记caffe的安装和open_nsfw项目的部署
open_nsfw_web 安装文档
caffe的python环境安装前准备
所需文件列表
~ require_softwear/
Anaconda2-4.3.1-Linux-x86_64.sh
boost_1_63_0.tar.gz
caffe-master.zip
gflags-2.1.2.tar.gz
glog-0.3.4.tar.gz
lmdb-LMDB_0.9.18.tar.gz
open_nsfw_web.zip
opencv.tar.gz
protobuf-cpp-3.1.0.tar.gz
Python-2.7.8.tar.xz
setuptools-1.4.2.tar.gz
查看系统版本
[root@host132 ~]# cat /proc/version
Linux version 2.6.32-504.12.2.el6.x86_64 ([email protected]) (gcc version 4.4.7 20120313 (Red Hat 4.4.7-11) (GCC) ) #1 SMP Wed Mar 11 22:03:14 UTC 2015
版本要求
- OpenCV >= 2.4 including 3.0
- Python >= 2.7
- Boost >= 1.55
安装前准备
将require_softwear目录下的所有文件放到/opt目录下面
caffe的python环境安装
安装Python2.7
1 进入/opt目录，解压Python-2.7.8.tar.xz
[root@host132 ~]# tar -xvf Python-2.7.8.tar.xz
如果出现下面信息：
[root@host132 opt]# tar -xvf Python-2.7.8.tar.xz
tar (child): xz：无法 exec: 没有那个文件或目录
tar (child): Error is not recoverable: exiting now
tar: Child returned status 2
tar: Error is not recoverable: exiting now
说明需要安装 xz，安装 xz ：
[root@host132 opt]# yum install xz
然后执行以下命令安装Python-2.7.8
[root@host132 opt]# tar -xvf Python-2.7.8.tar.xz
[root@host132 opt]# cd Python-2.7.8
[root@host132 Python-2.7.8]# ./configure --prefix=/usr/local/Python2.7 --enable-shared cFLAGS=-fPIC
[root@host132 Python-2.7.8]# make
[root@host132 Python-2.7.8]# make install
安装完成后，查看当前的python
[root@host132 Python-2.7.8]# python
Python 2.6.6 (r266:84292, Jan 22 2014, 09:42:36)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
\>>>
发现指现的依然是旧版本(Python2.6.6版本)，建立新版本的软链接：
[root@host132 ~]# ln -s /usr/local/Python2.7/bin/python2.7 /usr/bin/python
ln: 创建符号链接 "/usr/bin/python": 文件已存在
发现链接已存在，当然是旧版本的啦。这时mv掉/usr/bin/python:
[root@host132 ~]# mv /usr/bin/python /tmp
[root@host132 ~]# ln -s /usr/local/Python2.7/bin/python2.7 /usr/bin/python
此时再查看一下python，发现：
[root@host132 ~]# python
python: error while loading shared libraries: libpython2.7.so.1.0: cannot open shared object file: No such file or directory
这是因为编译Python时加了--enable-shared ，此时在/etc/ld.so.conf中加入/usr/local/Python2.7/lib
[root@host132 ~]# vim /etc/ld.so.conf
include ld.so.conf.d/*.conf
/usr/local/Python2.7/lib
然后执行以下命令：
[root@host132 ~]# /sbin/ldconfig
这时再查看python，发现是新版本了：
[root@host132 ~]# python
Python 2.7.8 (default, Apr 18 2017, 17:09:58)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-11)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
\>>>
修复yum
这个时候运行yum，会得到如下信息：
[root@host132 ~]# yum
There was a problem importing one of the Python modules
required to run yum. The error leading to this problem was:
No module named yum
Please install a package which provides this module, or
verify that the module is installed correctly.
It's possible that the above module doesn't match the
current version of Python, which is:
2.7.8 (default, Apr 18 2017, 17:09:58)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-11)]
If you cannot solve this problem yourself, please go to
the yum faq at:
http://yum.baseurl.org/wiki/Faq
这是因为yum与python2.7不兼容的缘故，此时需设置yum继续使用python2.6：
[root@host132 ~]# which yum
/usr/bin/yum
[root@host132 ~]# vim /usr/bin/yum
把第一行的
#!/usr/bin/python
修改为
#!/usr/bin/python2.6
安装pip
[root@host132 python]# curl https://bootstrap.pypa.io/get-pip.py \| python -
pip安装后路径为/usr/local/Python2.7/bin/pip
把
export PATH="/usr/local/Python2.7/bin:$PATH"
添加到/etc/profile文件的最后一行，然后使环境变量生效：
[root@host132 python]# source /etc/profile
安装Anaconda
[root@host132 ~]# cd /opt
[root@host132 ~]# bash Anaconda2-4.3.1-Linux-x86_64.sh
一路按回车默认安装，默认安装路径为/root/anaconda2
在/etc/ld.so.conf中加入/root/anaconda2/lib
[root@host132 ~]# vim /etc/ld.so.conf
include ld.so.conf.d/*.conf
/usr/local/Python2.7/lib
/root/anaconda2/lib
然后执行以下命令：
[root@host132 ~]# /sbin/ldconfig
安装ATLAS
[root@host132 opt]# yum install atlas-devel
安装glog
[root@host132 opt]# tar -xvf glog-0.3.4.tar.gz
[root@host132 opt]# cd glog-0.3.4
[root@host132 glog-0.3.4]# ./configure
[root@host132 glog-0.3.4]# make && make install
安装cmake
[root@host132 build]# yum install cmake
安装gflags
[root@host132 opt]# tar -xvf gflags-2.1.2.tar.gz
[root@host132 opt]# cd gflags-2.1.2
[root@host132 gflags-2.1.2]# mkdir build && cd build
[root@host132 build]# export CXXFLAGS="-fPIC" && cmake .. && make VERBOSE=1
[root@host132 build]# make && make install
安装lmdb
[root@host132 opt]# tar -xvf lmdb-LMDB_0.9.18.tar.gz
[root@host132 opt]# cd lmdb-LMDB_0.9.18/libraries/liblmdb
[root@host132 liblmdb]# make && make install
安装opencv2
[root@host132 opt]# tar -xvf opencv.tar.gz
[root@host132 opt]# cd opencv
[root@host132 opencv]# mkdir release && cd release
[root@host132 release]# cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local ..
[root@host132 release]# make
[root@host132 release]# make install
安装boost
[root@host132 opt]# tar -xvf boost_1_63_0.tar.gz
[root@host132 opt]# cd boost_1_63_0
[root@host132 boost_1_63_0]# ./bootstrap.sh --prefix=/usr/local
[root@host132 boost_1_63_0]# ./b2
[root@host132 boost_1_63_0]# ./b2 install
安装protobuf
[root@host132 opt]# tar -xvf protobuf-cpp-3.1.0.tar.gz
[root@host132 opt]# cd protobuf-3.1.0/
[root@host132 protobuf-3.1.0]# yum install autoconf automake libtool curl make g++ unzip
[root@host132 protobuf-3.1.0]# ./autogen.sh
[root@host132 protobuf-3.1.0]# ./configure
[root@host132 protobuf-3.1.0]# make
[root@host132 protobuf-3.1.0]# make check
[root@host132 protobuf-3.1.0]# make install
[root@host132 protobuf-3.1.0]# ldconfig
安装hdf5
[root@host132 caffe-master]# yum install hdf5-devel
安装epel-release
[root@host132 caffe-master]# yum install epel-release
安装leveldb-devel
[root@host132 caffe-master]# yum install leveldb-devel
安装snappy-devel
[root@host132 caffe-master]# yum install snappy-devel
安装caffe
[root@host132 opt]# unzip caffe-master.zip
如果没有unzip，先安装unzip：
[root@host132 opt]# yum install unzip
然后再
[root@host132 opt]# unzip caffe-master.zip
[root@host132 opt]# cd caffe-master
[root@host132 caffe-master]# cp Makefile.config.example Makefile.config
修改Makefile.config配置文件
[root@host132 caffe-master]# vim Makefile.config
把
# CPU_ONLY := 1
注释去掉，修改后为
CPU_ONLY := 1
把
# BLAS_INCLUDE := /path/to/your/blas
# BLAS_LIB := /path/to/your/blas
注释去掉，并修改为
BLAS_INCLUDE := /usr/include
BLAS_LIB := /usr/lib64/atlas
把
# ANACONDA_HOME := $(HOME)/anaconda
# PYTHON_INCLUDE := $(ANACONDA_HOME)/include \
# $(ANACONDA_HOME)/include/python2.7 \
# $(ANACONDA_HOME)/lib/python2.7/site-packages/numpy/core/include
注释去掉，并把anaconda改为anaconda2,修改后为
ANACONDA_HOME := $(HOME)/anaconda2
PYTHON_INCLUDE := $(ANACONDA_HOME)/include \
$(ANACONDA_HOME)/include/python2.7 \
$(ANACONDA_HOME)/lib/python2.7/site-packages/numpy/core/include
开始编译：
[root@host132 caffe-master]# make all
[root@host132 caffe-master]# make test
[root@host132 caffe-master]# make runtest
此时提示：
[root@host132 caffe-master]# make runtest
.build_release/tools/caffe
.build_release/tools/caffe: error while loading shared libraries: libglog.so.0: cannot open shared object file: No such file or directory
make: *** [runtest] 错误 127
找到libglog.so.0文件的位置
[root@host132 caffe-master]# find / -name libglog.so.0
/opt/glog-0.3.4/.libs/libglog.so.0
/usr/local/lib/libglog.so.0
把
/usr/local/lib
加入到/etc/ld.so.conf文件中
文件/etc/ld.so.conf修改前为：
include ld.so.conf.d/*.conf
/usr/local/Python2.7/lib
修改后为：
include ld.so.conf.d/*.conf
/usr/local/Python2.7/lib
/usr/local/lib
然后执行以下命令：
[root@host132 caffe-master]# /sbin/ldconfig
继续执行下面命令编译caffe
[root@host132 caffe-master]# make runtest
安装caffe的pycaffe接口模块
[root@host132 caffe-master]# cd python
[root@host132 python]# for req in $(cat requirements.txt); do pip install $req; done
如果出现以下信息
...
...
/usr/bin/ld: cannot find -lpython2.7
collect2: ld 返回 1
error: command 'g++' failed with exit status 1
...
...
则执行
[root@host132 python]# ln -s /usr/local/Python2.7/lib/python2.7/config/libpython2.7.a /usr/local/lib/
[root@host132 python]# for req in $(cat requirements.txt); do pip install $req; done
[root@host132 caffe-master]# make pycaffe
完成后测试caffe模块
[root@host132 caffe-master]# python
Python 2.7.8 (default, Apr 18 2017, 17:09:58)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-11)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
\>>> import caffe
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ImportError: No module named caffe
\>>>
提示没有caffe模块，是因为没有把caffe模块加入到环境变量。
把caffe模块的路径添加到当前用用的环境变量~/.bash_profile文件里面
把
export PYTHONPATH=/opt/caffe-master/python:$PYTHONPATH
添加到~/.bash_profile文件的最后一行。
使环境变量生效
[root@host132 caffe-master]# source ~/.bash_profile
再次测试caffe的python模块是否已安装成功
[root@host132 python]# python
Python 2.7.8 (default, Apr 18 2017, 17:09:58)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-11)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
\>>> import caffe
\>>>
信息提示已安装成功。
至此，caffe的python环境安装工作已结束。
open_nsfwweb程序的安装
[root@host132 opt]# cd /opt/
[root@host132 opt]# unzip open_nsfw_web.zip
[root@host132 open_nsfw_web]# pip install -r requirements.txt
启动web服务
[root@host132 open_nsfw_web]# python app.py &
测试web服务
[root@host132 ~]# curl http://127.0.0.1:5000/classify?imageurl=test_data/1.jpg
返回
{
"imageurl": "test_data/1.jpg",
"result": true,
"score": "0.788449168205",
"times": "0.368"
}
说明web服务已启动成功。
web服务接口说明
http://xx.xx.xx.xx:5000?classify?imageurl=图片路径
正确时返回格式为：
{
"imageurl": "test_data/1.jpg",
"result": true,
"score": "0.788449168205",
"times": "0.363"
}
错误时返回格式为：
{
"err": "check image path and only support png,bmp,jpg,jpe,jpeg,gif format.",
"imageurl": "test_data/21.jpg",
"result": false
}
2.
http://xx.xx.xx.xx:5000