LOCATION ${MKLDNN_LIB})
ADD_DEPENDENCIES(mkldnn ${MKLDNN_PROJECT})
MESSAGE(STATUS "Mkldnn library: ${MKLDNN_LIB}")
MESSAGE(STATUS "MKLDNN library: ${MKLDNN_LIB}")
add_definitions(-DPADDLE_USE_MKLDNN)
LIST(APPEND external_project_dependencies mkldnn)
cmake/external/openblas.cmake
浏览文件 @ 048cc30e
......@@ -45,15 +45,14 @@ IF(NOT ${CBLAS_FOUND})
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV8 BINARY=64 USE_THREAD=0)
ENDIF()
ELSEIF(IOS)
# FIXME(liuyiqun): support multiple architectures
SET(OPENBLAS_COMMIT "b5c96fcfcdc82945502a2303116a64d89985daf5")
SET(OPENBLAS_CC "${OPENBLAS_CC} ${CMAKE_C_FLAGS} -isysroot ${CMAKE_OSX_SYSROOT}")
IF(CMAKE_OSX_ARCHITECTURES MATCHES "armv7")
SET(OPENBLAS_CC "${OPENBLAS_CC} -arch armv7")
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV7 ARM_SOFTFP_ABI=1 USE_THREAD=0)
ELSEIF(CMAKE_OSX_ARCHITECTURES MATCHES "arm64")
IF(CMAKE_OSX_ARCHITECTURES MATCHES "arm64")
SET(OPENBLAS_COMMIT "b5c96fcfcdc82945502a2303116a64d89985daf5")
SET(OPENBLAS_CC "${OPENBLAS_CC} ${CMAKE_C_FLAGS} -isysroot ${CMAKE_OSX_SYSROOT}")
SET(OPENBLAS_CC "${OPENBLAS_CC} -arch arm64")
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV8 BINARY=64 USE_THREAD=0 CROSS_SUFFIX=${CROSS_SUFFIX})
ELSE()
MESSAGE(FATAL_ERROR "OpenBLAS only support arm64 architectures on iOS. "
"You can set IOS_USE_VECLIB_FOR_BLAS=ON or USE_EIGEN_FOR_BLAS=ON to use other blas library instead.")
ENDIF()
ELSEIF(RPI)
# use hardfp
......
cmake/external/protobuf.cmake
浏览文件 @ 048cc30e
......@@ -15,7 +15,18 @@
INCLUDE(ExternalProject)
# Always invoke `FIND_PACKAGE(Protobuf)` for importing function protobuf_generate_cpp
FIND_PACKAGE(Protobuf QUIET)
SET(PROTOBUF_FOUND "OFF")
macro(UNSET_VAR VAR_NAME)
UNSET(${VAR_NAME} CACHE)
UNSET(${VAR_NAME})
endmacro()
UNSET_VAR(PROTOBUF_INCLUDE_DIR)
UNSET_VAR(PROTOBUF_FOUND)
UNSET_VAR(PROTOBUF_PROTOC_EXECUTABLE)
UNSET_VAR(PROTOBUF_PROTOC_LIBRARY)
UNSET_VAR(PROTOBUF_LITE_LIBRARY)
UNSET_VAR(PROTOBUF_LIBRARY)
UNSET_VAR(PROTOBUF_INCLUDE_DIR)
UNSET_VAR(Protobuf_PROTOC_EXECUTABLE)
if(NOT COMMAND protobuf_generate_python) # before cmake 3.4, protobuf_genrerate_python is not defined.
function(protobuf_generate_python SRCS)
......@@ -110,7 +121,6 @@ macro(PROMPT_PROTOBUF_LIB)
# 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 ${dep})
......@@ -128,11 +138,11 @@ endmacro()
set(PROTOBUF_ROOT "" CACHE PATH "Folder contains protobuf")
if (NOT "${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)
find_path(PROTOBUF_INCLUDE_DIR google/protobuf/message.h PATHS ${PROTOBUF_ROOT}/include NO_DEFAULT_PATH)
find_library(PROTOBUF_LIBRARY protobuf PATHS ${PROTOBUF_ROOT}/lib NO_DEFAULT_PATH)
find_library(PROTOBUF_LITE_LIBRARY protobuf-lite PATHS ${PROTOBUF_ROOT}/lib NO_DEFAULT_PATH)
find_library(PROTOBUF_PROTOC_LIBRARY protoc PATHS ${PROTOBUF_ROOT}/lib NO_DEFAULT_PATH)
find_program(PROTOBUF_PROTOC_EXECUTABLE protoc PATHS ${PROTOBUF_ROOT}/bin NO_DEFAULT_PATH)
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()
......
cmake/external/warpctc.cmake
浏览文件 @ 048cc30e
......@@ -12,6 +12,10 @@
# See the License for the specific language governing permissions and
# limitations under the License.
IF(MOBILE_INFERENCE)
return()
ENDIF()
INCLUDE(ExternalProject)
SET(WARPCTC_SOURCES_DIR ${THIRD_PARTY_PATH}/warpctc)
......
cmake/external/zlib.cmake
浏览文件 @ 048cc30e
......@@ -50,6 +50,8 @@ ExternalProject_Add(
)
LIST(APPEND external_project_dependencies zlib)
ADD_LIBRARY(zlib_target STATIC IMPORTED GLOBAL)
SET_PROPERTY(TARGET zlib_target PROPERTY IMPORTED_LOCATION ${ZLIB_LIBRARIES})
IF(WITH_C_API)
INSTALL(DIRECTORY ${ZLIB_INCLUDE_DIR} DESTINATION third_party/zlib)
......
cmake/flags.cmake
浏览文件 @ 048cc30e
......@@ -149,58 +149,3 @@ endforeach()
foreach(flag ${GPU_COMMON_FLAGS})
safe_set_nvflag(${flag})
endforeach()
set(CUDA_PROPAGATE_HOST_FLAGS OFF)
# Release/Debug flags set by cmake. Such as -O3 -g -DNDEBUG etc.
# So, don't set these flags here.
LIST(APPEND CUDA_NVCC_FLAGS -std=c++11)
LIST(APPEND CUDA_NVCC_FLAGS --use_fast_math)
if(CMAKE_BUILD_TYPE STREQUAL "Debug")
LIST(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_DEBUG})
elseif(CMAKE_BUILD_TYPE STREQUAL "Release")
LIST(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_RELEASE})
elseif(CMAKE_BUILD_TYPE STREQUAL "RelWithDebInfo")
LIST(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_RELWITHDEBINFO})
elseif(CMAKE_BUILD_TYPE STREQUAL "MinSizeRel")
LIST(APPEND CUDA_NVCC_FLAGS ${CMAKE_CXX_FLAGS_MINSIZEREL})
endif()
function(specify_cuda_arch cuda_version cuda_arch)
if(${cuda_version} VERSION_GREATER "8.0")
foreach(capability 61 62)
if(${cuda_arch} STREQUAL ${capability})
list(APPEND __arch_flags " -gencode arch=compute_${cuda_arch},code=sm_${cuda_arch}")
endif()
endforeach()
elseif(${cuda_version} VERSION_GREATER "7.0" and ${cuda_arch} STREQUAL "53")
list(APPEND __arch_flags " -gencode arch=compute_${cuda_arch},code=sm_${cuda_arch}")
endif()
endfunction()
# Common gpu architectures: Kepler, Maxwell
foreach(capability 30 35 50)
list(APPEND __arch_flags " -gencode arch=compute_${capability},code=sm_${capability}")
endforeach()
if (CUDA_VERSION VERSION_GREATER "7.0" OR CUDA_VERSION VERSION_EQUAL "7.0")
list(APPEND __arch_flags " -gencode arch=compute_52,code=sm_52")
endif()
# Modern gpu architectures: Pascal
if (CUDA_VERSION VERSION_GREATER "8.0" OR CUDA_VERSION VERSION_EQUAL "8.0")
list(APPEND __arch_flags " -gencode arch=compute_60,code=sm_60")
list(APPEND CUDA_NVCC_FLAGS --expt-relaxed-constexpr)
endif()
# Custom gpu architecture
set(CUDA_ARCH)
if(CUDA_ARCH)
specify_cuda_arch(${CUDA_VERSION} ${CUDA_ARCH})
endif()
set(CUDA_NVCC_FLAGS ${__arch_flags} ${CUDA_NVCC_FLAGS})
cmake/generic.cmake
浏览文件 @ 048cc30e
......@@ -459,11 +459,58 @@ function(py_test TARGET_NAME)
if(WITH_TESTING)
set(options STATIC static SHARED shared)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(py_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
set(multiValueArgs SRCS DEPS ARGS)
cmake_parse_arguments(py_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
add_test(NAME ${TARGET_NAME}
COMMAND env PYTHONPATH=${PADDLE_PYTHON_BUILD_DIR}/lib-python
python2 ${py_test_SRCS}
${PYTHON_EXECUTABLE} -u ${py_test_SRCS} ${py_test_ARGS}
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
endif()
endfunction()
# grpc_library generate grpc code using grpc_cpp_plugin and protoc
# then build the generated protobuf code and grpc code with your
# implementation source codes together. Use SRCS argument for your
# implementation source files and PROTO argument for your .proto
# files.
#
# Usage: grpc_library(my_target SRCS my_client.cc PROTO my_target.proto DEPS my_dep)
function(grpc_library TARGET_NAME)
set(oneValueArgs PROTO)
set(multiValueArgs SRCS DEPS)
set(options "")
cmake_parse_arguments(grpc_library "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
message(STATUS "generating grpc ${grpc_library_PROTO}")
get_filename_component(ABS_PROTO ${grpc_library_PROTO} ABSOLUTE)
get_filename_component(PROTO_WE ${grpc_library_PROTO} NAME_WE)
get_filename_component(PROTO_PATH ${ABS_PROTO} PATH)
protobuf_generate_cpp(grpc_proto_srcs grpc_proto_hdrs "${ABS_PROTO}")
set(grpc_grpc_srcs "${CMAKE_CURRENT_BINARY_DIR}/${PROTO_WE}.grpc.pb.cc")
set(grpc_grpc_hdrs "${CMAKE_CURRENT_BINARY_DIR}/${PROTO_WE}.grpc.pb.h")
cc_library("${TARGET_NAME}_proto" SRCS "${grpc_proto_srcs}")
add_custom_command(
OUTPUT "${grpc_grpc_srcs}" "${grpc_grpc_hdrs}"
COMMAND ${PROTOBUF_PROTOC_EXECUTABLE}
ARGS --grpc_out "${CMAKE_CURRENT_BINARY_DIR}" -I "${PROTO_PATH}"
--plugin=protoc-gen-grpc="${GRPC_CPP_PLUGIN}" "${ABS_PROTO}"
DEPENDS "${ABS_PROTO}" ${PROTOBUF_PROTOC_EXECUTABLE} extern_grpc)
# FIXME(typhoonzero): grpc generated code do not generate virtual-dtor, mark it
# as compiler warnings instead of error. Should try remove the warnings also.
set_source_files_properties(
${grpc_grpc_srcs}
PROPERTIES
COMPILE_FLAGS "-Wno-error=non-virtual-dtor -Wno-error=delete-non-virtual-dtor")
cc_library("${TARGET_NAME}_grpc" SRCS "${grpc_grpc_srcs}")
set_source_files_properties(
${grpc_library_SRCS}
PROPERTIES
COMPILE_FLAGS "-Wno-error=non-virtual-dtor -Wno-error=delete-non-virtual-dtor")
cc_library("${TARGET_NAME}" SRCS "${grpc_library_SRCS}" DEPS "${TARGET_NAME}_grpc" "${TARGET_NAME}_proto" "${grpc_library_DEPS}")
endfunction()
cmake/util.cmake
浏览文件 @ 048cc30e
......@@ -115,8 +115,8 @@ function(link_paddle_exe TARGET_NAME)
target_link_libraries(${TARGET_NAME} log)
endif(ANDROID)
if(WITH_MKLDNN AND WITH_MKLML AND MKLDNN_IOMP_DIR)
target_link_libraries(${TARGET_NAME} "-L${MKLDNN_IOMP_DIR} -liomp5 -Wl,--as-needed")
if(WITH_MKLML AND MKLML_LIB_DIR AND MKLML_IOMP_LIB)
target_link_libraries(${TARGET_NAME} "-L${MKLML_LIB_DIR} -liomp5 -Wl,--as-needed")
endif()
add_dependencies(${TARGET_NAME} ${external_project_dependencies})
......@@ -168,17 +168,3 @@ function(create_resources res_file output_file)
COMMAND python ARGS ${PADDLE_SOURCE_DIR}/cmake/make_resource.py ${res_file} ${output_file}
DEPENDS ${res_file} ${PADDLE_SOURCE_DIR}/cmake/make_resource.py)
endfunction()
# Create a python unittest using run_python_tests.sh,
# which takes care of making correct running environment
function(add_python_test TEST_NAME)
foreach(arg ${ARGN})
get_filename_component(py_fn ${arg} NAME_WE)
set(TRG_NAME ${TEST_NAME}_${py_fn})
add_test(NAME ${TRG_NAME}
COMMAND env PYTHONPATH=${PADDLE_PYTHON_PACKAGE_DIR}
python2 ${arg}
WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
endforeach()
endfunction()
doc/api/v2/config/layer.rst
浏览文件 @ 048cc30e
......@@ -54,7 +54,7 @@ img_conv
.. _api_v2.layer_context_projection:
context_projection
context_projection
------------------
.. autoclass:: paddle.v2.layer.context_projection
:noindex:
......@@ -70,7 +70,7 @@ Image Pooling Layer
img_pool
--------
.. autoclass:: paddle.v2.layer.img_pool
:noindex:
:noindex:
spp
---
......@@ -104,7 +104,7 @@ sum_to_one_norm
---------------
.. autoclass:: paddle.v2.layer.sum_to_one_norm
:noindex:
cross_channel_norm
------------------
.. autoclass:: paddle.v2.layer.cross_channel_norm
......@@ -114,7 +114,7 @@ row_l2_norm
-----------
.. autoclass:: paddle.v2.layer.row_l2_norm
:noindex:
Recurrent Layers
================
......@@ -335,6 +335,16 @@ bilinear_interp
.. autoclass:: paddle.v2.layer.bilinear_interp
:noindex:
dot_prod
---------
.. autoclass:: paddle.v2.layer.dot_prod
:noindex:
out_prod
--------
.. autoclass:: paddle.v2.layer.out_prod
:noindex:
power
-----
.. autoclass:: paddle.v2.layer.power
......@@ -372,6 +382,11 @@ cos_sim
.. autoclass:: paddle.v2.layer.cos_sim
:noindex:
l2_distance
-----------
.. autoclass:: paddle.v2.layer.l2_distance
:noindex:
trans
-----
.. autoclass:: paddle.v2.layer.trans
......@@ -400,6 +415,13 @@ multiplex
.. autoclass:: paddle.v2.layer.multiplex
:noindex:
Factorization Machine Layer
============================
factorization_machine
---------------------
.. autoclass:: paddle.v2.layer.factorization_machine
:noindex:
Slicing and Joining Layers
==========================
......
doc/design/float16.md
浏览文件 @ 048cc30e
......@@ -28,6 +28,51 @@ The goal of float16 is to serve as a key for the executor to find and run the co
- [Eigen](https://github.com/RLovelett/eigen) >= 3.3 supports float16 calculation on both GPU and CPU using the `Eigen::half` class. It is mostly useful for Nvidia GPUs because of the overloaded arithmetic operators using cuda intrinsics. It falls back to using software emulation on CPU for calculation and there is no special treatment to ARM processors.
- [ARM compute library](https://github.com/ARM-software/ComputeLibrary) >= 17.02.01 supports NEON FP16 kernels (requires ARMv8.2-A CPU).
### CUDA version issue
There are currently three versions of CUDA that supports `__half` data type, namely, CUDA 7.5, 8.0, and 9.0.
CUDA 7.5 and 8.0 define `__half` as a simple struct that has a `uint16_t` data (see [`cuda_fp16.h`](https://github.com/ptillet/isaac/blob/9212ab5a3ddbe48f30ef373f9c1fb546804c7a8c/include/isaac/external/CUDA/cuda_fp16.h)) as follows:
```
typedef struct __align__(2) {
unsigned short x;
} __half;
typedef __half half;
```
This struct does not define any overloaded arithmetic operators. So you have to directly use `__hadd` instead of `+` to correctly add two half types:
```
__global__ void Add() {
half a, b, c;
c = __hadd(a, b); // correct
c = a + b; // compiler error: no operator "+" matches these operands
}
```
CUDA 9.0 provides a major update to the half data type. The related code can be found in the updated [`cuda_fp16.h`](https://github.com/ptillet/isaac/blob/master/include/isaac/external/CUDA/cuda_fp16.h) and the newly added [`cuda_fp16.hpp`](https://github.com/ptillet/isaac/blob/master/include/isaac/external/CUDA/cuda_fp16.hpp).
Essentially, CUDA 9.0 renames the original `__half` type in 7.5 and 8.0 as `__half_raw`, and defines a new `__half` class type that has constructors, conversion operators, and also provides overloaded arithmetic operators such as follows:
```
typedef struct __CUDA_ALIGN__(2) {
unsigned short x;
} __half_raw;
struct __CUDA_ALIGN__(2) __half {
protected:
unsigned short __x;
public:
// constructors and conversion operators from/to
// __half_raw and other built-in data types
}
typedef __half half;
__device__ __forceinline__
__half operator+(const __half &lh, const __half &rh) {
return __hadd(lh, rh);
}
// Other overloaded operators
```
This new design makes `c = a + b` work correctly for CUDA half data type.
## Implementation
The float16 class holds a 16-bit `uint16_t` data internally.
......
doc/design/mkldnn/README.MD
浏览文件 @ 048cc30e
......@@ -36,13 +36,13 @@ Figure 1. PaddlePaddle on IA.
我们把集成方案大致分为了如下几个方面。
### CMake
我们会在`CMakeLists.txt`中会添加`WITH_MKLDNN`的选项，当设置这个值为`ON`的时候会启用编译MKL-DNN功能。同时会自动开启OpenMP用于提高MKL-DNN的性能。
我们会在`CMakeLists.txt`中会给用户添加一个`WITH_MKL`的开关，他是负责`WITH_MKLML`和`WITH_MKLDNN`的总开关。
同时，我们会引入`WITH_MKLML`选项，用于选择是否使用MKL-DNN自带的MKLML安装包。这个安装包可以独立于MKL-DNN使用，但是建议在开启MKL-DNN的同时也打开MKLML的开关，这样才能发挥最好的性能。
当打开`WITH_MKL`时，会开启MKLML的功能，作为PaddlePaddle的CBLAS和LAPACK库，同时会开启Intel OpenMP用于提高MKLML的性能。 如果系统支持AVX2指令集及以上，同时会开启MKL-DNN功能。
所以，我们会在`cmake/external`目录新建`mkldnn.cmake`和`mklml.cmake`文件，它们会在编译PaddlePaddle的时候下载对应的软件包，并放到PaddlePaddle的third party目录中。
当关闭`WITH_MKL`时，MKLML和MKL-DNN功能会同时关闭。
**备注**：当`WITH_MKLML=ON`的时候，会优先使用这个包作为PaddlePaddle的CBLAS和LAPACK库，所以会稍微改动`cmake/cblas.cmake`中的逻辑。
所以，我们会在`cmake/external`目录新建`mkldnn.cmake`和`mklml.cmake`文件，它们会在编译PaddlePaddle的时候下载对应的软件包，并放到PaddlePaddle的third party目录中。
### Layers
所有MKL-DNN相关的C++ layers，都会按照PaddlePaddle的目录结构存放在
......
doc/design/reader/README.md
浏览文件 @ 048cc30e
# Python Data Reader Design Doc
At training and testing time, PaddlePaddle programs need to read data. To ease the users' work to write data reading code, we define that
During the training and testing phases, PaddlePaddle programs need to read data. To help the users write code that performs reading input data, we define the following:
- A *reader* is a function that reads data (from file, network, random number generator, etc) and yields data items.
- A *reader creator* is a function that returns a reader function.
- A *reader decorator* is a function, which accepts one or more readers, and returns a reader.
- A *batch reader* is a function that reads data (from *reader*, file, network, random number generator, etc) and yields a batch of data items.
- A *reader*: A function that reads data (from file, network, random number generator, etc) and yields the data items.
- A *reader creator*: A function that returns a reader function.
- A *reader decorator*: A function, which takes in one or more readers, and returns a reader.
- A *batch reader*: A function that reads data (from *reader*, file, network, random number generator, etc) and yields a batch of data items.
and provide function which converts reader to batch reader, frequently used reader creators and reader decorators.
and also provide a function which can convert a reader to a batch reader, frequently used reader creators and reader decorators.
## Data Reader Interface
Indeed, *data reader* doesn't have to be a function that reads and yields data items. It can be any function with no parameter that creates a iterable (anything can be used in `for x in iterable`):
*Data reader* doesn't have to be a function that reads and yields data items. It can just be any function without any parameters that creates an iterable (anything can be used in `for x in iterable`) as follows:
```
iterable = data_reader()
```
Element produced from the iterable should be a **single** entry of data, **not** a mini batch. That entry of data could be a single item, or a tuple of items. Item should be of [supported type](http://www.paddlepaddle.org/doc/ui/data_provider/pydataprovider2.html?highlight=dense_vector#input-types) (e.g., numpy 1d array of float32, int, list of int)
The item produced from the iterable should be a **single** entry of data and **not** a mini batch. The entry of data could be a single item or a tuple of items. Item should be of one of the [supported types](http://www.paddlepaddle.org/doc/ui/data_provider/pydataprovider2.html?highlight=dense_vector#input-types) (e.g., numpy 1d array of float32, int, list of int etc.)
An example implementation for single item data reader creator:
An example implementation for single item data reader creator is as follows:
```python
def reader_creator_random_image(width, height):
......@@ -29,7 +29,7 @@ def reader_creator_random_image(width, height):
return reader
```
An example implementation for multiple item data reader creator:
An example implementation for multiple item data reader creator is as follows:
```python
def reader_creator_random_image_and_label(width, height, label):
def reader():
......@@ -40,9 +40,10 @@ def reader_creator_random_image_and_label(width, height, label):
## Batch Reader Interface
*batch reader* can be any function with no parameter that creates a iterable (anything can be used in `for x in iterable`). The output of the iterable should be a batch (list) of data items. Each item inside the list must be a tuple.
*Batch reader* can be any function without any parameters that creates an iterable (anything can be used in `for x in iterable`). The output of the iterable should be a batch (list) of data items. Each item inside the list should be a tuple.
Here are some valid outputs:
Here are valid outputs:
```python
# a mini batch of three data items. Each data item consist three columns of data, each of which is 1.
[(1, 1, 1),
......@@ -58,20 +59,22 @@ Here are valid outputs:
Please note that each item inside the list must be a tuple, below is an invalid output:
```python
# wrong, [1,1,1] needs to be inside a tuple: ([1,1,1],).
# Otherwise it's ambiguous whether [1,1,1] means a single column of data [1, 1, 1],
# or three column of datas, each of which is 1.
# Otherwise it is ambiguous whether [1,1,1] means a single column of data [1, 1, 1],
# or three columns of data, each of which is 1.
[[1,1,1],
[2,2,2],
[3,3,3]]
```
It's easy to convert from reader to batch reader:
It is easy to convert from a reader to a batch reader:
```python
mnist_train = paddle.dataset.mnist.train()
mnist_train_batch_reader = paddle.batch(mnist_train, 128)
```
Also easy to create custom batch reader:
It is also straight forward to create a custom batch reader:
```python
def custom_batch_reader():
while True:
......@@ -85,7 +88,8 @@ mnist_random_image_batch_reader = custom_batch_reader
## Usage
batch reader, mapping from item(s) read to data layer, batch size and number of total pass will be passed into `paddle.train`:
Following is how we can use the reader with PaddlePaddle:
The batch reader, a mapping from item(s) to data layer, the batch size and the number of total passes will be passed into `paddle.train` as follows:
```python
# two data layer is created:
......@@ -99,13 +103,13 @@ paddle.train(batch_reader, {"image":0, "label":1}, 128, 10, ...)
## Data Reader Decorator
*Data reader decorator* takes a single or multiple data reader, returns a new data reader. It is similar to a [python decorator](https://wiki.python.org/moin/PythonDecorators), but it does not use `@` syntax.
The *Data reader decorator* takes in a single reader or multiple data readers and returns a new data reader. It is similar to a [python decorator](https://wiki.python.org/moin/PythonDecorators), but it does not use `@` in the syntax.
Since we have a strict interface for data readers (no parameter, return a single data item). Data reader can be used flexiable via data reader decorators. Following are a few examples:
Since we have a strict interface for data readers (no parameters and return a single data item), a data reader can be used in a flexible way using data reader decorators. Following are a few examples:
### Prefetch Data
Since reading data may take time and training can not proceed without data. It is generally a good idea to prefetch data.
Since reading data may take some time and training can not proceed without data, it is generally a good idea to prefetch the data.
Use `paddle.reader.buffered` to prefetch data:
......@@ -117,9 +121,9 @@ buffered_reader = paddle.reader.buffered(paddle.dataset.mnist.train(), 100)
### Compose Multiple Data Readers
For example, we want to use a source of real images (reusing mnist dataset), and a source of random images as input for [Generative Adversarial Networks](https://arxiv.org/abs/1406.2661).
For example, if we want to use a source of real images (say reusing mnist dataset), and a source of random images as input for [Generative Adversarial Networks](https://arxiv.org/abs/1406.2661).
We can do:
We can do the following :
```python
def reader_creator_random_image(width, height):
......@@ -139,13 +143,13 @@ false_reader = reader_creator_bool(False)
reader = paddle.reader.compose(paddle.dataset.mnist.train(), data_reader_creator_random_image(20, 20), true_reader, false_reader)
# Skipped 1 because paddle.dataset.mnist.train() produces two items per data entry.
# And we don't care second item at this time.
# And we don't care about the second item at this time.
paddle.train(paddle.batch(reader, 128), {"true_image":0, "fake_image": 2, "true_label": 3, "false_label": 4}, ...)
```
### Shuffle
Given shuffle buffer size `n`, `paddle.reader.shuffle` will return a data reader that buffers `n` data entries and shuffle them before a data entry is read.
Given the shuffle buffer size `n`, `paddle.reader.shuffle` returns a data reader that buffers `n` data entries and shuffles them before a data entry is read.
Example:
```python
......@@ -154,21 +158,21 @@ reader = paddle.reader.shuffle(paddle.dataset.mnist.train(), 512)
## Q & A
### Why reader return only a single entry, but not a mini batch?
### Why does a reader return only a single entry, and not a mini batch?
Always returning a single entry make reusing existing data readers much easier (e.g., if existing reader return not a single entry but 3 entries, training code will be more complex because it need to handle cases like batch size 2).
Returning a single entry makes reusing existing data readers much easier (for example, if an existing reader returns 3 entries instead if a single entry, the training code will be more complicated because it need to handle cases like a batch size 2).
We provide function `paddle.batch` to turn (single entry) reader into batch reader.
We provide a function: `paddle.batch` to turn (a single entry) reader into a batch reader.
### Why do we need batch reader, isn't train take reader and batch_size as arguments sufficient?
### Why do we need a batch reader, isn't is sufficient to give the reader and batch_size as arguments during training ?
In most of the case, train taking reader and batch_size as arguments would be sufficent. However sometimes user want to customize order of data entries inside a mini batch. Or even change batch size dynamically.
In most of the cases, it would be sufficient to give the reader and batch_size as arguments to the train method. However sometimes the user wants to customize the order of data entries inside a mini batch, or even change the batch size dynamically. For these cases using a batch reader is very efficient and helpful.
### Why use a dictionary but not a list to provide mapping?
### Why use a dictionary instead of a list to provide mapping?
We decided to use dictionary (`{"image":0, "label":1}`) instead of list (`["image", "label"]`) is because that user can easily resue item (e.g., using `{"image_a":0, "image_b":0, "label":1}`) or skip item (e.g., using `{"image_a":0, "label":2}`).
Using a dictionary (`{"image":0, "label":1}`) instead of a list (`["image", "label"]`) gives the advantage that the user can easily reuse the items (e.g., using `{"image_a":0, "image_b":0, "label":1}`) or even skip an item (e.g., using `{"image_a":0, "label":2}`).
### How to create custom data reader creator
### How to create a custom data reader creator ?
```python
def image_reader_creator(image_path, label_path, n):
......@@ -192,7 +196,7 @@ paddle.train(paddle.batch(reader, 128), {"image":0, "label":1}, ...)
### How is `paddle.train` implemented
An example implementation of paddle.train could be:
An example implementation of paddle.train is:
```python
def train(batch_reader, mapping, batch_size, total_pass):
......
doc/design/refactor/distributed_architecture.md
浏览文件 @ 048cc30e
......@@ -2,106 +2,70 @@
## Abstract
PaddlePaddle v0.10.0 uses the "trainer-parameter server"
architecture. We run multiple replicated instances of trainers (runs
the same code written by the user) and parameter servers for
distributed training. This architecture served us well, but has some
limitations:
PaddlePaddle version 0.10.0 uses the "trainer-parameter server" architecture. We run multiple instances of trainers (where each trainer runs the same model) and parameter servers for distributed training. This architecture serves well, but has few limitations:
1. Need to write special code to handle tasks which should only be run
by a single trainer. E.g., initializing model and saving model.
1. There is a need to write special code that handles tasks which should only be run on a single trainer. E.g., initializing the model, saving the model etc.
2. Model parallelism is hard: need to write if-else branches conditioned
on the trainer ID to partition model onto each trainer, and manually
write the inter-model-shard communication code.
2. Model parallelism is hard: It would need all the if-else branches conditioned on the trainer ID to partition the model onto the trainers, and eventually manually writing out the inter-model-shard communication code to communicate between different trainers.
3. The user can not directly specify the parameter update rule: need
to modify the parameter server C++ code and compile a new
binary. This adds complication for researchers: A lot of extra
effort is required. Besides, the training job submission program
may not allow running arbitrary binaries.
3. The user can not directly specify the parameter update rule: This would need to modify the parameter server code and compile a new binary. This makes things more complicated for researchers: A lot of extra effort is required to make this work. Besides, the training job submission program may not allow running arbitrary binaries.
This design doc discusses PaddlePaddle's new distributed training
architecture that addresses the above limitations.
This design doc discusses PaddlePaddle's new distributed training architecture that addresses the above mentioned limitations.
## Analysis
We will assume the user writes the trainer program by Python, the same
analysis holds if the trainer program is written in C++.
The assumption is that the user writes the trainer program in either Python or C++.
### Limitation 1
If we look at the Python code that the user writes, there are two
kinds of functionalities:
There are two basic functionalities in the trainer program:
- The training logic such as load / save model and print log.
- The neural network definition such as the definition of the data
layer, the fully connected layer, the cost function and the
1. The training logic such as loading / saving the model and printing out the logs.
2. The neural network definition such as the definition of the data layer, the fully connected layer, the cost function and the
optimizer.
When we training with PaddlePaddle v0.10.0 distributedly, multiple
replicated Python instances are running on different nodes: both the
training logic and the neural network computation is replicated.
When we train using PaddlePaddle v0.10.0 in a distributed fashion, multiple instances of the same Python code are run on different nodes, hence both: the
training logic as well as the neural network computation logic, is replicated.
The tasks that should only run once all belong to the training logic,
if we only replicate the neural network computation, but do **not**
replicate the training logic, the limitation could be solved.
The tasks that only need to be run once belong to the training logic. Hence if we only replicate the neural network computation part, and do **not**
replicate the training logic, the limitation mentioned above can be avoided.
### Limitation 2
Model parallelism means running a single model on multiple nodes by
partitioning the model onto different nodes and managing the
inter-model-shard communications.
Model parallelism means that a single model is partitioned into different components and each node runs one of the component separately. This comes at the extra cost of managing the
inter-model-shard communication between nodes.
PaddlePaddle should be able to modify the nerual network computation
definition to support model parallelism automatically. However, the
computation is only specified in Python code, and PaddlePaddle can not
modify Python code.
PaddlePaddle should ideally be able to modify the neural network computation and figure out the support for model parallelism automatically. However, the
computation is only specified in Python code which sits outside of PaddlePaddle, hence PaddlePaddle can not support the feature in this setup.
Just like compiler uses a intermediate representation (IR) so that
programmer does not need to manually optimize their code in most of
the cases - the compiler will optimize the IR:
Similar to how a compiler uses an intermediate representation (IR) so that the programmer does not need to manually optimize their code for most of the cases, we can have an intermediate representation in PaddlePaddle as well. The compiler optimizes the IR as follows:
<img src="src/compiler.png"/>
We can have our own IR too: PaddlePaddle can support model parallel by
converting the IR so the user no longer need to manually do it in
Python:
PaddlePaddle can support model parallelism by converting the IR so that the user no longer needs to manually perform the computation and operations in the Python component:
<img src="src/paddle-compile.png"/>
The IR for PaddlePaddle after refactor is called `Block`, it specifies
the computation dependency graph and the variables used in the
computation.
The IR for PaddlePaddle after refactoring is called a `Block`, it specifies the computation dependency graph and the variables used in the computation.
### Limitation 3
The user can not directly specify the parameter update rule for the
parameter server because the parameter server does not use the same
computation definition as the trainer. Instead, the update rule is
baked in the parameter server. The user can not specify the update
rule in the same way of specifying the trainer computation.
The user can not directly specify the parameter update rule for the parameter server in the Python module, since the parameter server does not use the same computation definition as the trainer. Instead, the update rule is baked inside the parameter server. The user can not specify the update rule explicitly.
This could be fixed by making the parameter server run the same
computation definition as the trainer. For a detailed explanation,
please
see
This could be fixed by making the parameter server run the same computation definition as the trainer (the user's Python module). For a detailed explanation, refer to this document -
[Design Doc: Operation Graph Based Parameter Server](./dist_train.md)
## Distributed Training Architecture
The new distributed training architecture can address the above
limitations. Below is the illustration:
The revamped distributed training architecture can address the above discussed limitations. Below is the illustration of how it does so:
<img src="src/distributed_architecture.png"/>
The architecture includes major components: *PaddlePaddle Python*,
*PaddlePaddle converter* and *PaddlePaddle runtime*:
The major components in the architecture are: *PaddlePaddle Python*, *PaddlePaddle converter* and *PaddlePaddle runtime*.
### PaddlePaddle Python
PaddlePaddle Python is the Python library that user's Python trainer
invoke to build the neural network topology, start training, etc.
PaddlePaddle Python is the Python library that user's Python code invokes, to read the data. build the neural network topology, start training, etc.
```Python
paddle.init()
......@@ -117,102 +81,60 @@ for i in range(1000):
print cost_val
```
The code above is a typical Python trainer code, the neural network
topology is built using helper functions such as
`paddle.layer.fc`. The training is done by calling `session.eval`
iteratively.
The above code is what a typical Python trainer code is, the neural network topology is built using the helper functions such as `paddle.layer.fc`. Training is done by calling `session.eval` iteratively.
#### session.eval
As shown in the graph, `session.eval` sends the IR and the evaluation
inputs/targets to the PaddlePaddle cluster for evaluation. The
targets can be any variable in the computation graph. When the target
is the `optimizer` variable, the neural network will be optimized
once. When the target is the `cost` variable, `session.eval` returns
the cost value.
As shown in the graph, `session.eval` sends the IR and the evaluation inputs or targets to the PaddlePaddle cluster for evaluation.
The targets can be any variable in the computation graph. When the target is say, the `optimizer` variable, the neural network will be optimized once. When the target is the `cost` variable, `session.eval` returns the cost value. Based on what the target is, an appropriate action is taken.
The Python `session` is a wrapper of the C++ `Session` class. For more
information about `Session`, please
see [Design Doc: Session](./session.md).
The Python `session` is a wrapper of the C++ `Session` class. For more information about `Session`, refer to this document - [Design Doc: Session](./session.md).
### PaddlePaddle Converter
PaddlePaddle converter automatically converts the IR in the request
(IR and evaluation inputs/targets) from PaddlePaddle Python to new
partitioned IRs and dispatch the new IRs and evaluation inputs/targets
to different PaddlePaddle runtimes. Below are the steps:
The PaddlePaddle converter automatically converts the IR in the request (IR and evaluation inputs/targets) from PaddlePaddle Python to partitioned IRs and dispatches the new IRs and evaluation inputs/targets to different PaddlePaddle runtimes. Below are the steps that are followed :
1. Add `feed` OP that feeds the eval inputs, and `fetch` OP that
fetches the eval targets to the IR.
1. Add a `feed` OP that feeds the eval inputs, and a `fetch` OP that fetches the eval targets to the IR.
1. Extract a new computation (sub)graph with `feed` and `fetch` OP as
the boundary. The runtime does not need to run the OP that is not
dependent by the `fetch` OP.
2. Extract a new computation (sub)graph with the `feed` and `fetch` OPs as the boundary. The runtime does not need to run the OP that is not dependent on the `fetch` OP.
1. Optimizes the computation graph.
3. Optimize the computation graph.
1. Place the OPs in the graph onto different devices on different
PaddlePaddle runtime according to a placement algorithm and device
constraint specified by the user.
4. Place the OPs in the graph onto different devices on different PaddlePaddle runtime according to a placement algorithm and the device constraints specified by the user.
1. Partition the graph according to runtime boundaries and add `send` /
`recv` OP pair on the runtime boundaries.
5. Partition the graph according to runtime boundaries and add `send` / `recv` OP pair on the runtime boundaries.
1. Dispatch the partitioned graph to different PaddlePaddle runtimes.
6. Dispatch the partitioned graph to different PaddlePaddle runtimes.
7. PaddlePaddle runtimes with the `fetch` OP reports evaluation results back to the converter, the converter reports the evaluation results back to the PaddlePaddle Python.
1. PaddlePaddle runtimes with the `fetch` OP reports evaluation
results back to the converter, the convert reports the evaluation
results back to the PaddlePaddle Python.
The output IRs will be cached to optimize the conversion latency.
#### Placement Algorithm
Our first implementation will only support "trainer-parameter server"
placement: the parameters, initializers, and optimizers are placed on
the PaddlePaddle runtimes with the parameter server role. And
everything else will be placed on the PaddlePaddle runtimes with the
trainer role. This has the same functionality of our
"trainer-parameter server" architecture of PaddlePaddle v0.10.0, but
is more general and flexible.
Our first implementation will only support "trainer-parameter server" placement: the parameters, initializers, and optimizers are all placed on the PaddlePaddle runtimes with the parameter server role. Everything else will be placed on the PaddlePaddle runtimes with the trainer role. This has the same functionality as the "trainer-parameter server" architecture of PaddlePaddle v0.10.0, but is more generic and flexible.
In the future, we will implement the general placement algorithm,
which makes placements according to the input IR, and a model of
device computation time and device communication time. Model
parallelism requires the general placement algorithm.
In the future, a more general placement algorithm should be implemented, which makes placements according to the input IR, and a model of device computation time and device communication time. Model parallelism requires the generic placement algorithm.
### PaddlePaddle Runtime
The PaddlePaddle runtime owns multiple devices (e.g., CPUs, GPUs) and
runs the IR. The runtime does not need to do OP placement since it's
already done by the converter.
The PaddlePaddle runtime owns multiple devices (e.g., CPUs, GPUs) and runs the IR. The runtime does not need to do OP placement since it is already done by the converter.
### Local Training Architecture
The local training architecture will be the same as the distributed
training architecture, the differences are everything runs locally,
and there is just one PaddlePaddle runtime:
The local training architecture will be the same as the distributed training architecture, the difference is that everything runs locally, and there is just one PaddlePaddle runtime:
<img src="src/local_architecture.png"/>
### Training Data
In PaddlePaddle v0.10.0, training data is typically read
with [data reader](../reader/README.md) from Python. This approach is
no longer efficient when training distributedly since the Python
process no longer runs on the same node with the trainer processes,
the Python reader will need to read from the distributed filesystem
(assuming it has the access) and send to the trainers, doubling the
network traffic.
When doing distributed training, the user can still use Python data
reader: the training data are sent with `session.eval`. However should
be used for debugging purpose only. The users are encouraged to use
the read data OPs.
In PaddlePaddle v0.10.0, training data is typically read with a [data reader](../reader/README.md) from Python. This approach is no longer efficient when training in a distributed fashion since the Python process no longer runs on the same node with the trainer processes. The Python reader will need to read from the distributed filesystem (assuming it has the required access) and send to the trainers, doubling the network traffic.
When doing distributed training, the user can still use Python data reader: the training data are sent with `session.eval`. However this should be used for debugging purpose only. The users are encouraged to use the read data OPs.
## References:
......
doc/getstarted/basic_usage/index_cn.rst 已删除 100644 → 0
浏览文件 @ 2488b81f
经典的线性回归任务
==================
PaddlePaddle是源于百度的一个深度学习平台。这份简短的介绍将向你展示如何利用PaddlePaddle来解决一个经典的线性回归问题。
任务简介
--------
我们展示如何用PaddlePaddle解决 `单变量的线性回归 <https://www.baidu.com/s?wd=单变量线性回归>`_ 问题。线性回归的输入是一批点 `(x, y)` ，其中 `y = wx + b + ε`， 而 ε 是一个符合高斯分布的随机变量。线性回归的输出是从这批点估计出来的参数 `w` 和 `b` 。
一个例子是房产估值。我们假设房产的价格（y）是其大小（x）的一个线性函数，那么我们可以通过收集市场上房子的大小和价格，用来估计线性函数的参数w 和 b。
准备数据
-----------
假设变量 `x` 和 `y` 的真实关系为： `y = 2x + 0.3 + ε`，这里展示如何使用观测数据来拟合这一线性关系。首先，Python代码将随机产生2000个观测点，作为线性回归的输入。下面脚本符合PaddlePaddle期待的读取数据的Python程序的模式。
.. code-block:: python
# dataprovider.py
from paddle.trainer.PyDataProvider2 import *
import random
# 定义输入数据的类型: 2个浮点数
@provider(input_types=[dense_vector(1), dense_vector(1)],use_seq=False)
def process(settings, input_file):
for i in xrange(2000):
x = random.random()
yield [x], [2*x+0.3]
训练模型
-----------
为了还原 `y = 2x + 0.3`，我们先从一条随机的直线 `y' = wx + b` 开始，然后利用观测数据调整 `w` 和 `b` 使得 `y'` 和 `y` 的差距不断减小，最终趋于接近。这个过程就是模型的训练过程，而 `w` 和 `b` 就是模型的参数，即我们的训练目标。
在PaddlePaddle里，该模型的网络配置如下。
.. code-block:: python
# trainer_config.py
from paddle.trainer_config_helpers import *
# 1. 定义数据来源，调用上面的process函数获得观测数据
data_file = 'empty.list'
with open(data_file, 'w') as f: f.writelines(' ')
define_py_data_sources2(train_list=data_file, test_list=None,
module='dataprovider', obj='process',args={})
# 2. 学习算法。控制如何改变模型参数 w 和 b
settings(batch_size=12, learning_rate=1e-3, learning_method=MomentumOptimizer())
# 3. 神经网络配置
x = data_layer(name='x', size=1)
y = data_layer(name='y', size=1)
# 线性计算网络层: ȳ = wx + b
ȳ = fc_layer(input=x, param_attr=ParamAttr(name='w'), size=1, act=LinearActivation(), bias_attr=ParamAttr(name='b'))
# 计算误差函数，即 ȳ 和真实 y 之间的距离
cost = square_error_cost(input= ȳ, label=y)
outputs(cost)
这段简短的配置展示了PaddlePaddle的基本用法：
- 第一部分定义了数据输入。一般情况下，PaddlePaddle先从一个文件列表里获得数据文件地址，然后交给用户自定义的函数（例如上面的 `process`函数）进行读入和预处理从而得到真实输入。本文中由于输入数据是随机生成的不需要读输入文件，所以放一个空列表（`empty.list`）即可。
- 第二部分主要是选择学习算法，它定义了模型参数改变的规则。PaddlePaddle提供了很多优秀的学习算法，这里使用一个基于momentum的随机梯度下降(SGD)算法，该算法每批量(batch)读取12个采样数据进行随机梯度计算来更新更新。
- 最后一部分是神经网络的配置。由于PaddlePaddle已经实现了丰富的网络层，所以很多时候你需要做的只是定义正确的网络层并把它们连接起来。这里使用了三种网络单元：
- **数据层**：数据层 `data_layer` 是神经网络的入口，它读入数据并将它们传输到接下来的网络层。这里数据层有两个，分别对应于变量 `x` 和 `y`。
- **全连接层**：全连接层 `fc_layer` 是基础的计算单元，这里利用它建模变量之间的线性关系。计算单元是神经网络的核心，PaddlePaddle支持大量的计算单元和任意深度的网络连接，从而可以拟合任意的函数来学习复杂的数据关系。
- **回归误差代价层**：回归误差代价层 `square_error_cost` 是众多误差代价函数层的一种，它们在训练过程作为网络的出口，用来计算模型的误差，是模型参数优化的目标函数。
定义了网络结构并保存为 `trainer_config.py` 之后，运行以下训练命令：
.. code-block:: bash
paddle train --config=trainer_config.py --save_dir=./output --num_passes=30
PaddlePaddle将在观测数据集上迭代训练30轮，并将每轮的模型结果存放在 `./output` 路径下。从输出日志可以看到，随着轮数增加误差代价函数的输出在不断的减小，这意味着模型在训练数据上不断的改进，直到逼近真实解：` y = 2x + 0.3 `
模型检验
-----------
训练完成后，我们希望能够检验模型的好坏。一种常用的做法是用学习的模型对另外一组测试数据进行预测，评价预测的效果。在这个例子中，由于已经知道了真实答案，我们可以直接观察模型的参数是否符合预期来进行检验。
PaddlePaddle将每个模型参数作为一个numpy数组单独存为一个文件，所以可以利用如下方法读取模型的参数。
.. code-block:: python
import numpy as np
import os
def load(file_name):
with open(file_name, 'rb') as f:
f.read(16) # skip header for float type.
return np.fromfile(f, dtype=np.float32)
print 'w=%.6f, b=%.6f' % (load('output/pass-00029/w'), load('output/pass-00029/b'))
# w=1.999743, b=0.300137
.. image:: ./parameters.png
:align: center
:scale: 80 %
从图中可以看到，虽然 `w` 和 `b` 都使用随机值初始化，但在起初的几轮训练中它们都在快速逼近真实值，并且后续仍在不断改进，使得最终得到的模型几乎与真实模型一致。
这样，我们用PaddlePaddle解决了单变量线性回归问题， 包括数据输入、模型训练和最后的结果验证。
doc/getstarted/basic_usage/index_en.rst 已删除 100644 → 0
浏览文件 @ 2488b81f
Simple Linear Regression
========================
PaddlePaddle is a deep learning platform open-sourced by Baidu. With PaddlePaddle, you can easily train a classic neural network within a couple lines of configuration, or you can build sophisticated models that provide state-of-the-art performance on difficult learning tasks like sentiment analysis, machine translation, image caption and so on.
Problem Background
------------------
Now, to give you a hint of what using PaddlePaddle looks like, let's start with a fundamental learning problem - `simple linear regression <https://en.wikipedia.org/wiki/Simple_linear_regression>`_: you have observed a set of two-dimensional data points of ``X`` and ``Y``, where ``X`` is an explanatory variable and ``Y`` is corresponding dependent variable, and you want to recover the underlying correlation between ``X`` and ``Y``. Linear regression can be used in many practical scenarios. For example, ``X`` can be a variable about house size, and ``Y`` a variable about house price. You can build a model that captures relationship between them by observing real estate markets.
Prepare the Data
-----------------
Suppose the true relationship can be characterized as ``Y = 2X + 0.3``, let's see how to recover this pattern only from observed data. Here is a piece of python code that feeds synthetic data to PaddlePaddle. The code is pretty self-explanatory, the only extra thing you need to add for PaddlePaddle is a definition of input data types.
.. code-block:: python
# dataprovider.py
from paddle.trainer.PyDataProvider2 import *
import random
# define data types of input: 2 real numbers
@provider(input_types=[dense_vector(1), dense_vector(1)],use_seq=False)
def process(settings, input_file):
for i in xrange(2000):
x = random.random()
yield [x], [2*x+0.3]
Train a NeuralNetwork
----------------------
To recover this relationship between ``X`` and ``Y``, we use a neural network with one layer of linear activation units and a square error cost layer. Don't worry if you are not familiar with these terminologies, it's just saying that we are starting from a random line ``Y' = wX + b`` , then we gradually adapt ``w`` and ``b`` to minimize the difference between ``Y'`` and ``Y``. Here is what it looks like in PaddlePaddle:
.. code-block:: python
# trainer_config.py
from paddle.trainer_config_helpers import *
# 1. read data. Suppose you saved above python code as dataprovider.py
data_file = 'empty.list'
with open(data_file, 'w') as f: f.writelines(' ')
define_py_data_sources2(train_list=data_file, 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 = square_error_cost(input=y_predict, label=y)
outputs(cost)
Some of the most fundamental usages of PaddlePaddle are demonstrated:
- The first part shows how to feed data into PaddlePaddle. In general cases, PaddlePaddle reads raw data from a list of files, and then do some user-defined process to get real input. In this case, we only need to create a placeholder file since we are generating synthetic data on the fly.
- The second part describes learning algorithm. It defines in what ways adjustments are made to model parameters. PaddlePaddle provides a rich set of optimizers, but a simple momentum based optimizer will suffice here, and it processes 12 data points each time.
- Finally, the network configuration. It usually is as simple as "stacking" layers. Three kinds of layers are used in this configuration:
- **Data Layer**: a network always starts with one or more data layers. They provide input data to the rest of the network. In this problem, two data layers are used respectively for ``X`` and ``Y``.
- **FC Layer**: FC layer is short for Fully Connected Layer, which connects all the input units to current layer and does the actual computation specified as activation function. Computation layers like this are the fundamental building blocks of a deeper model.
- **Cost Layer**: in training phase, cost layers are usually the last layers of the network. They measure the performance of current model, and provide guidence to adjust parameters.
Now that everything is ready, you can train the network with a simple command line call:
.. code-block:: bash
paddle train --config=trainer_config.py --save_dir=./output --num_passes=30
This means that PaddlePaddle will train this network on the synthectic dataset for 30 passes, and save all the models under path ``./output``. You will see from the messages printed out during training phase that the model cost is decreasing as time goes by, which indicates we are getting a closer guess.
Evaluate the Model
-------------------
Usually, a different dataset that left out during training phase should be used to evalute the models. However, we are lucky enough to know the real answer: ``w=2, b=0.3``, thus a better option is to check out model parameters directly.
In PaddlePaddle, training is just to get a collection of model parameters, which are ``w`` and ``b`` in this case. Each parameter is saved in an individual file in the popular ``numpy`` array format. Here is the code that reads parameters from last pass.
.. code-block:: python
import numpy as np
import os
def load(file_name):
with open(file_name, 'rb') as f:
f.read(16) # skip header for float type.
return np.fromfile(f, dtype=np.float32)
print 'w=%.6f, b=%.6f' % (load('output/pass-00029/w'), load('output/pass-00029/b'))
# w=1.999743, b=0.300137
.. image:: parameters.png
:align: center
Although starts from a random guess, you can see that value of ``w`` changes quickly towards 2 and ``b`` changes quickly towards 0.3. In the end, the predicted line is almost identical with real answer.
There, you have recovered the underlying pattern between ``X`` and ``Y`` only from observed data.
doc/getstarted/build_and_install/build_from_source_cn.rst 0 → 100644
浏览文件 @ 048cc30e
从源码编译
======================
.. _build_step:
编译方法
----------------
PaddlePaddle主要使用 `CMake <https://cmake.org>`_ 以及GCC, G++作为编译工具。
我们推荐您使用PaddlePaddle Docker编译环境镜像完成编译，这样可以免去单独安装编译依赖的步骤，可选的不同编译环境Docker镜像
可以在 `这里 <https://hub.docker.com/r/paddlepaddle/paddle_manylinux_devel/tags/>`_ 找到。
如果您选择不使用Docker镜像，则需要在本机安装下面章节列出的 `编译依赖`_ 之后才能开始编译的步骤。
编译PaddlePaddle，需要执行：
.. code-block:: bash
git clone https://github.com/PaddlePaddle/Paddle.git
cd Paddle
# 如果使用Docker编译环境，执行下面的命令编译CPU-Only的二进制
docker run -it -v $PWD:/paddle -e "WITH_GPU=OFF" -e "WITH_TESTING=OFF" paddlepaddle/paddle_manylinux_devel:cuda8.0_cudnn5 bash -x paddle/scripts/docker/build.sh
# 如果不使用Docker编译环境，执行下面的命令
mkdir build
cd build
cmake -DWITH_GPU=OFF -DWITH_TESTING=OFF ..
make
编译完成后会在build/python/dist目录下生成输出的whl包，可以选在在当前机器安装也可以拷贝到目标机器安装：
.. code-block:: bash
pip install python/dist/*.whl
.. _run_test:
执行单元测试
----------------
如果您期望在编译完成后立即执行所有的单元测试，可以按照下面的方法：
使用Docker的情况下，设置 :code:`RUN_TEST=ON` 和 :code:`WITH_TESTING=ON` 就会在完成编译之后，立即执行单元测试。
开启 :code:`WITH_GPU=ON` 可以指定同时执行GPU上的单元测试。
.. code-block:: bash
docker run -it -v $PWD:/paddle -e "WITH_GPU=OFF" -e "WITH_TESTING=ON" -e "RUN_TEST=ON" paddlepaddle/paddle_manylinux_devel:cuda8.0_cudnn5 bash -x paddle/scripts/docker/build.sh
如果不使用Docker，可以执行ctest命令即可：
.. code-block:: bash
mkdir build
cd build
cmake -DWITH_GPU=OFF -DWITH_TESTING=OFF ..
make
ctest
# 指定执行其中一个单元测试 test_mul_op
ctest -R test_mul_op
.. _compile_deps:
编译依赖
----------------
PaddlePaddle编译需要使用到下面的依赖（包含但不限于），其他的依赖软件，会自动在编译时下载。
.. csv-table:: PaddlePaddle编译依赖
:header: "依赖", "版本", "说明"
:widths: 10, 15, 30
"CMake", ">=3.5", ""
"GCC", "4.8.2", "推荐使用CentOS的devtools2"
"Python", "2.7.x", "依赖libpython2.7.so"
"pip", ">=9.0", ""
"numpy", "", ""
"SWIG", ">=2.0", ""
"Go", ">=1.8", "可选"
.. _build_options:
编译选项
----------------
PaddlePaddle的编译选项，包括生成CPU/GPU二进制文件、链接何种BLAS库等。
用户可在调用cmake的时候设置它们，详细的cmake使用方法可以参考
`官方文档 <https://cmake.org/cmake-tutorial>`_ 。
在cmake的命令行中，通过使用 ``-D`` 命令设置该类编译选项，例如：
.. code-block:: bash
cmake .. -DWITH_GPU=OFF
.. csv-table:: 编译选项说明
:header: "选项", "说明", "默认值"
:widths: 1, 7, 2
"WITH_GPU", "是否支持GPU", "ON"
"WITH_C_API", "是否仅编译CAPI", "OFF"
"WITH_DOUBLE", "是否使用双精度浮点数", "OFF"
"WITH_DSO", "是否运行时动态加载CUDA动态库，而非静态加载CUDA动态库。", "ON"
"WITH_AVX", "是否编译含有AVX指令集的PaddlePaddle二进制文件", "ON"
"WITH_PYTHON", "是否内嵌PYTHON解释器", "ON"
"WITH_STYLE_CHECK", "是否编译时进行代码风格检查", "ON"
"WITH_TESTING", "是否开启单元测试", "ON"
"WITH_DOC", "是否编译中英文文档", "OFF"
"WITH_SWIG_PY", "是否编译PYTHON的SWIG接口，该接口可用于预测和定制化训练", "Auto"
"WITH_GOLANG", "是否编译go语言的可容错parameter server", "ON"
"WITH_MKL", "是否使用MKL数学库，如果为否则是用OpenBLAS", "ON"
BLAS
+++++
PaddlePaddle支持 `MKL <https://software.intel.com/en-us/intel-mkl>`_ 和
`OpenBlAS <http://www.openblas.net/>`_ 两种BLAS库。默认使用MKL。如果使用MKL并且机器含有AVX2指令集，
还会下载MKL-DNN数学库，详细参考 `这里 <https://github.com/PaddlePaddle/Paddle/tree/develop/doc/design/mkldnn#cmake>`_ 。
如果关闭MKL，则会使用OpenBLAS作为BLAS库。
CUDA/cuDNN
+++++++++++
PaddlePaddle在编译时/运行时会自动找到系统中安装的CUDA和cuDNN库进行编译和执行。
使用参数 :code:`-DCUDA_ARCH_NAME=Auto` 可以指定开启自动检测SM架构，加速编译。
PaddlePaddle可以使用cuDNN v5.1之后的任何一个版本来编译运行，但尽量请保持编译和运行使用的cuDNN是同一个版本。
我们推荐使用最新版本的cuDNN。
编译选项的设置
++++++++++++++
PaddePaddle通过编译时指定路径来实现引用各种BLAS/CUDA/cuDNN库。cmake编译时，首先在系统路径（ :code:`/usr/lib:/usr/local/lib` ）中搜索这几个库，同时也会读取相关路径变量来进行搜索。 通过使用 ``-D`` 命令可以设置，例如
.. code-block:: bash
cmake .. -DWITH_GPU=ON -DWITH_TESTING=OFF -DCUDNN_ROOT=/opt/cudnnv5
**注意：这几个编译选项的设置，只在第一次cmake的时候有效。如果之后想要重新设置，推荐清理整个编译目录（** :code:`rm -rf` ）**后，再指定。**
doc/getstarted/build_and_install/build_from_source_en.md 已删除 100644 → 0
浏览文件 @ 2488b81f
Installing from Sources
==========================
* [1. Download and Setup](#download)
* [2. Requirements](#requirements)
* [3. Build on Ubuntu](#ubuntu)
* [4. Build on Centos](#centos)
## <span id="download">Download and Setup</span>
You can download PaddlePaddle from the [github source](https://github.com/PaddlePaddle/Paddle).
```bash
git clone https://github.com/PaddlePaddle/Paddle paddle
cd paddle
```
## <span id="requirements">Requirements</span>
To compile the source code, your computer must be equipped with the following dependencies.
- **Compiler**: GCC >= 4.8 or Clang >= 3.3 (AppleClang >= 5.1) and gfortran compiler
- **CMake**: CMake >= 3.0 (at least CMake 3.4 on Mac OS X)
- **BLAS**: MKL, OpenBlas or ATLAS
- **Python**: only support Python 2.7
- **Go**
**Note:** For CUDA 7.0 and CUDA 7.5, GCC 5.0 and up are not supported!
For CUDA 8.0, GCC versions later than 5.3 are not supported!
### Options
PaddlePaddle supports some build options.
<html>
<table>
<thead>
<tr>
<th scope="col" class="left">Optional</th>
<th scope="col" class="left">Description</th>
</tr>
</thead>
<tbody>
<tr><td class="left">WITH_GPU</td><td class="left">Compile PaddlePaddle with NVIDIA GPU</td></tr>
<tr><td class="left">WITH_AVX</td><td class="left">Compile PaddlePaddle with AVX intrinsics</td></tr>
<tr><td class="left">WITH_DSO</td><td class="left">Compile PaddlePaddle with dynamic linked CUDA</td></tr>
<tr><td class="left">WITH_TESTING</td><td class="left">Compile PaddlePaddle with unit testing</td></tr>
<tr><td class="left">WITH_SWIG_PY</td><td class="left">Compile PaddlePaddle with inference api</td></tr>
<tr><td class="left">WITH_STYLE_CHECK</td><td class="left">Compile PaddlePaddle with style check</td></tr>
<tr><td class="left">WITH_PYTHON</td><td class="left">Compile PaddlePaddle with python interpreter</td></tr>
<tr><td class="left">WITH_DOUBLE</td><td class="left">Compile PaddlePaddle with double precision</td></tr>
<tr><td class="left">WITH_RDMA</td><td class="left">Compile PaddlePaddle with RDMA support</td></tr>
<tr><td class="left">WITH_TIMER</td><td class="left">Compile PaddlePaddle with stats timer</td></tr>
<tr><td class="left">WITH_PROFILER</td><td class="left">Compile PaddlePaddle with GPU profiler</td></tr>
<tr><td class="left">WITH_DOC</td><td class="left">Compile PaddlePaddle with documentation</td></tr>
<tr><td class="left">WITH_COVERAGE</td><td class="left">Compile PaddlePaddle with code coverage</td></tr>
<tr><td class="left">COVERALLS_UPLOAD</td><td class="left">Package code coverage data to coveralls</td></tr>
<tr><td class="left">ON_TRAVIS</td><td class="left">Exclude special unit test on Travis CI</td></tr>
</tbody>
</table>
</html>
**Note:**
- The GPU version works best with Cuda Toolkit 8.0 and cuDNN v5.
- Other versions like Cuda Toolkit 7.0, 7.5 and cuDNN v3, v4 are also supported.
- **To utilize cuDNN v5, Cuda Toolkit 7.5 is prerequisite and vice versa.**
As a simple example, consider the following:
1. **BLAS Dependencies(optional)**
CMake will search BLAS libraries from the system. If not found, OpenBLAS will be downloaded, built and installed automatically.
To utilize preinstalled BLAS， you can simply specify MKL, OpenBLAS or ATLAS via `MKL_ROOT`, `OPENBLAS_ROOT` or `ATLAS_ROOT`.
```bash
# specify MKL
cmake .. -DMKL_ROOT=<mkl_path>
# or specify OpenBLAS
cmake .. -DOPENBLAS_ROOT=<openblas_path>
```
2. **Doc Dependencies(optional)**
To generate PaddlePaddle's documentation, install dependencies and set `-DWITH_DOC=ON` as follows:
```bash
pip install 'sphinx>=1.4.0'
pip install sphinx_rtd_theme recommonmark
# install doxygen on Ubuntu
sudo apt-get install doxygen
# install doxygen on Mac OS X
brew install doxygen
# active docs in cmake
cmake .. -DWITH_DOC=ON`
```
## <span id="ubuntu">Build on Ubuntu 14.04</span>
### Install Dependencies
- **Paddle Dependencies**
```bash
# necessary
sudo apt-get update
sudo apt-get install -y git curl gcc g++ gfortran make build-essential automake
sudo apt-get install -y python python-pip python-numpy libpython-dev bison
sudo pip install 'protobuf==3.1.0.post1'
# Install Go
# You can follow https://golang.org/doc/install for a detailed explanation.
wget -O go.tgz https://storage.googleapis.com/golang/go1.8.1.linux-amd64.tar.gz && \
tar -C $HOME -xzf go.tgz && \
mkdir $HOME/gopath && \
rm go.tgz
# Setup environment variables
export GOROOT=$HOME/go
export GOPATH=$HOME/gopath
export PATH=$PATH:$GOROOT/bin
# install cmake 3.4
curl -sSL https://cmake.org/files/v3.4/cmake-3.4.1.tar.gz | tar -xz && \
cd cmake-3.4.1 && ./bootstrap && make -j4 && sudo make install && \
cd .. && rm -rf cmake-3.4.1
```
- **GPU Dependencies (optional)**
To build GPU version, you will need the following installed:
1. a CUDA-capable GPU
2. A supported version of Linux with a GCC compiler and toolchain
3. NVIDIA CUDA Toolkit (available at http://developer.nvidia.com/cuda-downloads)
4. NVIDIA cuDNN Library (available at https://developer.nvidia.com/cudnn)
The CUDA development environment relies on tight integration with the host development environment,
including the host compiler and C runtime libraries, and is therefore only supported on
distribution versions that have been qualified for this CUDA Toolkit release.
After downloading cuDNN library, issue the following commands:
```bash
sudo tar -xzf cudnn-7.5-linux-x64-v5.1.tgz -C /usr/local
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
```
Then you need to set LD\_LIBRARY\_PATH, PATH environment variables in ~/.bashrc.
```bash
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
export PATH=/usr/local/cuda/bin:$PATH
```
### Build and Install
As usual, the best option is to create build folder under paddle project directory.
```bash
mkdir build && cd build
```
Finally, you can build and install PaddlePaddle:
```bash
# you can add build option here, such as:
cmake .. -DCMAKE_INSTALL_PREFIX=<path to install>
# please use sudo make install, if you want to install PaddlePaddle into the system
make -j `nproc` && make install
# set PaddlePaddle installation path in ~/.bashrc
export PATH=<path to install>/bin:$PATH
# install PaddlePaddle Python modules.
sudo pip install <path to install>/opt/paddle/share/wheels/*.whl
```
## <span id="centos">Build on Centos 7</span>
### Install Dependencies
- **CPU Dependencies**
```bash
# necessary
sudo yum update
sudo yum install -y epel-release
sudo yum install -y make cmake3 python-devel python-pip gcc-gfortran swig git
sudo pip install wheel numpy
sudo pip install 'protobuf>=3.0.0'
```
- **GPU Dependencies (optional)**
To build GPU version, you will need the following installed:
1. a CUDA-capable GPU
2. A supported version of Linux with a GCC compiler and toolchain
3. NVIDIA CUDA Toolkit (available at http://developer.nvidia.com/cuda-downloads)
4. NVIDIA cuDNN Library (available at https://developer.nvidia.com/cudnn)
The CUDA development environment relies on tight integration with the host development environment,
including the host compiler and C runtime libraries, and is therefore only supported on
distribution versions that have been qualified for this CUDA Toolkit release.
After downloading cuDNN library, issue the following commands:
```bash
sudo tar -xzf cudnn-7.5-linux-x64-v5.1.tgz -C /usr/local
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
```
Then you need to set LD\_LIBRARY\_PATH, PATH environment variables in ~/.bashrc.
```bash
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
export PATH=/usr/local/cuda/bin:$PATH
```
### Build and Install
As usual, the best option is to create build folder under paddle project directory.
```bash
mkdir build && cd build
```
Finally, you can build and install PaddlePaddle:
```bash
# you can add build option here, such as:
cmake3 .. -DCMAKE_INSTALL_PREFIX=<path to install>
# please use sudo make install, if you want to install PaddlePaddle into the system
make -j `nproc` && make install
# set PaddlePaddle installation path in ~/.bashrc
export PATH=<path to install>/bin:$PATH
# install PaddlePaddle Python modules.
sudo pip install <path to install>/opt/paddle/share/wheels/*.whl
```
doc/getstarted/build_and_install/build_from_source_en.rst 0 → 100644
浏览文件 @ 048cc30e
Build from Sources
==========================
.. _build_step:
How To Build
----------------
PaddlePaddle mainly uses `CMake <https://cmake.org>`_ and GCC, G++ as compile
tools. We recommend you to use our pre-built Docker image to run the build
to avoid installing dependencies by yourself. We have several build environment
Docker images `here <https://hub.docker.com/r/paddlepaddle/paddle_manylinux_devel/tags/>`_ .
If you choose not to use Docker image for your build, you need to install the
below `Compile Dependencies`_ before run the build.
Then run:
.. code-block:: bash
git clone https://github.com/PaddlePaddle/Paddle.git
cd Paddle
# run the following command to build a CPU-Only binaries if you are using docker
docker run -it -v $PWD:/paddle -e "WITH_GPU=OFF" -e "WITH_TESTING=OFF" paddlepaddle/paddle_manylinux_devel:cuda8.0_cudnn5 bash -x paddle/scripts/docker/build.sh
# else run these commands
mkdir build
cd build
cmake -DWITH_GPU=OFF -DWITH_TESTING=OFF ..
make
When the compile finishes, you can get the output whl package under
build/python/dist, then you can choose to install the whl on local
machine or copy it to the target machine.
.. code-block:: bash
pip install python/dist/*.whl
.. _run_test:
Run Tests
----------------
If you wish to run the tests, you may follow the below steps:
When using Docker, set :code:`RUN_TEST=ON` and :code:`WITH_TESTING=ON` will run test immediately after the build.
Set :code:`WITH_GPU=ON` Can also run tests on GPU.
.. code-block:: bash
docker run -it -v $PWD:/paddle -e "WITH_GPU=OFF" -e "WITH_TESTING=ON" -e "RUN_TEST=ON" paddlepaddle/paddle_manylinux_devel:cuda8.0_cudnn5 bash -x paddle/scripts/docker/build.sh
If you don't use Docker, just run ctest will start the tests:
.. code-block:: bash
mkdir build
cd build
cmake -DWITH_GPU=OFF -DWITH_TESTING=ON ..
make
ctest
# run a single test like test_mul_op
ctest -R test_mul_op
.. _compile_deps:
Compile Dependencies
----------------
PaddlePaddle need the following dependencies when compiling, other dependencies
will be downloaded automatically.
.. csv-table:: PaddlePaddle Compile Dependencies
:header: "Dependency", "Version", "Description"
:widths: 10, 15, 30
"CMake", ">=3.5", ""
"GCC", "4.8.2", "Recommend devtools2 for CentOS"
"Python", "2.7.x", "Need libpython2.7.so"
"pip", ">=9.0", ""
"numpy", "", ""
"SWIG", ">=2.0", ""
"Go", ">=1.8", "Optional"
.. _build_options:
Build Options
----------------
Build options include whether build binaries for CPU or GPU, which BLAS
library to use etc. You may pass these settings when running cmake.
For detailed cmake tutorial please refer to `here <https://cmake.org/cmake-tutorial>`_ 。
.. _build_options_bool:
Bool Type Options
----------------
You can add :code:`-D` argument to pass such options, like:
.. code-block:: bash
cmake .. -DWITH_GPU=OFF
.. csv-table:: Bool Type Options
:header: "Option", "Description", "Default"
:widths: 1, 7, 2
"WITH_GPU", "Build with GPU support", "ON"
"WITH_C_API", "Build only CAPI", "OFF"
"WITH_DOUBLE", "Build with double precision", "OFF"
"WITH_DSO", "Dynamically load CUDA libraries", "ON"
"WITH_AVX", "Build with AVX support", "ON"
"WITH_PYTHON", "Build with integrated Python interpreter", "ON"
"WITH_STYLE_CHECK", "Check code style when building", "ON"
"WITH_TESTING", "Build unit tests", "ON"
"WITH_DOC", "Build documentaions", "OFF"
"WITH_SWIG_PY", "Build Python SWIG interface for V2 API", "Auto"
"WITH_GOLANG", "Build fault-tolerant parameter server written in go", "ON"
"WITH_MKL", "Use MKL as BLAS library, else use OpenBLAS", "ON"
BLAS
+++++
PaddlePaddle supports `MKL <https://software.intel.com/en-us/intel-mkl>`_ and
`OpenBlAS <http://www.openblas.net/>`_ as BLAS library。By default it uses MKL.
If you are using MKL and your machine supports AVX2, MKL-DNN will also be downloaded
and used, for more `details <https://github.com/PaddlePaddle/Paddle/tree/develop/doc/design/mkldnn#cmake>`_ .
If you choose not to use MKL, then OpenBlAS will be used.
CUDA/cuDNN
+++++++++++
PaddlePaddle will automatically find CUDA and cuDNN when compiling and running.
parameter :code:`-DCUDA_ARCH_NAME=Auto` can be used to detect SM architecture
automatically in order to speed up the build.
PaddlePaddle can build with any version later than cuDNN v5.1, and we intend to
keep on with latest cuDNN versions. Be sure to run with the same version of cuDNN
you built.
Pass Compile Options
++++++++++++++
You can pass compile options to use intended BLAS/CUDA/Cudnn libraries.
When running cmake command, it will search system paths like
:code:`/usr/lib:/usr/local/lib` and then search paths that you
passed to cmake, i.e.
.. code-block:: bash
cmake .. -DWITH_GPU=ON -DWITH_TESTING=OFF -DCUDNN_ROOT=/opt/cudnnv5
**NOTE: These options only take effect when running cmake for the first time, you need to clean the cmake cache or clean the build directory (** :code:`rm -rf` **) if you want to change it.**
doc/getstarted/build_and_install/cmake/build_from_source_cn.rst 已删除 100644 → 0
浏览文件 @ 2488b81f
PaddlePaddle的编译选项
======================
PaddlePaddle的编译选项，包括生成CPU/GPU二进制文件、链接何种BLAS库等。用户可在调用cmake的时候设置它们，详细的cmake使用方法可以参考 `官方文档 <https://cmake.org/cmake-tutorial>`_ 。
Bool型的编译选项
----------------
用户可在cmake的命令行中，通过使用 ``-D`` 命令设置该类编译选项，例如
.. code-block:: bash
cmake .. -DWITH_GPU=OFF
.. csv-table:: Bool型的编译选项
:widths: 1, 7, 2
:file: compile_options.csv
BLAS/CUDA/Cudnn的编译选项
--------------------------
BLAS
+++++
PaddlePaddle支持以下任意一种BLAS库：`MKL <https://software.intel.com/en-us/intel-mkl>`_ ，`ATLAS <http://math-atlas.sourceforge.net/>`_ ，`OpenBlAS <http://www.openblas.net/>`_ 和 `REFERENCE BLAS <http://www.netlib.org/blas/>`_ 。
.. csv-table:: BLAS路径相关的编译选项
:widths: 1, 2, 7
:file: cblas_settings.csv
CUDA/Cudnn
+++++++++++
PaddlePaddle可以使用cudnn v2之后的任何一个版本来编译运行，但尽量请保持编译和运行使用的cudnn是同一个版本。 我们推荐使用最新版本的cudnn v5.1。
编译选项的设置
++++++++++++++
PaddePaddle通过编译时指定路径来实现引用各种BLAS/CUDA/Cudnn库。cmake编译时，首先在系统路径(/usr/lib\:/usr/local/lib)中搜索这几个库，同时也会读取相关路径变量来进行搜索。 通过使用 ``-D`` 命令可以设置，例如
.. code-block:: bash
cmake .. -DMKL_ROOT=/opt/mkl/ -DCUDNN_ROOT=/opt/cudnnv5
注意：这几个编译选项的设置，只在第一次cmake的时候有效。如果之后想要重新设置，推荐清理整个编译目录（``rm -rf``）后，再指定。
doc/getstarted/build_and_install/cmake/cblas_settings.csv 已删除 100644 → 0
浏览文件 @ 2488b81f
编译选项,描述,注意
MKL_ROOT,MKL的路径,${MKL_ROOT}/include下需要包含mkl.h，${MKL_ROOT}/lib目录下需要包含mkl_core，mkl_sequential和mkl_intel_lp64三个库。
ATLAS_ROOT,ATLAS的路径,${ATLAS_ROOT}/include下需要包含cblas.h，${ATLAS_ROOT}/lib下需要包含cblas和atlas两个库。
OPENBLAS_ROOT,OpenBLAS的路径,${OPENBLAS_ROOT}/include下需要包含cblas.h，${OPENBLAS_ROOT}/lib下需要包含openblas库。
REFERENCE_CBLAS_ROOT,REFERENCE BLAS的路径,${REFERENCE_CBLAS_ROOT}/include下需要包含cblas.h，${REFERENCE_CBLAS_ROOT}/lib下需要包含cblas库。
\ No newline at end of file
doc/getstarted/build_and_install/cmake/compile_options.csv 已删除 100644 → 0
浏览文件 @ 2488b81f
选项,说明,默认值
WITH_GPU,是否支持GPU。,取决于是否寻找到CUDA工具链
WITH_DOUBLE,是否使用双精度浮点数。,否
WITH_DSO,是否运行时动态加载CUDA动态库，而非静态加载CUDA动态库。,是
WITH_AVX,是否编译含有AVX指令集的PaddlePaddle二进制文件,是
WITH_PYTHON,是否内嵌PYTHON解释器。方便今后的嵌入式移植工作。,是
WITH_STYLE_CHECK,是否编译时进行代码风格检查,是
WITH_RDMA,是否开启RDMA,否
WITH_TIMER,是否开启计时功能。如果开启会导致运行略慢，打印的日志变多，但是方便调试和测Benchmark,否
WITH_TESTING,是否开启单元测试,取决于是否寻找到GTEST
WITH_DOC,是否编译中英文文档,否
WITH_SWIG_PY,是否编译PYTHON的SWIG接口，该接口可用于预测和定制化训练,取决于是否寻找到SWIG
\ No newline at end of file
doc/getstarted/build_and_install/docker_install_cn.rst
浏览文件 @ 048cc30e
PaddlePaddle的Docker容器使用方式
使用Docker安装运行
================================
PaddlePaddle目前唯一官方支持的运行的方式是Docker容器。因为Docker能在所有主要操作系统（包括Linux，Mac OS X和Windows）上运行。 请注意，您需要更改 `Dockers设置 <https://github.com/PaddlePaddle/Paddle/issues/627>`_ 才能充分利用Mac OS X和Windows上的硬件资源。
使用Docker安装和运行PaddlePaddle可以无需考虑依赖环境即可运行。并且也可以在Windows的docker中运行。
您可以在 `Docker官网 <https://docs.docker.com/get-started/>`_ 获得基本的Docker安装和使用方法。
Docker使用入门
------------------------------
几个基础的概念帮助理解和使用Docker：
如果您在使用Windows，可以参考
`这篇 <https://docs.docker.com/toolbox/toolbox_install_windows/>`_
教程，完成在Windows上安装和使用Docker。
- *镜像*：一个Docker镜像是一个打包好的软件。它包含了这个软件本身和它所依赖的运行环境。PaddlePaddle的Docker镜像就包含了PaddlePaddle的Python库以及其依赖的多个Python库。这样我们可以直接在Docker中运行需要的程序而不需要安装后在执行。可以执行：
在了解Docker的基本使用方法之后，即可开始下面的步骤：
.. code-block:: bash
.. _docker_pull:
docker images
获取PaddlePaddle的Docker镜像
------------------------------
来列出当前系统中的所有镜像，同样可以执行：
执行下面的命令获取最新的PaddlePaddle Docker镜像
.. code-block:: bash
docker pull paddlepaddle/paddle:0.10.0
来下载Docker镜像，paddlepaddle/paddle是从官方镜像源Dockerhub.com下载的，推荐国内用户使用docker.paddlepaddle.org/paddle下载。
docker pull paddlepaddle/paddle
- *容器*： 如果说一个Docker镜像就是一个程序，那容器就是这个程序运行时产生的“进程”。
实际上，一个容器就是一个操作系统的进程，但是是运行在独立的进程空间，文件系统以及网络之上。
可以执行：
对于国内用户，我们提供了加速访问的镜像源：
.. code-block:: bash
docker run paddlepaddle/paddle:0.10.0
docker pull docker.paddlepaddle.org/paddle
来使用一个镜像启动一个容器。
- 默认情况下，Docker容器会运行在独立的文件系统空间之上，我们无法在Docker容器中
访问到主机上的文件。可以通过*挂载Volume*的方式，将主机上的文件或目录挂载到
Docker容器中。下面的命令把当前目录挂载到了容器中的 /data 目录下，容器使用
debian镜像，并且启动后执行 :code:`ls /data`。
下载GPU版本的Docker镜像：
.. code-block:: bash
docker run --rm -v $(pwd):/data debian ls /data
PaddlePaddle发布的Docker镜像使用说明
------------------------------
我们把PaddlePaddle的编译环境打包成一个镜像，称为开发镜像，里面涵盖了
PaddlePaddle需要的所有编译工具。把编译出来的PaddlePaddle也打包成一个镜
像，称为生产镜像，里面涵盖了PaddlePaddle运行所需的所有环境。每次
PaddlePaddle发布新版本的时候都会发布对应版本的生产镜像以及开发镜像。运
行镜像包括纯CPU版本和GPU版本以及其对应的非AVX版本。我们会在
`dockerhub.com <https://hub.docker.com/r/paddlepaddle/paddle/tags/>`_
和国内镜像`docker.paddlepaddle.org` 提供最新
的Docker镜像，可以在"tags"标签下找到最新的Paddle镜像版本。
**注意：为了方便在国内的开发者下载Docker镜像，我们提供了国内的镜像服务器供大家使用。如果您在国内，请把文档里命令中的paddlepaddle/paddle替换成docker.paddlepaddle.org/paddle。**
1. 开发镜像：:code:`paddlepaddle/paddle:0.10.0-dev`
这个镜像包含了Paddle相关的开发工具以及编译和运行环境。用户可以使用开发镜像代替配置本地环境，完成开发，编译，发布，
文档编写等工作。由于不同的Paddle的版本可能需要不同的依赖和工具，所以如果需要自行配置开发环境需要考虑版本的因素。
开发镜像包含了以下工具：
- gcc/clang
- nvcc
- Python
- sphinx
- woboq
- sshd
很多开发者会使用远程的安装有GPU的服务器工作，用户可以使用ssh登录到这台服务器上并执行 :code:`docker exec`进入开发镜像并开始工作，
也可以在开发镜像中启动一个SSHD服务，方便开发者直接登录到镜像中进行开发:
以交互容器方式运行开发镜像：
.. code-block:: bash
docker run -it --rm -v $(pwd):/paddle paddlepaddle/paddle:0.10.0-dev /bin/bash
或者，可以以后台进程方式运行容器：
.. code-block:: bash
docker run -d -p 2202:22 -p 8888:8888 -v $(pwd):/paddle paddlepaddle/paddle:0.10.0-dev /usr/sbin/sshd -D
然后用密码 :code:`root` SSH进入容器：
.. code-block:: bash
ssh -p 2202 root@localhost
SSH方式的一个优点是我们可以从多个终端进入容器。比如，一个终端运行vi，另一个终端运行Python。另一个好处是我们可以把PaddlePaddle容器运行在远程服务器上，并在笔记本上通过SSH与其连接。
2. 生产镜像：根据CPU、GPU和非AVX区分了如下4个镜像：
- GPU/AVX：:code:`paddlepaddle/paddle:<version>-gpu`
- GPU/no-AVX：:code:`paddlepaddle/paddle:<version>-gpu-noavx`
- CPU/AVX：:code:`paddlepaddle/paddle:<version>`
- CPU/no-AVX：:code:`paddlepaddle/paddle:<version>-noavx`
纯CPU镜像以及GPU镜像都会用到AVX指令集，但是2008年之前生产的旧电脑不支持AVX。以下指令能检查Linux电脑是否支持AVX：
.. code-block:: bash
if cat /proc/cpuinfo | grep -i avx; then echo Yes; else echo No; fi
如果输出是No，就需要选择使用no-AVX的镜像
**注：在0.10.0之后的版本，PaddlePaddle都可以自动判断硬件是否支持AVX，所以无需判断AVX即可使用**
docker pull paddlepaddle/paddle:latest-gpu
docker pull docker.paddlepaddle.org/paddle:latest-gpu
以上方法在GPU镜像里也能用，只是请不要忘记提前在物理机上安装GPU最新驱动。
为了保证GPU驱动能够在镜像里面正常运行，我们推荐使用[nvidia-docker](https://github.com/NVIDIA/nvidia-docker)来运行镜像。
选择下载使用不同的BLAS库的Docker镜像：
.. code-block:: bash
nvidia-docker run -it --rm paddledev/paddle:0.10.0-gpu /bin/bash
.. code-block:: bash
注意: 如果使用nvidia-docker存在问题，你也许可以尝试更老的方法，具体如下，但是我们并不推荐这种方法。：
# 默认是使用MKL的镜像
docker pull paddlepaddle/paddle
# 使用OpenBLAS的镜像
docker pull paddlepaddle/paddle:latest-openblas
.. code-block:: bash
下载指定版本的Docker镜像，可以从 `DockerHub网站 <https://hub.docker.com/r/paddlepaddle/paddle/tags/>`_ 获取可选的tag，并执行下面的命令：
export CUDA_SO="$(\ls /usr/lib64/libcuda* | xargs -I{} echo '-v {}:{}') $(\ls /usr/lib64/libnvidia* | xargs -I{} echo '-v {}:{}')"
export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}')
docker run ${CUDA_SO} ${DEVICES} -it paddledev/paddle:0.10.0-gpu
.. code-block:: bash
3. 运行以及发布您的AI程序
docker pull paddlepaddle/paddle:[tag]
# 比如：
docker pull docker.paddlepaddle.org/paddle:0.10.0-gpu
假设您已经完成了一个AI训练的python程序 :code:`a.py`，这个程序是您在开发机上使用开发镜像完成开发。此时您可以运行这个命令在开发机上进行测试运行：
.. _docker_run:
.. code-block:: bash
在Docker中执行PaddlePaddle训练程序
------------------------------
docker run -it -v $PWD:/work paddle /work/a.py
假设您已经在当前目录（比如在/home/work）编写了一个PaddlePaddle的程序 :code:`train.py` （可以参考
`PaddlePaddleBook <http://www.paddlepaddle.org/docs/develop/book/01.fit_a_line/index.cn.html>`_
编写），就可以使用下面的命令开始执行训练：
如果要使用GPU，请运行：
.. code-block:: bash
.. code-block:: bash
cd /home/work
docker run -it -v $PWD:/work paddlepaddle/paddle /work/train.py
上述命令中， :code:`-it` 参数说明容器已交互式运行； :code:`-v $PWD:/work`
指定将当前路径（Linux中$PWD变量会展开为当前路径的绝对路径）挂载到容器内部的 :code:`/work`
目录； :code:`paddlepaddle/paddle` 指定需要使用的容器； 最后 :code:`/work/train.py`
为容器内执行的命令，即运行训练程序。
nvidia-docker run -it -v $PWD:/work paddle /work/a.py
当然，您也可以进入到Docker容器中，以交互式的方式执行或调试您的代码：
.. code-block:: bash
docker run -it -v $PWD:/work paddlepaddle/paddle /bin/bash
cd /work
python train.py
这里`a.py`包含的所有依赖假设都可以在Paddle的运行容器中。如果需要包含更多的依赖、或者需要发布您的应用的镜像，可以编写`Dockerfile`使用`FROM paddledev/paddle:0.10.0`
创建和发布自己的AI程序镜像。
**注：PaddlePaddle Docker镜像为了减小体积，默认没有安装vim，您可以在容器中执行** :code:`apt-get install -y vim` **安装后，在容器中编辑代码。**
运行PaddlePaddle Book
---------------------
.. _docker_run_book:
Jupyter Notebook是一个开源的web程序，大家可以通过它制作和分享带有代码、公式、图表、文字的交互式文档。用户可以通过网页浏览文档。
使用Docker启动PaddlePaddle Book教程
------------------------------
使用Docker可以快速在本地启动一个包含了PaddlePaddle官方Book教程的Jupyter Notebook，可以通过网页浏览。
PaddlePaddle Book是为用户和开发者制作的一个交互式的Jupyter Notebook。
如果您想要更深入了解deep learning，PaddlePaddle Book一定是您最好的选择。
大家可以通过它阅读教程，或者制作和分享带有代码、公式、图表、文字的交互式文档。
我们提供可以直接运行PaddlePaddle Book的Docker镜像，直接运行：
.. code-block:: bash
.. code-block:: bash
docker run -p 8888:8888 paddlepaddle/book
docker run -p 8888:8888 paddlepaddle/book
然后在浏览器中输入以下网址：
.. code-block:: text
.. code-block:: text
http://localhost:8888/
http://localhost:8888/
就这么简单，享受您的旅程！
通过Docker容器开发PaddlePaddle
------------------------------
开发人员可以在Docker开发镜像中开发PaddlePaddle。这样开发人员可以以一致的方式在不同的平台上工作 - Linux，Mac OS X和Windows。
.. _docker_run_gpu:
1. 制作PaddlePaddle开发镜像
PaddlePaddle每次发布新版本都会发布对应的开发镜像供开发者直接使用。这里介绍如生成造这个开发镜像。
生成Docker镜像的方式有两个，一个是直接把一个容器转换成镜像，另一个是创建Dockerfile并运行docker build指令按照Dockerfile生成镜像。第一个方法的好处是简单快捷，适合自己实验，可以快速迭代。第二个方法的好处是Dockerfile可以把整个生成流程描述很清楚，其他人很容易看懂镜像生成过程，持续集成系统也可以简单地复现这个过程。我们采用第二个方法。Dockerfile位于PaddlePaddle repo的根目录。生成生产镜像只需要运行：
.. code-block:: bash
git clone https://github.com/PaddlePaddle/Paddle.git
cd Paddle
docker build -t paddle:dev .
docker build这个命令的-t指定了生成的镜像的名字，这里我们用paddle:dev。到此，PaddlePaddle开发镜像就被构建完毕了。
使用Docker执行GPU训练
------------------------------
2. 制作PaddlePaddle生产镜像
为了保证GPU驱动能够在镜像里面正常运行，我们推荐使用
`nvidia-docker <https://github.com/NVIDIA/nvidia-docker>`_ 来运行镜像。
请不要忘记提前在物理机上安装GPU最新驱动。
生产镜像的生成分为两步，第一步是运行：
.. code-block:: bash
.. code-block:: bash
docker run -v $(pwd):/paddle -e "WITH_GPU=OFF" -e "WITH_AVX=OFF" -e "WITH_TEST=ON" paddle:dev
nvidia-docker run -it -v $PWD:/work paddledev/paddle:latest-gpu /bin/bash
以上命令会编译PaddlePaddle，生成运行程序，以及生成创建生产镜像的Dockerfile。所有生成的的文件都在build目录下。“WITH_GPU”控制生成的生产镜像是否支持GPU，“WITH_AVX”控制生成的生产镜像是否支持AVX，”WITH_TEST“控制是否生成单元测试。
**注: 如果没有安装nvidia-docker，可以尝试以下的方法，将CUDA库和Linux设备挂载到Docker容器内：**
第二步是运行：
.. code-block:: bash
.. code-block:: bash
docker build -t paddle:prod -f build/Dockerfile ./build
export CUDA_SO="$(\ls /usr/lib64/libcuda* | xargs -I{} echo '-v {}:{}') $(\ls /usr/lib64/libnvidia* | xargs -I{} echo '-v {}:{}')"
export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}')
docker run ${CUDA_SO} ${DEVICES} -it paddledev/paddle:latest-gpu
以上命令会按照生成的Dockerfile把生成的程序拷贝到生产镜像中并做相应的配置，最终生成名为paddle:prod的生产镜像。
**关于AVX：**
3. 运行单元测试
AVX是一种CPU指令集，可以加速PaddlePaddle的计算。最新的PaddlePaddle Docker镜像默认
是开启AVX编译的，所以，如果您的电脑不支持AVX，需要单独
`编译 <./build_from_source_cn.rst>`_ PaddlePaddle为no-avx版本。
运行以下指令：
以下指令能检查Linux电脑是否支持AVX：
.. code-block:: bash
docker run -it -v $(pwd):/paddle paddle:dev bash -c "cd /paddle/build && ctest"
文档
----
Paddle的Docker开发镜像带有一个通过 `woboq code browser
<https://github.com/woboq/woboq_codebrowser>`_ 生成的HTML版本的C++源代码，便于用户浏览C++源码。
只要在Docker里启动PaddlePaddle的时候给它一个名字，就可以再运行另一个Nginx Docker镜像来服务HTML代码：
.. code-block:: bash
docker run -d --name paddle-cpu-doc paddle:0.10.0-dev
docker run -d --volumes-from paddle-cpu-doc -p 8088:80 nginx
if cat /proc/cpuinfo | grep -i avx; then echo Yes; else echo No; fi
接着我们就能够打开浏览器在 http://localhost:8088/paddle/ 浏览代码。
如果输出是No，就需要选择使用no-AVX的镜像
doc/getstarted/build_and_install/docker_install_en.rst
浏览文件 @ 048cc30e
PaddlePaddle in Docker Containers
Run in Docker Containers
=================================
Docker container is currently the only officially-supported way to
running PaddlePaddle. This is reasonable as Docker now runs on all
major operating systems including Linux, Mac OS X, and Windows.
Please be aware that you will need to change `Dockers settings
<https://github.com/PaddlePaddle/Paddle/issues/627>`_ to make full use
of your hardware resource on Mac OS X and Windows.
Run PaddlePaddle in Docker container so that you don't need to care about
runtime dependencies, also you can run under Windows system. You can get
tutorials at `here <https://docs.docker.com/get-started/>`_ .
Working With Docker
-------------------
If you are using Windows, please refer to
`this <https://docs.docker.com/toolbox/toolbox_install_windows/>`_
tutorial to start running docker under windows.
Docker is simple as long as we understand a few basic concepts:
After you've read above tutorials you may proceed the following steps.
- *image*: A Docker image is a pack of software. It could contain one or more programs and all their dependencies. For example, the PaddlePaddle's Docker image includes pre-built PaddlePaddle and Python and many Python packages. We can run a Docker image directly, other than installing all these software. We can type
.. _docker_pull:
.. code-block:: bash
docker images
Pull PaddlePaddle Docker Image
------------------------------
to list all images in the system. We can also run
Run the following command to download the latest Docker images:
.. code-block:: bash
docker pull paddlepaddle/paddle:0.10.0
to download a Docker image, paddlepaddle/paddle in this example,
from Dockerhub.com.
docker pull paddlepaddle/paddle
- *container*: considering a Docker image a program, a container is a
"process" that runs the image. Indeed, a container is exactly an
operating system process, but with a virtualized filesystem, network
port space, and other virtualized environment. We can type
For users in China, we provide a faster mirror:
.. code-block:: bash
docker run paddlepaddle/paddle:0.10.0
docker pull docker.paddlepaddle.org/paddle
to start a container to run a Docker image, paddlepaddle/paddle in this example.
- By default docker container have an isolated file system namespace,
we can not see the files in the host file system. By using *volume*,
mounted files in host will be visible inside docker container.
Following command will mount current dirctory into /data inside
docker container, run docker container from debian image with
command :code:`ls /data`.
Download GPU version images:
.. code-block:: bash
docker run --rm -v $(pwd):/data debian ls /data
Usage of CPU-only and GPU Images
----------------------------------
We package PaddlePaddle's compile environment into a Docker image,
called the develop image, it contains all compiling tools that
PaddlePaddle needs. We package compiled PaddlePaddle program into a
Docker image as well, called the production image, it contains all
runtime environment that running PaddlePaddle needs. For each version
of PaddlePaddle, we release both of them. Production image includes
CPU-only version and a CUDA GPU version and their no-AVX versions.
We put the docker images on `dockerhub.com
<https://hub.docker.com/r/paddlepaddle/paddle/tags/>`_. You can find the
latest versions under "tags" tab at dockerhub.com.
** NOTE: If you are in China, you can use our Docker image registry mirror to speed up the download process. To use it, please replace all paddlepaddle/paddle in the commands to docker.paddlepaddle.org/paddle.**
1. development image :code:`paddlepaddle/paddle:<version>-dev`
This image has packed related develop tools and runtime
environment. Users and developers can use this image instead of
their own local computer to accomplish development, build,
releasing, document writing etc. While different version of paddle
may depends on different version of libraries and tools, if you
want to setup a local environment, you must pay attention to the
versions. The development image contains:
- gcc/clang
- nvcc
- Python
- sphinx
- woboq
- sshd
Many developers use servers with GPUs, they can use ssh to login to
the server and run :code:`docker exec` to enter the docker
container and start their work. Also they can start a development
docker image with SSHD service, so they can login to the container
and start work.
2. Production images, this image might have multiple variants:
- GPU/AVX：:code:`paddlepaddle/paddle:<version>-gpu`
- GPU/no-AVX：:code:`paddlepaddle/paddle:<version>-gpu-noavx`
- CPU/AVX：:code:`paddlepaddle/paddle:<version>`
- CPU/no-AVX：:code:`paddlepaddle/paddle:<version>-noavx`
Please be aware that the CPU-only and the GPU images both use the
AVX instruction set, but old computers produced before 2008 do not
support AVX. The following command checks if your Linux computer
supports AVX:
.. code-block:: bash
if cat /proc/cpuinfo | grep -i avx; then echo Yes; else echo No; fi
**NOTE：versions after 0.10.0 will automatically detect system AVX support, so manual detect is not needed in this case.**
To run the CPU-only image as an interactive container:
.. code-block:: bash
docker run -it --rm paddlepaddle/paddle:0.10.0 /bin/bash
Above method work with the GPU image too -- the recommended way is
using `nvidia-docker <https://github.com/NVIDIA/nvidia-docker>`_.
Please install nvidia-docker first following this `tutorial
<https://github.com/NVIDIA/nvidia-docker#quick-start>`_.
Now you can run a GPU image:
.. code-block:: bash
nvidia-docker run -it --rm paddlepaddle/paddle:0.10.0-gpu /bin/bash
Train Model Using Python API
----------------------------
Our official docker image provides a runtime for PaddlePaddle
programs. The typical workflow will be as follows:
Create a directory as workspace:
.. code-block:: bash
mkdir ~/workspace
Edit a PaddlePaddle python program using your favourite editor
.. code-block:: bash
emacs ~/workspace/example.py
Run the program using docker:
.. code-block:: bash
docker run --rm -v ~/workspace:/workspace paddlepaddle/paddle:0.10.0 python /workspace/example.py
Or if you are using GPU for training:
docker pull paddlepaddle/paddle:latest-gpu
docker pull docker.paddlepaddle.org/paddle:latest-gpu
.. code-block:: bash
Choose between different BLAS version:
nvidia-docker run --rm -v ~/workspace:/workspace paddlepaddle/paddle:0.10.0-gpu python /workspace/example.py
Above commands will start a docker container by running :code:`python
/workspace/example.py`. It will stop once :code:`python
/workspace/example.py` finishes.
Another way is to tell docker to start a :code:`/bin/bash` session and
run PaddlePaddle program interactively:
.. code-block:: bash
docker run -it -v ~/workspace:/workspace paddlepaddle/paddle:0.10.0 /bin/bash
# now we are inside docker container
cd /workspace
python example.py
Running with GPU is identical:
.. code-block:: bash
nvidia-docker run -it -v ~/workspace:/workspace paddlepaddle/paddle:0.10.0-gpu /bin/bash
# now we are inside docker container
cd /workspace
python example.py
Develop PaddlePaddle or Train Model Using C++ API
---------------------------------------------------
We will be using PaddlePaddle development image since it contains all
compiling tools and dependencies.
.. code-block:: bash
1. Build PaddlePaddle develop image
# image using MKL by default
docker pull paddlepaddle/paddle
# image using OpenBLAS
docker pull paddlepaddle/paddle:latest-openblas
Use following command to build PaddlePaddle develop image:
.. code-block:: bash
If you want to use legacy versions, choose a tag from
`DockerHub <https://hub.docker.com/r/paddlepaddle/paddle/tags/>`_
and run:
git clone https://github.com/PaddlePaddle/Paddle.git && cd Paddle
docker build -t paddle:dev .
2. Build PaddlePaddle production image
.. code-block:: bash
There are two steps for building production image, the first step is to run:
docker pull paddlepaddle/paddle:[tag]
# i.e.
docker pull docker.paddlepaddle.org/paddle:0.10.0-gpu
.. code-block:: bash
.. _docker_run:
docker run -v $(pwd):/paddle -e "WITH_GPU=OFF" -e "WITH_AVX=OFF" -e "WITH_TEST=ON" paddle:dev
Launch your training program in Docker
------------------------------
The above command will compile PaddlePaddle and create a Dockerfile for building production image. All the generated files are in the build directory. "WITH_GPU" controls if the generated production image supports GPU. "WITH_AVX" controls if the generated production image supports AVX. "WITH_TEST" controls if the unit test will be generated.
Assume that you have already written a PaddlePaddle program
named :code:`train.py` under directory :code:`/home/work` (refer to
`PaddlePaddleBook <http://www.paddlepaddle.org/docs/develop/book/01.fit_a_line/index.cn.html>`_
for more samples), then run the following command:
The second step is to run:
.. code-block:: bash
.. code-block:: bash
cd /home/work
docker run -it -v $PWD:/work paddlepaddle/paddle /work/train.py
docker build -t paddle:prod -f build/Dockerfile ./build
In the above command, :code:`-it` means run the container interactively;
:code:`-v $PWD:/work` means mount the current directory ($PWD will expand
to current absolute path in Linux) under :code:`/work` in the container.
:code:`paddlepaddle/paddle` to specify image to use; finnally
:code:`/work/train.py` is the command to run inside docker.
The above command will generate the production image by copying the compiled PaddlePaddle program into the image.
Also, you can go into the container shell, run or debug your code
interactively:
3. Run unit test
.. code-block:: bash
docker run -it -v $PWD:/work paddlepaddle/paddle /bin/bash
cd /work
python train.py
Following command will run unit test:
**NOTE: We did not install vim in the default docker image to reduce the image size, you can run** :code:`apt-get install -y vim` **to install it if you need to edit python files.**
.. code-block:: bash
docker run -it -v $(pwd):/paddle paddle:dev bash -c "cd /paddle/build && ctest"
.. _docker_run_book:
PaddlePaddle Book
------------------
The Jupyter Notebook is an open-source web application that allows
you to create and share documents that contain live code, equations,
visualizations and explanatory text in a single browser.
PaddlePaddle Book is an interactive Jupyter Notebook for users and developers.
We already exposed port 8888 for this book. If you want to
You can create a container serving PaddlePaddle Book using Jupyter Notebook in
one minute using Docker. PaddlePaddle Book is an interactive Jupyter Notebook
for users and developers.If you want to
dig deeper into deep learning, PaddlePaddle Book definitely is your best choice.
We provide a packaged book image, simply issue the command:
.. code-block:: bash
.. code-block:: bash
docker run -p 8888:8888 paddlepaddle/book
docker run -p 8888:8888 paddlepaddle/book
Then, you would back and paste the address into the local browser:
.. code-block:: text
.. code-block:: text
http://localhost:8888/
http://localhost:8888/
That's all. Enjoy your journey!
.. _docker_run_gpu:
Documentation
-------------
Train with Docker with GPU
------------------------------
Paddle Docker images include an HTML version of C++ source code
generated using `woboq code browser
<https://github.com/woboq/woboq_codebrowser>`_. This makes it easy
for users to browse and understand the C++ source code.
We recommend using
`nvidia-docker <https://github.com/NVIDIA/nvidia-docker>`_
to run GPU training jobs. Please ensure you have latest
GPU driver installed before move on.
As long as we give the Paddle Docker container a name, we can run an
additional Nginx Docker container to serve the volume from the Paddle
container:
.. code-block:: bash
.. code-block:: bash
nvidia-docker run -it -v $PWD:/work paddledev/paddle:latest-gpu /bin/bash
docker run -d --name paddle-cpu-doc paddle:<version>
docker run -d --volumes-from paddle-cpu-doc -p 8088:80 nginx
**NOTE: If you don't have nvidia-docker installed, try the following method to mount CUDA libs and devices into the container.**
.. code-block:: bash
Then we can direct our Web browser to the HTML version of source code
at http://localhost:8088/paddle/
export CUDA_SO="$(\ls /usr/lib64/libcuda* | xargs -I{} echo '-v {}:{}') $(\ls /usr/lib64/libnvidia* | xargs -I{} echo '-v {}:{}')"
export DEVICES=$(\ls /dev/nvidia* | xargs -I{} echo '--device {}:{}')
docker run ${CUDA_SO} ${DEVICES} -it paddledev/paddle:latest-gpu
**About AVX:**
AVX is a kind of CPU instruction can accelerate PaddlePaddle's calculations.
The latest PaddlePaddle Docker image turns AVX on by default, so, if your
computer doesn't support AVX, you'll probably need to
`build <./build_from_source_en.rst>`_ with :code:`WITH_AVX=OFF`.
The following command will tell you whether your computer supports AVX.
.. code-block:: bash
if cat /proc/cpuinfo | grep -i avx; then echo Yes; else echo No; fi
doc/getstarted/build_and_install/index_cn.rst
浏览文件 @ 048cc30e
......@@ -6,12 +6,13 @@
安装流程
++++++++
PaddlePaddle提供Docker镜像来部署环境。
PaddlePaddle提供pip和Docker的安装方式：
.. toctree::
:maxdepth: 1
docker_install_cn.rst
pip_install_cn.rst
docker_install_cn.rst
编译流程
......@@ -19,9 +20,14 @@ PaddlePaddle提供Docker镜像来部署环境。
.. warning::
编译流程主要推荐高级用户查看，普通用户请走安装流程。
建议直接使用上述安装流程，方便快速安装。只有在遇到需要独立定制的二进制时才需要编译。
.. toctree::
:maxdepth: 1
cmake/build_from_source_cn.rst
build_from_source_cn.rst
常见问题解答
++++++++++
`常见问题解答 <http://www.paddlepaddle.org/docs/develop/documentation/zh/faq/build_and_install/index_cn.html>`_
doc/getstarted/build_and_install/index_en.rst
浏览文件 @ 048cc30e
Install and Build
=================
Install PaddlePaddle
----------------------
.. _install_steps:
.. toctree::
:maxdepth: 1
Install Steps
++++++++
You can choose either pip or Docker to complete your install:
.. toctree::
:maxdepth: 1
pip_install_en.rst
docker_install_en.rst
docker_install_en.rst
Build from Source
-----------------
.. warning::
Please use :code:`docker` image to install paddle. The building guide is used for hacking or contributing PaddlePaddle source code.
We recommend to directly install via above installation steps, you'll only need to build PaddlePaddle from source when you need a modifed binary.
.. toctree::
:maxdepth: 1
build_from_source_en.md
FAQ
++++++++++
`FAQ <http://www.paddlepaddle.org/docs/develop/documentation/zh/faq/build_and_install/index_en.html>`_
doc/getstarted/build_and_install/pip_install_cn.rst 0 → 100644
浏览文件 @ 048cc30e
使用pip安装
================================
PaddlePaddle可以使用常用的Python包管理工具
`pip <https://pip.pypa.io/en/stable/installing/>`_
完成安装，并可以在大多数主流的Linux操作系统以及MacOS上执行。
.. _pip_install:
使用pip安装
------------------------------
执行下面的命令即可在当前机器上安装PaddlePaddle的运行时环境，并自动下载安装依赖软件。
.. code-block:: bash
pip install paddlepaddle
如果需要安装支持GPU的版本，需要执行：
.. code-block:: bash
pip install paddlepaddle-gpu
如果需要获取并安装最新的（开发分支）PaddlePaddle，可以从我们的CI系统中下载最新的whl安装包和c-api开发包并安装，
您可以从下面的表格中找到需要的版本：
如果在点击下面链接时出现如下登陆界面，点击“Log in as guest”即可开始下载：
.. image:: paddleci.png
:scale: 50 %
:align: center
.. csv-table:: 各个版本最新的whl包
:header: "版本说明", "cp27-cp27mu", "cp27-cp27mu", "C-API"
:widths: 1, 3, 3, 3
"cpu_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddle.tgz>`_"
"cpu_avx_openblas", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "暂无"
"cuda7.5_cudnn5_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
"cuda8.0_cudnn5_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
"cuda8.0_cudnn7_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
.. _pip_dependency:
运行环境依赖
------------------------------
PaddlePaddle安装包由于不仅仅包含.py程序，而且包含了C++编写的部分，所以我们确保发布的二进制包可以支持主流的Linux操作系统，比如CentOS 6以上，Ubuntu 14.04以上，MacOS 10.12以上。
PaddlePaddle发布的安装包会尽量对齐 `manylinux1 <https://www.python.org/dev/peps/pep-0513/#the-manylinux1-policy>`_ 标准，通常使用CentOS 5作为编译环境。但由于CUDA库通常需要CentOS 6以上，而且CentOS 5即将停止维护，所以我们默认使用CentOS 6作为标准编译环境。
.. csv-table:: PaddlePaddle环境依赖
:header: "依赖", "版本", "说明"
:widths: 10, 15, 30
"操作系统", "Linux, MacOS", "CentOS 6以上，Ubuntu 14.04以上，MacOS 10.12以上"
"Python", "2.7.x", "暂时不支持Python3"
"libc.so", "GLIBC_2.7", "glibc至少包含GLIBC_2.7以上的符号"
"libstdc++.so", "GLIBCXX_3.4.11, CXXABI_1.3.3", "至少包含GLIBCXX_3.4.11, CXXABI_1.3.3以上的符号"
"libgcc_s.so", "GCC_3.3", "至少包含GCC_3.3以上的符号"
.. _pip_faq:
安装常见问题和解决方法
------------------------------
- paddlepaddle*.whl is not a supported wheel on this platform.
出现这个问题的主要原因是，没有找到和当前系统匹配的paddlepaddle安装包。请检查Python版本是否为2.7系列。另外最新的pip官方源中的安装包默认是manylinux1标准，需要使用最新的pip (>9.0.0) 才可以安装。可以使用下面的命令更新您的pip：
.. code-block:: bash
pip install --upgrade pip
如果仍然存在问题，可以执行：
.. code-block:: bash
python -c "import pip; print(pip.pep425tags.get_supported())"
获取当前系统支持的安装包格式，并检查和需安装的包是否匹配。pypi安装包可以在 `这个 <https://pypi.python.org/pypi/paddlepaddle/0.10.5>`_ 链接中找到。
如果系统支持的是 linux_x86_64 而安装包是 manylinux1_x86_64 ，需要升级pip版本到最新； 如果系统支持 manylinux1_x86_64 而安装包（本地）是 linux_x86_64 ，可以重命名这个whl包为 manylinux1_x86_64 再安装。
\ No newline at end of file
doc/getstarted/build_and_install/pip_install_en.rst 0 → 100644
浏览文件 @ 048cc30e
Install Using pip
================================
You can use current widely used Python package management
tool `pip <https://pip.pypa.io/en/stable/installing/>`_
to install PaddlePaddle. This method can be used in
most of current Linux systems or MacOS.
.. _pip_install:
Install Using pip
------------------------------
Run the following command to install PaddlePaddle on the current
machine, it will also download requirements.
.. code-block:: bash
pip install paddlepaddle
If you wish to install GPU version, just run:
.. code-block:: bash
pip install paddlepaddle-gpu
If you wish to install the latest develop branch PaddlePaddle,
you can download the latest whl package from our CI system. Access
the below links, log in as guest, then click at the "Artifact"
tab, you'll find the download link of whl packages.
If the links below shows up the login form, just click "Log in as guest" to start the download:
.. image:: paddleci.png
:scale: 50 %
:align: center
.. csv-table:: whl package of each version
:header: "version", "cp27-cp27mu", "cp27-cp27mu", "C-API"
:widths: 1, 3, 3, 3
"cpu_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/paddle.tgz>`_"
"cpu_avx_openblas", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "Not Available"
"cuda7.5_cudnn5_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
"cuda8.0_cudnn5_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
"cuda8.0_cudnn7_avx_mkl", "`paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27mu-linux_x86_64.whl>`_", "`paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddlepaddle-0.10.0-cp27-cp27m-linux_x86_64.whl>`_", "`paddle.tgz <http://guest@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/paddle.tgz>`_"
.. _pip_dependency:
Runtime Dependency
------------------------------
PaddlePaddle installation packages (whl) does not only contain .py files,
but also binaries built from C++ code. We ensure that PaddlePaddle can
run on current mainline Linux distributions, like CentOS 6, Ubuntu 14.04
and MacOS 10.12.
PaddlePaddle whl packages are trying to satisfy
`manylinux1 <https://www.python.org/dev/peps/pep-0513/#the-manylinux1-policy>`_
standard, which uses CentOS 5 as default build environment. But CUDA libraries
seems only run on CentOS 6 at least, also, CentOS 5 is about to end its lifetime,
so we use CentOS 6 as default build environment.
.. csv-table:: PaddlePaddle Runtime Deps
:header: "Dependency", "version", "description"
:widths: 10, 15, 30
"OS", "Linux, MacOS", "CentOS 6 or later，Ubuntu 14.04 or later，MacOS 10.12 or later"
"Python", "2.7.x", "Currently Python3 is not supported"
"libc.so", "GLIBC_2.7", "glibc at least include GLIBC_2.7 symbols"
"libstdc++.so", "GLIBCXX_3.4.11, CXXABI_1.3.3", "At least include GLIBCXX_3.4.11, CXXABI_1.3.3 symbols"
"libgcc_s.so", "GCC_3.3", "At least include GCC_3.3 symbols"
.. _pip_faq:
FAQ
------------------------------
- paddlepaddle*.whl is not a supported wheel on this platform.
The main cause of this issue is that your current platform is
not supported. Please check that you are using Python 2.7 series.
Besides, pypi only supports manylinux1 standard, you'll need to
upgrade your pip to >9.0.0. Then run the below command:
.. code-block:: bash
pip install --upgrade pip
If the problem still exists, run the following command:
.. code-block:: bash
python -c "import pip; print(pip.pep425tags.get_supported())"
Then you'll get supported package suffixes, then check if it matches
the file name of the whl package. You can find default whl package at
`here <https://pypi.python.org/pypi/paddlepaddle/0.10.5>`_
If your system supports linux_x86_64 but the whl package is manylinux1_x86_64,
you'll need to update pip to the latest version; If your system supports
manylinux1_x86_64 but the whl package is linux_x86_64 you can rename the
file to manylinux1_x86_64 suffix and then install.
doc/getstarted/index_cn.rst
浏览文件 @ 048cc30e
新手入门
============
.. _quick_install:
快速安装
++++++++
PaddlePaddle支持使用pip快速安装，目前支持CentOS 6以上, Ubuntu 14.04以及MacOS 10.12，并安装有Python2.7。
执行下面的命令完成快速安装：
.. code-block:: bash
pip install paddlepaddle
如果需要安装支持GPU的版本，需要执行：
.. code-block:: bash
pip install paddlepaddle-gpu
更详细的安装和编译方法参考：
.. toctree::
:maxdepth: 1
build_and_install/index_cn.rst
concepts/use_concepts_cn.rst
- `深度学习入门课程 <http://book.paddlepaddle.org/index.cn.html>`_
.. _quick_start:
快速开始
++++++++
创建一个 housing.py 并粘贴此Python代码：
.. code-block:: python
import paddle.v2 as paddle
# Initialize PaddlePaddle.
paddle.init(use_gpu=False, trainer_count=1)
# Configure the neural network.
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
# Infer using provided test data.
probs = paddle.infer(
output_layer=y_predict,
parameters=paddle.dataset.uci_housing.model(),
input=[item for item in paddle.dataset.uci_housing.test()()])
for i in xrange(len(probs)):
print 'Predicted price: ${:,.2f}'.format(probs[i][0] * 1000)
执行 :code:`python housing.py` 瞧！ 它应该打印出预测住房数据的清单。
.. toctree::
:maxdepth: 1
concepts/use_concepts_cn.rst
doc/getstarted/index_en.rst
浏览文件 @ 048cc30e
GET STARTED
============
.. _quick_install:
Quick Install
----------------------
You can use pip to install PaddlePaddle with a single command, supports
CentOS 6 above, Ubuntu 14.04 above or MacOS 10.12, with Python 2.7 installed.
Simply run the following command to install:
.. code-block:: bash
pip install paddlepaddle
If you need to install GPU version, run:
.. code-block:: bash
pip install paddlepaddle-gpu
For more details about installation and build:
.. toctree::
:maxdepth: 1
build_and_install/index_en.rst
- `Deep Learning 101 <http://book.paddlepaddle.org/index.html>`_
.. _quick_start:
Quick Start
++++++++
Create a new file called housing.py, and paste this Python
code:
.. code-block:: python
import paddle.v2 as paddle
# Initialize PaddlePaddle.
paddle.init(use_gpu=False, trainer_count=1)
# Configure the neural network.
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
# Infer using provided test data.
probs = paddle.infer(
output_layer=y_predict,
parameters=paddle.dataset.uci_housing.model(),
input=[item for item in paddle.dataset.uci_housing.test()()])
for i in xrange(len(probs)):
print 'Predicted price: ${:,.2f}'.format(probs[i][0] * 1000)
Run :code:`python housing.py` and voila! It should print out a list of predictions
for the test housing data.
doc/howto/dev/write_docs_cn.rst
浏览文件 @ 048cc30e
......@@ -72,7 +72,7 @@ PaddlePaddle的文档构建有三种方式。
cd TO_YOUR_PADDLE_CLONE_PATH
mkdir -p build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Debug -DWITH_GPU=OFF -DWITH_MKLDNN=OFF -DWITH_MKLML=OFF -DWITH_DOC=ON
cmake .. -DCMAKE_BUILD_TYPE=Debug -DWITH_GPU=OFF -DWITH_MKL=OFF -DWITH_DOC=ON
make gen_proto_py
make paddle_docs paddle_docs_cn
......
doc/howto/index_en.rst
浏览文件 @ 048cc30e
......@@ -18,7 +18,6 @@ Development
.. toctree::
:maxdepth: 1
dev/build_en.rst
dev/new_layer_en.rst
dev/contribute_to_paddle_en.md
dev/write_docs_en.rst
......
doc/howto/optimization/cpu_profiling.md 0 → 100644
浏览文件 @ 048cc30e
此教程会介绍如何使用Python的cProfile包，与Python库yep，google perftools来运行性能分析(Profiling)与调优。
运行性能分析可以让开发人员科学的，有条不紊的对程序进行性能优化。性能分析是性能调优的基础。因为在程序实际运行中，真正的瓶颈可能和程序员开发过程中想象的瓶颈相去甚远。
性能优化的步骤，通常是循环重复若干次『性能分析 --> 寻找瓶颈 ---> 调优瓶颈 --> 性能分析确认调优效果』。其中性能分析是性能调优的至关重要的量化指标。
Paddle提供了Python语言绑定。用户使用Python进行神经网络编程，训练，测试。Python解释器通过`pybind`和`swig`调用Paddle的动态链接库，进而调用Paddle C++部分的代码。所以Paddle的性能分析与调优分为两个部分:
* Python代码的性能分析
* Python与C++混合代码的性能分析
## Python代码的性能分析
### 生成性能分析文件
Python标准库中提供了性能分析的工具包，[cProfile](https://docs.python.org/2/library/profile.html)。生成Python性能分析的命令如下:
```bash
python -m cProfile -o profile.out main.py
```
其中`-o`标识了一个输出的文件名，用来存储本次性能分析的结果。如果不指定这个文件，`cProfile`会打印一些统计信息到`stdout`。这不方便我们进行后期处理(进行`sort`, `split`, `cut`等等)。
### 查看性能分析文件
当main.py运行完毕后，性能分析结果文件`profile.out`就生成出来了。我们可以使用[cprofilev](https://github.com/ymichael/cprofilev)来查看性能分析结果。`cprofilev`是一个Python的第三方库。使用它会开启一个HTTP服务，将性能分析结果以网页的形式展示出来。
使用`pip install cprofilev`安装`cprofilev`工具。安装完成后，使用如下命令开启HTTP服务
```bash
cprofilev -a 0.0.0.0 -p 3214 -f profile.out main.py
```
其中`-a`标识HTTP服务绑定的IP。使用`0.0.0.0`允许外网访问这个HTTP服务。`-p`标识HTTP服务的端口。`-f`标识性能分析的结果文件。`main.py`标识被性能分析的源文件。
访问对应网址，即可显示性能分析的结果。性能分析结果格式如下:
```text
ncalls tottime percall cumtime percall filename:lineno(function)
1 0.284 0.284 29.514 29.514 main.py:1(<module>)
4696 0.128 0.000 15.748 0.003 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/executor.py:20(run)
4696 12.040 0.003 12.040 0.003 {built-in method run}
1 0.144 0.144 6.534 6.534 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/__init__.py:14(<module>)
```
每一列的含义是:
| 列名 | 含义 |
| --- | --- |
| ncalls | 函数的调用次数 |
| tottime | 函数实际使用的总时间。该时间去除掉本函数调用其他函数的时间 |
| percall | tottime的每次调用平均时间 |
| cumtime | 函数总时间。包含这个函数调用其他函数的时间 |
| percall | cumtime的每次调用平均时间 |
| filename:lineno(function) | 文件名, 行号，函数名 |
### 寻找性能瓶颈
通常`tottime`和`cumtime`是寻找瓶颈的关键指标。这两个指标代表了某一个函数真实的运行时间。
将性能分析结果按照tottime排序，效果如下:
```text
4696 12.040 0.003 12.040 0.003 {built-in method run}
300005 0.874 0.000 1.681 0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/dataset/mnist.py:38(reader)
107991 0.676 0.000 1.519 0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:219(__init__)
4697 0.626 0.000 2.291 0.000 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:428(sync_with_cpp)
1 0.618 0.618 0.618 0.618 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/__init__.py:1(<module>)
```
可以看到最耗时的函数是C++端的`run`函数。这需要联合我们第二节`Python`与`C++`混合代码的性能分析来进行调优。而`sync_with_cpp`函数的总共耗时很长，每次调用的耗时也很长。于是我们可以点击`sync_with_cpp`的详细信息，了解其调用关系。
```text
Called By:
Ordered by: internal time
List reduced from 4497 to 2 due to restriction <'sync_with_cpp'>
Function was called by...
ncalls tottime cumtime
/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:428(sync_with_cpp) <- 4697 0.626 2.291 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:562(sync_with_cpp)
/home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:562(sync_with_cpp) <- 4696 0.019 2.316 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:487(clone)
1 0.000 0.001 /home/yuyang/perf_test/.env/lib/python2.7/site-packages/paddle/v2/fluid/framework.py:534(append_backward)
Called:
Ordered by: internal time
List reduced from 4497 to 2 due to restriction <'sync_with_cpp'>
```
通常观察热点函数间的调用关系，和对应行的代码，就可以了解到问题代码在哪里。当我们做出性能修正后，再次进行性能分析(profiling)即可检查我们调优后的修正是否能够改善程序的性能。
## Python与C++混合代码的性能分析
### 生成性能分析文件
C++的性能分析工具非常多。常见的包括`gprof`, `valgrind`, `google-perftools`。但是调试Python中使用的动态链接库与直接调试原始二进制相比增加了很多复杂度。幸而Python的一个第三方库`yep`提供了方便的和`google-perftools`交互的方法。于是这里使用`yep`进行Python与C++混合代码的性能分析
使用`yep`前需要安装`google-perftools`与`yep`包。ubuntu下安装命令为
```bash
apt install libgoogle-perftools-dev
pip install yep
```
安装完毕后，我们可以通过
```bash
python -m yep -v main.py
```
生成性能分析文件。生成的性能分析文件为`main.py.prof`。
命令行中的`-v`指定在生成性能分析文件之后，在命令行显示分析结果。我们可以在命令行中简单的看一下生成效果。因为C++与Python不同，编译时可能会去掉调试信息，运行时也可能因为多线程产生混乱不可读的性能分析结果。为了生成更可读的性能分析结果，可以采取下面几点措施:
1. 编译时指定`-g`生成调试信息。使用cmake的话，可以将CMAKE_BUILD_TYPE指定为`RelWithDebInfo`。
2. 编译时一定要开启优化。单纯的`Debug`编译性能会和`-O2`或者`-O3`有非常大的差别。`Debug`模式下的性能测试是没有意义的。
3. 运行性能分析的时候，先从单线程开始，再开启多线程，进而多机。毕竟单线程调试更容易。可以设置`OMP_NUM_THREADS=1`这个环境变量关闭openmp优化。
### 查看性能分析文件
在运行完性能分析后，会生成性能分析结果文件。我们可以使用[pprof](https://github.com/google/pprof)来显示性能分析结果。注意，这里使用了用`Go`语言重构后的`pprof`，因为这个工具具有web服务界面，且展示效果更好。
安装`pprof`的命令和一般的`Go`程序是一样的，其命令如下:
```bash
go get github.com/google/pprof
```
进而我们可以使用如下命令开启一个HTTP服务:
```bash
pprof -http=0.0.0.0:3213 `which python` ./main.py.prof
```
这行命令中，`-http`指开启HTTP服务。`which python`会产生当前Python二进制的完整路径，进而指定了Python可执行文件的路径。`./main.py.prof`输入了性能分析结果。
访问对应的网址，我们可以查看性能分析的结果。结果如下图所示:
![result](./pprof_1.png)
### 寻找性能瓶颈
与寻找Python代码的性能瓶颈类似，寻找Python与C++混合代码的性能瓶颈也是要看`tottime`和`cumtime`。而`pprof`展示的调用图也可以帮助我们发现性能中的问题。
例如下图中，
![kernel_perf](./pprof_2.png)
在一次训练中，乘法和乘法梯度的计算占用2%-4%左右的计算时间。而`MomentumOp`占用了17%左右的计算时间。显然，`MomentumOp`的性能有问题。
在`pprof`中，对于性能的关键路径都做出了红色标记。先检查关键路径的性能问题，再检查其他部分的性能问题，可以更有次序的完成性能的优化。
## 总结
至此，两种性能分析的方式都介绍完毕了。希望通过这两种性能分析的方式，Paddle的开发人员和使用人员可以有次序的，科学的发现和解决性能问题。
doc/mobile/cross_compiling_for_android_cn.md
浏览文件 @ 048cc30e
# 构建Android平台上的PaddlePaddle库
# Android平台编译指南
用户可通过如下两种方式，交叉编译Android平台上适用的PaddlePaddle库：
- 基于Docker容器的编译方式
......
doc/mobile/cross_compiling_for_ios_cn.md
浏览文件 @ 048cc30e
# 构建iOS平台上的PaddlePaddle库
# iOS平台编译指南
交叉编译iOS平台上适用的PaddlePaddle库，需要在MacOS系统上进行。本文的将介绍在MacOS上，从源码交叉编译iOS平台上适用的PaddlePaddle库。
## 准备交叉编译环境
......@@ -25,7 +25,7 @@ iOS平台可选配置参数：
- `IOS_PLATFORM`，可设置为`OS/SIMULATOR`，默认值为`OS`。
- `OS`，构建目标为`arm`架构的iPhone或者iPad等物理设备。
- `SIMULATOR`，构建目标为`x86`架构的模拟器平台。
- `IOS_ARCH`，目标架构。针对不同的`IOS_PLATFORM`，可设置的目标架构如下表所示：
- `IOS_ARCH`，目标架构。针对不同的`IOS_PLATFORM`，可设置的目标架构如下表所示，默认编译所有架构：
<table class="docutils">
<colgroup>
......@@ -41,11 +41,11 @@ iOS平台可选配置参数：
<tbody valign="top">
<tr class="row-even">
<td>OS</td>
<td>armv7, armv7s, arm64 (默认)</td>
<td>armv7, armv7s, arm64 </td>
</tr>
<tr class="row-odd">
<td>SIMULATOR</td>
<td>i386, x86_64 (默认)</td>
<td>i386, x86_64 </td>
</tr>
</tbody>
</table>
......@@ -66,7 +66,7 @@ iOS平台可选配置参数：
```bash
cmake -DCMAKE_SYSTEM_NAME=iOS \
-DIOS_PLATFORM=OS \
-DIOS_ARCH="arm64" \
-DIOS_ARCH="armv7;arm64" \
-DIOS_ENABLE_BITCODE=ON \
-DIOS_USE_VECLIB_FOR_BLAS=ON \
-DCMAKE_INSTALL_PREFIX=your/path/to/install \
......@@ -112,6 +112,6 @@ $ make install
- `lib`目录，其中包含PaddlePaddle的C-API静态库
- `third_party`目录，其中包含所依赖的所有第三方库
注意，不同架构的PaddlePaddle库建议安装到不同的目录下，然后使用`lipo`工具将多个静态库合并成一个支持多个架构的fat库。
注意，如果PaddlePaddle库需要同时支持真机和模拟器，则需要分别编译真机和模拟器版本，然后使用`lipo`工具合并fat库。
自此，PaddlePaddle库已经安装完成，用户可将合成的fat库用于深度学习相关的iOS App中，调用方法见C-API文档。
doc/mobile/cross_compiling_for_raspberry_cn.md
浏览文件 @ 048cc30e
# 构建Raspberry Pi平台上的PaddlePaddle库
# Raspberry Pi平台编译指南
通常有两个方法来构建基于 Rasspberry Pi 的版本：
......
paddle/capi/Matrix.cpp
浏览文件 @ 048cc30e
......@@ -55,7 +55,7 @@ paddle_error paddle_matrix_set_row(paddle_matrix mat,
}
PD_API paddle_error paddle_matrix_set_value(paddle_matrix mat,
paddle_real* value) {
paddle_real* value) {
if (mat == nullptr || value == nullptr) return kPD_NULLPTR;
auto ptr = cast(mat);
if (ptr->mat == nullptr) return kPD_NULLPTR;
......@@ -75,7 +75,7 @@ PD_API paddle_error paddle_matrix_set_value(paddle_matrix mat,
}
PD_API paddle_error paddle_matrix_get_value(paddle_matrix mat,
paddle_real* result) {
paddle_real* result) {
if (mat == nullptr || result == nullptr) return kPD_NULLPTR;
auto ptr = cast(mat);
if (ptr->mat == nullptr) return kPD_NULLPTR;
......
paddle/capi/examples/model_inference/dense/main.c
浏览文件 @ 048cc30e
#include <paddle/capi.h>
#include <time.h>
#include "../common/common.h"
#define CONFIG_BIN "./trainer_config.bin"
......@@ -27,20 +28,19 @@ int main() {
CHECK(paddle_arguments_resize(in_args, 1));
// Create input matrix.
paddle_matrix mat = paddle_matrix_create(/* sample_num */ 10,
paddle_matrix mat = paddle_matrix_create(/* sample_num */ 1,
/* size */ 784,
/* useGPU */ false);
srand(time(0));
std::vector<paddle_real> input;
input.resize(784 * 10);
paddle_real* array;
// Get First row.
CHECK(paddle_matrix_get_row(mat, 0, &array));
for (int i = 0; i < input.size(); ++i) {
input[i] = rand() / ((float)RAND_MAX);
for (int i = 0; i < 784; ++i) {
array[i] = rand() / ((float)RAND_MAX);
}
// Set value for the input matrix
CHECK(paddle_matrix_set_value(mat, input.data()));
CHECK(paddle_arguments_set_value(in_args, 0, mat));
......@@ -53,17 +53,18 @@ int main() {
CHECK(paddle_arguments_get_value(out_args, 0, prob));
std::std::vector<paddle_real> result;
int height;
int width;
uint64_t height;
uint64_t width;
CHECK(paddle_matrix_get_shape(prob, &height, &width);
result.resize(height * width);
CHECK(paddle_matrix_get_value(prob, result.data()));
CHECK(paddle_matrix_get_shape(prob, &height, &width));
CHECK(paddle_matrix_get_row(prob, 0, &array));
printf("Prob: ");
printf("Prob: \n");
for (int i = 0; i < height * width; ++i) {
printf("%.2f ", result[i]);
printf("%.4f ", array[i]);
if ((i + 1) % width == 0) {
printf("\n");
}
}
printf("\n");
......
paddle/capi/matrix.h
浏览文件 @ 048cc30e
......@@ -79,7 +79,7 @@ PD_API paddle_error paddle_matrix_set_row(paddle_matrix mat,
* @note value should contain enough element of data to init the mat
*/
PD_API paddle_error paddle_matrix_set_value(paddle_matrix mat,
paddle_real* value);
paddle_real* value);
/**
* @brief PDMatGetRow Get raw row buffer from matrix
......@@ -93,14 +93,14 @@ PD_API paddle_error paddle_matrix_get_row(paddle_matrix mat,
paddle_real** rawRowBuffer);
/**
* @brief copy data from the matrix
* @brief copy data from the matrix
* @param [in] mat Target matrix
* @param [out] result pointer to store the matrix data
* @param [out] result pointer to store the matrix data
* @return paddle_error
* @note the space of the result should allocated before invoke this API
*/
PD_API paddle_error paddle_matrix_get_value(paddle_matrix mat,
paddle_real* result);
paddle_real* result);
/**
* @brief PDMatCreateNone Create None Matrix
* @return
......
paddle/cuda/include/hl_gpu.h
浏览文件 @ 048cc30e
......@@ -25,7 +25,9 @@ limitations under the License. */
#include "hl_matrix.h"
#include "hl_sequence.h"
#include "hl_sparse.h"
#ifndef PADDLE_MOBILE_INFERENCE
#include "hl_warpctc_wrap.h"
#endif
#ifdef HPPL_STUB_FUNC
#include "stub/hl_aggregate_stub.h"
......
paddle/framework/CMakeLists.txt
浏览文件 @ 048cc30e
......@@ -6,7 +6,10 @@ cc_test(ddim_test SRCS ddim_test.cc DEPS ddim)
nv_test(dim_test SRCS dim_test.cu DEPS ddim)
cc_library(tensor SRCS tensor.cc DEPS ddim place paddle_memory device_context)
cc_test(tensor_test SRCS tensor_test.cc DEPS tensor)
cc_test(tensor_util_test SRCS tensor_util_test.cc DEPS tensor)
cc_test(eigen_test SRCS eigen_test.cc DEPS tensor)
cc_library(lod_tensor SRCS lod_tensor.cc DEPS ddim place tensor framework_proto)
......@@ -51,10 +54,6 @@ cc_library(executor SRCS executor.cc DEPS op_registry device_context scope frame
cc_library(prune SRCS prune.cc DEPS framework_proto)
cc_test(prune_test SRCS prune_test.cc DEPS op_info prune recurrent_op device_context)
cc_library(tensor_array SRCS tensor_array.cc DEPS lod_tensor)
cc_test(tensor_array_test SRCS tensor_array_test.cc DEPS tensor_array place)
cc_test(var_type_inference_test SRCS var_type_inference_test.cc DEPS op_registry
proto_desc)
cc_library(selected_rows SRCS selected_rows.cc DEPS tensor)
......
paddle/framework/backward.cc
浏览文件 @ 048cc30e
......@@ -22,7 +22,6 @@
#include "paddle/framework/block_desc.h"
#include "paddle/framework/op_registry.h"
#include "paddle/operators/dynamic_recurrent_op.h"
#include "paddle/operators/net_op.h"
namespace paddle {
......@@ -218,21 +217,6 @@ static std::unique_ptr<OperatorBase> BackwardRecursive(
return false;
});
// process recurrent gradient op as a special operator.
if (forwardOp.Type() == "dynamic_recurrent") {
// NOTE clean up cycle call somewhere (RNN's stepnet constains itself),
// or this will result in infinite loop.
const auto& rnnop =
*static_cast<const operators::DynamicRecurrentOp*>(&forwardOp);
auto rnn_grad_op =
static_cast<operators::DynamicRecurrentGradientOp*>(grad_op.get());
const auto& stepnet_op =
*static_cast<const OperatorBase*>(&rnnop.rnn.GetStepUnit());
// create stepnet's gradient op
rnn_grad_op->rnn.SetStepUnit(
BackwardRecursive(stepnet_op, no_grad_names, grad_to_var, uniq_id));
}
if (net->ops_.empty()) { // Current no aux op is added to network
return grad_op;
}
......@@ -513,21 +497,16 @@ ParamGradInfoMap AppendBackward(
const int root_block_idx = 0;
auto root_block = program_desc.MutableBlock(root_block_idx);
// insert fill one op for target
// TODO(qiao) add some check to the target.
std::string fill_one_op_out = GradVarName(target.Name());
std::vector<int64_t> target_shape_desc = target.Shape();
std::vector<int> target_shape;
std::transform(target_shape_desc.begin(), target_shape_desc.end(),
std::back_inserter(target_shape),
[](int64_t dim) { return static_cast<int>(dim); });
bool is_scalar = target.Shape() == std::vector<int64_t>{1};
PADDLE_ENFORCE(is_scalar, "target should be scalar");
VLOG(3) << "backward from loss=" << target.Name()
<< " data_type=" << target.GetDataType();
std::unique_ptr<OpDescBind> fill_one_op(
new OpDescBind("fill_constant", {}, {{"Out", {fill_one_op_out}}},
{{"shape", target_shape},
{{"shape", std::vector<int>{1}},
{"value", static_cast<float>(1.0)},
{"data_type", target.GetDataType()}}));
{"dtype", target.GetDataType()}}));
// infer var type of fill_one_op
fill_one_op->InferVarType(root_block);
......
paddle/framework/backward_test.cc
浏览文件 @ 048cc30e
......@@ -508,6 +508,7 @@ TEST(Backward, simple_single_op) {
op->SetOutput("Out", {"out"});
auto target = f::VarDescBind("out");
target.SetShape({1});
auto var_to_grad = AppendBackward(program, target, {});
ASSERT_EQ(block->AllOps().size(), 3UL);
......@@ -544,6 +545,7 @@ TEST(Backward, default_attribute) {
op->CheckAttrs();
auto target = f::VarDescBind("out");
target.SetShape({1});
AppendBackward(program, target, {});
ASSERT_EQ(block->AllOps().size(), 3UL);
......@@ -581,6 +583,7 @@ TEST(Backward, simple_mult_op) {
op3->SetOutput("Out", {"out3"});
auto target = f::VarDescBind("out3");
target.SetShape({1});
size_t forward_len = block->AllOps().size();
auto var_to_grad = AppendBackward(program, target, {});
......@@ -670,6 +673,7 @@ TEST(Backward, intermedia_var_no_grad) {
op4->SetOutput("Out", {"out4"});
auto target = f::VarDescBind("out4");
target.SetShape({1});
size_t forward_len = block->AllOps().size();
auto var_to_grad = AppendBackward(program, target, {"out3"});
......@@ -730,6 +734,7 @@ TEST(Backward, var_no_grad) {
op2->SetOutput("Z", {"z2"});
auto target = f::VarDescBind("z2");
target.SetShape({1});
size_t forward_len = block->AllOps().size();
auto var_to_grad = AppendBackward(program, target, {"z1"});
......@@ -810,6 +815,7 @@ TEST(Backward, shared_var) {
op3->SetOutput("Out", {"out3"});
auto target = f::VarDescBind("out3");
target.SetShape({1});
size_t forward_len = block->AllOps().size();
auto var_to_grad = AppendBackward(program, target, {});
......@@ -888,6 +894,7 @@ TEST(Backward, half_backward) {
op1->SetOutput("Out", {"out"});
auto target = f::VarDescBind("out");
target.SetShape({1});
size_t forward_len = block->AllOps().size();
auto var_to_grad = AppendBackward(program, target, {"b"});
f::OpDescBind *fill_op = block->AllOps()[forward_len];
......
paddle/framework/data_type.h
浏览文件 @ 048cc30e
......@@ -46,6 +46,8 @@ inline std::type_index ToTypeIndex(DataType type) {
return typeid(int);
case DataType::INT64:
return typeid(int64_t);
case DataType::BOOL:
return typeid(bool);
default:
PADDLE_THROW("Not support type %d", type);
}
......@@ -66,6 +68,9 @@ inline void VisitDataType(DataType type, Visitor visitor) {
case DataType::INT64:
visitor.template operator()<int64_t>();
break;
case DataType::BOOL:
visitor.template operator()<bool>();
break;
default:
PADDLE_THROW("Not supported");
}
......
paddle/framework/executor.cc
浏览文件 @ 048cc30e
......@@ -120,7 +120,7 @@ void Executor::Run(const ProgramDescBind& pdesc, Scope* scope, int block_id,
for (auto& op_desc : block.AllOps()) {
auto op = paddle::framework::OpRegistry::CreateOp(*op_desc);
VLOG(10) << op->DebugString();
VLOG(3) << op->DebugString();
op->Run(*local_scope, *device);
}
if (create_local_scope) {
......
paddle/framework/lod_tensor.cc
浏览文件 @ 048cc30e
......@@ -13,6 +13,8 @@
limitations under the License. */
#include "paddle/framework/lod_tensor.h"
#include "paddle/framework/data_type.h"
#include "paddle/framework/framework.pb.h"
#include "paddle/memory/memcpy.h"
#include "paddle/memory/memory.h"
......@@ -27,11 +29,11 @@
namespace paddle {
namespace framework {
std::ostream& operator<<(std::ostream& os, const LoD& lod) {
std::ostream &operator<<(std::ostream &os, const LoD &lod) {
os << "{";
for (auto& v : lod) {
for (auto &v : lod) {
os << "{";
for (auto& i : v) {
for (auto &i : v) {
os << i << ",";
}
os << "}";
......@@ -41,7 +43,7 @@ std::ostream& operator<<(std::ostream& os, const LoD& lod) {
return os;
}
LoD SliceLevels(const LoD& in, size_t level_begin, size_t level_end) {
LoD SliceLevels(const LoD &in, size_t level_begin, size_t level_end) {
LoD new_lod;
new_lod.reserve(level_end - level_begin);
for (size_t i = level_begin; i < level_end; i++) {
......@@ -53,7 +55,7 @@ LoD SliceLevels(const LoD& in, size_t level_begin, size_t level_end) {
return new_lod;
}
LoD SliceInLevel(const LoD& in, size_t level, size_t elem_begin,
LoD SliceInLevel(const LoD &in, size_t level, size_t elem_begin,
size_t elem_end) {
PADDLE_ENFORCE_LT(level, in.size());
PADDLE_ENFORCE_LT(elem_end, in[level].size());
......@@ -64,9 +66,9 @@ LoD SliceInLevel(const LoD& in, size_t level, size_t elem_begin,
res[0].assign(in[level].begin() + elem_begin,
in[level].begin() + elem_end + 1);
for (size_t lvl = 1; lvl < res.size(); lvl++) {
const auto& in_level = in[level + lvl];
const auto& above_level = res[lvl - 1];
auto& out_level = res[lvl];
const auto &in_level = in[level + lvl];
const auto &above_level = res[lvl - 1];
auto &out_level = res[lvl];
out_level.assign(in_level.begin() + above_level.front(),
in_level.begin() + above_level.back() + 1);
}
......@@ -74,33 +76,33 @@ LoD SliceInLevel(const LoD& in, size_t level, size_t elem_begin,
// to make the first offset equals 0, all the elements minus the first
// element
size_t front = res[lvl].front();
for (auto& ele : res[lvl]) {
for (auto &ele : res[lvl]) {
ele -= front;
}
}
return res;
}
LoD ToAbsOffset(const LoD& in) {
LoD ToAbsOffset(const LoD &in) {
// the lowest level stores relative offsets
if (in.empty() || in.size() == 1) return in;
LoD result = in;
for (int level = result.size() - 2; level >= 0; level--) {
for (auto& ele : result[level]) {
for (auto &ele : result[level]) {
ele = result[level + 1][ele];
}
}
return result;
}
bool operator==(const LoD& a, const LoD& b) {
bool operator==(const LoD &a, const LoD &b) {
if (a.size() != b.size()) {
return false;
}
for (size_t i = 0; i < a.size(); i++) {
const auto& a_level = a[i];
const auto& b_level = b[i];
const auto &a_level = a[i];
const auto &b_level = b[i];
if (a_level.size() != b_level.size()) {
return false;
}
......@@ -151,7 +153,7 @@ void LoDTensor::ShrinkInLevel(size_t level, size_t elem_begin,
}
using LoDAndOffset = std::pair<LoD, std::pair<size_t, size_t>>;
LoDAndOffset GetSubLoDAndAbsoluteOffset(const LoD& lod, size_t start_idx,
LoDAndOffset GetSubLoDAndAbsoluteOffset(const LoD &lod, size_t start_idx,
size_t end_idx, size_t start_level) {
LoD sub_lod;
......@@ -170,7 +172,7 @@ LoDAndOffset GetSubLoDAndAbsoluteOffset(const LoD& lod, size_t start_idx,
return LoDAndOffset{sub_lod, {start_idx, end_idx}};
}
void AppendLoD(LoD* lod, const LoD& lod_length) {
void AppendLoD(LoD *lod, const LoD &lod_length) {
PADDLE_ENFORCE(
lod->empty() || lod->size() == lod_length.size(),
"The lod_length should has the same size with the appended lod.");
......@@ -178,12 +180,139 @@ void AppendLoD(LoD* lod, const LoD& lod_length) {
*lod = LoD(lod_length.size(), std::vector<size_t>({0}));
}
for (size_t i = 0; i < lod->size(); ++i) {
auto& level = (*lod)[i];
auto &level = (*lod)[i];
for (size_t len : lod_length[i]) {
level.push_back(level.back() + len);
}
}
}
void SerializeToStream(std::ostream &os, const LoDTensor &tensor,
const platform::DeviceContext &dev_ctx) {
// TODO(typhoonzero): serialize to ostream
{ // the 1st field, uint32_t version
constexpr uint32_t version = 0;
os.write(reinterpret_cast<const char *>(&version), sizeof(version));
}
{ // the 2nd field, tensor description
// int32_t size
// void* protobuf message
framework::TensorDesc desc;
desc.set_data_type(framework::ToDataType(tensor.type()));
auto dims = framework::vectorize(tensor.dims());
auto *pb_dims = desc.mutable_dims();
pb_dims->Resize(static_cast<int>(dims.size()), 0);
std::copy(dims.begin(), dims.end(), pb_dims->begin());
int32_t size = desc.ByteSize();
os.write(reinterpret_cast<const char *>(&size), sizeof(size));
auto out = desc.SerializeAsString();
os.write(out.data(), size);
}
{ // the 3rd field, tensor data
uint64_t size = tensor.memory_size();
auto *data_ptr = tensor.data<void>();
PADDLE_ENFORCE(size < std::numeric_limits<std::streamsize>::max(),
"Index overflow when writing tensor");
if (platform::is_gpu_place(tensor.place())) {
#ifdef PADDLE_WITH_CUDA
constexpr size_t kBufSize = 1024 * 1024 * 64; // 64MB
std::unique_ptr<char[]> buf(new char[kBufSize]);
auto &gpu_dev_ctx =
static_cast<const platform::CUDADeviceContext &>(dev_ctx);
platform::CPUPlace cpu;
uintptr_t data = reinterpret_cast<uintptr_t>(data_ptr);
while (size != 0) {
size_t size_to_write = std::min(kBufSize, static_cast<size_t>(size));
memory::Copy(cpu, buf.get(),
boost::get<platform::GPUPlace>(tensor.place()),
reinterpret_cast<const void *>(data), size_to_write,
gpu_dev_ctx.stream());
gpu_dev_ctx.Wait();
os.write(buf.get(), size_to_write);
data += size_to_write;
size -= size_to_write;
}
#else
PADDLE_THROW("Unexpected branch");
#endif
} else {
os.write(static_cast<const char *>(data_ptr),
static_cast<std::streamsize>(size));
}
}
{ // the 4th field, lod information
// uint64_t lod_level
// uint64_t lod_level_1 size in byte.
// int* lod_level_1 data
// ...
auto lod = tensor.lod();
uint64_t size = lod.size();
os.write(reinterpret_cast<const char *>(&size), sizeof(size));
for (auto &each : lod) {
size = each.size() * sizeof(framework::LoD::value_type::value_type);
os.write(reinterpret_cast<const char *>(&size), sizeof(size));
os.write(reinterpret_cast<const char *>(each.data()),
static_cast<std::streamsize>(size));
}
}
}
void DeserializeFromStream(std::istream &is, LoDTensor *tensor) {
uint32_t version;
is.read(reinterpret_cast<char *>(&version), sizeof(version));
PADDLE_ENFORCE_EQ(version, 0U, "Only version 0 is supported");
framework::TensorDesc desc;
{ // int32_t size
// proto buffer
int32_t size;
is.read(reinterpret_cast<char *>(&size), sizeof(size));
std::unique_ptr<char[]> buf(new char[size]);
is.read(reinterpret_cast<char *>(buf.get()), size);
PADDLE_ENFORCE(desc.ParseFromArray(buf.get(), size),
"Cannot parse tensor desc");
}
{ // read tensor
std::vector<int64_t> dims;
dims.reserve(static_cast<size_t>(desc.dims().size()));
std::copy(desc.dims().begin(), desc.dims().end(), std::back_inserter(dims));
tensor->Resize(framework::make_ddim(dims));
void *buf;
platform::Place cpu = platform::CPUPlace();
switch (desc.data_type()) {
case framework::FP32:
buf = tensor->mutable_data<float>(cpu);
break;
case framework::FP64:
buf = tensor->mutable_data<double>(cpu);
break;
case framework::INT32:
buf = tensor->mutable_data<int>(cpu);
break;
case framework::INT64:
buf = tensor->mutable_data<int64_t>(cpu);
break;
default:
PADDLE_THROW("DataType %d not supported", desc.data_type());
}
is.read(static_cast<char *>(buf), tensor->memory_size());
}
{ // read lod
uint64_t lod_level;
is.read(reinterpret_cast<char *>(&lod_level), sizeof(lod_level));
auto &lod = *tensor->mutable_lod();
lod.resize(lod_level);
for (uint64_t i = 0; i < lod_level; ++i) {
uint64_t size;
is.read(reinterpret_cast<char *>(&size), sizeof(size));
std::vector<size_t> tmp(size / sizeof(size_t));
is.read(reinterpret_cast<char *>(tmp.data()),
static_cast<std::streamsize>(size));
lod[i] = tmp;
}
}
}
} // namespace framework
} // namespace paddle
paddle/framework/lod_tensor.h
浏览文件 @ 048cc30e
......@@ -24,6 +24,7 @@
#include <glog/logging.h>
#include "paddle/framework/ddim.h"
#include "paddle/framework/tensor.h"
#include "paddle/framework/tensor_util.h"
#include "paddle/platform/enforce.h"
#include "paddle/platform/place.h"
......@@ -175,9 +176,9 @@ LoDTensor LodExpand(const LoDTensor& source, const LoD& lod, size_t level,
PADDLE_ENFORCE_EQ(num_instances, lod_level.size() - 1);
for (size_t ins = 0; ins < num_instances; ins++) {
for (size_t elem = lod_level[ins]; elem < lod_level[ins + 1]; elem++) {
tensor.Slice(elem, elem + 1)
.CopyFrom(source.Slice(ins, ins + 1), platform::CPUPlace(),
platform::CPUDeviceContext());
auto slice = tensor.Slice(elem, elem + 1);
CopyFrom(source.Slice(ins, ins + 1), platform::CPUPlace(),
platform::CPUDeviceContext(), &slice);
}
}
return tensor;
......@@ -188,5 +189,14 @@ std::pair<LoD, std::pair<size_t, size_t>> GetSubLoDAndAbsoluteOffset(
void AppendLoD(LoD* lod, const LoD& lod_length);
/*
* Serialize/Desiralize LoDTensor to std::ostream
* You can pass ofstream or ostringstream to serilize to file
* or to a in memory string. GPU tensor will be copied to CPU.
*/
void SerializeToStream(std::ostream& os, const LoDTensor& tensor,
const platform::DeviceContext& dev_ctx);
void DeserializeFromStream(std::istream& is, LoDTensor* tensor);
} // namespace framework
} // namespace paddle
paddle/framework/prune.cc
浏览文件 @ 048cc30e
......@@ -26,6 +26,8 @@ namespace framework {
const std::string kFeedOpType = "feed";
const std::string kFetchOpType = "fetch";
const std::string kDropOutOpType = "dropout";
const std::string kBatchNormOpType = "batch_norm";
bool HasDependentVar(const OpDesc& op_desc,
const std::set<std::string>& dependent_vars) {
......@@ -106,5 +108,26 @@ void Prune(const ProgramDesc& input, ProgramDesc* output) {
prune_impl(input, output, 0);
}
void inference_optimize_impl(const ProgramDesc& input, ProgramDesc* output,
int block_id) {
*output = input;
auto* op_field = output->mutable_blocks(block_id)->mutable_ops();
for (auto& op_desc : *op_field) {
if (op_desc.type() == kDropOutOpType ||
op_desc.type() == kBatchNormOpType) {
for (auto& attr : *op_desc.mutable_attrs()) {
if (attr.name() == "is_test") {
attr.set_b(true);
break;
}
}
}
}
}
void InferenceOptimize(const ProgramDesc& input, ProgramDesc* output) {
inference_optimize_impl(input, output, 0);
}
} // namespace framework
} // namespace paddle
paddle/framework/prune.h
浏览文件 @ 048cc30e
......@@ -22,5 +22,7 @@ namespace framework {
void Prune(const ProgramDesc& input, ProgramDesc* output);
void InferenceOptimize(const ProgramDesc& input, ProgramDesc* output);
} // namespace framework
} // namespace paddle
paddle/framework/tensor.h
浏览文件 @ 048cc30e
......@@ -89,34 +89,6 @@ class Tensor {
/*! The internal of two tensors share the same memory block. */
inline Tensor& ShareDataWith(const Tensor& src);
/**
* @brief Copy the content of external tensor to a new place.
*
* @param[in] src The external tensor.
* @param[in] dst_place The dst place.
* @param[in] ctx The device context contains device resources.
*
* @note CopyFrom supports CPU <-> GPU, GPU <-> GPU.
*/
// TODO(qijun): https://github.com/PaddlePaddle/Paddle/issues/4647
// Remove `CopyFrom` and `CopyFromVector` from Tensor interface
// and make them global functions
inline void CopyFrom(const Tensor& src, const platform::Place& dst_place,
const platform::DeviceContext& ctx);
/**
* @brief Copy the content of an external vector to a tensor.
*
* @param[in] src The external tensor.
* @param[in] ctx The device context contains device resources.
*
* * @note CopyFromVector assumes that the tensor has been resized
* before invoking.
*/
template <typename T>
inline void CopyFromVector(const std::vector<T>& src,
const platform::DeviceContext& ctx);
/**
* @brief Return a sub-tensor of the given tensor.
*
......@@ -141,7 +113,6 @@ class Tensor {
size_t memory_size() const;
private:
inline void check_memory_size() const;
private:
......
paddle/framework/tensor_array.cc 已删除 100644 → 0
浏览文件 @ 2488b81f
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */
#include "paddle/framework/tensor_array.h"
#include <glog/logging.h>
#include <algorithm>
#include <limits>
#include "paddle/framework/eigen.h"
namespace paddle {
namespace framework {
namespace detail {
/*
* Offer an iterator over the length-sorted lod-tensor's top level. The top
* level of a lod-tensor stores batch-size of sequences, each top-level sequence
* may contains several lower-level sequences, sort top-level lod by the numbers
* of lower-level sequences in descending order, so that during RNN's running,
* the batch-size will keep decreasing, the short sentences will end at the tail
* of each batch.
*
* Let's take a simple lod-tensor for example
*
* |(0) |(1) top-level has two instances
* ||| ||||| lower-level
*
* sort by lower-level's length
*
* |(1) |(0)
* ||||| |||
*
* when RNN runs, it get 5 batches (equals the number of elements the longest
* sequence has)
*
* |||||
* |||
*
* the first three batches has two elements, the last two elements just has 1
* element each.
*/
struct DynamicBatchUnpacker {
using value_type = float;
DynamicBatchUnpacker(const LoDTensor& source, size_t level,
bool descend = true)
: source(&source), level(level) {
BuildLengthSortedMeta(descend);
}
LoDTensor GetBatch(size_t index);
std::vector<DySeqMeta> meta;
LoDTensor const* source;
size_t level;
protected:
void BuildLengthSortedMeta(bool descend);
};
LoDTensor PackDynamicBatch(const std::vector<LoDTensor>& source,
const std::vector<DySeqMeta>& meta, const LoD& lod,
size_t level);
std::vector<size_t> GenDyBatchIndice(const DySeqMetaBatch& meta, int batch_id) {
// collect indice need to copy to the batch
std::vector<size_t> indice;
for (const auto& seq : meta) {
size_t id = seq.begin + batch_id;
if (id >= seq.end) break;
indice.push_back(id);
}
return indice;
}
} // namespace detail
const LoDTensor& TensorArray::Read(size_t index) const {
PADDLE_ENFORCE_LE(index, MAX_SIZE, "index[%d] too large", index);
if (index >= size()) {
values_.resize(index + 1);
}
return values_[index];
}
void TensorArray::Write(size_t index, const LoDTensor& value) {
PADDLE_ENFORCE_LE(index, MAX_SIZE, "index[%d] too large", index);
if (index >= size()) {
values_.resize(index + 1);
}
values_[index].set_lod(value.lod());
values_[index].Resize(value.dims());
values_[index].mutable_data<value_type>(value.place());
values_[index].CopyFrom(value, value.place(), platform::CPUDeviceContext());
}
void TensorArray::WriteShared(size_t index, const LoDTensor& value) {
PADDLE_ENFORCE_LE(index, MAX_SIZE, "index[%d] too large", index);
if (index >= size()) {
values_.resize(index + 1);
}
values_[index].set_lod(value.lod());
values_[index].ShareDataWith(value);
}
LoDTensor TensorArray::Pack(size_t level, const std::vector<DySeqMeta>& meta,
const LoD& lod) const {
return detail::PackDynamicBatch(values_, meta, lod, level);
}
DySeqMetaBatch TensorArray::Unpack(const LoDTensor& source, int level,
bool length_desend) {
detail::DynamicBatchUnpacker unpacker(source, level,
length_desend /*descend*/);
// find max length of all the sequences
size_t max_length = 0;
for (const auto& seq : unpacker.meta) {
max_length = std::max(max_length, seq.end - seq.begin);
}
// write batches to values
for (size_t batch_id = 0; batch_id < max_length; batch_id++) {
Write(batch_id, unpacker.GetBatch(batch_id));
}
PADDLE_ENFORCE(!unpacker.meta.empty());
return unpacker.meta;
}
LoDTensor TensorArray::LodPack(size_t level) const {
PADDLE_ENFORCE_GT(size(), 0UL, "no time step exists");
// the levels should be no less than 2
LoDTensor merged;
const LoDTensor *pre, *cur;
pre = &Read(0);
for (size_t step = 1; step < size(); step++) {
cur = &Read(step);
PADDLE_ENFORCE_GT(cur->NumLevels(), 0);
PADDLE_ENFORCE_GT(pre->NumLevels(), 0);
PADDLE_ENFORCE_EQ(pre->NumLevels(), cur->NumLevels());
PADDLE_ENFORCE_EQ(pre->NumElements(level), cur->NumElements(level));
merged = LodPackTwo(*pre, *cur, level);
pre = &merged;
}
return merged;
}
/*
* NOTE currently, only the lowest level supports packing.
* The lowest LoD will be changed, while the relative offsets in levels above
* stay unchanged.
*
* previous step : [0] [1] [3]
* current step: [0 1 2] [2 3] []
* packed to
* [0 0] [0 1] [0 2] [1 2] [1 3] [3]
*/
LoDTensor TensorArray::LodPackTwo(const LoDTensor& pre, const LoDTensor& cur,
size_t level) const {
PADDLE_ENFORCE_EQ(pre.NumLevels(), cur.NumLevels());
PADDLE_ENFORCE_EQ(pre.NumLevels(), level + 1,
"Only the lowest LoD level supports pack temporarily.");
// calculate the result tensor's shape first
size_t num_instances = 0;
for (size_t elem = 0; elem < pre.NumElements(level); elem++) {
size_t prefix_size = pre.NumElements(level, elem);
size_t num_candidates = cur.NumElements(level, elem);
if (num_candidates > 0) {
num_instances += num_candidates * (prefix_size + 1);
} else {
num_instances += prefix_size;
}
}
auto res_dims = pre.dims();
res_dims[0] = num_instances;
LoDTensor result;
result.Resize(res_dims);
result.mutable_data<value_type>(cur.place());
Vector<size_t> last_lod_level;
// copy data
size_t index = 0;
last_lod_level.push_back(index);
for (size_t elem = 0; elem < pre.NumElements(level); elem++) {
size_t prefix_size = pre.NumElements(level, elem);
size_t num_candidates = cur.NumElements(level, elem);
// slice the prefix Tensor
LoDTensor prefix = pre;
prefix.ShrinkInLevel(level, elem, elem + 1);
LoDTensor candidate = cur;
if (num_candidates > 0) {
candidate.ShrinkInLevel(level, elem, elem + 1);
} else { // just push prefix
result.Slice(index, index + prefix_size)
.CopyFrom(prefix, result.place(), platform::CPUDeviceContext());
index += prefix_size;
last_lod_level.push_back(index);
}
for (size_t candi = 0; candi < num_candidates; candi++) {
// TODO(superjom) support GPU
result.Slice(index, index + prefix_size)
.CopyFrom(prefix, result.place(), platform::CPUDeviceContext());
index += prefix_size;
// copy candidate record
result.Slice(index, index + 1)
.CopyFrom(candidate.Slice(candi, candi + 1), result.place(),
platform::CPUDeviceContext());
index++;
last_lod_level.push_back(index);
}
}
// update lod
auto lod = cur.lod();
lod.back() = last_lod_level;
result.set_lod(lod);
return result;
}
/*
* source [0 1 2] [3 4] [5 6 7] will be transformd to a list of LoDTensors such
* as
* [0 3 5] [1 4 6] [2 7] with 1-level LoDs:
* - [0 1 2 3]
* - [0 1 2 3]
* - [0 1 1 2], the [1,1) here means the second sequence is empty
*
* NOTE Unpack a LoDTensor in this approach may result in a big LoD.
*/
void TensorArray::LodUnpack(const LoDTensor& source, size_t level) {
PADDLE_ENFORCE_EQ(level, source.NumLevels() - 1,
"only the lowest LoD level supports unpack.");
const size_t non_empty_instances = source.dims()[0];
size_t index = 0;
Vector<size_t> lowest_lod_level;
lowest_lod_level.push_back(index);
for (size_t step = 0; step < non_empty_instances; step++) {
size_t num_instances = 0;
for (size_t id = 0; id < source.NumElements(level); id++) {
auto instance = source;
instance.ShrinkInLevel(level, id, id + 1);
if (static_cast<size_t>(instance.dims()[0]) > step) {
num_instances++;
index++;
}
lowest_lod_level.push_back(index);
}
// create tensor for this time step
LoDTensor tensor;
auto dims = source.dims();
dims[0] = num_instances;
// set lod
auto lod = source.lod();
lod.back() = lowest_lod_level;
tensor.set_lod(lod);
index = 0;
for (size_t id = 0; id < source.NumElements(level); id++) {
auto instance = source;
instance.ShrinkInLevel(level, id, id + 1);
if (static_cast<size_t>(instance.dims()[0]) > step) {
// copy this instance
tensor.Slice(index, index + 1)
.CopyFrom(instance.Slice(step, step + 1), tensor.place(),
platform::CPUDeviceContext());
index++;
}
}
Write(step, tensor);
}
}
LoDTensor TensorArray::Stack() const {
LoDTensor result;
if (size() == 0) return result;
const auto& first_dims = values_.front().dims();
// check all the values have the same shape
// TODO(superjom) check the same dtypes
for (size_t idx = 1; idx < size(); idx++) {
const auto& value_dims = values_[idx].dims();
PADDLE_ENFORCE_EQ(first_dims, value_dims);
}
// copy
auto result_dims = vectorize(first_dims);
result_dims.insert(result_dims.begin(), size());
result.Resize(make_ddim(result_dims));
result.mutable_data<value_type>(platform::CPUPlace());
for (size_t idx = 0; idx < size(); idx++) {
result.Slice(idx, idx + 1)
.CopyFrom(Read(idx), platform::CPUPlace(),
platform::CPUDeviceContext());
}
return result;
}
void TensorArray::Unstack(const LoDTensor& source) const {
Unstack(source, false /*data_shared*/);
}
void TensorArray::UnstackShared(const LoDTensor& source) const {
Unstack(source, true /*data_shared*/);
}
void TensorArray::Unstack(const LoDTensor& source, bool data_shared) const {
size_t first_dim = source.dims()[0];
DDim value_dims = slice_ddim(source.dims(), 1, source.dims().size());
PADDLE_ENFORCE_GT(first_dim, 0,
"source should have some data to be unstacked");
values_.resize(first_dim);
for (size_t elem = 0; elem < first_dim; elem++) {
// create a new value
auto& value = values_[elem];
if (data_shared) {
// share memory
value.ShareDataWith(source.Slice(elem, elem + 1));
} else {
// copy
value.Resize(value_dims);
value.CopyFrom(source.Slice(elem, elem + 1), platform::CPUPlace(),
platform::CPUDeviceContext());
}
}
}
size_t TensorArray::size() const { return values_.size(); }
namespace detail {
void DynamicBatchUnpacker::BuildLengthSortedMeta(bool descend) {
PADDLE_ENFORCE(meta.empty(), "duplicate build meta");
// collect meta for each sequence in some level
auto lod = SliceLevels(source->lod(), level, level + 1)[0];
for (size_t seq_id = 0; seq_id < lod.size() - 1; seq_id++) {
DySeqMeta seq_meta({lod[seq_id], lod[seq_id + 1], seq_id});
meta.push_back(seq_meta);
}
PADDLE_ENFORCE_GT(meta.size(), 0, "meta is empty");
// sort by length
sort(meta.begin(), meta.end(),
[descend](const DySeqMeta& a, const DySeqMeta& b) {
bool a_ge_b = (a.end - a.begin) > (b.end - b.begin);
return descend ? a_ge_b : !a_ge_b;
});
}
LoDTensor DynamicBatchUnpacker::GetBatch(size_t index) {
PADDLE_ENFORCE(!meta.empty(), "should build meta first");
LoDTensor result;
auto indice = detail::GenDyBatchIndice(meta, index);
PADDLE_ENFORCE(!indice.empty(), "invalid batch at %d", index);
// copy the indice of records in LoDTensor
auto record_dims = slice_ddim(source->dims(), 1, source->dims().size());
auto record_dims_vec = vectorize(record_dims);
record_dims_vec.insert(record_dims_vec.begin(), indice.size());
result.Resize(make_ddim(record_dims_vec));
result.mutable_data<value_type>(platform::CPUPlace());
for (size_t i = 0; i < indice.size(); i++) {
auto index = indice[i];
auto target = result.Slice(i, i + 1);
auto slice = source->Slice(index, index + 1);
target.CopyFrom(slice, platform::CPUPlace(), platform::CPUDeviceContext());
}
return result;
}
// TODO(supejom) to cache lod if reasonable
LoDTensor PackDynamicBatch(const std::vector<LoDTensor>& source,
const std::vector<DySeqMeta>& meta, const LoD& lod,
size_t level) {
PADDLE_ENFORCE(!source.empty());
PADDLE_ENFORCE(!meta.empty());
PADDLE_ENFORCE(!lod.empty());
LoDTensor result;
// init result space
auto record_dims = slice_ddim(source[0].dims(), 1, source[0].dims().size());
auto record_dims_vec = vectorize(record_dims);
auto height = lod[level].back();
record_dims_vec.insert(record_dims_vec.begin(), height);
result.Resize(make_ddim(record_dims_vec));
result.mutable_data<float>(platform::CPUPlace());
for (size_t batch_id = 0; batch_id < source.size(); batch_id++) {
for (size_t seq_id = 0; seq_id < meta.size(); seq_id++) {
const auto& seq_meta = meta[seq_id];
// source is source[batch_id][seq_id]
// target is result[index]
auto index = seq_meta.begin + batch_id;
if (index >= seq_meta.end) break;
auto source_ = source[batch_id].Slice(seq_id, seq_id + 1);
auto target = result.Slice(index, index + 1);
target.CopyFrom(source_, platform::CPUPlace(),
platform::CPUDeviceContext());
}
}
result.set_lod(lod);
return result;
}
} // namespace detail
} // namespace framework
} // namespace paddle
paddle/framework/tensor_array.h 已删除 100644 → 0
浏览文件 @ 2488b81f
/* 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. */
#pragma once
#include <vector>
#include "paddle/framework/lod_tensor.h"
namespace paddle {
namespace framework {
/*
* DyBatchSeqPosition stores indices of the basic element in tensor. It is used
* after lod-tensor's re-assembling, its info can be used to recover the order
* in original lod-tensor.
*/
struct DySeqMeta {
DySeqMeta(size_t begin, size_t end, size_t ori_idx)
: begin(begin), end(end), ori_idx(ori_idx) {}
size_t begin;
size_t end; // not included
size_t ori_idx;
};
using DySeqMetaBatch = std::vector<DySeqMeta>;
/*
* Extract the indices of instances.
*/
std::vector<size_t> GenDyBatchIndice(const DySeqMetaBatch &metas, int batch_id);
/*
* TensorArray is a C-array-like array of tensors, it is meant to be used with
* dynamic iteration primitives such as while_loop. It is used to segment inputs
* and store states in all time steps.
*
* By providing some methods similar to a C++ array, the difinition of some
* state-based dynamic models such as RNN cound be more natural and highly
* flexible.
*/
class TensorArray {
public:
using value_type = float;
// max number of values allowed to store.
const size_t MAX_SIZE{100000};
/*
* Read the value at location `index` in the `TensorArray`.
*/
const LoDTensor &Read(size_t index) const;
/*
* Write value into the index of the TensorArray.
*/
void Write(size_t index, const LoDTensor &value);
/*
* Write value into the index of the TensorArray, with memory shared.
*/
void WriteShared(size_t index, const LoDTensor &value);
/*
* Recover the original LoD-arranged LoDTensor with the `values`, `level` and
* `indice_map`.
*/
LoDTensor Pack(size_t level, const DySeqMetaBatch &meta,
const LoD &lod) const;
/*
* Split LoDTensor in some `level` and write the generated batches to
* `values`, if set `desend`, will sort by length in descending order else in
* ascending order.
*/
DySeqMetaBatch Unpack(const LoDTensor &source, int level, bool length_desend);
/*
* Pack an array of LoDTensors to a LoDTensor.
*/
LoDTensor LodPack(size_t level) const;
/*
* Unpack a LoDTensor to an array of LoDTensors.
*/
void LodUnpack(const LoDTensor &source, size_t level);
/*
* Pack the values into a tensor with rank one higher than each tensor in
* values.
*/
LoDTensor Stack() const;
/*
* Unstacks the given division of a rank-`R` tensor into rank-`(R-1)` tensors.
*/
void Unstack(const LoDTensor &source) const;
/*
* Unstacks the given division of a rank-`R` tensor into rank-`(R-1)` tensors,
* with memory of tensors shared.
*/
void UnstackShared(const LoDTensor &source) const;
/*
* Return the number of values.
*/
size_t size() const;
protected:
void Unstack(const LoDTensor &source, bool data_shared) const;
LoDTensor LodPackTwo(const LoDTensor &pre, const LoDTensor &cur,
size_t level) const;
private:
mutable std::vector<LoDTensor> values_;
}; // class TensorArray
} // namespace framework
} // namespace paddle
paddle/framework/tensor_array_test.cc 已删除 100644 → 0
浏览文件 @ 2488b81f
/* Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserve.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License. */
#include "paddle/framework/tensor_array.h"
#include <gtest/gtest.h>
namespace paddle {
namespace framework {
class TensorArrayTester : public ::testing::Test {
protected:
void SetUp() override {
LoDTensor source;
source.Resize(make_ddim({batch_size, dim}));
int* data = source.mutable_data<int>(platform::CPUPlace());
for (int i = 0; i < 16 * 32; i++) {
data[i] = i;
}
ta.Unstack(source);
}
TensorArray ta;
const int batch_size = 16;
const int dim = 32;
};
TEST_F(TensorArrayTester, Read) {
for (int i = 0; i < batch_size; i++) {
const auto& tensor = ta.Read(i);
ASSERT_EQ(tensor.dims()[0], 1);
ASSERT_EQ(tensor.dims()[1], dim);
}
}
TEST_F(TensorArrayTester, Write) {
LoDTensor source;
source.Resize(make_ddim({1, dim}));
for (int i = 0; i < dim; i++) {
*(source.mutable_data<int>(platform::CPUPlace()) + i) = i;
}
ta.Write(2, source);
const auto& tensor = ta.Read(2);
for (int i = 0; i < dim; i++) {
EXPECT_EQ(*(tensor.data<int>() + i), *(source.data<int>() + i));
}
}
TEST_F(TensorArrayTester, WriteShared) {
LoDTensor source;
source.Resize(make_ddim({1, dim}));
for (int i = 0; i < dim; i++) {
*(source.mutable_data<int>(platform::CPUPlace()) + i) = i;
}
ta.WriteShared(2, source);
const auto& tensor = ta.Read(2);
for (int i = 0; i < dim; i++) {
EXPECT_EQ(*(tensor.data<int>() + i), *(source.data<int>() + i));
}
EXPECT_EQ(source.data<int>(), tensor.data<int>());
}
class TensorArrayPackTester : public ::testing::Test {
protected:
virtual void SetUp() override {
lod.push_back(std::vector<size_t>{0, 2, 9, 13});
source.set_lod(lod);
source.Resize(make_ddim({13, 128}));
source.mutable_data<int>(platform::CPUPlace());
// content of each setence: 0 1 2 3 4
const auto& level = lod.front();
for (size_t i = 0; i < level.size() - 1; i++) {
size_t begin = level[i];
size_t end = level[i + 1];
for (size_t j = begin; j < end; j++) {
auto record = source.Slice(j, j + 1);
for (int dim = 0; dim < 128; dim++) {
record.mutable_data<int>(platform::CPUPlace())[dim] = j - begin;
}
}
}
// unpack
meta = ta.Unpack(source, 0, true);
}
LoD lod;
TensorArray ta;
LoDTensor source;
std::vector<DySeqMeta> meta;
};
TEST_F(TensorArrayPackTester, Unpack) {
ASSERT_EQ(ta.size(), 7UL);
const auto& t0 = ta.Read(0);
const auto& t1 = ta.Read(1);
ASSERT_EQ(t0.data<int>()[0], int(0));
ASSERT_EQ(t1.data<int>()[0], int(1));
}
TEST_F(TensorArrayPackTester, Pack) {
LoDTensor packed = ta.Pack(0, meta, lod);
}
TEST_F(TensorArrayTester, size) {
ASSERT_EQ(ta.size(), static_cast<size_t>(batch_size));
}
TEST(TensorArray, LodPack) {
// three time steps, each step stores a LoDTensors
// - [0] [1]
// - [2 3], [4 5]
// - [6 7] [] [8], [9, 10]
// try to get a LoDTensor with content:
// - [0 2 6]
// - [0 2 7]
// - [0 3]
// - [1 4 8]
// - [1 5 9]
// - [1 5 10]
std::array<LoDTensor, 3> tensors;
tensors[0].Resize(make_ddim({2, 1}));
tensors[1].Resize(make_ddim({4, 1}));
tensors[2].Resize(make_ddim({5, 1}));
int index = 0;
for (auto& t : tensors) {
t.mutable_data<int>(platform::CPUPlace());
for (int i = 0; i < t.dims()[0]; i++) {
t.data<int>()[i] = index;
index++;
}
}
std::array<LoD, 3> lods;
std::vector<std::vector<size_t>> levels{
{0, 1, 2}, {0, 2, 4}, {0, 2, 2, 3, 5}};
for (int i = 0; i < 3; i++) {
lods[i].emplace_back(levels[i].begin(), levels[i].end());
}
TensorArray ta;
for (int i = 0; i < 3; i++) {
tensors[i].set_lod(lods[i]);
ta.Write(i, tensors[i]);
}
auto merged = ta.LodPack(0);
std::vector<int> target_tensor_data{{0, 2, 6, // 0
0, 2, 7, // 1
0, 3, // 2
1, 4, 8, // 3
1, 5, 9, // 5
1, 5, 10}};
EXPECT_EQ(merged.dims()[0], (int)target_tensor_data.size());
for (size_t i = 0; i < target_tensor_data.size(); i++) {
EXPECT_EQ(target_tensor_data[i], merged.data<int>()[i]);
}
}
} // namespace framework
} // namespace paddle
paddle/framework/tensor_impl.h
浏览文件 @ 048cc30e
......@@ -150,84 +150,6 @@ inline Tensor& Tensor::ShareDataWith(const Tensor& src) {
return *this;
}
inline void Tensor::CopyFrom(const Tensor& src,
const platform::Place& dst_place,
const platform::DeviceContext& ctx) {
src.check_memory_size();
Resize(src.dims());
auto src_place = src.holder_->place();
auto src_ptr = src.data<void>();
auto dst_ptr = mutable_data(dst_place, src.type());
auto size = src.numel() * SizeOfType(src.type());
if (platform::is_cpu_place(src_place) && platform::is_cpu_place(dst_place)) {
memory::Copy(boost::get<platform::CPUPlace>(dst_place), dst_ptr,
boost::get<platform::CPUPlace>(src_place), src_ptr, size);
}
#ifdef PADDLE_WITH_CUDA
else if (platform::is_gpu_place(src_place) &&
platform::is_cpu_place(dst_place)) {
auto src_gpu_place = boost::get<platform::GPUPlace>(src_place);
auto dst_cpu_place = boost::get<platform::CPUPlace>(dst_place);
auto ctx_place = ctx.GetPlace();
PADDLE_ENFORCE(platform::is_gpu_place(ctx_place));
auto ctx_gpu_place = boost::get<platform::GPUPlace>(ctx_place);
PADDLE_ENFORCE_EQ(src_gpu_place, ctx_gpu_place);
memory::Copy(
dst_cpu_place, dst_ptr, src_gpu_place, src_ptr, size,
reinterpret_cast<const platform::CUDADeviceContext&>(ctx).stream());
} else if (platform::is_cpu_place(src_place) &&
platform::is_gpu_place(dst_place)) {
auto src_cpu_place = boost::get<platform::CPUPlace>(src_place);
auto dst_gpu_place = boost::get<platform::GPUPlace>(dst_place);
auto ctx_place = ctx.GetPlace();
PADDLE_ENFORCE(platform::is_gpu_place(ctx_place));
auto ctx_gpu_place = boost::get<platform::GPUPlace>(ctx_place);
PADDLE_ENFORCE_EQ(dst_gpu_place, ctx_gpu_place);
memory::Copy(
dst_gpu_place, dst_ptr, src_cpu_place, src_ptr, size,
reinterpret_cast<const platform::CUDADeviceContext&>(ctx).stream());
} else if (platform::is_gpu_place(src_place) &&
platform::is_gpu_place(dst_place)) {
auto src_gpu_place = boost::get<platform::GPUPlace>(src_place);
auto dst_gpu_place = boost::get<platform::GPUPlace>(dst_place);
auto ctx_place = ctx.GetPlace();
PADDLE_ENFORCE(platform::is_gpu_place(ctx_place));
auto ctx_gpu_place = boost::get<platform::GPUPlace>(ctx_place);
PADDLE_ENFORCE_EQ(src_gpu_place, ctx_gpu_place);
memory::Copy(
dst_gpu_place, dst_ptr, src_gpu_place, src_ptr, size,
reinterpret_cast<const platform::CUDADeviceContext&>(ctx).stream());
}
#endif
}
template <typename T>
inline void Tensor::CopyFromVector(const std::vector<T>& src,
const platform::DeviceContext& ctx) {
auto dst_place = ctx.GetPlace();
auto src_ptr = static_cast<const void*>(src.data());
platform::CPUPlace src_place;
auto dst_ptr = static_cast<void*>(mutable_data<T>(dst_place));
auto size = src.size() * sizeof(T);
if (platform::is_cpu_place(dst_place)) {
memory::Copy(boost::get<platform::CPUPlace>(dst_place), dst_ptr, src_place,
src_ptr, size);
}
#ifdef PADDLE_WITH_CUDA
else if (platform::is_gpu_place(dst_place)) {
memory::Copy(
boost::get<platform::GPUPlace>(dst_place), dst_ptr, src_place, src_ptr,
size,
reinterpret_cast<const platform::CUDADeviceContext&>(ctx).stream());
}
#endif
}
inline Tensor Tensor::Slice(int begin_idx, int end_idx) const {
check_memory_size();
PADDLE_ENFORCE_GE(begin_idx, 0,
......
paddle/framework/tensor_test.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/framework/tensor_util.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/framework/tensor_util_test.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/CMakeLists.txt
浏览文件 @ 048cc30e
......@@ -73,7 +73,6 @@ if(MOBILE_INFERENCE)
list(REMOVE_ITEM GSERVER_SOURCES
dataproviders/DataProvider.cpp
dataproviders/MultiDataProvider.cpp
dataproviders/ProtoDataProvider.cpp
dataproviders/PyDataProvider2.cpp
dataproviders/PyDataProvider.cpp)
......
paddle/gserver/activations/ActivationFunction.cpp
浏览文件 @ 048cc30e
......@@ -212,6 +212,37 @@ Error __must_check backward(Argument& act) {
}
END_DEFINE_ACTIVATION(sequence_softmax)
/*
* @brief SoftSign Activation.
* \f[
* f(z) = \frac{z}{1 + |z|}
* \f]
*/
BEGIN_DEFINE_ACTIVATION(softsign)
private:
MatrixPtr denominator_;
Error __must_check forward(Argument& act) {
size_t height = act.value->getHeight();
size_t width = act.value->getWidth();
Matrix::resizeOrCreate(
denominator_, height, width, false, useGpu(act.deviceId));
denominator_->assign(*act.value);
denominator_->abs2();
denominator_->add(1.);
act.value->dotDiv(*act.value, *denominator_);
return Error();
}
Error __must_check backward(Argument& act) {
denominator_->square2();
denominator_->scalarDiv(*denominator_, 1.);
act.grad->dotMul(*act.grad, *denominator_);
return Error();
}
END_DEFINE_ACTIVATION(softsign)
/**
* @brief Relu Activation.
* forward. y = max(0, z)
......
paddle/gserver/dataproviders/DataProvider.cpp
浏览文件 @ 048cc30e
......@@ -16,8 +16,8 @@ limitations under the License. */
#include <unistd.h>
#include <algorithm>
#include "ProtoDataProvider.h"
#include "paddle/utils/Logging.h"
#include "paddle/utils/Stat.h"
#include "paddle/utils/StringUtil.h"
#include "paddle/utils/Util.h"
......@@ -164,8 +164,6 @@ DataProvider* DataProvider::create(const DataConfig& config,
REGISTER_DATA_PROVIDER(simple, SimpleDataProvider);
REGISTER_DATA_PROVIDER(dummy, DummyDataProvider);
REGISTER_DATA_PROVIDER(proto, ProtoDataProvider);
REGISTER_DATA_PROVIDER(proto_sequence, ProtoSequenceDataProvider);
int64_t DataProvider::getNextBatch(int64_t size, DataBatch* batch) {
int64_t batchSize = doubleBuffer_ ? getNextBatchFromBuffer(size, batch)
......
paddle/gserver/layers/BatchNormBaseLayer.cpp
浏览文件 @ 048cc30e
......@@ -41,6 +41,7 @@ bool BatchNormBaseLayer::init(const LayerMap& layerMap,
useGlobalStats_ = config_.use_global_stats();
}
movingAvgFraction_ = config_.moving_average_fraction();
epsilon_ = config_.epsilon();
weight_.reset(new Weight(1, channels_, parameters_[0]));
movingMean_.reset(new Weight(1, channels_, parameters_[1]));
......
paddle/gserver/layers/BatchNormBaseLayer.h
浏览文件 @ 048cc30e
......@@ -94,6 +94,8 @@ protected:
bool useGlobalStats_;
// use to compute moving mean and variance.
real movingAvgFraction_;
// Epsilon is a small random noise used in batch normalization for stability.
real epsilon_;
};
} // namespace paddle
paddle/gserver/layers/BatchNormalizationLayer.h
浏览文件 @ 048cc30e
......@@ -39,9 +39,6 @@ public:
void backward(const UpdateCallback& callback = nullptr) override;
protected:
/// Epsilon value used in the batch normalization formula.
static const real EPS;
/// Load pre-calculated mean and std.
void setMeanAndStd();
......
paddle/gserver/layers/DotProdLayer.cpp 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/layers/FactorizationMachineLayer.cpp 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/layers/FactorizationMachineLayer.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/layers/L2DistanceLayer.cpp 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/layers/L2DistanceLayer.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/layers/MKLDNNConcatLayer.cpp 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/layers/MKLDNNConcatLayer.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/layers/MKLDNNLayer.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/tests/CMakeLists.txt
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/tests/MKLDNNTester.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/gserver/tests/proto_files.txt 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/gserver/tests/sequence_recurrent_group.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/math/CpuSparseMatrix.cpp
浏览文件 @ 048cc30e
此差异已折叠。
paddle/math/CpuSparseMatrix.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/memory/CMakeLists.txt
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/CMakeLists.txt
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/activation_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/activation_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adadelta_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adadelta_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adadelta_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adagrad_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adam_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adam_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adam_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adamax_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adamax_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/adamax_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/array_operator.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/assign_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/batch_norm_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/beam_search_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/cast_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/cast_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/conv_cudnn_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/conv_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/conv_op.cu.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/conv_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/detail/CMakeLists.txt 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/detail/recv_impl.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/detail/send_impl.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/detail/send_recv.proto 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/detail/send_recv_impl.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/detail/simple_block_queue.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/dropout_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/dropout_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/dropout_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/dynamic_recurrent_op.cc 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/operators/dynamic_recurrent_op.h 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/operators/dynamic_recurrent_op_test.cc 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/operators/expand_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/feed_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/fetch_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/ftrl_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/ftrl_op.cu 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/ftrl_op.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/gru_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/gru_unit_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/gru_unit_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/huber_loss_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/load_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/lod_reset_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/log_loss_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/log_loss_op.cu 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/log_loss_op.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/logical_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/logical_op.cu 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/logical_op.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/gru_compute.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/im2col.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/im2col.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/maxouting.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/maxouting.cu 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/maxouting.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/pooling.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/pooling.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/pooling.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/unpooling.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/unpooling.cu 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/unpooling.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/math/vol2col.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/max_sequence_len_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/maxout_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/maxout_op.cu.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/maxout_op.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/multiplex_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/nccl_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/nccl_op_test.cu.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/pool_cudnn_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/pool_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/pool_op.cu.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/recurrent_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/recv_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/reshape_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/rnn/recurrent_op_utils.cc 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/operators/rnn/recurrent_op_utils.h 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/operators/roi_pool_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/roi_pool_op.cu 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/roi_pool_op.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/save_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/scale_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/scale_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/send_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/send_recv_op_test.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/sequence_slice_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/sequence_slice_op.cu 0 → 100755
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/sequence_slice_op.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/sgd_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/softmax_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/sum_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/sum_op.cu
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/sum_op.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/tensor.save 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/unpool_op.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/unpool_op.cu.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/unpool_op.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/operators/while_op.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/CMakeLists.txt
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/cuda_helper.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/cuda_profiler.h 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/cudnn_helper.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/dynload/cudnn.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/dynload/cudnn.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/enforce.cc 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
paddle/platform/enforce.h
浏览文件 @ 048cc30e
此差异已折叠。
paddle/pybind/CMakeLists.txt
浏览文件 @ 048cc30e
此差异已折叠。
paddle/pybind/protobuf.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/pybind/pybind.cc
浏览文件 @ 048cc30e
此差异已折叠。
paddle/scripts/docker/README.md
浏览文件 @ 048cc30e
此差异已折叠。
paddle/scripts/docker/build.sh
浏览文件 @ 048cc30e
此差异已折叠。
paddle/scripts/submit_local.sh.in
浏览文件 @ 048cc30e
此差异已折叠。
paddle/scripts/travis/build_doc.sh
浏览文件 @ 048cc30e
此差异已折叠。
paddle/trainer/Trainer.cpp
浏览文件 @ 048cc30e
此差异已折叠。
paddle/trainer/tests/CMakeLists.txt
浏览文件 @ 048cc30e
此差异已折叠。
paddle/trainer/tests/chunking.conf 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/compare_sparse_data 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/data_bin_part 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/gen_proto_data.py 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/mnist.list 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/mnist_bin_part 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/test.txt 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/test_CompareTwoOpts.cpp 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/test_files.txt 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/train.list 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/train.txt 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/train_files.txt 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
paddle/trainer/tests/train_sparse.list 已删除 100644 → 0
浏览文件 @ 2488b81f
此差异已折叠。
proto/ModelConfig.proto
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/__init__.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/__init__.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/evaluator.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/executor.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/framework.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/io.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/layers.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/nets.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/optimizer.py
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/param_attr.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/profiler.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_ftrl_op.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_log_loss_op.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_logical_op.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_maxout_op.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_mnist_if_else_op.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_profiler.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_roi_pool_op.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_sequence_slice_op.py 0 → 100644
浏览文件 @ 048cc30e
此差异已折叠。
python/paddle/v2/fluid/tests/test_unpool_op.py 0 → 100644
浏览文件 @ 048cc30e
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