Skip to content

P
Paddle

Crayon鑫 / Paddle
与 Fork 源项目一致

Fork自 PaddlePaddle / Paddle

通知 1
Star 1
Fork 0
P
Paddle
提交 bfe7e242 编写于 作者: W wanghaoshuang
浏览文件
操作

Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into blockexpand

上级 92baa885 38c61053
展开全部 隐藏空白更改
内联 并排
Showing with 37647 addition and 15370 deletion

要显示的变更太多。

To preserve performance only 1000 of 1000+ files are displayed.
文本差异 电子邮件补丁
.clang-format
浏览文件 @ bfe7e242
......@@ -25,4 +25,3 @@ AllowAllParametersOfDeclarationOnNextLine: true
BinPackParameters: false
BinPackArguments: false
...
.copyright.hook 0 → 100644
浏览文件 @ bfe7e242
from __future__ import absolute_import
from __future__ import print_function
from __future__ import unicode_literals
import argparse
import io, re
import sys, os
import subprocess
import platform
COPYRIGHT = '''
Copyright (c) 2016 PaddlePaddle Authors. All Rights 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.
'''
LANG_COMMENT_MARK = None
NEW_LINE_MARK = None
COPYRIGHT_HEADER = None
if platform.system() == "Windows":
NEW_LINE_MARK = "\r\n"
else:
NEW_LINE_MARK = '\n'
COPYRIGHT_HEADER = COPYRIGHT.split(NEW_LINE_MARK)[1]
p = re.search('(\d{4})', COPYRIGHT_HEADER).group(0)
process = subprocess.Popen(["date", "+%Y"], stdout=subprocess.PIPE)
date, err = process.communicate()
date = date.decode("utf-8").rstrip("\n")
COPYRIGHT_HEADER = COPYRIGHT_HEADER.replace(p, date)
def generate_copyright(template, lang='C'):
if lang == 'Python':
LANG_COMMENT_MARK = '#'
else:
LANG_COMMENT_MARK = "//"
lines = template.split(NEW_LINE_MARK)
ans = LANG_COMMENT_MARK + COPYRIGHT_HEADER + NEW_LINE_MARK
for lino, line in enumerate(lines):
if lino == 0 or lino == 1 or lino == len(lines) - 1: continue
ans += LANG_COMMENT_MARK + line + NEW_LINE_MARK
return ans
def lang_type(filename):
if filename.endswith(".py"):
return "Python"
elif filename.endswith(".h"):
return "C"
elif filename.endswith(".hpp"):
return "C"
elif filename.endswith(".cc"):
return "C"
elif filename.endswith(".cpp"):
return "C"
elif filename.endswith(".cu"):
return "C"
elif filename.endswith(".cuh"):
return "C"
elif filename.endswith(".go"):
return "C"
elif filename.endswith(".proto"):
return "C"
else:
print("Unsupported filetype")
exit(0)
def main(argv=None):
parser = argparse.ArgumentParser(
description='Checker for copyright declaration.')
parser.add_argument('filenames', nargs='*', help='Filenames to check')
args = parser.parse_args(argv)
retv = 0
for filename in args.filenames:
first_line = io.open(filename).readline()
if "COPYRIGHT" in first_line.upper() : continue
original_contents = io.open(filename).read()
new_contents = generate_copyright(
COPYRIGHT, lang_type(filename)) + original_contents
print('Auto Insert Copyright Header {}'.format(filename))
retv = 1
with io.open(filename, 'w') as output_file:
output_file.write(new_contents)
return retv
if __name__ == '__main__':
exit(main())
.gitignore
浏览文件 @ bfe7e242
......@@ -28,3 +28,4 @@ cmake_install.cmake
paddle/.timestamp
python/paddlepaddle.egg-info/
paddle/pybind/pybind.h
python/paddle/version.py
.pre-commit-config.yaml
浏览文件 @ bfe7e242
......@@ -31,3 +31,11 @@
- id: go-fmt
types:
- go
- repo: local
hooks:
- id: copyright_checker
name: copyright_checker
entry: python ./.copyright.hook
language: system
files: \.(c|cc|cxx|cpp|cu|h|hpp|hxx|proto|py)$
exclude: (?!.*third_party)^.*$ | (?!.*book)^.*$
.travis.yml
浏览文件 @ bfe7e242
......@@ -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 @ bfe7e242
......@@ -20,6 +20,10 @@ set(PADDLE_BINARY_DIR ${CMAKE_CURRENT_BINARY_DIR})
include(system)
project(paddle CXX C Go)
message(STATUS "CXX compiler: ${CMAKE_CXX_COMPILER}, version: "
"${CMAKE_CXX_COMPILER_ID} ${CMAKE_CXX_COMPILER_VERSION}")
message(STATUS "C compiler: ${CMAKE_C_COMPILER}, version: "
"${CMAKE_C_COMPILER_ID} ${CMAKE_C_COMPILER_VERSION}")
find_package(Sphinx)
if(NOT CMAKE_CROSSCOMPILING)
......@@ -54,7 +58,9 @@ option(WITH_C_API "Compile PaddlePaddle with C-API(Prediction)" OFF)
option(WITH_GOLANG "Compile PaddlePaddle with GOLANG" OFF)
option(GLIDE_INSTALL "Download and install go dependencies " ON)
option(USE_NNPACK "Compile PaddlePaddle with NNPACK library" OFF)
option(WITH_DISTRIBUTE "Compile with grpc distributed support" OFF)
option(USE_EIGEN_FOR_BLAS "Use matrix multiplication in Eigen" OFF)
option(WITH_ARM_FP16 "Use half precision support on armv8.2-a cpu" OFF)
# CMAKE_BUILD_TYPE
if(NOT CMAKE_BUILD_TYPE)
......@@ -67,9 +73,6 @@ if(ANDROID OR IOS)
if(ANDROID)
if(${CMAKE_SYSTEM_VERSION} VERSION_LESS "16")
message(FATAL_ERROR "Unsupport standalone toolchains with Android API level lower than 16")
elseif(${CMAKE_SYSTEM_VERSION} VERSION_LESS "21")
# TODO: support glog for Android api 16 ~ 19 in the future
message(WARNING "Using the unofficial git repository <https://github.com/Xreki/glog.git> instead")
endif()
endif()
......@@ -83,6 +86,8 @@ if(ANDROID OR IOS)
"Disable RDMA when cross-compiling for Android and iOS" FORCE)
set(WITH_MKL OFF CACHE STRING
"Disable MKL when cross-compiling for Android and iOS" FORCE)
set(WITH_GOLANG OFF CACHE STRING
"Disable golang when cross-compiling for Android and iOS" FORCE)
# Compile PaddlePaddle mobile inference library
if (NOT WITH_C_API)
......@@ -133,6 +138,8 @@ include(external/any) # download libn::any
include(external/eigen) # download eigen3
include(external/pybind11) # download pybind11
include(external/nccl)
include(external/cares)
include(external/grpc)
include(cudnn) # set cudnn libraries, must before configure
include(configure) # add paddle env configuration
......@@ -194,6 +201,10 @@ if(WITH_GOLANG)
endif(WITH_GOLANG)
set(PADDLE_PYTHON_BUILD_DIR "${CMAKE_CURRENT_BINARY_DIR}/python/build")
SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
SET(CMAKE_C_FLAGS_RELWITHDEBINFO "-O3 -g -DNDEBUG")
add_subdirectory(paddle)
if(WITH_PYTHON)
add_subdirectory(python)
......
Dockerfile
浏览文件 @ bfe7e242
......@@ -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
......
README.md
浏览文件 @ bfe7e242
......@@ -2,8 +2,8 @@
[![Build Status](https://travis-ci.org/PaddlePaddle/Paddle.svg?branch=develop)](https://travis-ci.org/PaddlePaddle/Paddle)
[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://doc.paddlepaddle.org/develop/doc/)
[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://doc.paddlepaddle.org/develop/doc_cn/)
[![Documentation Status](https://img.shields.io/badge/docs-latest-brightgreen.svg?style=flat)](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/index_en.html)
[![Documentation Status](https://img.shields.io/badge/中文文档-最新-brightgreen.svg)](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/index_cn.html)
[![Coverage Status](https://coveralls.io/repos/github/PaddlePaddle/Paddle/badge.svg?branch=develop)](https://coveralls.io/github/PaddlePaddle/Paddle?branch=develop)
[![Release](https://img.shields.io/github/release/PaddlePaddle/Paddle.svg)](https://github.com/PaddlePaddle/Paddle/releases)
[![License](https://img.shields.io/badge/license-Apache%202-blue.svg)](LICENSE)
......@@ -36,7 +36,8 @@ Please refer to our [release announcement](https://github.com/PaddlePaddle/Paddl
examples:
- Optimized math operations through SSE/AVX intrinsics, BLAS libraries
(e.g. MKL, ATLAS, cuBLAS) or customized CPU/GPU kernels.
(e.g. MKL, OpenBLAS, cuBLAS) or customized CPU/GPU kernels.
- Optimized CNN networks through MKL-DNN library.
- Highly optimized recurrent networks which can handle **variable-length**
sequence without padding.
- Optimized local and distributed training for models with high dimensional
......@@ -61,32 +62,32 @@ Please refer to our [release announcement](https://github.com/PaddlePaddle/Paddl
## Installation
It is recommended to check out the
[Docker installation guide](http://doc.paddlepaddle.org/develop/doc/getstarted/build_and_install/docker_install_en.html)
[Docker installation guide](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/docker_install_en.html)
before looking into the
[build from source guide](http://doc.paddlepaddle.org/develop/doc/getstarted/build_and_install/build_from_source_en.html).
[build from source guide](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/build_and_install/build_from_source_en.html).
## Documentation
We provide [English](http://doc.paddlepaddle.org/develop/doc/) and
[Chinese](http://doc.paddlepaddle.org/doc_cn/) documentation.
We provide [English](http://www.paddlepaddle.org/docs/develop/documentation/en/getstarted/index_en.html) and
[Chinese](http://www.paddlepaddle.org/docs/develop/documentation/zh/getstarted/index_cn.html) documentation.
- [Deep Learning 101](http://book.paddlepaddle.org/index.html)
- [Deep Learning 101](http://www.paddlepaddle.org/docs/develop/book/01.fit_a_line/index.html)
You might want to start from this online interactive book that can run in a Jupyter Notebook.
- [Distributed Training](http://doc.paddlepaddle.org/develop/doc/howto/usage/cluster/cluster_train_en.html)
- [Distributed Training](http://www.paddlepaddle.org/docs/develop/documentation/en/howto/usage/cluster/cluster_train_en.html)
You can run distributed training jobs on MPI clusters.
- [Distributed Training on Kubernetes](http://doc.paddlepaddle.org/develop/doc/howto/usage/k8s/k8s_en.html)
- [Distributed Training on Kubernetes](http://www.paddlepaddle.org/docs/develop/documentation/en/howto/usage/cluster/k8s_en.html)
You can also run distributed training jobs on Kubernetes clusters.
- [Python API](http://doc.paddlepaddle.org/develop/doc/api/index_en.html)
- [Python API](http://www.paddlepaddle.org/docs/develop/documentation/en/api/index_en.html)
Our new API enables much shorter programs.
- [How to Contribute](http://doc.paddlepaddle.org/develop/doc/howto/dev/contribute_to_paddle_en.html)
- [How to Contribute](http://www.paddlepaddle.org/docs/develop/documentation/en/howto/dev/contribute_to_paddle_en.html)
We appreciate your contributions!
......
RELEASE.cn.md
浏览文件 @ bfe7e242
# v0.11.0版本
## PaddlePaddle Fluid
- PaddlePaddle发布版本v0.11.0包含一个新的特性*PaddlePaddle Fluid*. Fluid 是设计用来让用户像Pytorch和Tensorflow Eager Execution一样执行程序。在这些系统中，不再有*模型*这个概念，应用也不再包含一个用于描述Operator图或者一系列层的符号描述，而是像通用程序那样描述训练或者预测的过程。而Fluid与PyTorch或Eager Execution的区别在于Fluid不依赖Python提供的控制流，例如 if-else-then或者for，而是提供了基于C++实现的控制流并暴露了对应的用with语法实现的Python接口。例如：
https://github.com/PaddlePaddle/Paddle/blob/3df78ed2a98d37f7ae6725894cc7514effd5664b/python/paddle/v2/fluid/tests/test_while_op.py#L36-L44
- 在v0.11.0版本中，我们提供了一个C++类`Executor`用于运行一个Fluid程序。Executor类似一个解释器。在未来的版本中，我们将提升和优化Executor成为一个调试器，就像GDB。并可能提供一些编译器，这个编译器会读取一个上文所描述的应用然后编译成一个等价的
源代码，这个源代码可以被nvcc编译成可以使用CUDA的二进制，或者被icc编译成可以充分利用Intel CPU的二进制。
## 新特点
* 发布 `PaddlePaddle Fluid`。
* 增加了用于模型预测的C-API。
* 用Fluid API实现了一个简单的GAN的例子。
* 增加了关于性能调优的文档。
* 为`paddle.v2.dataset`下载数据集提供了重试机制.
* C++中使用protobuf-lite替换protobuf减少了二进制的大小。
* 发布了新特性 [Elastic Deep Learning (EDL)](https://github.com/PaddlePaddle/cloud/tree/develop/doc/autoscale/experiment).
* 基于Bazel API利用cmake实现了一个的新的构建系统函数库。
* 当使用编译选项`WITH_MKL=ON`时自动下载和编译Intel® [MKLML](https://github.com/01org/mkl-dnn/releases/download/v0.11/mklml_lnx_2018.0.1.20171007.tgz) 函数库.
* [Intel® MKL-DNN on PaddlePaddle](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/design/mkldnn):
- 完成了 11个 MKL-DNN 层: Convolution, Fully connectivity, Pooling, ReLU, Tanh, ELU, Softmax, BatchNorm, AddTo, Concat, LRN。
- 完成了 3个 MKL-DNN 网络: VGG-19, ResNet-50, GoogleNet
- 基于Intel Skylake 6148 CPU的[性能测试](https://github.com/PaddlePaddle/Paddle/blob/develop/benchmark/IntelOptimizedPaddle.md) : 相对于MKLML有2~3倍的训练加速。
* 增加 [softsign activation](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/activation.html#softsign)
* 增加 [dot product layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#dot-prod)
* 增加 [L2 distance layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#l2-distance)
* 增加 [sub-nested sequence layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#sub-nested-seq)
* 增加 [kmax sequence score layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#kmax-sequence-score)
* 增加 [sequence slice layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#seq-slice)
* 增加 [row convolution layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#row-conv)
* 增加移动端友好的网页
## 改进
* 使用一个Python`whl`包即可安装.
* [V2 API可以实现用户定制化评估](https://github.com/PaddlePaddle/models/tree/develop/ltr#训练过程中输出自定义评估指标)。
* 将 `PADDLE_ONLY_CPU` 改为 `PADDLE_WITH_GPU`, 因为我们会支持多种设备。
* 删除了有一些bug的BarrierStat。
* 清理和删除了paddle::Parameter中未使用的函数。
* 删除了ProtoDataProvider。
* Huber loss同时支持回归和分类。
* 为sequence pooling 层增加`stride`参数。
* v2 API自动使用cudnn batch normalization。
* 可以使用一个固定的参数名共享BN层的参数。
* 2D convolution operation支持variable-dimension input特性。
* 重构cmake中关于CUDA的部分并实现自动检测GPU架构的功能。
* 优化网页导航。
## 错误修复
* 修复ROI pooling的Bug. cc9a761
* 修复当label是dense vector是AUC变成0的问题. #5274
* 修复WarpCTC 层的Bug.
# v0.10.0版本
我们非常高兴发布了PaddlePaddle V0.10.0版，并开发了新的[Python API](http://research.baidu.com/paddlepaddles-new-api-simplifies-deep-learning-programs/)。
......
RELEASE.md
浏览文件 @ bfe7e242
# Release v0.11.0
## PaddlePaddle Fluid
- Release 0.11.0 includes a new feature *PaddlePaddle Fluid*. Fluid is
designed to allow users to program like PyTorch and TensorFlow Eager Execution.
In these systems, there is no longer the concept *model* and applications
do not include a symbolic description of a graph of operators nor a sequence
of layers. Instead, applications look exactly like a usual program that
describes a process of training or inference. The difference between
Fluid and PyTorch or Eager Execution is that Fluid doesn't rely on Python's
control-flow, `if-then-else` nor `for`. Instead, Fluid provides its
C++ implementations and their Python binding using the `with` statement. For an example
https://github.com/PaddlePaddle/Paddle/blob/3df78ed2a98d37f7ae6725894cc7514effd5664b/python/paddle/v2/fluid/tests/test_while_op.py#L36-L44
- In 0.11.0, we provides a C++ class `Executor` to run a Fluid program.
Executor works like an interpreter. In future version, we will improve
`Executor` into a debugger like GDB, and we might provide some compilers,
which, for example, takes an application like the above one, and outputs
an equivalent C++ source program, which can be compiled using
[`nvcc`](http://docs.nvidia.com/cuda/cuda-compiler-driver-nvcc/index.html)
to generate binaries that use CUDA, or using
[`icc`](https://software.intel.com/en-us/c-compilers) to generate binaries
that make full use of Intel CPUs.
## New Features
* Release `PaddlePaddle Fluid`.
* Add C-API for model inference
* Use fluid API to create a simple GAN demo.
* Add develop guide about performance tunning.
* Add retry when download `paddle.v2.dataset`.
* Linking protobuf-lite not protobuf in C++. Reduce the binary size.
* Feature [Elastic Deep Learning (EDL)](https://github.com/PaddlePaddle/cloud/tree/develop/doc/autoscale/experiment) released.
* A new style cmake functions for Paddle. It is based on Bazel API.
* Automatically download and compile with Intel® [MKLML](https://github.com/01org/mkl-dnn/releases/download/v0.11/mklml_lnx_2018.0.1.20171007.tgz) library as CBLAS when build `WITH_MKL=ON`.
* [Intel® MKL-DNN on PaddlePaddle](https://github.com/PaddlePaddle/Paddle/tree/develop/doc/design/mkldnn):
- Complete 11 MKL-DNN layers: Convolution, Fully connectivity, Pooling, ReLU, Tanh, ELU, Softmax, BatchNorm, AddTo, Concat, LRN.
- Complete 3 MKL-DNN networks: VGG-19, ResNet-50, GoogleNet
- [Benchmark](https://github.com/PaddlePaddle/Paddle/blob/develop/benchmark/IntelOptimizedPaddle.md) on Intel Skylake 6148 CPU: 2~3x training speedup compared with MKLML.
* Add the [`softsign` activation](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/activation.html#softsign).
* Add the [dot product layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#dot-prod).
* Add the [L2 distance layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#l2-distance).
* Add the [sub-nested sequence layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#sub-nested-seq).
* Add the [kmax sequence score layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#kmax-sequence-score).
* Add the [sequence slice layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#seq-slice).
* Add the [row convolution layer](http://www.paddlepaddle.org/docs/develop/documentation/zh/api/v2/config/layer.html#row-conv)
* Add mobile friendly webpages.
## Improvements
* Build and install using a single `whl` package.
* [Custom evaluating in V2 API](https://github.com/PaddlePaddle/models/tree/develop/ltr#训练过程中输出自定义评估指标).
* Change `PADDLE_ONLY_CPU` to `PADDLE_WITH_GPU`, since we will support many kinds of devices.
* Remove buggy BarrierStat.
* Clean and remove unused functions in paddle::Parameter.
* Remove ProtoDataProvider.
* Huber loss supports both regression and classification.
* Add the `stride` parameter for sequence pooling layers.
* Enable v2 API use cudnn batch normalization automatically.
* The BN layer's parameter can be shared by a fixed the parameter name.
* Support variable-dimension input feature for 2D convolution operation.
* Refine cmake about CUDA to automatically detect GPU architecture.
* Improved website navigation.
## Bug Fixes
* Fix bug in ROI pooling. cc9a761
* Fix AUC is zero when label is dense vector. #5274
* Fix bug in WarpCTC layer.
# Release v0.10.0
We are glad to release version 0.10.0. In this version, we are happy to release the new
......
adversarial/README.md 0 → 100644
浏览文件 @ bfe7e242
# Advbox
Advbox is a Python toolbox to create adversarial examples that fool neural networks. It requires Python and paddle.
## How to use
1. train a model and save it's parameters. (like fluid_mnist.py)
2. load the parameters which is trained in step1, then reconstruct the model.(like mnist_tutorial_fgsm.py)
3. use advbox to generate the adversarial sample.
adversarial/advbox/__init__.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2017 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
A set of tools for generating adversarial example on paddle platform
"""
adversarial/advbox/attacks/base.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
"""
The base model of the model.
"""
from abc import ABCMeta, abstractmethod
class Attack(object):
"""
Abstract base class for adversarial attacks. `Attack` represent an adversarial attack
which search an adversarial example. subclass should implement the _apply() method.
Args:
model(Model): an instance of the class advbox.base.Model.
"""
__metaclass__ = ABCMeta
def __init__(self, model):
self.model = model
def __call__(self, image_label):
"""
Generate the adversarial sample.
Args:
image_label(list): The image and label tuple list with one element.
"""
adv_img = self._apply(image_label)
return adv_img
@abstractmethod
def _apply(self, image_label):
"""
Search an adversarial example.
Args:
image_batch(list): The image and label tuple list with one element.
"""
raise NotImplementedError
adversarial/advbox/attacks/gradientsign.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
"""
This module provide the attack method for FGSM's implement.
"""
from __future__ import division
import numpy as np
from collections import Iterable
from .base import Attack
class GradientSignAttack(Attack):
"""
This attack was originally implemented by Goodfellow et al. (2015) with the
infinity norm (and is known as the "Fast Gradient Sign Method"). This is therefore called
the Fast Gradient Method.
Paper link: https://arxiv.org/abs/1412.6572
"""
def _apply(self, image_label, epsilons=1000):
assert len(image_label) == 1
pre_label = np.argmax(self.model.predict(image_label))
min_, max_ = self.model.bounds()
gradient = self.model.gradient(image_label)
gradient_sign = np.sign(gradient) * (max_ - min_)
if not isinstance(epsilons, Iterable):
epsilons = np.linspace(0, 1, num=epsilons + 1)
for epsilon in epsilons:
adv_img = image_label[0][0].reshape(
gradient_sign.shape) + epsilon * gradient_sign
adv_img = np.clip(adv_img, min_, max_)
adv_label = np.argmax(self.model.predict([(adv_img, 0)]))
if pre_label != adv_label:
return adv_img
FGSM = GradientSignAttack
class IteratorGradientSignAttack(Attack):
"""
This attack was originally implemented by Alexey Kurakin(Google Brain).
Paper link: https://arxiv.org/pdf/1607.02533.pdf
"""
def _apply(self, image_label, epsilons=100, steps=10):
"""
Apply the iterative gradient sign attack.
Args:
image_label(list): The image and label tuple list of one element.
epsilons(list|tuple|int): The epsilon (input variation parameter).
steps(int): The number of iterator steps.
Return:
numpy.ndarray: The adversarail sample generated by the algorithm.
"""
assert len(image_label) == 1
pre_label = np.argmax(self.model.predict(image_label))
gradient = self.model.gradient(image_label)
min_, max_ = self.model.bounds()
if not isinstance(epsilons, Iterable):
epsilons = np.linspace(0, 1, num=epsilons + 1)
for epsilon in epsilons:
adv_img = image_label[0][0].reshape(gradient.shape)
for _ in range(steps):
gradient = self.model.gradient([(adv_img, image_label[0][1])])
gradient_sign = np.sign(gradient) * (max_ - min_)
adv_img = adv_img + epsilon * gradient_sign
adv_img = np.clip(adv_img, min_, max_)
adv_label = np.argmax(self.model.predict([(adv_img, 0)]))
if pre_label != adv_label:
return adv_img
adversarial/advbox/models/__init__.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2017 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Paddle model for target of attack
"""
adversarial/advbox/models/base.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
"""
The base model of the model.
"""
from abc import ABCMeta
import abc
abstractmethod = abc.abstractmethod
class Model(object):
"""
Base class of model to provide attack.
Args:
bounds(tuple): The lower and upper bound for the image pixel.
channel_axis(int): The index of the axis that represents the color channel.
preprocess(tuple): Two element tuple used to preprocess the input. First
substract the first element, then divide the second element.
"""
__metaclass__ = ABCMeta
def __init__(self, bounds, channel_axis, preprocess=None):
assert len(bounds) == 2
assert channel_axis in [0, 1, 2, 3]
if preprocess is None:
preprocess = (0, 1)
self._bounds = bounds
self._channel_axis = channel_axis
self._preprocess = preprocess
def bounds(self):
"""
Return the upper and lower bounds of the model.
"""
return self._bounds
def channel_axis(self):
"""
Return the channel axis of the model.
"""
return self._channel_axis
def _process_input(self, input_):
res = input_
sub, div = self._preprocess
if sub != 0:
res = input_ - sub
assert div != 0
if div != 1:
res /= div
return res
@abstractmethod
def predict(self, image_batch):
"""
Calculate the prediction of the image batch.
Args:
image_batch(numpy.ndarray): image batch of shape (batch_size, height, width, channels).
Return:
numpy.ndarray: predictions of the images with shape (batch_size, num_of_classes).
"""
raise NotImplementedError
@abstractmethod
def num_classes(self):
"""
Determine the number of the classes
Return:
int: the number of the classes
"""
raise NotImplementedError
@abstractmethod
def gradient(self, image_batch):
"""
Calculate the gradient of the cross-entropy loss w.r.t the image.
Args:
image_batch(list): The image and label tuple list.
Return:
numpy.ndarray: gradient of the cross-entropy loss w.r.t the image with
the shape (height, width, channel).
"""
raise NotImplementedError
adversarial/advbox/models/paddle.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from __future__ import absolute_import
import numpy as np
import paddle.v2 as paddle
import paddle.v2.fluid as fluid
from paddle.v2.fluid.framework import program_guard
from .base import Model
class PaddleModel(Model):
"""
Create a PaddleModel instance.
When you need to generate a adversarial sample, you should construct an instance of PaddleModel.
Args:
program(paddle.v2.fluid.framework.Program): The program of the model which generate the adversarial sample.
input_name(string): The name of the input.
logits_name(string): The name of the logits.
predict_name(string): The name of the predict.
cost_name(string): The name of the loss in the program.
"""
def __init__(self,
program,
input_name,
logits_name,
predict_name,
cost_name,
bounds,
channel_axis=3,
preprocess=None):
super(PaddleModel, self).__init__(
bounds=bounds, channel_axis=channel_axis, preprocess=preprocess)
if preprocess is None:
preprocess = (0, 1)
self._program = program
self._place = fluid.CPUPlace()
self._exe = fluid.Executor(self._place)
self._input_name = input_name
self._logits_name = logits_name
self._predict_name = predict_name
self._cost_name = cost_name
# gradient
loss = self._program.block(0).var(self._cost_name)
param_grads = fluid.backward.append_backward(
loss, parameter_list=[self._input_name])
self._gradient = dict(param_grads)[self._input_name]
def predict(self, image_batch):
"""
Predict the label of the image_batch.
Args:
image_batch(list): The image and label tuple list.
Return:
numpy.ndarray: predictions of the images with shape (batch_size, num_of_classes).
"""
feeder = fluid.DataFeeder(
feed_list=[self._input_name, self._logits_name],
place=self._place,
program=self._program)
predict_var = self._program.block(0).var(self._predict_name)
predict = self._exe.run(self._program,
feed=feeder.feed(image_batch),
fetch_list=[predict_var])
return predict
def num_classes(self):
"""
Calculate the number of classes of the output label.
Return:
int: the number of classes
"""
predict_var = self._program.block(0).var(self._predict_name)
assert len(predict_var.shape) == 2
return predict_var.shape[1]
def gradient(self, image_batch):
"""
Calculate the gradient of the loss w.r.t the input.
Args:
image_batch(list): The image and label tuple list.
Return:
list: The list of the gradient of the image.
"""
feeder = fluid.DataFeeder(
feed_list=[self._input_name, self._logits_name],
place=self._place,
program=self._program)
grad, = self._exe.run(self._program,
feed=feeder.feed(image_batch),
fetch_list=[self._gradient])
return grad
adversarial/fluid_mnist.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
"""
CNN on mnist data using fluid api of paddlepaddle
"""
import paddle.v2 as paddle
import paddle.v2.fluid as fluid
def mnist_cnn_model(img):
"""
Mnist cnn model
Args:
img(Varaible): the input image to be recognized
Returns:
Variable: the label prediction
"""
conv_pool_1 = fluid.nets.simple_img_conv_pool(
input=img,
num_filters=20,
filter_size=5,
pool_size=2,
pool_stride=2,
act='relu')
conv_pool_2 = fluid.nets.simple_img_conv_pool(
input=conv_pool_1,
num_filters=50,
filter_size=5,
pool_size=2,
pool_stride=2,
act='relu')
logits = fluid.layers.fc(input=conv_pool_2, size=10, act='softmax')
return logits
def main():
"""
Train the cnn model on mnist datasets
"""
img = fluid.layers.data(name='img', shape=[1, 28, 28], dtype='float32')
label = fluid.layers.data(name='label', shape=[1], dtype='int64')
logits = mnist_cnn_model(img)
cost = fluid.layers.cross_entropy(input=logits, label=label)
avg_cost = fluid.layers.mean(x=cost)
optimizer = fluid.optimizer.Adam(learning_rate=0.01)
optimizer.minimize(avg_cost)
accuracy = fluid.evaluator.Accuracy(input=logits, label=label)
BATCH_SIZE = 50
PASS_NUM = 3
ACC_THRESHOLD = 0.98
LOSS_THRESHOLD = 10.0
train_reader = paddle.batch(
paddle.reader.shuffle(
paddle.dataset.mnist.train(), buf_size=500),
batch_size=BATCH_SIZE)
place = fluid.CPUPlace()
exe = fluid.Executor(place)
feeder = fluid.DataFeeder(feed_list=[img, label], place=place)
exe.run(fluid.default_startup_program())
for pass_id in range(PASS_NUM):
accuracy.reset(exe)
for data in train_reader():
loss, acc = exe.run(fluid.default_main_program(),
feed=feeder.feed(data),
fetch_list=[avg_cost] + accuracy.metrics)
pass_acc = accuracy.eval(exe)
print("pass_id=" + str(pass_id) + " acc=" + str(acc) + " pass_acc="
+ str(pass_acc))
if loss < LOSS_THRESHOLD and pass_acc > ACC_THRESHOLD:
break
pass_acc = accuracy.eval(exe)
print("pass_id=" + str(pass_id) + " pass_acc=" + str(pass_acc))
fluid.io.save_params(
exe, dirname='./mnist', main_program=fluid.default_main_program())
print('train mnist done')
if __name__ == '__main__':
main()
adversarial/mnist_tutorial_fgsm.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
"""
FGSM demos on mnist using advbox tool.
"""
import paddle.v2 as paddle
import paddle.v2.fluid as fluid
import matplotlib.pyplot as plt
import numpy as np
from advbox.models.paddle import PaddleModel
from advbox.attacks.gradientsign import GradientSignAttack
def cnn_model(img):
"""
Mnist cnn model
Args:
img(Varaible): the input image to be recognized
Returns:
Variable: the label prediction
"""
#conv1 = fluid.nets.conv2d()
conv_pool_1 = fluid.nets.simple_img_conv_pool(
input=img,
num_filters=20,
filter_size=5,
pool_size=2,
pool_stride=2,
act='relu')
conv_pool_2 = fluid.nets.simple_img_conv_pool(
input=conv_pool_1,
num_filters=50,
filter_size=5,
pool_size=2,
pool_stride=2,
act='relu')
logits = fluid.layers.fc(input=conv_pool_2, size=10, act='softmax')
return logits
def main():
"""
Advbox demo which demonstrate how to use advbox.
"""
IMG_NAME = 'img'
LABEL_NAME = 'label'
img = fluid.layers.data(name=IMG_NAME, shape=[1, 28, 28], dtype='float32')
# gradient should flow
img.stop_gradient = False
label = fluid.layers.data(name=LABEL_NAME, shape=[1], dtype='int64')
logits = cnn_model(img)
cost = fluid.layers.cross_entropy(input=logits, label=label)
avg_cost = fluid.layers.mean(x=cost)
place = fluid.CPUPlace()
exe = fluid.Executor(place)
BATCH_SIZE = 1
train_reader = paddle.batch(
paddle.reader.shuffle(
paddle.dataset.mnist.train(), buf_size=500),
batch_size=BATCH_SIZE)
feeder = fluid.DataFeeder(
feed_list=[IMG_NAME, LABEL_NAME],
place=place,
program=fluid.default_main_program())
fluid.io.load_params(
exe, "./mnist/", main_program=fluid.default_main_program())
# advbox demo
m = PaddleModel(fluid.default_main_program(), IMG_NAME, LABEL_NAME,
logits.name, avg_cost.name, (-1, 1))
att = GradientSignAttack(m)
for data in train_reader():
# fgsm attack
adv_img = att(data)
plt.imshow(n[0][0], cmap='Greys_r')
plt.show()
#np.save('adv_img', adv_img)
break
if __name__ == '__main__':
main()
benchmark/IntelOptimizedPaddle.md
浏览文件 @ bfe7e242
......@@ -2,28 +2,27 @@
Machine:
- Server
- Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz, 2 Sockets, 20 Cores per socket
- Laptop
- DELL XPS15-9560-R1745: i7-7700HQ 8G 256GSSD
- i5 MacBook Pro (Retina, 13-inch, Early 2015)
- Desktop
- i7-6700k
- Server: Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz, 2 Sockets, 20 Cores per socket
- Laptop: TBD
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
- OpenBLAS v0.2.20
PaddlePaddle:
- paddlepaddle/paddle:0.11.0 (for MKLML and MKL-DNN)
- MKL-DNN tag v0.11
- MKLML 2018.0.1.20171007
- paddlepaddle/paddle:0.11.0-openblas (for OpenBLAS)
- OpenBLAS v0.2.20
On each machine, we will test and compare the performance of training on single node using MKL-DNN / MKLML / OpenBLAS respectively.
## Benchmark Model
### Server
#### Training
Test on batch size 64, 128, 256 on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
Pay attetion that the speed below includes forward, backward and parameter update time. So we can not directly compare the data with the benchmark of caffe `time` [command](https://github.com/PaddlePaddle/Paddle/blob/develop/benchmark/caffe/image/run.sh#L9), which only contain forward and backward. The updating time of parameter would become very heavy when the weight size are large, especially on alexnet.
Input image size - 3 * 224 * 224, Time: images/second
......@@ -31,18 +30,83 @@ 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 |
<img src="figs/vgg-cpu-train.png" width="500">
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 |
<img src="figs/resnet-cpu-train.png" width="500">
- 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 |
<img src="figs/googlenet-cpu-train.png" width="500">
- AlexNet
| BatchSize | 64 | 128 | 256 |
|--------------|--------| ------ | -------|
| OpenBLAS | 45.62 | 72.79 | 107.22 |
| MKLML | 66.37 | 105.60 | 144.04 |
| MKL-DNN | 399.00 | 498.94 | 626.53 |
<img src="figs/alexnet-cpu-train.png" width="500">
#### Inference
Test on batch size 1, 2, 4, 8, 16 on Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz
- VGG-19
| BatchSize | 1 | 2 | 4 | 8 | 16 |
|-----------|-------|-------|-------|-------|-------|
| OpenBLAS | 1.10 | 1.96 | 3.62 | 3.63 | 2.25 |
| MKLML | 5.58 | 9.80 | 15.15 | 21.21 | 28.67 |
| MKL-DNN | 75.07 | 88.64 | 82.58 | 92.29 | 96.75 |
<img src="figs/vgg-cpu-infer.png" width="500">
- ResNet-50
| BatchSize | 1 | 2 | 4 | 8 | 16 |
|-----------|-------|--------|--------|--------|--------|
| OpenBLAS | 3.31 | 6.72 | 11.59 | 13.17 | 9.27 |
| MKLML | 6.33 | 12.02 | 22.88 | 40.53 | 63.09 |
| MKL-DNN | 107.83| 148.84 | 177.78 | 189.35 | 217.69 |
<img src="figs/resnet-cpu-infer.png" width="500">
- GoogLeNet
| BatchSize | 1 | 2 | 4 | 8 | 16 |
|-----------|--------|--------|--------|--------|--------|
| OpenBLAS | 12.06 | 23.56 | 34.48 | 36.45 | 23.12 |
| MKLML | 22.74 | 41.56 | 81.22 | 133.47 | 210.53 |
| MKL-DNN | 175.10 | 272.92 | 450.70 | 512.00 | 600.94 |
<img src="figs/googlenet-cpu-infer.png" width="500">
- AlexNet
| BatchSize | 1 | 2 | 4 | 8 | 16 |
|-----------|--------|--------|--------|--------|--------|
| OpenBLAS | 3.53 | 6.23 | 15.04 | 26.06 | 31.62 |
| MKLML | 21.32 | 36.55 | 73.06 | 131.15 | 192.77 |
| MKL-DNN | 442.91 | 656.41 | 719.10 | 847.68 | 850.51 |
<img src="figs/alexnet-cpu-infer.png" width="500">
### Laptop
TBD
### Desktop
TBD
benchmark/cluster/README.md 0 → 100644
浏览文件 @ bfe7e242
# Cluster Training Benchmark
## Setup
- Platform
- Kubernetes: v1.6.2
- Linux Kernel: v3.10.0
- Resource
- CPU: 10 Cores per Pod
- Memory: 5GB per Pod
- Docker Image
We use different base Docker Image to run the benchmark on Kubernetes:
- PaddlePaddle v2: paddlepaddle/paddle:0.11.0
- PaddlePaddle Fluid: paddlepaddle/paddle:[commit-id]
- TensorFlow: tensorflow/tensorflow:1.5.0-rc0
- Model
vgg16 is used in this benchmark.
## Cases
- Variable
- Batch Size of training data.
- PServer count of the training job.
- The number of trainers.
- Invariant
- The resource of trainer/pserver Pod.
### Measure the Performance for Different Batch Size
- PServer Count: 40
- Trainer Count: 100
- Metrics: mini-batch / sec
| Batch Size | 32 | 64 | 128 | 256 |
| -- | -- | -- | -- | -- |
| PaddlePaddle Fluid | - | - | - | - |
| PaddlePaddle v2 | - | - | - | - |
| TensorFlow | - | - | - | - |
### Measure the Performance for Different PServer Count
- Trainer Count: 100
- Batch Size: 64
- Metrics: mini-batch / sec
| PServer Count | 10 | 20 | 40 | 60 |
| -- | -- | -- | -- | -- |
| PaddlePaddle Fluid | - | - | - | - |
| PaddlePaddle v2 | - | - | - | - |
| TensorFlow | - | - | - | - |
### Measure Parallel Efficiency By Increasing Trainer Count
- PServer Count: 20
- Batch Size: 64
- Metrics:
$S = \div(T1, TN)$
which S is the ratio of T1 over TN, training time of 1 and N trainers.
The parallel efficiency is:
$E = \div(S, N)$
| Trainer Counter | 1 | 10 | 20 | 30 | 40 | 50 | 60 | 70 | 80 | 90 | 100 |
| -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- | -- |
| PaddlePaddle Fluid | - | - | - | - | - | - | - | - | - | - | - |
| PaddlePaddle v2 | - | - | - | - | - | - | - | - | - | - | - | - |
| TensorFlow | - | - | - | - | - | - | - | - | - | - | - | - | - |
## Reproduce the benchmark
TODO
benchmark/paddle/image/alexnet.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from paddle.trainer_config_helpers import *
......@@ -6,10 +19,24 @@ height = 227
width = 227
num_class = 1000
batch_size = get_config_arg('batch_size', int, 128)
gp = get_config_arg('layer_num', int, 1)
is_infer = get_config_arg("is_infer", bool, False)
num_samples = get_config_arg('num_samples', int, 2560)
args = {'height': height, 'width': width, 'color': True, 'num_class': num_class}
args = {
'height': height,
'width': width,
'color': True,
'num_class': num_class,
'is_infer': is_infer,
'num_samples': num_samples
}
define_py_data_sources2(
"train.list", None, module="provider", obj="process", args=args)
"train.list" if not is_infer else None,
"test.list" if is_infer else None,
module="provider",
obj="process",
args=args)
settings(
batch_size=batch_size,
......@@ -31,7 +58,7 @@ net = img_pool_layer(input=net, pool_size=3, stride=2)
# conv2
net = img_conv_layer(
input=net, filter_size=5, num_filters=256, stride=1, padding=2, groups=1)
input=net, filter_size=5, num_filters=256, stride=1, padding=2, groups=gp)
net = img_cmrnorm_layer(input=net, size=5, scale=0.0001, power=0.75)
net = img_pool_layer(input=net, pool_size=3, stride=2)
......@@ -40,11 +67,11 @@ net = img_conv_layer(
input=net, filter_size=3, num_filters=384, stride=1, padding=1)
# conv4
net = img_conv_layer(
input=net, filter_size=3, num_filters=384, stride=1, padding=1, groups=1)
input=net, filter_size=3, num_filters=384, stride=1, padding=1, groups=gp)
# conv5
net = img_conv_layer(
input=net, filter_size=3, num_filters=256, stride=1, padding=1, groups=1)
input=net, filter_size=3, num_filters=256, stride=1, padding=1, groups=gp)
net = img_pool_layer(input=net, pool_size=3, stride=2)
net = fc_layer(
......@@ -59,6 +86,9 @@ net = fc_layer(
layer_attr=ExtraAttr(drop_rate=0.5))
net = fc_layer(input=net, size=1000, act=SoftmaxActivation())
lab = data_layer('label', num_class)
loss = cross_entropy(input=net, label=lab)
outputs(loss)
if is_infer:
outputs(net)
else:
lab = data_layer('label', num_class)
loss = cross_entropy(input=net, label=lab)
outputs(loss)
benchmark/paddle/image/googlenet.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from paddle.trainer_config_helpers import *
......@@ -6,10 +19,23 @@ width = 224
num_class = 1000
batch_size = get_config_arg('batch_size', int, 128)
use_gpu = get_config_arg('use_gpu', bool, True)
is_infer = get_config_arg("is_infer", bool, False)
num_samples = get_config_arg('num_samples', int, 2560)
args = {'height': height, 'width': width, 'color': True, 'num_class': num_class}
args = {
'height': height,
'width': width,
'color': True,
'num_class': num_class,
'is_infer': is_infer,
'num_samples': num_samples
}
define_py_data_sources2(
"train.list", None, module="provider", obj="process", args=args)
"train.list" if not is_infer else None,
"test.list" if is_infer else None,
module="provider",
obj="process",
args=args)
settings(
batch_size=batch_size,
......@@ -146,7 +172,6 @@ def inception(name, input, channels, \
return cat
lab = data_layer(name="label", size=1000)
data = data_layer(name="input", size=3 * height * width)
# stage 1
......@@ -224,6 +249,10 @@ pool5 = img_pool_layer(
dropout = dropout_layer(name="dropout", input=pool5, dropout_rate=0.4)
out3 = fc_layer(
name="output3", input=dropout, size=1000, act=SoftmaxActivation())
loss3 = cross_entropy(name='loss3', input=out3, label=lab)
outputs(loss3)
if is_infer:
outputs(out3)
else:
lab = data_layer(name="label", size=num_class)
loss3 = cross_entropy(name='loss3', input=out3, label=lab)
outputs(loss3)
benchmark/paddle/image/plotlog.py 0 → 100644
浏览文件 @ bfe7e242
# Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import sys
import argparse
import matplotlib.pyplot as plt
def parse_args():
parser = argparse.ArgumentParser('Parse Log')
parser.add_argument(
'--file_path', '-f', type=str, help='the path of the log file')
parser.add_argument(
'--sample_rate',
'-s',
type=float,
default=1.0,
help='the rate to take samples from log')
parser.add_argument(
'--log_period', '-p', type=int, default=1, help='the period of log')
args = parser.parse_args()
return args
def parse_file(file_name):
loss = []
error = []
with open(file_name) as f:
for i, line in enumerate(f):
line = line.strip()
if not line.startswith('pass'):
continue
line_split = line.split(' ')
if len(line_split) != 5:
continue
loss_str = line_split[2][:-1]
cur_loss = float(loss_str.split('=')[-1])
loss.append(cur_loss)
err_str = line_split[3][:-1]
cur_err = float(err_str.split('=')[-1])
error.append(cur_err)
accuracy = [1.0 - err for err in error]
return loss, accuracy
def sample(metric, sample_rate):
interval = int(1.0 / sample_rate)
if interval > len(metric):
return metric[:1]
num = len(metric) / interval
idx = [interval * i for i in range(num)]
metric_sample = [metric[id] for id in idx]
return metric_sample
def plot_metric(metric,
batch_id,
graph_title,
line_style='b-',
line_label='y',
line_num=1):
plt.figure()
plt.title(graph_title)
if line_num == 1:
plt.plot(batch_id, metric, line_style, label=line_label)
else:
for i in range(line_num):
plt.plot(batch_id, metric[i], line_style[i], label=line_label[i])
plt.xlabel('batch')
plt.ylabel(graph_title)
plt.legend()
plt.savefig(graph_title + '.jpg')
plt.close()
def main():
args = parse_args()
assert args.sample_rate > 0. and args.sample_rate <= 1.0, "The sample rate should in the range (0, 1]."
loss, accuracy = parse_file(args.file_path)
batch = [args.log_period * i for i in range(len(loss))]
batch_sample = sample(batch, args.sample_rate)
loss_sample = sample(loss, args.sample_rate)
accuracy_sample = sample(accuracy, args.sample_rate)
plot_metric(loss_sample, batch_sample, 'loss', line_label='loss')
plot_metric(
accuracy_sample,
batch_sample,
'accuracy',
line_style='g-',
line_label='accuracy')
if __name__ == '__main__':
main()
benchmark/paddle/image/provider.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
import io, os
import random
import numpy as np
......@@ -13,14 +26,21 @@ def initHook(settings, height, width, color, num_class, **kwargs):
settings.data_size = settings.height * settings.width * 3
else:
settings.data_size = settings.height * settings.width
settings.slots = [dense_vector(settings.data_size), integer_value(1)]
settings.is_infer = kwargs.get('is_infer', False)
settings.num_samples = kwargs.get('num_samples', 2560)
if settings.is_infer:
settings.slots = [dense_vector(settings.data_size)]
else:
settings.slots = [dense_vector(settings.data_size), integer_value(1)]
@provider(
init_hook=initHook, min_pool_size=-1, cache=CacheType.CACHE_PASS_IN_MEM)
def process(settings, file_list):
for i in xrange(1024):
for i in xrange(settings.num_samples):
img = np.random.rand(1, settings.data_size).reshape(-1, 1).flatten()
lab = random.randint(0, settings.num_class - 1)
yield img.astype('float32'), int(lab)
if settings.is_infer:
yield img.astype('float32')
else:
lab = random.randint(0, settings.num_class - 1)
yield img.astype('float32'), int(lab)
benchmark/paddle/image/resnet.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from paddle.trainer_config_helpers import *
......@@ -6,11 +19,23 @@ width = 224
num_class = 1000
batch_size = get_config_arg('batch_size', int, 64)
layer_num = get_config_arg("layer_num", int, 50)
is_test = get_config_arg("is_test", bool, False)
args = {'height': height, 'width': width, 'color': True, 'num_class': num_class}
is_infer = get_config_arg("is_infer", bool, False)
num_samples = get_config_arg('num_samples', int, 2560)
args = {
'height': height,
'width': width,
'color': True,
'num_class': num_class,
'is_infer': is_infer,
'num_samples': num_samples
}
define_py_data_sources2(
"train.list", None, module="provider", obj="process", args=args)
"train.list" if not is_infer else None,
"test.list" if is_infer else None,
module="provider",
obj="process",
args=args)
settings(
batch_size=batch_size,
......@@ -45,7 +70,10 @@ def conv_bn_layer(name,
act=LinearActivation(),
bias_attr=False)
return batch_norm_layer(
name=name + "_bn", input=tmp, act=active_type, use_global_stats=is_test)
name=name + "_bn",
input=tmp,
act=active_type,
use_global_stats=is_infer)
def bottleneck_block(name, input, num_filters1, num_filters2):
......@@ -207,7 +235,9 @@ elif layer_num == 152:
else:
print("Wrong layer number.")
lbl = data_layer(name="label", size=num_class)
loss = cross_entropy(name='loss', input=resnet, label=lbl)
inputs(img, lbl)
outputs(loss)
if is_infer:
outputs(resnet)
else:
lbl = data_layer(name="label", size=num_class)
loss = cross_entropy(name='loss', input=resnet, label=lbl)
outputs(loss)
benchmark/paddle/image/run_mkl_infer.sh 0 → 100755
浏览文件 @ bfe7e242
set -e
function clock_to_seconds() {
hours=`echo $1 | awk -F ':' '{print $1}'`
mins=`echo $1 | awk -F ':' '{print $2}'`
secs=`echo $1 | awk -F ':' '{print $3}'`
echo `awk 'BEGIN{printf "%.2f",('$secs' + '$mins' * 60 + '$hours' * 3600)}'`
}
function infer() {
unset OMP_NUM_THREADS MKL_NUM_THREADS OMP_DYNAMIC KMP_AFFINITY
topology=$1
layer_num=$2
bs=$3
use_mkldnn=$4
if [ $4 == "True" ]; then
thread=1
log="logs/infer-${topology}-${layer_num}-mkldnn-${bs}.log"
elif [ $4 == "False" ]; then
thread=`nproc`
if [ $thread -gt $bs ]; then
thread=$bs
fi
log="logs/infer-${topology}-${layer_num}-${thread}mklml-${bs}.log"
else
echo "Wrong input $4, use True or False."
exit 0
fi
models_in="models/${topology}-${layer_num}/pass-00000/"
if [ ! -d $models_in ]; then
echo "Training model ${topology}_${layer_num}"
paddle train --job=train \
--config="${topology}.py" \
--use_mkldnn=True \
--use_gpu=False \
--trainer_count=1 \
--num_passes=1 \
--save_dir="models/${topology}-${layer_num}" \
--config_args="batch_size=128,layer_num=${layer_num},num_samples=256" \
> /dev/null 2>&1
echo "Done"
fi
log_period=$((256 / bs))
paddle train --job=test \
--config="${topology}.py" \
--use_mkldnn=$use_mkldnn \
--use_gpu=False \
--trainer_count=$thread \
--log_period=$log_period \
--config_args="batch_size=${bs},layer_num=${layer_num},is_infer=True" \
--init_model_path=$models_in \
2>&1 | tee ${log}
# calculate the last 5 logs period time of 1280 samples,
# the time before are burning time.
start=`tail ${log} -n 7 | head -n 1 | awk -F ' ' '{print $2}' | xargs`
end=`tail ${log} -n 2 | head -n 1 | awk -F ' ' '{print $2}' | xargs`
start_sec=`clock_to_seconds $start`
end_sec=`clock_to_seconds $end`
fps=`awk 'BEGIN{printf "%.2f",(1280 / ('$end_sec' - '$start_sec'))}'`
echo "Last 1280 samples start: ${start}(${start_sec} sec), end: ${end}(${end_sec} sec;" >> ${log}
echo "FPS: $fps images/sec" 2>&1 | tee -a ${log}
}
if [ ! -f "train.list" ]; then
echo " " > train.list
fi
if [ ! -f "test.list" ]; then
echo " " > test.list
fi
if [ ! -d "logs" ]; then
mkdir logs
fi
if [ ! -d "models" ]; then
mkdir -p models
fi
# inference benchmark
for use_mkldnn in True False; do
for batchsize in 1 2 4 8 16; do
infer vgg 19 $batchsize $use_mkldnn
infer resnet 50 $batchsize $use_mkldnn
infer googlenet v1 $batchsize $use_mkldnn
infer alexnet 2 $batchsize $use_mkldnn
done
done
benchmark/paddle/image/run_mkldnn.sh → benchmark/paddle/image/run_mkl_train.sh
浏览文件 @ bfe7e242
......@@ -8,13 +8,13 @@ function train() {
use_mkldnn=$4
if [ $4 == "True" ]; then
thread=1
log="logs/${topology}-${layer_num}-mkldnn-${bs}.log"
log="logs/train-${topology}-${layer_num}-mkldnn-${bs}.log"
elif [ $4 == "False" ]; then
thread=`nproc`
# each trainer_count use only 1 core to avoid conflict
log="logs/${topology}-${layer_num}-${thread}mklml-${bs}.log"
log="logs/train-${topology}-${layer_num}-${thread}mklml-${bs}.log"
else
echo "Wrong input $3, use True or False."
echo "Wrong input $4, use True or False."
exit 0
fi
args="batch_size=${bs},layer_num=${layer_num}"
......@@ -28,19 +28,25 @@ function train() {
--test_period=100 \
--config_args=$args \
2>&1 | tee ${log}
avg_time=`tail ${log} -n 1 | awk -F ' ' '{print $8}' | sed 's/avg=//'`
fps=`awk 'BEGIN{printf "%.2f",('$bs' / '$avg_time' * 1000)}'`
echo "FPS: $fps images/sec" 2>&1 | tee -a ${log}
}
if [ ! -d "train.list" ]; then
if [ ! -f "train.list" ]; then
echo " " > train.list
fi
if [ ! -d "logs" ]; then
mkdir logs
fi
# training benchmark
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 googlenet v1 $batchsize $use_mkldnn
train alexnet 2 $batchsize $use_mkldnn
done
done
benchmark/paddle/image/run_openblas_infer.sh 0 → 100755
浏览文件 @ bfe7e242
set -e
function clock_to_seconds() {
hours=`echo $1 | awk -F ':' '{print $1}'`
mins=`echo $1 | awk -F ':' '{print $2}'`
secs=`echo $1 | awk -F ':' '{print $3}'`
echo `awk 'BEGIN{printf "%.2f",('$secs' + '$mins' * 60 + '$hours' * 3600)}'`
}
function infer() {
export OPENBLAS_MAIN_FREE=1
topology=$1
layer_num=$2
bs=$3
trainers=`nproc`
if [ $trainers -gt $bs ]; then
trainers=$bs
fi
log="logs/infer-${topology}-${layer_num}-${trainers}openblas-${bs}.log"
threads=$((`nproc` / trainers))
if [ $threads -eq 0 ]; then
threads=1
fi
export OPENBLAS_NUM_THREADS=$threads
models_in="models/${topology}-${layer_num}/pass-00000/"
if [ ! -d $models_in ]; then
echo "./run_mkl_infer.sh to save the model first"
exit 0
fi
log_period=$((32 / bs))
paddle train --job=test \
--config="${topology}.py" \
--use_mkldnn=False \
--use_gpu=False \
--trainer_count=$trainers \
--log_period=$log_period \
--config_args="batch_size=${bs},layer_num=${layer_num},is_infer=True,num_samples=256" \
--init_model_path=$models_in \
2>&1 | tee ${log}
# calculate the last 5 logs period time of 160(=32*5) samples,
# the time before are burning time.
start=`tail ${log} -n 7 | head -n 1 | awk -F ' ' '{print $2}' | xargs`
end=`tail ${log} -n 2 | head -n 1 | awk -F ' ' '{print $2}' | xargs`
start_sec=`clock_to_seconds $start`
end_sec=`clock_to_seconds $end`
fps=`awk 'BEGIN{printf "%.2f",(160 / ('$end_sec' - '$start_sec'))}'`
echo "Last 160 samples start: ${start}(${start_sec} sec), end: ${end}(${end_sec} sec;" >> ${log}
echo "FPS: $fps images/sec" 2>&1 | tee -a ${log}
}
if [ ! -f "train.list" ]; then
echo " " > train.list
fi
if [ ! -f "test.list" ]; then
echo " " > test.list
fi
if [ ! -d "logs" ]; then
mkdir logs
fi
# inference benchmark
for batchsize in 1 2 4 8 16; do
infer vgg 19 $batchsize
infer resnet 50 $batchsize
infer googlenet v1 $batchsize
infer alexnet 2 $batchsize
done
benchmark/paddle/image/run_openblas_train.sh 0 → 100755
浏览文件 @ bfe7e242
set -e
function train() {
export OPENBLAS_NUM_THREADS=1
topology=$1
layer_num=$2
bs=$3
thread=`nproc`
# each trainer_count use only 1 core to avoid conflict
log="logs/train-${topology}-${layer_num}-${thread}openblas-${bs}.log"
args="batch_size=${bs},layer_num=${layer_num}"
config="${topology}.py"
paddle train --job=time \
--config=$config \
--use_mkldnn=False \
--use_gpu=False \
--trainer_count=$thread \
--log_period=3 \
--test_period=30 \
--config_args=$args \
2>&1 | tee ${log}
avg_time=`tail ${log} -n 1 | awk -F ' ' '{print $8}' | sed 's/avg=//'`
fps=`awk 'BEGIN{printf "%.2f",('$bs' / '$avg_time' * 1000)}'`
echo "FPS: $fps images/sec" 2>&1 | tee -a ${log}
}
if [ ! -f "train.list" ]; then
echo " " > train.list
fi
if [ ! -d "logs" ]; then
mkdir logs
fi
# training benchmark
for batchsize in 64 128 256; do
train vgg 19 $batchsize
train resnet 50 $batchsize
train googlenet v1 $batchsize
train alexnet 2 $batchsize
done
benchmark/paddle/image/smallnet_mnist_cifar.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from paddle.trainer_config_helpers import *
......
benchmark/paddle/image/vgg.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from paddle.trainer_config_helpers import *
......@@ -6,10 +19,23 @@ width = 224
num_class = 1000
batch_size = get_config_arg('batch_size', int, 64)
layer_num = get_config_arg('layer_num', int, 19)
is_infer = get_config_arg("is_infer", bool, False)
num_samples = get_config_arg('num_samples', int, 2560)
args = {'height': height, 'width': width, 'color': True, 'num_class': num_class}
args = {
'height': height,
'width': width,
'color': True,
'num_class': num_class,
'is_infer': is_infer,
'num_samples': num_samples
}
define_py_data_sources2(
"train.list", None, module="provider", obj="process", args=args)
"train.list" if not is_infer else None,
"test.list" if is_infer else None,
module="provider",
obj="process",
args=args)
settings(
batch_size=batch_size,
......@@ -98,6 +124,9 @@ elif layer_num == 19:
else:
print("Wrong layer number.")
lab = data_layer('label', num_class)
loss = cross_entropy(input=vgg, label=lab)
outputs(loss)
if is_infer:
outputs(vgg)
else:
lab = data_layer('label', num_class)
loss = cross_entropy(input=vgg, label=lab)
outputs(loss)
benchmark/paddle/rnn/provider.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
import io, os
import random
import numpy as np
......
benchmark/paddle/rnn/rnn.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from paddle.trainer_config_helpers import *
......
benchmark/tensorflow/image/alexnet.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from six.moves import xrange # pylint: disable=redefined-builtin
from datetime import datetime
import math
......
benchmark/tensorflow/image/alexnet_multi_gpu.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from six.moves import xrange # pylint: disable=redefined-builtin
from datetime import datetime
import math
......
benchmark/tensorflow/image/googlenet.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from six.moves import xrange
from datetime import datetime
import math
......
benchmark/tensorflow/image/googlenet_multi_gpu.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from six.moves import xrange # pylint: disable=redefined-builtin
from datetime import datetime
import math
......
benchmark/tensorflow/image/smallnet_mnist_cifar.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from six.moves import xrange # pylint: disable=redefined-builtin
from datetime import datetime
import math
......
benchmark/tensorflow/rnn/reader.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
import os.path
import io
import numpy as np
......
benchmark/tensorflow/rnn/rnn.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from six.moves import xrange # pylint: disable=redefined-builtin
import math
......
benchmark/tensorflow/rnn/rnn_multi_gpu.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
#!/usr/bin/env python
from six.moves import xrange # pylint: disable=redefined-builtin
import re
......
cmake/cblas.cmake
浏览文件 @ bfe7e242
......@@ -3,7 +3,7 @@
# It will search MKLML, atlas, OpenBlas, reference-cblas in order.
#
# If any cblas implementation found, the following variable will be set.
# CBLAS_PROVIDER # one of MKLML, ATLAS, OPENBLAS, REFERENCE
# CBLAS_PROVIDER # one of MKLML, OPENBLAS, REFERENCE
# CBLAS_INC_DIR # the include directory for cblas.
# CBLAS_LIBS # a list of libraries should be linked by paddle.
# # Each library should be full path to object file.
......@@ -17,7 +17,7 @@ if(WITH_MKLML AND MKLML_INC_DIR AND MKLML_LIB)
set(CBLAS_INC_DIR ${MKLML_INC_DIR})
set(CBLAS_LIBRARIES ${MKLML_LIB})
add_definitions(-DPADDLE_USE_MKLML)
add_definitions(-DPADDLE_WITH_MKLML)
add_definitions(-DLAPACK_FOUND)
message(STATUS "Found cblas and lapack in MKLML "
......@@ -25,42 +25,6 @@ if(WITH_MKLML AND MKLML_INC_DIR AND MKLML_LIB)
return()
endif()
## Then find atlas.
set(ATLAS_ROOT $ENV{ATLAS_ROOT} CACHE PATH "Folder contains Atlas")
set(ATLAS_INCLUDE_SEARCH_PATHS
${ATLAS_ROOT}/include
/usr/include
/usr/include/atlas)
set(ATLAS_LIB_SEARCH_PATHS
${ATLAS_ROOT}/lib
/usr/lib
/usr/lib/blas/atlas
/usr/lib/atlas
/usr/lib/atlas-base # special for ubuntu 14.04.
)
find_path(ATLAS_INC_DIR NAMES cblas.h
PATHS ${ATLAS_INCLUDE_SEARCH_PATHS})
find_path(ATLAS_CLAPACK_INC_DIR NAMES clapack.h
PATHS ${ATLAS_INCLUDE_SEARCH_PATHS})
find_library(ATLAS_CBLAS_LIB NAMES cblas libcblas.so.3
PATHS ${ATLAS_LIB_SEARCH_PATHS})
find_library(ATLAS_CLAPACK_LIB NAMES lapack_atlas liblapack_atlas.so.3
PATHS ${ATLAS_LIB_SEARCH_PATHS})
if(ATLAS_CLAPACK_INC_DIR AND ATLAS_INC_DIR AND ATLAS_CBLAS_LIB AND ATLAS_CLAPACK_LIB)
set(CBLAS_FOUND ON)
set(CBLAS_PROVIDER ATLAS)
set(CBLAS_INC_DIR ${ATLAS_INC_DIR} ${ATLAS_CLAPACK_INC_DIR})
set(CBLAS_LIBRARIES ${ATLAS_CLAPACK_LIB} ${ATLAS_CBLAS_LIB})
add_definitions(-DPADDLE_USE_ATLAS)
add_definitions(-DLAPACK_FOUND)
message(STATUS "Found ATLAS (include: ${ATLAS_INC_DIR}, library: ${CBLAS_LIBRARIES})")
message(STATUS "Found lapack in ATLAS (include: ${ATLAS_CLAPACK_INC_DIR})")
return()
endif()
## Then find openblas.
set(OPENBLAS_ROOT $ENV{OPENBLAS_ROOT} CACHE PATH "Folder contains Openblas")
set(OPENBLAS_INCLUDE_SEARCH_PATHS
......
cmake/configure.cmake
浏览文件 @ bfe7e242
......@@ -24,6 +24,11 @@ if(WITH_DOUBLE)
add_definitions(-DPADDLE_TYPE_DOUBLE)
endif(WITH_DOUBLE)
if(WITH_ARM_FP16)
add_definitions(-DPADDLE_ARM_FP16)
add_definitions("-march=armv8.2-a+fp16+simd")
endif(WITH_ARM_FP16)
if(WITH_TESTING)
add_definitions(-DPADDLE_WITH_TESTING)
endif(WITH_TESTING)
......
cmake/external/cares.cmake 0 → 100644
浏览文件 @ bfe7e242
# 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 OR NOT WITH_DISTRIBUTE)
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 -j8
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/eigen.cmake
浏览文件 @ bfe7e242
......@@ -19,7 +19,7 @@ ExternalProject_Add(
if (${CMAKE_VERSION} VERSION_LESS "3.3.0")
set(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/eigen3_dummy.c)
file(WRITE ${dummyfile} "const char * dummy_eigen3 = \"${dummyfile}\";")
file(WRITE ${dummyfile} "const char *dummy_eigen3 = \"${dummyfile}\";")
add_library(eigen3 STATIC ${dummyfile})
else()
add_library(eigen3 INTERFACE)
......
cmake/external/gflags.cmake
浏览文件 @ bfe7e242
......@@ -28,15 +28,8 @@ INCLUDE_DIRECTORIES(${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/glog.cmake
浏览文件 @ bfe7e242
......@@ -26,12 +26,21 @@ ENDIF(WIN32)
INCLUDE_DIRECTORIES(${GLOG_INCLUDE_DIR})
IF(ANDROID AND ${CMAKE_SYSTEM_VERSION} VERSION_LESS "21")
# Using the unofficial glog for Android API < 21
SET(GLOG_REPOSITORY "https://github.com/Xreki/glog.git")
SET(GLOG_TAG "8a547150548b284382ccb6582408e9140ff2bea8")
ELSE()
SET(GLOG_REPOSITORY "https://github.com/google/glog.git")
SET(GLOG_TAG "v0.3.5")
ENDIF()
ExternalProject_Add(
extern_glog
${EXTERNAL_PROJECT_LOG_ARGS}
DEPENDS gflags
GIT_REPOSITORY "https://github.com/google/glog.git"
GIT_TAG v0.3.5
GIT_REPOSITORY ${GLOG_REPOSITORY}
GIT_TAG ${GLOG_TAG}
PREFIX ${GLOG_SOURCES_DIR}
UPDATE_COMMAND ""
CMAKE_ARGS -DCMAKE_CXX_COMPILER=${CMAKE_CXX_COMPILER}
......
cmake/external/grpc.cmake 0 → 100644
浏览文件 @ bfe7e242
# 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 OR NOT WITH_DISTRIBUTE)
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 HAS_SYSTEM_PROTOBUF=false -s -j8 static grpc_cpp_plugin | sed "s/-Werror//g" | sh)
ELSE()
SET(BUILD_CMD make HAS_SYSTEM_PROTOBUF=false -s -j8 static grpc_cpp_plugin)
ENDIF()
ExternalProject_Add(
extern_grpc
DEPENDS protobuf zlib
GIT_REPOSITORY "https://github.com/grpc/grpc.git"
GIT_TAG "v1.8.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 ${BUILD_CMD}
INSTALL_COMMAND make prefix=${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(${GRPC_INCLUDE_DIR})
ADD_DEPENDENCIES(grpc++_unsecure extern_grpc)
cmake/external/mkldnn.cmake
浏览文件 @ bfe7e242
......@@ -63,9 +63,30 @@ ExternalProject_Add(
-DMKLROOT:PATH=${MKLML_ROOT}
)
ADD_LIBRARY(mkldnn SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET mkldnn PROPERTY IMPORTED_LOCATION ${MKLDNN_LIB})
ADD_DEPENDENCIES(mkldnn ${MKLDNN_PROJECT})
ADD_LIBRARY(shared_mkldnn SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET shared_mkldnn PROPERTY IMPORTED_LOCATION ${MKLDNN_LIB})
ADD_DEPENDENCIES(shared_mkldnn ${MKLDNN_PROJECT})
MESSAGE(STATUS "MKLDNN library: ${MKLDNN_LIB}")
add_definitions(-DPADDLE_USE_MKLDNN)
LIST(APPEND external_project_dependencies mkldnn)
add_definitions(-DPADDLE_WITH_MKLDNN)
LIST(APPEND external_project_dependencies shared_mkldnn)
# generate a static dummy target to track mkldnn dependencies
# for cc_library(xxx SRCS xxx.c DEPS mkldnn)
SET(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/mkldnn_dummy.c)
FILE(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
ADD_LIBRARY(mkldnn STATIC ${dummyfile})
TARGET_LINK_LIBRARIES(mkldnn ${MKLDNN_LIB} ${MKLML_LIB} ${MKLML_IOMP_LIB})
ADD_DEPENDENCIES(mkldnn ${MKLDNN_PROJECT})
# copy the real so.0 lib to install dir
# it can be directly contained in wheel or capi
SET(MKLDNN_SHARED_LIB ${MKLDNN_INSTALL_DIR}/libmkldnn.so.0)
ADD_CUSTOM_COMMAND(OUTPUT ${MKLDNN_SHARED_LIB}
COMMAND cp ${MKLDNN_LIB} ${MKLDNN_SHARED_LIB}
DEPENDS mkldnn)
ADD_CUSTOM_TARGET(mkldnn_shared_lib ALL DEPENDS ${MKLDNN_SHARED_LIB})
IF(WITH_C_API)
INSTALL(FILES ${MKLDNN_SHARED_LIB} DESTINATION lib)
ENDIF()
cmake/external/mklml.cmake
浏览文件 @ bfe7e242
......@@ -66,3 +66,7 @@ ADD_LIBRARY(mklml SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET mklml PROPERTY IMPORTED_LOCATION ${MKLML_LIB})
ADD_DEPENDENCIES(mklml ${MKLML_PROJECT})
LIST(APPEND external_project_dependencies mklml)
IF(WITH_C_API)
INSTALL(FILES ${MKLML_LIB} ${MKLML_IOMP_LIB} DESTINATION lib)
ENDIF()
cmake/external/openblas.cmake
浏览文件 @ bfe7e242
......@@ -30,23 +30,21 @@ IF(NOT ${CBLAS_FOUND})
CACHE FILEPATH "openblas library." FORCE)
SET(OPENBLAS_CC "${CMAKE_C_COMPILER} -Wno-unused-but-set-variable -Wno-unused-variable")
SET(OPENBLAS_COMMIT "v0.2.20")
IF(CMAKE_CROSSCOMPILING)
SET(OPTIONAL_ARGS HOSTCC=${HOST_C_COMPILER})
GET_FILENAME_COMPONENT(CROSS_SUFFIX ${CMAKE_C_COMPILER} DIRECTORY)
SET(CROSS_SUFFIX ${CROSS_SUFFIX}/)
IF(ANDROID)
# arm_soft_fp_abi branch of OpenBLAS to support softfp
# https://github.com/xianyi/OpenBLAS/tree/arm_soft_fp_abi
SET(OPENBLAS_COMMIT "b5c96fcfcdc82945502a2303116a64d89985daf5")
IF(ANDROID_ABI MATCHES "^armeabi(-v7a)?$")
# use softfp
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV7 ARM_SOFTFP_ABI=1 USE_THREAD=0)
ELSEIF(ANDROID_ABI STREQUAL "arm64-v8a")
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV8 BINARY=64 USE_THREAD=0)
ENDIF()
ELSEIF(IOS)
IF(CMAKE_OSX_ARCHITECTURES MATCHES "arm64")
SET(OPENBLAS_COMMIT "b5c96fcfcdc82945502a2303116a64d89985daf5")
SET(OPENBLAS_CC "${OPENBLAS_CC} ${CMAKE_C_FLAGS} -isysroot ${CMAKE_OSX_SYSROOT}")
SET(OPENBLAS_CC "${OPENBLAS_CC} -arch arm64")
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV8 BINARY=64 USE_THREAD=0 CROSS_SUFFIX=${CROSS_SUFFIX})
......@@ -56,14 +54,12 @@ IF(NOT ${CBLAS_FOUND})
ENDIF()
ELSEIF(RPI)
# use hardfp
SET(OPENBLAS_COMMIT "v0.2.20")
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} TARGET=ARMV7 USE_THREAD=0)
ENDIF()
ELSE()
IF(APPLE)
SET(OPENBLAS_CC "${CMAKE_C_COMPILER} -isysroot ${CMAKE_OSX_SYSROOT}")
ENDIF()
SET(OPENBLAS_COMMIT "v0.2.20")
SET(OPTIONAL_ARGS "")
IF(CMAKE_SYSTEM_PROCESSOR MATCHES "^x86(_64)?$")
SET(OPTIONAL_ARGS DYNAMIC_ARCH=1 NUM_THREADS=64)
......@@ -113,12 +109,8 @@ INCLUDE_DIRECTORIES(${CBLAS_INC_DIR})
# FIXME(gangliao): generate cblas target to track all high performance
# linear algebra libraries for cc_library(xxx SRCS xxx.c DEPS cblas)
SET(dummyfile ${CMAKE_CURRENT_BINARY_DIR}/cblas_dummy.c)
FILE(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
IF("${CBLAS_PROVIDER}" STREQUAL "MKLML")
ADD_LIBRARY(cblas SHARED ${dummyfile})
ELSE()
ADD_LIBRARY(cblas STATIC ${dummyfile})
ENDIF()
FILE(WRITE ${dummyfile} "const char *dummy_cblas = \"${dummyfile}\";")
ADD_LIBRARY(cblas STATIC ${dummyfile})
TARGET_LINK_LIBRARIES(cblas ${CBLAS_LIBRARIES})
IF(NOT ${CBLAS_FOUND})
......
cmake/external/protobuf.cmake
浏览文件 @ bfe7e242
......@@ -15,7 +15,18 @@
INCLUDE(ExternalProject)
# Always invoke `FIND_PACKAGE(Protobuf)` for importing function protobuf_generate_cpp
FIND_PACKAGE(Protobuf QUIET)
SET(PROTOBUF_FOUND "OFF")
macro(UNSET_VAR VAR_NAME)
UNSET(${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 ${PROTOBUF_ROOT}.")
SET_PROTOBUF_VERSION()
......@@ -178,14 +188,26 @@ FUNCTION(build_protobuf TARGET_NAME BUILD_FOR_HOST)
SET(OPTIONAL_CACHE_ARGS "-DZLIB_ROOT:STRING=${ZLIB_ROOT}")
ENDIF()
SET(PROTOBUF_REPO "https://github.com/google/protobuf.git")
SET(PROTOBUF_TAG "9f75c5aa851cd877fb0d93ccc31b8567a6706546")
IF(MOBILE_INFERENCE)
# The reason why the official version is not used is described in
# https://github.com/PaddlePaddle/Paddle/issues/6114
SET(PROTOBUF_REPO "https://github.com/qingqing01/protobuf.git")
SET(PROTOBUF_TAG "v3.2.0")
IF(NOT BUILD_FOR_HOST)
SET(OPTIONAL_ARGS ${OPTIONAL_ARGS} "-Dprotobuf_BUILD_PROTOC_BINARIES=OFF")
ENDIF()
ENDIF()
ExternalProject_Add(
${TARGET_NAME}
${EXTERNAL_PROJECT_LOG_ARGS}
PREFIX ${PROTOBUF_SOURCES_DIR}
UPDATE_COMMAND ""
DEPENDS zlib
GIT_REPOSITORY "https://github.com/google/protobuf.git"
GIT_TAG "9f75c5aa851cd877fb0d93ccc31b8567a6706546"
GIT_REPOSITORY ${PROTOBUF_REPO}
GIT_TAG ${PROTOBUF_TAG}
CONFIGURE_COMMAND
${CMAKE_COMMAND} ${PROTOBUF_SOURCES_DIR}/src/${TARGET_NAME}/cmake
${OPTIONAL_ARGS}
......@@ -203,7 +225,11 @@ FUNCTION(build_protobuf TARGET_NAME BUILD_FOR_HOST)
)
ENDFUNCTION()
SET(PROTOBUF_VERSION 3.1)
IF(NOT MOBILE_INFERENCE)
SET(PROTOBUF_VERSION 3.1)
ELSE()
SET(PROTOBUF_VERSION 3.2)
ENDIF()
IF(CMAKE_CROSSCOMPILING)
build_protobuf(protobuf_host TRUE)
LIST(APPEND external_project_dependencies protobuf_host)
......@@ -227,9 +253,9 @@ IF(NOT PROTOBUF_FOUND)
IF(WITH_C_API)
INSTALL(DIRECTORY ${PROTOBUF_INCLUDE_DIR} DESTINATION third_party/protobuf)
IF(ANDROID)
INSTALL(FILES ${PROTOBUF_LIBRARY} DESTINATION third_party/protobuf/lib/${ANDROID_ABI})
INSTALL(FILES ${PROTOBUF_LITE_LIBRARY} DESTINATION third_party/protobuf/lib/${ANDROID_ABI})
ELSE()
INSTALL(FILES ${PROTOBUF_LIBRARY} DESTINATION third_party/protobuf/lib)
INSTALL(FILES ${PROTOBUF_LITE_LIBRARY} DESTINATION third_party/protobuf/lib)
ENDIF()
ENDIF()
......
cmake/external/warpctc.cmake
浏览文件 @ bfe7e242
......@@ -63,7 +63,7 @@ ExternalProject_Add(
MESSAGE(STATUS "warp-ctc library: ${WARPCTC_LIBRARIES}")
INCLUDE_DIRECTORIES(${WARPCTC_INCLUDE_DIR})
ADD_LIBRARY(warpctc STATIC IMPORTED GLOBAL)
ADD_LIBRARY(warpctc SHARED IMPORTED GLOBAL)
SET_PROPERTY(TARGET warpctc PROPERTY IMPORTED_LOCATION ${WARPCTC_LIBRARIES})
ADD_DEPENDENCIES(warpctc extern_warpctc)
......
cmake/external/zlib.cmake
浏览文件 @ bfe7e242
......@@ -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 @ bfe7e242
......@@ -111,6 +111,8 @@ set(COMMON_FLAGS
-Wno-error=sign-compare
-Wno-error=unused-local-typedefs
-Wno-error=parentheses-equality # Warnings in pybind11
-Wno-error=ignored-attributes # Warnings in Eigen, gcc 6.3
-Wno-error=terminate # Warning in PADDLE_ENFORCE
)
set(GPU_COMMON_FLAGS
......
cmake/generic.cmake
浏览文件 @ bfe7e242
......@@ -120,7 +120,7 @@ function(merge_static_libs TARGET_NAME)
DEPENDS ${libs})
# Generate dummy staic lib
file(WRITE ${target_SRCS} "const char *dummy = \"${target_SRCS}\";")
file(WRITE ${target_SRCS} "const char *dummy_${TARGET_NAME} = \"${target_SRCS}\";")
add_library(${TARGET_NAME} STATIC ${target_SRCS})
target_link_libraries(${TARGET_NAME} ${libs_deps})
......@@ -160,7 +160,7 @@ function(merge_static_libs TARGET_NAME)
DEPENDS ${libs} ${target_OBJS})
# Generate dummy staic lib
file(WRITE ${target_SRCS} "const char *dummy = \"${target_SRCS}\";")
file(WRITE ${target_SRCS} "const char *dummy_${TARGET_NAME} = \"${target_SRCS}\";")
add_library(${TARGET_NAME} STATIC ${target_SRCS})
target_link_libraries(${TARGET_NAME} ${libs_deps})
......@@ -227,8 +227,8 @@ function(cc_test TARGET_NAME)
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(cc_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
add_executable(${TARGET_NAME} ${cc_test_SRCS})
target_link_libraries(${TARGET_NAME} ${cc_test_DEPS} gtest gtest_main)
add_dependencies(${TARGET_NAME} ${cc_test_DEPS} gtest gtest_main)
target_link_libraries(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
add_dependencies(${TARGET_NAME} ${cc_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
add_test(NAME ${TARGET_NAME} COMMAND ${TARGET_NAME} WORKING_DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR})
endif()
endfunction(cc_test)
......@@ -288,8 +288,8 @@ function(nv_test TARGET_NAME)
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(nv_test "${options}" "${oneValueArgs}" "${multiValueArgs}" ${ARGN})
cuda_add_executable(${TARGET_NAME} ${nv_test_SRCS})
target_link_libraries(${TARGET_NAME} ${nv_test_DEPS} gtest gtest_main)
add_dependencies(${TARGET_NAME} ${nv_test_DEPS} gtest gtest_main)
target_link_libraries(${TARGET_NAME} ${nv_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
add_dependencies(${TARGET_NAME} ${nv_test_DEPS} paddle_gtest_main paddle_memory gtest gflags)
add_test(${TARGET_NAME} ${TARGET_NAME})
endif()
endfunction(nv_test)
......@@ -324,7 +324,7 @@ function(go_library TARGET_NAME)
)
# Add dummy code to support `make target_name` under Terminal Command
file(WRITE ${dummyfile} "const char * dummy = \"${dummyfile}\";")
file(WRITE ${dummyfile} "const char *dummy_${TARGET_NAME} = \"${dummyfile}\";")
if (go_library_SHARED OR go_library_shared)
add_library(${TARGET_NAME} SHARED ${dummyfile})
else()
......@@ -459,11 +459,58 @@ function(py_test TARGET_NAME)
if(WITH_TESTING)
set(options STATIC static SHARED shared)
set(oneValueArgs "")
set(multiValueArgs SRCS DEPS)
cmake_parse_arguments(py_test "${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(
${grpc_grpc_srcs}
PROPERTIES
COMPILE_FLAGS "-Wno-non-virtual-dtor -Wno-error=non-virtual-dtor -Wno-error=delete-non-virtual-dtor")
cc_library("${TARGET_NAME}_grpc" SRCS "${grpc_grpc_srcs}")
set_source_files_properties(
${grpc_library_SRCS}
PROPERTIES
COMPILE_FLAGS "-Wno-non-virtual-dtor -Wno-error=non-virtual-dtor -Wno-error=delete-non-virtual-dtor")
cc_library("${TARGET_NAME}" SRCS "${grpc_library_SRCS}" DEPS "${TARGET_NAME}_grpc" "${TARGET_NAME}_proto" "${grpc_library_DEPS}")
endfunction()
cmake/make_resource.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
import os
import re
import sys
......
cmake/util.cmake
浏览文件 @ bfe7e242
......@@ -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/index_cn.rst
浏览文件 @ bfe7e242
......@@ -7,3 +7,4 @@ API
模型配置 <v2/model_configs.rst>
数据访问 <v2/data.rst>
训练与应用 <v2/run_logic.rst>
v2/fluid.rst
doc/api/index_en.rst
浏览文件 @ bfe7e242
......@@ -7,3 +7,4 @@ API
v2/model_configs.rst
v2/data.rst
v2/run_logic.rst
v2/fluid.rst
doc/api/v1/data_provider/src/mnist_config.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from paddle.trainer_config_helpers import *
define_py_data_sources2(
......
doc/api/v1/data_provider/src/mnist_provider.dict.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from paddle.trainer.PyDataProvider2 import *
......
doc/api/v1/data_provider/src/sentimental_config.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from paddle.trainer_config_helpers import *
dictionary = dict()
......
doc/api/v1/data_provider/src/sentimental_provider.py
浏览文件 @ bfe7e242
# Copyright (c) 2018 PaddlePaddle Authors. All Rights Reserve.
#
#Licensed under the Apache License, Version 2.0 (the "License");
#you may not use this file except in compliance with the License.
#You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
#Unless required by applicable law or agreed to in writing, software
#distributed under the License is distributed on an "AS IS" BASIS,
#WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
#See the License for the specific language governing permissions and
#limitations under the License.
from paddle.trainer.PyDataProvider2 import *
......
doc/api/v2/config/activation.rst
浏览文件 @ bfe7e242
......@@ -99,3 +99,10 @@ STanh
.. automodule:: paddle.v2.activation
:members: STanh
:noindex:
SoftSign
========
.. automodule:: paddle.v2.activation
:members: SoftSign
:noindex:
doc/api/v2/config/layer.rst
浏览文件 @ bfe7e242
......@@ -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
================
......@@ -252,6 +252,11 @@ first_seq
.. autoclass:: paddle.v2.layer.first_seq
:noindex:
sub_seq
---------
.. autoclass:: paddle.v2.layer.sub_seq
:noindex:
concat
------
.. autoclass:: paddle.v2.layer.concat
......@@ -415,6 +420,13 @@ multiplex
.. autoclass:: paddle.v2.layer.multiplex
:noindex:
Factorization Machine Layer
============================
factorization_machine
---------------------
.. autoclass:: paddle.v2.layer.factorization_machine
:noindex:
Slicing and Joining Layers
==========================
......@@ -460,7 +472,7 @@ lambda_cost
:noindex:
square_error_cost
--------
-----------------
.. autoclass:: paddle.v2.layer.square_error_cost
:noindex:
......@@ -526,7 +538,7 @@ Miscs
=====
dropout
--------------
--------
.. autoclass:: paddle.v2.layer.dropout
:noindex:
......
doc/api/v2/fluid.rst 0 → 100644
浏览文件 @ bfe7e242
======================
Fluid
======================
.. toctree::
:maxdepth: 1
fluid/layers.rst
fluid/data_feeder.rst
fluid/executor.rst
fluid/initializer.rst
fluid/evaluator.rst
fluid/nets.rst
fluid/optimizer.rst
fluid/param_attr.rst
fluid/profiler.rst
fluid/regularizer.rst
fluid/io.rst
doc/api/v2/fluid/data_feeder.rst 0 → 100644
浏览文件 @ bfe7e242
===========
DataFeeder
===========
DataFeeder
-----------
.. automodule:: paddle.v2.fluid.data_feeder
:members: DataFeeder
:noindex:
doc/api/v2/fluid/evaluator.rst 0 → 100644
浏览文件 @ bfe7e242
===========
Evaluator
===========
Evaluator
-----------
.. automodule:: paddle.v2.fluid.evaluator
:members: Evaluator
:noindex:
doc/api/v2/fluid/executor.rst 0 → 100644
浏览文件 @ bfe7e242
===========
Executor
===========
Executor
-----------
.. automodule:: paddle.v2.fluid.executor
:members: Executor
:noindex:
doc/api/v2/fluid/initializer.rst 0 → 100644
浏览文件 @ bfe7e242
===========
Initializer
===========
Initializer
-----------
.. automodule:: paddle.v2.fluid.initializer
:members: Initializer
:noindex:
ConstantInitializer
-------------------
.. automodule:: paddle.v2.fluid.initializer
:members: ConstantInitializer
:noindex:
UniformInitializer
------------------
.. automodule:: paddle.v2.fluid.initializer
:members: UniformInitializer
:noindex:
NormalInitializer
-----------------
.. automodule:: paddle.v2.fluid.initializer
:members: NormalInitializer
:noindex:
XavierInitializer
-----------------
.. automodule:: paddle.v2.fluid.initializer
:members: XavierInitializer
:noindex:
MSRAInitializer
---------------
.. automodule:: paddle.v2.fluid.initializer
:members: MSRAInitializer
:noindex:
doc/api/v2/fluid/io.rst 0 → 100644
浏览文件 @ bfe7e242
===========
IO
===========
is_parameter
-----------
.. autofunction:: paddle.v2.fluid.io.is_parameter
:noindex:
doc/api/v2/fluid/layers.rst 0 → 100644
浏览文件 @ bfe7e242
==========
Layers
==========
fc
---
.. autofunction:: paddle.v2.fluid.layers.fc
:noindex:
embedding
---------
.. autofunction:: paddle.v2.fluid.layers.embedding
:noindex:
dynamic_lstm
------------
.. autofunction:: paddle.v2.fluid.layers.dynamic_lstm
:noindex:
data
----
.. autofunction:: paddle.v2.fluid.layers.data
:noindex:
mean
----
.. autofunction:: paddle.v2.fluid.layers.mean
:noindex:
mul
---
.. autofunction:: paddle.v2.fluid.layers.mul
:noindex:
elementwise_add
---------------
.. autofunction:: paddle.v2.fluid.layers.elementwise_add
:noindex:
elementwise_sub
---------------
.. autofunction:: paddle.v2.fluid.layers.elementwise_sub
:noindex:
elementwise_mul
---------------
.. autofunction:: paddle.v2.fluid.layers.elementwise_mul
:noindex:
elementwise_div
---------------
.. autofunction:: paddle.v2.fluid.layers.elementwise_div
:noindex:
dropout
-------
.. autofunction:: paddle.v2.fluid.layers.dropout
:noindex:
reshape
--------
.. autofunction:: paddle.v2.fluid.layers.reshape
:noindex:
sigmoid
---------
.. autofunction:: paddle.v2.fluid.layers.sigmoid
:noindex:
scale
---------
.. autofunction:: paddle.v2.fluid.layers.scale
:noindex:
transpose
---------
.. autofunction:: paddle.v2.fluid.layers.transpose
:noindex:
sigmoid_cross_entropy_with_logits
---------------------------------
.. autofunction:: paddle.v2.fluid.layers.esigmoid_cross_entropy_with_logits
:noindex:
cast
----
.. autofunction:: paddle.v2.fluid.layers.cast
:noindex:
concat
-------
.. autofunction:: paddle.v2.fluid.layers.concat
:noindex:
sums
----
.. autofunction:: paddle.v2.fluid.layers.sums
:noindex:
linear_chain_crf
----------------
.. autofunction:: paddle.v2.fluid.layers.linear_chain_crf
:noindex:
assign
-------
.. autofunction:: paddle.v2.fluid.layers.embedding
:noindex:
split_lod_tensor
----------------
.. autofunction:: paddle.v2.fluid.layers.split_lod_tensor
:noindex:
merge_lod_tensor
----------------
.. autofunction:: paddle.v2.fluid.layers.merge_lod_tensor
:noindex:
cos_sim
--------
.. autofunction:: paddle.v2.fluid.layers.cos_sim
:noindex:
cross_entropy
-------------
.. autofunction:: paddle.v2.fluid.layers.cross_entropy
:noindex:
square_error_cost
-----------------
.. autofunction:: paddle.v2.fluid.layers.square_error_cost
:noindex:
accuracy
---------
.. autofunction:: paddle.v2.fluid.layers.accuracy
:noindex:
sequence_conv
-------------
.. autofunction:: paddle.v2.fluid.layers.sequence_conv
:noindex:
conv2d
------
.. autofunction:: paddle.v2.fluid.layers.conv2d
:noindex:
sequence_pool
-------------
.. autofunction:: paddle.v2.fluid.layers.sequence_pool
:noindex:
sequence_first_step
-------------------
.. autofunction:: paddle.v2.fluid.layers.sequence_first_step
:noindex:
sequence_last_step
------------------
.. autofunction:: paddle.v2.fluid.layers.sequence_last_step
:noindex:
pool2d
------
.. autofunction:: paddle.v2.fluid.layers.pool2d
:noindex:
batch_norm
----------
.. autofunction:: paddle.v2.fluid.layers.batch_norm
:noindex:
beam_search_decode
------------------
.. autofunction:: paddle.v2.fluid.layers.beam_search_decode
:noindex:
lod_rank_table
--------------
.. autofunction:: paddle.v2.fluid.layers.lod_rank_table
:noindex:
max_sequence_len
----------------
.. autofunction:: paddle.v2.fluid.layers.max_sequence_len
:noindex:
topk
-----
.. autofunction:: paddle.v2.fluid.layers.topk
:noindex:
lod_tensor_to_array
-------------------
.. autofunction:: paddle.v2.fluid.layers.lod_tensor_to_array
:noindex:
array_to_lod_tensor
-------------------
.. autofunction:: paddle.v2.fluid.layers.array_to_lod_tensor
:noindex:
fill_constant
-------------
.. autofunction:: paddle.v2.fluid.layers.fill_constant
:noindex:
fill_constant_batch_size_like
-----------------------------
.. autofunction:: paddle.v2.fluid.layers.fill_constant_batch_size_like
:noindex:
ones
----
.. autofunction:: paddle.v2.fluid.layers.ones
:noindex:
zeros
-----
.. autofunction:: paddle.v2.fluid.layers.zeros
:noindex:
increment
---------
.. autofunction:: paddle.v2.fluid.layers.increment
:noindex:
array_write
-----------
.. autofunction:: paddle.v2.fluid.layers.array_write
:noindex:
create_array
------------
.. autofunction:: paddle.v2.fluid.layers.create_array
:noindex:
less_than
---------
.. autofunction:: paddle.v2.fluid.layers.less_than
:noindex:
array_read
----------
.. autofunction:: paddle.v2.fluid.layers.array_read
:noindex:
shrink_memory
--------------
.. autofunction:: paddle.v2.fluid.layers.shrink_memory
:noindex:
array_length
-------------
.. autofunction:: paddle.v2.fluid.layers.array_length
:noindex:
conv2d_transpose
----------------
.. autofunction:: paddle.v2.fluid.layers.conv2d_transpose
:noindex:
sequence_expand
---------------
.. autofunction:: paddle.v2.fluid.layers.sequence_expand
:noindex:
gru_unit
--------
.. autofunction:: paddle.v2.fluid.layers.gru_unit
:noindex:
lstm_unit
---------
.. autofunction:: paddle.v2.fluid.layers.lstm_unit
:noindex:
sequence_softmax
----------------
.. autofunction:: paddle.v2.fluid.layers.sequence_softmax
:noindex:
reduce_sum
----------
.. autofunction:: paddle.v2.fluid.layers.reduce_sum
:noindex:
reduce_mean
-----------
.. autofunction:: paddle.v2.fluid.layers.reduce_mean
:noindex:
reduce_max
----------
.. autofunction:: paddle.v2.fluid.layers.reduce_max
:noindex:
reduce_min
----------
.. autofunction:: paddle.v2.fluid.layers.reduce_min
:noindex:
split
-----
.. autofunction:: paddle.v2.fluid.layers.split
:noindex:
logsigmoid
----------
.. autofunction:: paddle.v2.fluid.layers.logsigmoid
:noindex:
exp
---
.. autofunction:: paddle.v2.fluid.layers.exp
:noindex:
relu
----
.. autofunction:: paddle.v2.fluid.layers.relu
:noindex:
tanh
----
.. autofunction:: paddle.v2.fluid.layers.tanh
:noindex:
tanh_shrink
-----------
.. autofunction:: paddle.v2.fluid.layers.tanh_shrink
:noindex:
softshrink
----------
.. autofunction:: paddle.v2.fluid.layers.softshrink
:noindex:
sqrt
----
.. autofunction:: paddle.v2.fluid.layers.sqrt
:noindex:
abs
----
.. autofunction:: paddle.v2.fluid.layers.abs
:noindex:
ceil
----
.. autofunction:: paddle.v2.fluid.layers.ceil
:noindex:
floor
-----
.. autofunction:: paddle.v2.fluid.layers.floor
:noindex:
round
-----
.. autofunction:: paddle.v2.fluid.layers.round
:noindex:
reciprocal
----------
.. autofunction:: paddle.v2.fluid.layers.reciprocal
:noindex:
log
---
.. autofunction:: paddle.v2.fluid.layers.log
:noindex:
square
------
.. autofunction:: paddle.v2.fluid.layers.square
:noindex:
softplus
--------
.. autofunction:: paddle.v2.fluid.layers.softplus
:noindex:
softsign
---------
.. autofunction:: paddle.v2.fluid.layers.softsign
:noindex:
brelu
-----
.. autofunction:: paddle.v2.fluid.layers.brelu
:noindex:
leaky_relu
----------
.. autofunction:: paddle.v2.fluid.layers.leaky_relu
:noindex:
soft_relu
---------
.. autofunction:: paddle.v2.fluid.layers.soft_relu
:noindex:
elu
----
.. autofunction:: paddle.v2.fluid.layers.elu
:noindex:
relu6
-----
.. autofunction:: paddle.v2.fluid.layers.relu6
:noindex:
pow
----
.. autofunction:: paddle.v2.fluid.layers.pow
:noindex:
hard_shrink
-----------
.. autofunction:: paddle.v2.fluid.layers.hard_shrink
:noindex:
thresholded_relu
----------------
.. autofunction:: paddle.v2.fluid.layers.thresholded_relu
:noindex:
hard_sigmoid
-------------
.. autofunction:: paddle.v2.fluid.layers.hard_sigmoid
:noindex:
swish
------
.. autofunction:: paddle.v2.fluid.layers.swish
:noindex:
doc/api/v2/fluid/nets.rst 0 → 100644
浏览文件 @ bfe7e242
===========
Nets
===========
simple_img_conv_pool
--------------------
.. autofunction:: paddle.v2.fluid.nets.simple_img_conv_pool
:noindex:
img_conv_group
---------------
.. autofunction:: paddle.v2.fluid.nets.img_conv_group
:noindex:
sequence_conv_pool
------------------
.. autofunction:: paddle.v2.fluid.nets.sequence_conv_pool
:noindex:
glu
---
.. autofunction:: paddle.v2.fluid.nets.glu
:noindex:
doc/api/v2/fluid/optimizer.rst 0 → 100644
浏览文件 @ bfe7e242
===========
Optimizer
===========
Optimizer
-----------
.. automodule:: paddle.v2.fluid.optimizer
:members: Optimizer
:noindex:
SGDOptimizer
-----------
.. automodule:: paddle.v2.fluid.optimizer
:members: SGDOptimizer
:noindex:
MomentumOptimizer
-----------------
.. automodule:: paddle.v2.fluid.optimizer
:members: MomentumOptimizer
:noindex:
AdagradOptimizer
----------------
.. automodule:: paddle.v2.fluid.optimizer
:members: AdagradOptimizer
:noindex:
AdamOptimizer
-------------
.. automodule:: paddle.v2.fluid.optimizer
:members: AdamOptimizer
:noindex:
AdamaxOptimizer
-----------
.. automodule:: paddle.v2.fluid.optimizer
:members: AdamaxOptimizer
:noindex:
DecayedAdagradOptimizer
-----------------------
.. automodule:: paddle.v2.fluid.optimizer
:members: DecayedAdagradOptimizer
:noindex:
doc/api/v2/fluid/param_attr.rst 0 → 100644
浏览文件 @ bfe7e242
===========
ParamAttr
===========
ParamAttr
-----------
.. automodule:: paddle.v2.fluid.param_attr
:members: ParamAttr
:noindex:
doc/api/v2/fluid/profiler.rst 0 → 100644
浏览文件 @ bfe7e242
===========
Profiler
===========
Profiler
-----------
.. autofunction:: paddle.v2.fluid.profiler.cuda_profiler
:noindex:
doc/api/v2/fluid/regularizer.rst 0 → 100644
浏览文件 @ bfe7e242
===========
Regularizer
===========
WeightDecayRegularizer
----------------------
.. automodule:: paddle.v2.fluid.regularizer
:members: WeightDecayRegularizer
:noindex:
L2DecayRegularizer
------------------
.. automodule:: paddle.v2.fluid.regularizer
:members: L2DecayRegularizer
:noindex:
L1DecayRegularizer
-------------------
.. automodule:: paddle.v2.fluid.regularizer
:members: L1DecayRegularizer
doc/design/backward.md 0 → 100644
浏览文件 @ bfe7e242
# Backward Building
## Motivation
In Neural Network, most models are solved by the backpropagation algorithm(known as **BP**) at present. Technically, BP calculates the gradient of the loss function, then propagates it back through the networks following the chain rule. However, when configuring the model structure, users do not need to define the backward part. So a mechanism is required by the framework which can complete the model's backward part automatically according to the given forward part.
When implementing a specific `op`, the developer is also asked to implement its backward version, called `grad_op`. A `grad_op` takes gradients of its corresponding `op`'s outputs, and calculate gradients of the `op`'s inputs. During the building of a model's backward part, the framework creates each forward `op`'s `grad_op`, and then string them together in reverse order of forwarding part. In this way, gradients spread from the end to the beginning of the model, in another word, from the loss to parameters.
## Challenges
The motivation of backward building is apparent. However, implementation it correctly is not so easy. In the **Fluid** design, a deep learning model is described by `Program`, `Block`, `Op` and `Variable`. The `Block` itself can be nested. It means that the `op`s and `variable`s are scattered across different blocks rather than all be gathered in a single graph. Our backward building algorithm shall visit blocks in recursive order and be able to insert `grad_op`s and new created `variable`s into the right place.
## Usage
Although the whole algorithm is comprised of many functions, only one is exposed as API:
```python
def append_backward(loss, parameter_list=None, no_grad_set=None):
"""
Append backward part to main_program
Args:
loss(Variable): The variable generated by the cost function.
parameter_list(list): Parameters that need to be updated by optimizers.
If None, it means all parameters need to be updated.
no_grad_set(set): Variables that have no gradients in Block 0.
If None, the set will be generated inside the function and
contains all variables with `step_gradient=True` from all blocks.
Return:
(list[Variable]): list of (parameters, gradients) pair.
"""
```
By invoking this API, the framework appends backward part of the program where the `loss` is. It takes three arguments. `loss` means the final loss value. It must be a scalar and is usually the output of the loss layer. It is also where the gradient generated and backpropagation starts. `parameter_list` marks all parameters needs updating. If it's `None`, all parameter will be updated by optimizers. `no_grad_set` marks variables without gradient. if all outputs of some `grad_op` are in `no_grad_set`, the `grad_op` will not be run.
This API will be invoked automatically before optimizer building.
As a result, in most cases, users do not need to invoke the API by themselves to append backward part.
## Implementation
The implementation of backward building algorithm is in `backward.py` file. The whole algorithm can be divided into two independent parts: creating `grad_op`s and creating new variables.
### Creating `grad_op`s
The creating of `grad_op`s is implemented by:
```python
def _append_backward_ops_(target,
block,
target_block,
no_grad_dict,
grad_to_var):
"""
Create all grad ops, and insert them into given block
Args:
target(Variable): the target variable of forward pass
block(Block): the block where forward ops are
target_block(Block): the block which is going to hold new generated grad ops
no_grad_dict(dict):
key(int) block index
val(set) a set of varibale names. These varibales have no gradient
grad_to_var(dict)(output argument):
key(str): grad variable name
val(str): corresponding forward variable name
"""
```
Given a `block`, the function will traverses all `op`s in this block in reverse order, gets corresponding `grad_op` from the C++ core via `core.get_grad_op_desc()`, then append it to `target_block`.
However, some specific `op`(e.g. `while_op`, `if_else_op`) can hold its own sub-block. For these sub-blocks contains `op`s as well, the `grad_op` creating should be recursive.
During the reverse traversal, we check each `op` whether it has an attribute named `sub_block`. If so, it means there is a sub-block and we need to deal with it first. After creating a new block whose father is the one in `op`'s attribute, we invoke `_append_backward_ops_()` recursively, assigning the new block to parameter `target_block` and the one in `op`'s attribute to `block`. The *pseudo-code* shows this process:
```
******* pseudo-code ********
for op in reversed(block.ops):
if op has an attribute named 'sub_block':
Get the sub-block(`s_block`) from op's attribute.
Create a new block(`grad_s_block`), whose father is `s_block`.
Invoke _append_backward_ops_(), with `block=s_block` and `target_block=grad_s_block`
Invoke `core.get_grad_op_desc()` to get op's grad_op.
Insert name correspondings between variables and their gradients of the grad_op to grad_to_var
Assign grad_s_block to grad_op as it's 'sub_block' attribute.
Append grad_op to current target_block.
```
The first invoking of `_append_backward_ops_()` is initiated by `append_backward()`, in which parameters `block` and `target_block` are all assigned with root block(the block with index 0).
### Corner Cases of `grad_op` Creating
In the previous section, we show the regular process of `grad_op` creating. However, in some corner cases, the conventional algorithm is not enough to get the correct result and appending handling is required. These additional processes run after the algorithm mentioned above and do some special adjusts on its output `grad_op`s.
#### Shared Variables
If a variable is read by more than one `op` in the forward pass, its gradient is likely to be written by more than one `grad_op`s in the next backward pass. To make the gradient result being the sum of all `grad_op`s' outputs instead of the last running one, we assign each output with a temporary variable and then add a `sum_op` to add them up.
For the debug convenience, if the final gradient name is `w@GRAD`, it's corresponding temporary variables will be named as `w@GRAD@RENAME@0`, `w@GRAD@RENAME@1`...
See function `_addup_repetitive_outputs_` in `backward.py` for implementation details.
#### No Gradient Variables
In our framework, variables can be marked as *no_gradient*, it means that the gradient of this variable is unnecessary and can be considered as zero in model training. Apparently, when all the outputs of some `grad_op` are marked as *no_gradient*, the `grad_op` itself can be skipped in backward pass.
Another situation is all the gradient inputs of some `grad_op` are marked as *no_gradient*, which means all of them can be considered as zeros. For `grad_op`s are in essence the propagation of gradients, all the outputs are definitely zeros when all gradient inputs are zeros. Therefore the `grad_op` can also be skipped.
It should be noted that all these zero gradients still need to be creating and initialized by something, otherwise following `grad_op`s who take these gradients as inputs take the risk of using uninitialized memory. In our code, we employ `fill_zeros_like_op` to initialize them as all zeros.
This features are implemented in function `_remove_no_grad_branch_`. It checks new created `grad_op`s one-by-one, removes who can be skipped and inserts `fill_zeros_like_op` when its necessary. We can get the `no_grad_set` from the `_append_backward_ops_` argument `no_grad_dict` or generate it on the fly by scanning all variables' `no_gradient` attribute(True or False).
### Creating Backward Variables
Up to now, we have completed all creating and adjusting jobs of `grad_op`s. However, backward variables have not been created. Now they are only represented by `grad_op`'s input and output arguments. The backward variable creating job will be done by:
```python
def _append_backward_vars_(block,
start_op_idx,
grad_to_var,
grad_info_map):
"""
Create new variables required by backward pass.
Args:
block(Block): the block where new variables will be created
start_op_idx(int): Only variables required by ops in block.ops[start_op_idx : ] will be created
grad_to_var(dict):
key(str): grad variable name
val(str): corresponding forward variable name
In most cases, this dict is generated by _append_backward_ops_()
grad_info_map(dict)(output argument):
key(str): forward variable name
val(tuple): a tuple of (str, int), str is the corresponding grad name, int is the block index
"""
```
Given a `block`, this function traverses all the `grad_op`s in it(The argument `start_op_idx` indicates where the grad_op sequence starts.) and creates all the uncreated outputs. The *pseudo-code* shows this process:
```
for op in block.ops[start_op_idx : ]:
if op has an attribute named 'sub_block':
Get the sub-block(`s_block`) from op's attribute.
Invoke _append_backward_vars_(), with `block=s_block`
for var_name in op.all_output_names():
if block.has_var_recursive(var_name) or var_name is the name of empty variable:
continue
create a new variable named 'var_name' in block
if grad_to_var.has_key(var_name):
set grad_info_map[grad_to_var[var_name]] as a tuple of (var_name. block)
do op's var type inference
do op's shape inference
```
doc/design/block.md
浏览文件 @ bfe7e242
......@@ -202,8 +202,8 @@ This `OpDesc` value is in the `ops` field of the `BlockDesc` value representing
During the generation of the Protobuf message, the Block should store VarDesc (the Protobuf message which describes Variable) and OpDesc (the Protobuf message which describes Operator).
VarDesc in a block should have its name scope to avoid local variables affect parent block's name scope.
Child block's name scopes should inherit the parent's so that OpDesc in child block can reference a VarDesc that stored in parent block. For example:
VarDesc in a block should have its name scope to avoid local variables affecting parent block's name scope.
Child block's name scopes should inherit the parent's so that OpDesc in child block can reference a VarDesc that is stored in the parent block. For example:
```python
a = pd.Variable(shape=[20, 20])
......@@ -291,10 +291,10 @@ public:
}
void Run(const framework::Scope& scope,
const platform::DeviceContext& dev_ctx) const override {
const platform::Place& place) const override {
PADDLE_ENFORCE(symbols_ready_, "operators and variables should be created first.");
for (auto& op : runtime_table_.ops()) {
op->Run(scope, dev_ctx);
op->Run(scope, place);
}
}
......
doc/design/concurrent_programming.md 0 → 100644
浏览文件 @ bfe7e242
# Design Doc: Concurrent Programming with Fluid
With PaddlePaddle Fluid, users describe a program other than a model. The program is a [`ProgramDesc`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/framework.proto) protobuf message. TensorFlow/MxNet/Caffe2 applications generate protobuf messages too, but their protobuf messages represent the model, a graph of operators, but not the program that trains/uses the model.
Many know that when we program TensorFlow, we can specify the device on which each operator runs. This allows us to create a concurrent/parallel AI application. An interesting questions is **how does a `ProgramDesc` represents a concurrent program?**
The answer relies on the fact that a `ProgramDesc` is similar to an abstract syntax tree (AST) that describes a program. So users just program a concurrent program that they do with any concurrent programming language, e.g., [Go](https://golang.org).
## An Analogy
The following table compares concepts in Fluid and Go
| Go | Fluid |
|----|-------|
|user-defined functions | [layers](https://github.com/PaddlePaddle/Paddle/tree/develop/python/paddle/v2/fluid) |
| control-flow and built-in functions | [intrinsics/operators](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/operators) |
| goroutines, channels | [class ThreadPool](https://github.com/PaddlePaddle/Paddle/tree/develop/paddle/framework/thread_pool.h) |
| runtime | [class Executor](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.h) |
## An Example Concurrent Program
To review all above concepts in an example, let us take a simple program and writes its distributed version.
Suppose that we want to parallelize a naive Fluid program (written in Go and calling Fluid's Go binding) that multiplies two tensors.
```go
import "fluid"
func paddlepaddle() {
X = fluid.read(...)
W = fluid.Tensor(...)
Y = fluid.mult(X, W)
}
```
Please be aware that the Fluid's Go binding provides the default `main` function, which calls the `paddlepaddle` function, which, in this case, is defined in above program and creates the following `ProgramDesc` message.
```protobuf
message ProgramDesc {
block[0] = Block {
vars = [X, W, Y],
ops = [
read(output = X)
assign(input = ..., output = W)
mult(input = {X, W}, output = Y)
],
}
}
```
Then, the default `main` function calls `fluid.run()`, which creates an instance of the [`class Executor`](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/executor.h) and calls `Executor.Run(block[0])`, where `block[0]` is the first and only block defined in above `ProgramDesc` message.
The default `main` function is defined as follows:
```go
func main() {
paddlepaddle()
fluid.run()
}
```
## The Concurrent Version
By parallelizing the above program, we could support very big tensor X by splitting into small pieces {x_1, x_2, ...} and sent each piece to worker process/node for parallel multiplication.
In this case, we can write a transpiler that takes a `ProgramDesc` message that represents the above example program and outputs two `ProgramDesc` messages, one for running on the master process/node, and the other one for worker processes/nodes.
### The Master Program
The master program could look like the following:
```protobuf
message ProgramDesc {
block[0] = Block {
vars = [X, L, Y],
ops = [
read(output = X)
kube_get_workers_addrs(output = L)
Y = tensor_array(len(L))
parallel_for(input = X, output = Y,
attrs = {L, block_id(1)}) # referring to block 1
]
}
block[1] = Block {
parent = 0,
vars = [x, y, index],
ops = [
slice(input = [X, index], output = x) # index is initialized by parallel_for
send(input = x, attrs = L[index])
recv(outputs = y, attrs = L[index])
assign(input = y, output = Y[index])
]
}
}
```
The equivalent Fluid program (calling the Go binding) is:
```go
func main() { //// block 0
X = fluid.read(...)
L = fluid.k8s.get_worker_addrs()
Y = fluid.tensor_array(len(L))
fluid.parallel_for(X, L,
func(index int) { //// block 1
x = X[index]
fluid.send(L[index], x)
y = fluid.recv(L[index])
Y[index] = y
})
}
```
An explanation of the above program:
- `fluid.k8s` is a package that provides access to Kubernetes API.
- `fluid.k8s.get_worker_addrs` returns the list of IP and ports of all pods of the current job except for the current one (the master pod).
- `fluid.tensor_array` creates a [tensor array](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/lod_tensor_array.h). `fluid.parallel_for` creates a `ParallelFor` intrinsic, which, when executed,
1. creates `len(L)` scopes, each for the concurrent running of the sub-block (block 1 in this case), and initializes a variable named "index" in the scope to an integer value in the range `[0, len(L)-1]`, and
2. creates `len(L)` threads by calling into the `ThreadPool` singleton, each thread
1. creates an Executor instance, and
2. calls `Executor.Run(block)`, where `block` is block 1 as explained above.
1. Please be aware that block 1 is a sub-block of block 0, so ops in block 1 could refer to variables defined in block 0.
### The Worker Program
The worker program looks like
```go
func main() {
W = Tensor(...)
x = fluid.listen_and_do(
fluid.k8s.self_addr(),
func(input Tensor) {
output = fluid.mult(input, W)
})
}
```
where
- `fluid.listen_and_do` creates a `ListenAndDo` intrinsic, which, when executed,
1. listens on the current pod's IP address, as returned by `fliud.k8s.self_addr()`,
2. once a connection is established,
1. creates a scope of two parameters, "input" and "output",
2. reads a [Fluid variable](https://github.com/PaddlePaddle/Paddle/blob/develop/paddle/framework/variable.h) and saves it into "input",
3. creates an Executor instance and calls `Executor.Run(block)`, where the block is generated by running the lambda specified as the second parameter of `fluid.listen_and_do`.
## Summarization
From the above example, we see that:
1. Fluid enables the imperative programming paradigm by:
1. letting users describe a program, but not a model (a sequence of layers, or a graph of operators), and
2. call the `fluid.run` function that runs the program implicitly.
1. The program is described as a `ProgramDesc` protobuf message.
2. Function `Executor.Run` takes a block, instead of a `ProgramDesc`, as its parameter.
3. `fluid.run` calls `Executor.Run` to run the first block in the `ProgramDesc` message.
4. `Executor.Run`'s implementation is extremely simple -- it doesn't plan the execution nor create threads; instead, it runs on the current thread and execute intrinsics/operators' `Run` method sequentially as they appear in the `Block.ops` array.
5. Intrinsics/operators' `Run` method might create threads. For example, the `ListenAndDo` operator creates a thread to handle each incoming request.
6. Threads are not necessarily OS thread; instead, they could be [green threads](https://en.wikipedia.org/wiki/Green_threads) managed by ThreadPool. Multiple green threads might run on the same OS thread. An example green threads is Go's [goroutines](https://tour.golang.org/concurrency/1).
doc/design/dist_refactor/distributed_architecture.md 0 → 100644
浏览文件 @ bfe7e242
# Design Doc: Distributed Training Architecture
## Abstract
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. 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: 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: 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 mentioned limitations.
## Analysis
The assumption is that the user writes the trainer program in either Python or C++.
### Limitation 1
There are two basic functionalities in the trainer program:
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 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 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 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 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.
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"/>
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 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 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 also run an IR, which can be different to the trainer side
For a detailed explanation, refer to this document -
[Design Doc: Parameter Server](./parameter_server.md)
## Distributed Training Architecture
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 major components are: *Python API*, *Distribute Transpiler* and *Remote Executor*.
### Python API
Python API is the Python library that user's Python code invokes, to read the data, build the neural network topology, and start training, etc.
```Python
images = fluid.layers.data(name='pixel', shape=[1, 28, 28], dtype='float32')
label = fluid.layers.data(name='label', shape=[1], dtype='int64')
...
predict = fluid.layers.fc(input=conv_pool_2, size=10, act="softmax")
cost = fluid.layers.cross_entropy(input=predict, label=label)
avg_cost = fluid.layers.mean(x=cost)
optimizer = fluid.optimizer.Adam(learning_rate=0.01)
optimizer.minimize(avg_cost)
train_reader = paddle.batch(
paddle.reader.shuffle(
paddle.dataset.mnist.train(), buf_size=500),
batch_size=BATCH_SIZE)
place = fluid.CPUPlace()
exe = fluid.Executor(place)
for pass_id in range(10):
for data in train_reader():
loss, acc = exe.run(trainer_prog,
feed=feeder.feed(data),
fetch_list=[avg_cost])
```
The code above is a typical local training program, the "Training Program" is built using helper functions such as
`fluid.layer.fc`. The training is done by calling `Executor.run`
iteratively.
For more details, the implementation of IR is [Program](../program.md), and `ProgramDesc` is the protobuf type.
[Executor](../executor.md) simply runs the `ProgramDesc`. For local training you generally use
`Executor` to run the program locally. For any kind of distributed training, you can use
`RemoteExecutor` to specify desired distributed training method with some optional arguments.
### Distributed Transpiler
The Distributed Transpiler automatically converts the IR (in protobuf format) to partitioned IRs. Then
the Remote Executor dispatches the new IRs to Remote Executors across the cluster.
Below are the steps that are followed :
1. User only need to change `Executor` to `RemoteExecutor` to change local program to distributed program.
1. `RemoteExecutor` calls `Distributed Transpiler` to "transpile" user's program to several IRs representing a
distributed training program:
1. Parse configurations from `RemoteExecutor`.
1. Determine the type of distributed program, can be DataParallelism, ModelParallelism or Streaming.
1. Partition the `ProgramDesc` according to type and add `send` / `recv` OP pair on the boundaries. Take
DataParallelism type for example, it removes the optimization operators and add a `send` OP to the
"trainer" role, then add the optimization operators to the parameter server role within the `recv` OP.
1. Dispatch the partitioned graph to different `RemoteExecutor` in the cluster.
1. `RemoteExecutor` on each node run the received `ProgramDesc` utill the end.
### RemoteExecutor
As shown in the graph, `RemoteExecutor.run` sends the IR to the cluster for Execution.
You can also use parameter `fetch_list` to interactively fetch variable back to local for
log printing.
The Python `RemoteExecutor` is derived from `Executor` class.
```python
exe = RemoteExecutor(
feed=feeder.feed(data),
fetch_list=[avg_cost],
job_desc=JobDesc(
jobname,
num_trainer,
num_pserver,
cpu_per_trainer,
gpu_per_trainer,
mem_per_trainer,
cpu_per_pserver,
mem_per_pserver
))
for data in train_reader():
loss, acc = exe.run(trainer_prog,
feed=feeder.feed(data),
fetch_list=[avg_cost])
```
`JobDesc` object describe the distributed job resource specification to run on
Cluster environment.
<img src="src/remote_executor.png"/>
`RemoteExecutor.run` sends the `ProgramDesc` and
[TrainingJob](https://github.com/PaddlePaddle/cloud/blob/develop/doc/autoscale/README.md#training-job-resource)
to a server in the cluster which executes `RemoteExecutor.listen`. This server is responsible
to start the final Kubernetes Jobs to run the different role of `ProgramDesc`.
### Placement Algorithm
Our first implementation will only support "trainer-parameter server" placement: the parameters, initializers, and optimizers are all placed on the PaddlePaddle runtimes with the parameter server role. Everything else will be placed on the PaddlePaddle runtimes with the trainer role. This has the same functionality as the "trainer-parameter server" architecture of PaddlePaddle v0.10.0, but is more generic and flexible.
In the future, a more general placement algorithm should be implemented, which makes placements according to the input IR, and a model of device computation time and device communication time. Model parallelism requires the generic placement algorithm.
### Local Training Architecture
The local training architecture will be the same as the distributed training architecture, the difference is that everything runs locally, and there is just one PaddlePaddle runtime:
<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 `Executor.run`. However, should
be used for debugging purpose only. The users are encouraged to use
the read data OPs.
## References:
[1] [TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems](https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/45166.pdf)
[2] [TensorFlow: A System for Large-Scale Machine Learning](https://www.usenix.org/system/files/conference/osdi16/osdi16-abadi.pdf)
doc/design/dist_refactor/multi_cpu.md 0 → 100644
浏览文件 @ bfe7e242
# Design Doc: Execute the Program with Multi CPU
## Abstract
This Design Doc propose an approach to make the user-defined Op graph
running with multi-CPU, we will use an auto transpiler to convert the user-defined
Op graph to a multi-CPU Op graph, and run `ParallelDo` Op to run the graph.
## Transpiler
<img src="src/multi-threads/single-thread@3x.png" width="300">
After converted:
<img src="src/multi-threads/multi-threads@3x.png" width="1000">
## Implement
- `Multi-CPU Transpiler` will convert the graph to a multi-CPU graph
which would be executed with multi-threads.
- `BlockingCounter` will `Init/Decrement` an atomic counter, and Blocking `Wait`
for the atomic counter become `0`:
```cpp
BlockingCounter bc(thread_count);
for (int i = 0; i < thread_count; ++i) {
thread_pool->Start([&bc] {bc.DecrementCount(); })
}
bc.Wait();
```
- `ParallelDo` Operator
- Initialize a thread pool which is a Singleton.
- Use a block id as the input, and create run the specify Block on independent scope
with multi-threads.
- Initialize a `BlockingCounter` instance and wait until all threads are done.
- `Split` Operator will split the Input Tensor into a TensorArray.
- `Merge` merge all the gradients which calculated in different threads
with `mean/sum/max/min...` method, and then run the Optimizer Op to optimize `W`.
## TODO
- Improve the optimizer stage with multi-threads, since we could
assign the parameters to the different threads and execute
optimizer with multi-threads.
doc/design/refactor/parameter_server.md → doc/design/dist_refactor/parameter_server.md
浏览文件 @ bfe7e242
# Design Doc: Operation Graph Based Parameter Server
# Design Doc: Parameter Server
## Abstract
......@@ -10,7 +10,7 @@ different purposes.
## Background
The previous implementations of the parameter server does not run a
subgraph. parameter initialization, optimizer computation, network
fluid sub-program. Parameter initialization, optimizer computation, network
communication and checkpointing are implemented twice on both the
trainer and the parameter server.
......@@ -23,10 +23,10 @@ server becomes a natural extension.
## Design
### Graph Converter
### Distributed Transpiler
The *graph converter* converts the user-defined operation (OP) graph
into subgraphs to be scheduled on different nodes with the following
The *Distributed Transpiler* converts the user-defined fluid program
into sub-programs to be scheduled on different nodes with the following
steps:
1. OP placement: the OPs will be placed on different nodes according
......@@ -34,7 +34,6 @@ steps:
time. Currently we will use a simple heuristic that puts parameter
varable on parameter server workers and everything else on trainer
workers.
1. Add communication OPs to enable the communication between nodes.
We will need these OPs: *Send*, *Recv*, *Enqueue*, *Dequeue*.
......@@ -48,8 +47,8 @@ After converting:
<img src="src/dist-graph.png" width="700"/>
1. The parameter variable W and it's optimizer subgraph are placed on the parameter server.
1. Operators are added to the subgraphs.
1. The parameter variable W and it's optimizer program are placed on the parameter server.
1. Operators are added to the program.
- *Send* sends data to the connected *Recv* operator. The
scheduler on the receive node will only schedule *Recv* operator
to run when the *Send* operator has ran (the *Send* OP will mark
......@@ -64,39 +63,30 @@ After converting:
### Benefits
- Model parallelism become easier to implement: it's an extension to
the trainer - parameter server approach. we already have the
communication OPs, but need to extend the graph converter's
placement functionality.
the trainer - parameter server approach. We can have several "Transpilers"
to achieve different goals.
- User-defined optimizer is easier to add - user can now express it as
a subgraph.
a sub-program.
- No more duplication logic inside the trainer and the parameter
server mentioned in the background section.
### Challenges
- It might be hard for the graph converter to cut a general graph
(without any hint for which subgraph is the optimizer). We may need
to label which subgraph inside the OP graph is the optimizer.
- It's important to balance the parameter shards of on multiple
parameter server. If a single parameter is very big (some
word-embedding, fully connected, softmax layer), we need to
automatically partition the single parameter onto different
parameter servers when possible (only element-wise optimizer depends
on the parameter variable).
- In the "Aync SGD" figure, the "W" variable on the parameter server
could be read and wrote concurrently. See
[here](https://github.com/PaddlePaddle/Paddle/pull/6394) for more
details about concurrent program in fluid.
### Discussion
- In the "Aync SGD" figure, the "W" variable on the parameter server
could be read and wrote concurrently, what is our locking strategy?
E.g., each variable have a lock cpp method to be invoked by every
OP, or, have a lock OP.
- Can the Enqueue OP be implemented under our current tensor design
(puts the input tensor into the queue tensor)?
- *Dequeue* OP will have variable numbers of output (depends on the
`min_count` attribute), does our current design support it? (similar
question for the *Add* OP)
......
doc/design/dist_refactor/src/local_architecture.graffle 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/dist_refactor/src/local_architecture.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/dist_refactor/src/multi-threads.graffle 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/dist_refactor/src/remote_executor.graffle 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/dist_refactor/src/remote_executor.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/error_clip.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/evaluator.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/executor.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/float16.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/fluid-compiler.graffle 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/fluid-compiler.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/fluid.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/control_flow_graph.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/dataflow_equations.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/deep_learning.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/multigpu_allreduce.graffle 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/multigpu_allreduce.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/multigpu_before_convert.graffle 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/multigpu_before_convert.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/images/profiler.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/kernel_hint_design.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/memory_optimization.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/image/engine.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/image/gradients.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/image/layers.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/image/matrix.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/image/overview.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/mkl_packed.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/mkldnn.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkl/mkldnn_fluid.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/mkldnn/README.MD 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/design/mkldnn/image/overview.png 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/design/operator_kernel_type.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/optimizer.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/paddle_nccl.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/profiler.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/python_api.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/reader/README.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/refactor/session.md 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/design/refactor/src/local_architecture.png 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/design/releasing_process.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/support_new_device.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/switch_kernel.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/design/var_desc.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/getstarted/basic_usage/index_cn.rst 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/getstarted/basic_usage/index_en.rst 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/getstarted/basic_usage/parameters.png 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/getstarted/build_and_install/cmake.png 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/getstarted/build_and_install/paddleci.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/getstarted/build_and_install/pip_install_cn.rst 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/getstarted/build_and_install/pip_install_en.rst 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/getstarted/concepts/src/infer.py 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/getstarted/index_cn.rst
浏览文件 @ bfe7e242
此差异已折叠。
doc/getstarted/index_en.rst
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/dev/build_cn.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/dev/build_en.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/dev/new_op_cn.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/dev/new_op_en.md
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/dev/write_docs_cn.rst
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/dev/write_docs_en.rst 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/index_cn.rst
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/index_en.rst
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/optimization/cpu_profiling.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/optimization/cpu_profiling_cn.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/optimization/pprof_1.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/optimization/pprof_2.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/read_source.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/capi/compile_paddle_lib_cn.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/capi/images/csr.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/capi/images/sequence_data.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/capi/images/workflow_of_CAPI.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/capi/index_cn.rst 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/capi/organization_of_the_inputs_cn.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/capi/workflow_of_capi_cn.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/cluster/fabric_cn.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/cluster/fabric_en.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/cluster/k8s_aws_cn.md 0 → 120000
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/cluster/openmpi_cn.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/cluster/openmpi_en.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/cluster/src/k8s-paddle-arch.png 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/howto/usage/k8s/k8s_basis_cn.md 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/howto/usage/k8s/src/k8s-paddle-arch.png 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
doc/mobile/cross_compiling_for_ios_en.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
doc/mobile/index_en.rst
浏览文件 @ bfe7e242
此差异已折叠。
paddle/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/api/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/api/test/testTrainConfig.py
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/Main.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/Matrix.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/config.h.in
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/error.cpp 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/error.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/examples/model_inference/dense/mnist_v2.py 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/gradient_machine.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/gradient_machine.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/main.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/capi/matrix.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/cuda/include/hl_cnn.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/cuda/src/avx_mathfun.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/cuda/src/hl_cuda_cnn.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/attribute.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/attribute.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/backward.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/backward.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/backward.md 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/framework/backward_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/block_desc.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/block_desc.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_device_transform.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_device_transform.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_device_transform_test.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_layout.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_layout_transform.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_layout_transform.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_transform.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_transform.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_type.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_type_transform.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/data_type_transform.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/ddim_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/details/cow_ptr.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/dim.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/dim_test.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/eigen_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/executor.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/executor.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/feed_fetch_type.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/framework.proto
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/init.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/init.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/init_test.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/library_type.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/lod_rank_table.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/lod_rank_table.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/lod_tensor.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/lod_tensor.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/lod_tensor_array.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/lod_tensor_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/lod_tensor_test.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_desc.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_desc.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_info.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_info.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_kernel_type.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_kernel_type_test.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_proto_maker.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_registry.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/op_registry.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/operator.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/operator.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/operator_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/program_desc.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/program_desc.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/prune.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/prune.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/prune_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/scope.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/scope.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/scope_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/selected_rows.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/selected_rows.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/shape_inference.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/shape_inference.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor.md
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor_array.cc 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/framework/tensor_array.h 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/framework/tensor_array_test.cc 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/framework/tensor_impl.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor_util.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor_util.cu 0 → 120000
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor_util.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/tensor_util_test.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/threadpool.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/threadpool.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/threadpool_test.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/type_defs.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/var_desc.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/var_desc.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/var_type.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/framework/variable.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/function/EigenGemm.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/function/GemmConvOp.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/function/Im2Col.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/function/Im2ColTest.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/FactorizationMachineLayer.cpp 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/FactorizationMachineLayer.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/MKLDNNLRNLayer.cpp 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/MKLDNNLRNLayer.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/MKLDNNLayer.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/MKLPackedRecurrentLayer.cpp 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/MKLPackedRecurrentLayer.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/MKLPackedWeight.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/PoolLayer.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/PoolLayer.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/layers/RecurrentLayer.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/gserver/tests/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/inference/CMakeLists.txt 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/inference/example.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/inference/inference.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/inference/inference.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/Allocator.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/CpuSparseMatrix.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/CpuSparseMatrix.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/MathFunctions.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/MathFunctions.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/Matrix.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/Matrix.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/NEONFunctions.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/NEONFunctions.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/SIMDFunctions.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/float16.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/tests/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/tests/test_float16.cpp 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/math/tests/test_float16.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/README.md
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/detail/memory_block.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/detail/meta_cache.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/detail/meta_cache.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/detail/meta_data.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/detail/meta_data.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/memcpy.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/memory.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/memory.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/memory/memory_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/accuracy_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/accuracy_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/accuracy_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/activation_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/activation_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/activation_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adadelta_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adadelta_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adadelta_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adagrad_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adagrad_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adagrad_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adam_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adam_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adam_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adamax_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adamax_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/adamax_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/array_operator.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/assign_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/assign_value_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/assign_value_op.cu.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/assign_value_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/auc_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/auc_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/batch_norm_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/batch_norm_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/beam_search_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/beam_search_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cast_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cast_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cast_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/chunk_eval_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/chunk_eval_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/clip_by_norm_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/clip_by_norm_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/clip_by_norm_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/clip_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/clip_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/clip_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/compare_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/compare_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/compare_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/concat_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/concat_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/concat_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cond_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cond_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/conv_cudnn_op.cc 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/operators/conv_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/conv_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/conv_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/conv_shift_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/conv_shift_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/conv_shift_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/conv_transpose_cudnn_op.cc 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/operators/cos_sim_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cos_sim_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cos_sim_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/crf_decoding_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/crf_decoding_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/crop_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/crop_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/crop_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/cross_entropy_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/CMakeLists.txt 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/grpc_client.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/grpc_client.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/grpc_server.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/grpc_server.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/safe_ref.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/send_recv.proto 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/sendrecvop_utils.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/sendrecvop_utils.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detail/simple_block_queue.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detection_output_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detection_output_op.cu.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/detection_output_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/dropout_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/dropout_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/dropout_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/dynamic_recurrent_op.cc 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/operators/dynamic_recurrent_op.h 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/operators/dynamic_recurrent_op_test.cc 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/operators/edit_distance_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/edit_distance_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/edit_distance_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/elementwise_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/expand_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/expand_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/expand_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/feed_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/fetch_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/fill_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/ftrl_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/ftrl_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/ftrl_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gather.cu.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gather_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gather_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gather_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/get_places_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gru_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gru_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gru_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gru_unit_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gru_unit_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/gru_unit_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/hinge_loss_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/hinge_loss_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/huber_loss_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/huber_loss_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/huber_loss_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/increment_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/is_empty_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/l1_norm_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/l1_norm_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/l1_norm_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/load_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lod_reset_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lod_reset_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lod_reset_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/log_loss_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/log_loss_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/log_loss_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/logical_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/logical_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/logical_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lookup_table_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lookup_table_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lookup_table_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lrn_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lrn_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lrn_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lstm_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lstm_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lstm_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lstm_unit_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lstm_unit_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/lstm_unit_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/cos_sim_functor.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/cos_sim_functor.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/cos_sim_functor.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/detection_util.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/gru_compute.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/im2col.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/im2col.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/im2col.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/matmul.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/maxouting.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/maxouting.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/maxouting.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/pooling.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/pooling.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/pooling.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/sequence_padding.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/sequence_padding.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/sequence_padding.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/sequence_padding_test.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/sequence_scale.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/sequence_scale.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/sequence_scale.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/softmax.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/softmax.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/softmax.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/unpooling.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/unpooling.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/unpooling.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/vol2col.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/vol2col.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/math/vol2col.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/matmul_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/matmul_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/matmul_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/max_sequence_len_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/maxout_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/maxout_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/maxout_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/mean_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/mean_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/mean_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/minus_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/minus_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/minus_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/momentum_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/momentum_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/momentum_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/mul_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/mul_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/mul_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/multiplex_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/multiplex_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/multiplex_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/nccl_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/nccl_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/nccl_op_test.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/nce_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/nce_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/net_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/net_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/net_op_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/norm_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/norm_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/norm_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/op_documentation/op_markdown_format.md 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/pad_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/pad_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/pad_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/parallel_do_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/pool_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/pool_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/pool_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/prelu_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/prelu_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/prelu_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/print_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/proximal_gd_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/proximal_gd_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/proximal_gd_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/rank_loss_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/rank_loss_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/rank_loss_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/recurrent_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/recv_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/reduce_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/reduce_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/reduce_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/reorder_lod_tensor_by_rank_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/reshape_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/reshape_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/reshape_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/rmsprop_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/rmsprop_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/rmsprop_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/rnn/recurrent_op_utils.cc 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/operators/rnn/recurrent_op_utils.h 已删除 100644 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/operators/roi_pool_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/roi_pool_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/roi_pool_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/row_conv_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/row_conv_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/row_conv_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/save_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/scale_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/scale_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/scale_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/scatter.cu.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/scatter_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/scatter_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/send_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/send_recv_op_test.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_conv_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_erase_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_erase_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_erase_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_expand_op.cu 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_pool_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_slice_op.cc 100755 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sequence_slice_op.h 100755 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sgd_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sgd_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sgd_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sign_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sign_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sign_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/softmax_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/softmax_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/softmax_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/split_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/split_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/split_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/spp_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/spp_op.cu.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/spp_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/strided_memcpy.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sum_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sum_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/sum_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/top_k_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/top_k_op.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/top_k_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/transpose_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/transpose_op.cu.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/transpose_op.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/unpool_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/unpool_op.cu.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/unpool_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/warpctc_op.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/warpctc_op.cu.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/warpctc_op.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/operators/while_op.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/adam_optimizer.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/adam_optimizer.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/lr_policy.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/optimizer.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/optimizer.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/sgd_optimizer.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/sgd_optimizer.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/optimizer/tensor.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/call_once.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/cpu_info_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/cuda_helper.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/cuda_profiler.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/cudnn_helper.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/device_context.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/device_context.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/dynload/cublas.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/dynload/cudnn.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/dynload/cudnn.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/dynload/nccl.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/dynload/nccl.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/dynload/warpctc.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/dynload/warpctc.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/enforce.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/for_range.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/gpu_info.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/gpu_info.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/hostdevice.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/mkldnn_helper.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/nccl_test.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/place.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/place.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/place_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/profiler.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/profiler.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/profiler_test.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/transform.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/transform_test.cu
浏览文件 @ bfe7e242
此差异已折叠。
paddle/platform/variant.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pserver/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/const_value.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/const_value.h 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/exception.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/exception.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/protobuf.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/pybind.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/pybind/tensor_py.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/scripts/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/scripts/check_env.sh 0 → 100755
浏览文件 @ bfe7e242
此差异已折叠。
paddle/scripts/cpplint.py
浏览文件 @ bfe7e242
此差异已折叠。
paddle/scripts/docker/README.md
浏览文件 @ bfe7e242
此差异已折叠。
paddle/scripts/docker/build.sh
浏览文件 @ bfe7e242
此差异已折叠。
paddle/scripts/submit_local.sh.in
浏览文件 @ bfe7e242
此差异已折叠。
paddle/scripts/tools/usage_stat/usage.sh 已删除 100755 → 0
浏览文件 @ 92baa885
此差异已折叠。
paddle/scripts/travis/build_doc.sh
浏览文件 @ bfe7e242
此差异已折叠。
paddle/string/piece.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/string/piece.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/string/piece_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/string/printf.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/string/printf_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/string/to_string.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/string/to_string_test.cc
浏览文件 @ bfe7e242
此差异已折叠。
paddle/testing/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/testing/paddle_gtest_main.cc 0 → 100644
浏览文件 @ bfe7e242
此差异已折叠。
paddle/trainer/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/trainer/Trainer.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/trainer/tests/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
paddle/utils/Flags.cpp
浏览文件 @ bfe7e242
此差异已折叠。
paddle/utils/Flags.h
浏览文件 @ bfe7e242
此差异已折叠。
paddle/utils/enable_virtualenv.py
浏览文件 @ bfe7e242
此差异已折叠。
proto/ModelConfig.proto
浏览文件 @ bfe7e242
此差异已折叠。
proto/OptimizerConfig.proto
浏览文件 @ bfe7e242
此差异已折叠。
python/.gitignore
浏览文件 @ bfe7e242
此差异已折叠。
python/CMakeLists.txt
浏览文件 @ bfe7e242
此差异已折叠。
python/paddle/__init__.py
浏览文件 @ bfe7e242
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