Skip to content

Commit bfecb57

Browse files
luotao1luotao1
authored
Merge pull request #11090 from luotao1/capi_doc
add build and install document of fluid inference library
2 parents 01bbe53 + 2007f63 commit bfecb57

File tree

5 files changed

+106
-60
lines changed

5 files changed

+106
-60
lines changed

‎doc/fluid/howto/index_cn.rst‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,5 +4,5 @@
44
.. toctree::
55
:maxdepth: 1
66

7+
inference/index_cn.rst
78
optimization/index_cn.rst
8-
inference/inference_support_in_fluid.md

‎doc/fluid/howto/index_en.rst‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -5,4 +5,3 @@ HOW TO
55
:maxdepth: 1
66

77
optimization/index_en.rst
8-
inference/inference_support_in_fluid.md

‎doc/fluid/howto/inference/build_and_install_lib_cn.rst‎

Lines changed: 96 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,96 @@
1+
安装与编译C++预测库
2+
===========================
3+
4+
直接下载安装
5+
-------------
6+
7+
====================== ========================================
8+
版本说明 C++预测库
9+
====================== ========================================
10+
cpu_avx_mkl `fluid.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxCp27cp27mu/.lastSuccessful/fluid.tgz>`_
11+
cpu_avx_openblas `fluid.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuAvxOpenblas/.lastSuccessful/fluid.tgz>`_
12+
cpu_noavx_openblas `fluid.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_CpuNoavxOpenblas/.lastSuccessful/fluid.tgz>`_
13+
cuda7.5_cudnn5_avx_mkl `fluid.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda75cudnn5cp27cp27mu/.lastSuccessful/fluid.tgz>`_
14+
cuda8.0_cudnn5_avx_mkl `fluid.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda80cudnn5cp27cp27mu/.lastSuccessful/fluid.tgz>`_
15+
cuda8.0_cudnn7_avx_mkl `fluid.tgz <https://guest:@paddleci.ngrok.io/repository/download/Manylinux1_Cuda8cudnn7cp27cp27mu/.lastSuccessful/fluid.tgz>`_
16+
====================== ========================================
17+
18+
从源码编译
19+
----------
20+
用户也可以从 PaddlePaddle 核心代码编译C++预测库，只需在编译时配制下面这些编译选项：
21+
22+
================= =========
23+
选项 值
24+
================= =========
25+
CMAKE_BUILD_TYPE Release
26+
FLUID_INSTALL_DIR 安装路径
27+
WITH_FLUID_ONLY ON（推荐）
28+
WITH_SWIG_PY OFF（推荐
29+
WITH_PYTHON OFF（推荐）
30+
WITH_GPU ON/OFF
31+
WITH_MKL ON/OFF
32+
================= =========
33+
34+
建议按照推荐值设置，以避免链接不必要的库。其它可选编译选项按需进行设定。
35+
36+
下面的代码片段从github拉取最新代码，配制编译选项（需要将PADDLE_ROOT替换为PaddlePaddle预测库的安装路径）：
37+
38+
.. code-block:: bash
39+
40+
pip install paddlepaddle-gpu
41+
PADDLE_ROOT=/path/of/capi
42+
git clone https://github.com/PaddlePaddle/Paddle.git
43+
cd Paddle
44+
mkdir build
45+
cd build
46+
cmake -DFLUID_INSTALL_DIR=$PADDLE_ROOT \
47+
-DCMAKE_BUILD_TYPE=Release \
48+
-DWITH_FLUID_ONLY=ON \
49+
-DWITH_SWIG_PY=OFF \
50+
-DWITH_PYTHON=OFF \
51+
-DWITH_MKL=OFF \
52+
-DWITH_GPU=OFF \
53+
..
54+
make
55+
make inference_lib_dist
56+
57+
成功编译后，使用C++预测库所需的依赖（包括：（1）编译出的PaddlePaddle预测库和头文件；（2）第三方链接库和头文件；（3）版本信息与编译选项信息）
58+
均会存放于PADDLE_ROOT目录中。目录结构如下：
59+
60+
.. code-block:: text
61+
62+
PaddleRoot/
63+
├── CMakeCache.txt
64+
├── paddle
65+
│   └── fluid
66+
│   ├── framework
67+
│   ├── inference
68+
│   ├── memory
69+
│   ├── platform
70+
│   ├── pybind
71+
│   └── string
72+
├── third_party
73+
│   ├── boost
74+
│   │   └── boost
75+
│   ├── eigen3
76+
│   │   ├── Eigen
77+
│   │   └── unsupported
78+
│   └── install
79+
│   ├── gflags
80+
│   ├── glog
81+
│   ├── mklml
82+
│   ├── protobuf
83+
│   ├── snappy
84+
│   ├── snappystream
85+
│   └── zlib
86+
└── version.txt
87+
88+
version.txt 中记录了该预测库的版本信息，包括Git Commit ID、使用OpenBlas或MKL数学库、CUDA/CUDNN版本号，如：
89+
90+
.. code-block:: text
91+
92+
GIT COMMIT ID: c95cd4742f02bb009e651a00b07b21c979637dc8
93+
WITH_MKL: ON
94+
WITH_GPU: ON
95+
CUDA version: 8.0
96+
CUDNN version: v5

‎doc/fluid/howto/inference/index_cn.rst‎

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,8 @@
1+
预测库
2+
------------
3+
4+
.. toctree::
5+
:maxdepth: 1
6+
7+
build_and_install_lib_cn.rst
8+
inference_support_in_fluid_cn.md

‎doc/fluid/howto/inference/inference_support_in_fluid.md‎ renamed to ‎doc/fluid/howto/inference/inference_support_in_fluid_cn.md‎

Lines changed: 1 addition & 58 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,8 @@
1-
# Fluid Inference使用指南
1+
# 使用指南
22

33
## 目录：
44

55
- Python Inference API
6-
- 编译Fluid Inference库
76
- Inference C++ API
87
- Inference实例
98
- Inference计算优化
@@ -55,62 +54,6 @@
5554
return [program, feed_target_names, fetch_targets]
5655
```
5756

58-
59-
## 编译Fluid Inference库
60-
61-
- **不需要额外的CMake选项**
62-
- 1、 配置CMake命令，更多配置请参考[源码编译PaddlePaddle](http://www.paddlepaddle.org/docs/develop/documentation/zh/build_and_install/build_from_source_cn.html)
63-
```bash
64-
$ git clone https://github.com/PaddlePaddle/Paddle.git
65-
$ cd Paddle
66-
$ mkdir build
67-
$ cd build
68-
$ cmake -DCMAKE_INSTALL_PREFIX=your/path/to/paddle_inference_lib \
69-
-DCMAKE_BUILD_TYPE=Release \
70-
-DWITH_PYTHON=ON \
71-
-DWITH_MKL=OFF \
72-
-DWITH_GPU=OFF \
73-
..
74-
```
75-
76-
- 2、 编译PaddlePaddle
77-
```bash
78-
$ make
79-
```
80-
81-
- 3、 部署。执行如下命令将PaddlePaddle Fluid Inference库部署到`your/path/to/paddle_inference_lib`目录。
82-
```bash
83-
$ make inference_lib_dist
84-
```
85-
86-
- 目录结构
87-
88-
```bash
89-
$ cd your/path/to/paddle_inference_lib
90-
$ tree
91-
.
92-
|-- paddle
93-
| `-- fluid
94-
| |-- framework
95-
| |-- inference
96-
| | |-- io.h
97-
| | `-- libpaddle_fluid.so
98-
| |-- memory
99-
| |-- platform
100-
| `-- string
101-
|-- third_party
102-
| |-- eigen3
103-
| `-- install
104-
| |-- gflags
105-
| |-- glog
106-
| `-- protobuf
107-
`-- ...
108-
```
109-
110-
假设`PADDLE_ROOT=your/path/to/paddle_inference_lib`。
111-
112-
113-
11457
## 链接Fluid Inference库
11558
- 示例项目([链接](https://github.com/luotao1/fluid_inference_example.git))
11659

0 commit comments

Comments
 (0)