ubuntu16.04安装caffe

1. 安装opencv

opencv只是用于读取图片的（参考网上），所以不需要安装cuda版的，不然后有很多问题。

(1) 从官网(http://opencv.org/downloads.html)下载Opencv,并将其解压到你要安装的位置，假设解压到了/home/opencv。

• unzip opencv-3.2.0.zip
• 将pencv-3.2.0改为opencv

(2) 安装前准备，创建编译文件夹

cd ~/opencv
mkdir build
cd build

(3) 配置

sudo apt install cmake
sudo cmake -D WITH_CUDA=OFF -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local …
WITH_CUDA=OFF表示不安装gpu版

(4) 编译

sudo make -j8 (-j8表示并行计算，根据自己电脑的配置进行设置，配置比较低的电脑可以将数字改小或不使用，直接输make)

(5) 安装
sudo make install # 安装

(6) 看opencv的版本
pkg-config --modversion opencv

可能错误：

1. 下载ippicv_linux_20151201.tgz的时候超时，请参考：ICV: Downloading ippicv_linux_20151201.tgz 超时
2. "libopencv_core.so.3.2: cannot open shared object file: No such file or direct，请参考
   OpenCV runtime error: "libopencv_core.so.3.2: cannot open shared object file: No such file or direct
3. 运行编译的文件时出现：
   test.cpp:(.text+0x46)：对‘cv::imread(cv::String const&, int)’未定义的引用
   test.cpp:(.text+0xbe)：对‘cv::imshow(cv::String const&, cv::_InputArray const&)’未定义的引用
   test.cpp:(.text+0xe6)：对‘cv::waitKey(int)’未定义的引用
   /tmp/ccSDVP3R.o：在函数‘cv::String::String(char const*)’中：
   test.cpp:(.text._ZN2cv6StringC2EPKc[_ZN2cv6StringC5EPKc]+0x4d)：对‘cv::String::allocate(unsigned long)’未定义的引用
   /tmp/ccSDVP3R.o：在函数‘cv::String::~String()’中：
   test.cpp:(.text._ZN2cv6StringD2Ev[_ZN2cv6StringD5Ev]+0x14)：对‘cv::String::deallocate()’未定义的引用
   /tmp/ccSDVP3R.o：在函数‘cv::String::operator=(cv::String const&)’中：
   test.cpp:(.text.ZN2cv6StringaSERKS0[ZN2cv6StringaSERKS0]+0x28)：对‘cv::String::deallocate()’未定义的引用
   /tmp/ccSDVP3R.o：在函数‘cv::Mat::~Mat()’中：
   test.cpp:(.text._ZN2cv3MatD2Ev[_ZN2cv3MatD5Ev]+0x39)：对‘cv::fastFree(void*)’未定义的引用
   /tmp/ccSDVP3R.o：在函数‘cv::Mat::release()’中：
   test.cpp:(.text._ZN2cv3Mat7releaseEv[_ZN2cv3Mat7releaseEv]+0x4b)：对‘cv::Mat::deallocate()’未定义的引用
   collect2: error: ld returned 1 exit status

编译的时候用下面的命令：

g++  opencvtest.cpp -o opencvtest  `pkg-config  --cflags --libs opencv` 

---

可以正常显示图片则表面安装成功了

2. 测试opencv

// 文件名保存为opencvtest.cpp，采用上面的方式进行编译，读取一张图片，并进行显示，能正常显示图片就可以
#include <opencv2/opencv.hpp>

using namespace cv;

using namespace std;

int main()
{

Mat img = imread("dazu.jpg");
if(img.empty())
{
cout<<"error";
return -1;
}
imshow("mypic",img);
waitKey();
return 0;
}

3. 安装caffe

转载自:https://blog.csdn.net/sinat_23853639/article/details/80990967 ，尊重原创，请看原文。

1. 安装相关依赖项

1 sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
2 sudo apt-get install --no-install-recommends libboost-all-dev
3 sudo apt-get install libopenblas-dev liblapack-dev libatlas-base-dev
4 sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

2. 下载caffe
   git clone https://github.com/BVLC/caffe.git
3. 修改Makefile.config

将 Makefile.config.example 文件复制一份并更名为 Makefile.config ，也可以在 caffe 目录下直接调用以下命令完成复制操作 ：

sudo cp Makefile.config.example Makefile.config

复制一份的原因是编译 caffe 时需要的是 Makefile.config 文件，而Makefile.config.example 只是caffe 给出的配置文件例子，不能用来编译 caffe。

然后修改 Makefile.config 文件，在 caffe 目录下打开该文件：

sudo gedit Makefile.config

修改 Makefile.config 文件内容：

1.应用 cudnn

将
#USE_CUDNN := 1
修改成： 
USE_CUDNN := 1

2.应用 opencv 版本

将
#OPENCV_VERSION := 3 
修改为： 
OPENCV_VERSION := 3

3.使用 python 接口

将
#WITH_PYTHON_LAYER := 1 
修改为 
WITH_PYTHON_LAYER := 1

4.修改 python 路径

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib 
修改为： 
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu /usr/lib/x86_64-linux-gnu/hdf5/serial     

5.然后修改 caffe 目录下的 Makefile 文件：

将：
NVCCFLAGS +=-ccbin=$(CXX) -Xcompiler-fPIC $(COMMON_FLAGS)
替换为：
NVCCFLAGS += -D_FORCE_INLINES -ccbin=$(CXX) -Xcompiler -fPIC $(COMMON_FLAGS)
 
将：
LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_hl hdf5
改为：
LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial

6.然后修改 /usr/local/cuda/include/host_config.h 文件 :（cuda9.0没有）

将
#error-- unsupported GNU version! gcc versions later than 4.9 are not supported!
改为
//#error-- unsupported GNU version! gcc versions later than 4.9 are not supported!
OK ，可以开始编译了，在 caffe 目录下执行 ：

make all -j8

这是如果之前的配置或安装出错，那么编译就会出现各种各样的问题，所以前面的步骤一定要细心。

---

4. 遇到的问题

1. libpng16.so.16：对‘inflateValidate@ZLIB_1.2.9’未定义的引用
   
   　1）GitHub上的解决方案，链接：https://github.com/BVLC/caffe/issues/6139
   
   在Makefile.config文件中添加如下一行:

LINKFLAGS := -Wl,-rpath,$(HOME)/anaconda3/lib

• 编译通过
  

2）fatal error: numpy/arrayobject.h: 没有那个文件或目录
解决：
sudo apt-get install python-numpy

3）python中import caffe，报错：ModuleNotFoundError: No module named ‘caffe’
解决办法：

vi ~/.bashrc 添加 
export PYTHONPATH=/xx/xx/caffe/python:$PYTHONPATH  然后source ~/.bashrc

4）ImportError: dynamic module does not define module export function (PyInit__caffe)

• caffe支持python2.7，所以在终端中运行python2.7
  

5）ImportError: /lib/x86_64-linux-gnu/libz.so.1: version `ZLIB_1.2.9’ not found (required by /home/deepblue/anaconda3/lib/libpng16.so.16)

参考自 https://www.cnblogs.com/zxj9487/p/11393546.html

zlib的版本不对，重装以下zlib

1 git clone --branch v1.2.9 https://github.com/madler/zlib
2 cd path/to/zlib    #这里黄色标注的换成自己的路径即可
3 ./configure
4 make
5 make install # you may add 'sudo'

6）ImportError: No module named skimage.io

sudo apt-get install python-skimage

7）ImportError: No module named google.protobuf.internal
protobuf安装的问题

• 首先通过which python 确定用的默认的python2.7还是conda中的python
• caffe好像只支持python2.7，所以需要将默认的python设置成usr/bin/python2.7

pip install protobuf

8） libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9’
collect2: error: ld returned 1 exit status

参考：https://blog.csdn.net/zhangruijerry/article/details/79724720,安装了
zlib1g_1.2.11就OK了。

具体问题：

/usr/local/lib/libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9'
collect2: error: ld returned 1 exit status
Makefile:619: recipe for target '.build_release/tools/upgrade_net_proto_binary.bin' failed
make: *** [.build_release/tools/upgrade_net_proto_binary.bin] Error 1
make: *** Waiting for unfinished jobs....
/usr/local/lib/libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9'
collect2: error: ld returned 1 exit status
Makefile:619: recipe for target '.build_release/tools/convert_annoset.bin' failed
make: *** [.build_release/tools/convert_annoset.bin] Error 1
/usr/local/lib/libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9'
collect2: error: ld returned 1 exit status
Makefile:619: recipe for target '.build_release/tools/create_label_map.bin' failed
make: *** [.build_release/tools/create_label_map.bin] Error 1
/usr/local/lib/libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9'
collect2: error: ld returned 1 exit status
Makefile:619: recipe for target '.build_release/tools/caffe.bin' failed
make: *** [.build_release/tools/caffe.bin] Error 1
/usr/local/lib/libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9'
collect2: error: ld returned 1 exit status
Makefile:619: recipe for target '.build_release/tools/get_image_size.bin' failed
make: *** [.build_release/tools/get_image_size.bin] Error 1
/usr/local/lib/libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9'
collect2: error: ld returned 1 exit status
Makefile:619: recipe for target '.build_release/tools/upgrade_solver_proto_text.bin' failed
make: *** [.build_release/tools/upgrade_solver_proto_text.bin] Error 1
/usr/local/lib/libopencv_imgcodecs.so: undefined reference to `inflateValidate@ZLIB_1.2.9'
collect2: error: ld returned 1 exit status
Makefile:619: recipe for target '.build_release/tools/convert_imageset.bin' failed
make: *** [.build_release/tools/convert_imageset.bin] Error 1

5. 测试

• import caffe
  

reference

1. Ubuntu 16.04下安装Caffe（GPU版本 GTX970