200字范文 > Ubuntu16.04 安装 CUDA CUDNN OpenCV 并用 Anaconda 配置 Tensorflow 和 Caffe 详细过程

Ubuntu16.04 安装 CUDA CUDNN OpenCV 并用 Anaconda 配置 Tensorflow 和 Caffe 详细过程

时间：2021-11-13 17:13:57

Ubuntu16.04 安装 CUDA、CUDNN、OpenCV 并用 Anaconda 配置 Tensorflow 和 Caffe 详细过程

1.配置环境

Ubuntu16.04 64-bit系统cuda_8.0.61_375.26_linux.run (链接:/s/1jHM96wY密码: cggg)cuda_8.0.61.2_linux.run (链接:/s/1qXASaEK密码: g2xr)cudnn-8.0-linux-x64-v6.0.tgz (链接:/s/1nvOfAdz密码: ma8t)opencv-3.2.0.zip（链接:/s/1slkHRaP密码: tfhi）ippicv_linux_1201.tgz (链接:/s/1b9sEOu密码: yuix)Anaconda3-5.0.1-Linux-x86_64.sh (链接:/s/1slJTIuL密码: 23z2)caffe-master.zip (链接:/s/1i44uCgX密码: m72x)

2.安装显卡驱动、CUDA 和 CUDNN

在Ubuntu系统中的 System Settings -> Software & Updates ->

Additional Drivers 安装相应的显卡和CPU驱动，重启后在 System Settings -> Details

中可以看到自己的显卡型号说明安装成功查看 CUDA 官方文档的 PRE-INSTALLATION ACTIONS 部分，一一验证是否符合安装条件在终端运行sudo sh cuda_8.0.61_375.26_linux --override，Ctrl + C 跳过协议部分，然后accept，下面只有安装驱动的时候选 no （上面已经安装过驱动），其他都是 yes运行sudo sh cuda_8.0.61.2_linux.run添加官方补丁添加环境变量，终端运行sudo gedit /etc/profile在文件末尾添加

export PATH=/usr/local/cuda-8.0/bin:$PATHexport LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64:$LD_LIBRARY_PATH

运行source /etc/profile检查驱动cat /proc/driver/nvidia/version检查CUDAnvcc -V解压 cudnn-8.0-v6.0 ，拷贝到 CUDA 安装路径

sudo cp cuda/include/cudnn.h /usr/local/cuda/includesudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*

3.编译安装 OpenCV3.2

安装相应的依赖

sudo apt-get install build-essentialsudo apt-get install cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-devsudo apt-get install python-dev python-numpy libtbb2 libtbb-dev libjpeg-dev libpng-dev libtiff-dev libjasper-dev libdc1394-22-dev # 处理图像所需的包sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libv4l-devsudo apt-get install libxvidcore-dev libx264-dev # 处理视频所需的包sudo apt-get install libatlas-base-dev gfortran # 优化opencv功能，此步显示找不到软件sudo apt-get install ffmpeg

解压 opencv3.2.0.zip

cd .../opencv-3.2.0mkdir buildcd buildcmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local ..

若出现 – ICV: Downloading ippicv_linux_1201.tgz… 错误，则复制

ippicv_linux_1201.tgz 到

opencv-3.2.0/3rdparty/ippicv/downloads/linux-808b791a6eac9ed78d32a7666804320e/

再继续执行 cmake编译并安装

make -j4 #四核运算 sudo make install

4.安装 Anaconda 和 Tensorflow

在终端运行sudo sh ./Anaconda3-5.0.1-Linux-x86_64.sh即可安装在终端运行如下命令即可安装最新版 Tensorflow

conda create -n tensorflow python=3.6source activate tensorflowpip install tensorflow-gpu

若要在终端使用 Tensorflow，则要先输入source activate tensorflow，退出时输入source deactivate若要在 Spyder 中使用 Tensorflow，则需要复制并替换路径

/home/senius/anaconda3/envs/tensorflow/lib/python3.6/site-packages

文件夹中的所有文件到 Anaconda 安装路径下的 lib/python3.6/site-packages 文件夹中若要在 PyCharm 中使用 Tensorflow，则需要在 Files -> Settings -> Project Interpreter 中添加 ~/anaconda3/envs/tensorflow/bin/python 即可

4.安装 Caffe 并配置 Python 接口

安装必要的依赖库

sudo apt-get install build-essentialsudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compilersudo apt-get install --no-install-recommends libboost-all-devsudo apt-get install libatlas-base-devsudo apt-get install python-devsudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

解压 caffe-master, 重命名 Makefile.config.sample 为 Makefile.config 并更改如下内容

a.使用 cudnn USE_CUDNN := 1b.使用的 opencv 版本是 3OPENCV_VERSION := 3c.使用 python 来编写 layerWITH_PYTHON_LAYER := 1d.重要的一项 :将# Whatever else you find you need goes here.下面的INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/includeLIBRARY_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这是因为ubuntu16.04的文件包含位置发生了变化，尤其是需要用到的hdf5的位置，所以需要更改这一路径e.删除 CUDA_ARCH := 的前两行,避免 CUDA 报错f. 修改 PYTHON_INCLUDE 路径# We need to be able to find Python.h and numpy/arrayobject.h.#PYTHON_INCLUDE := /usr/include/python2.7 \/usr/lib/python2.7/dist-packages/numpy/core/includeg. 修改 Anaconda 路径# Verify anaconda location, sometimes it's in root.ANACONDA_HOME := $(HOME)/anaconda3PYTHON_INCLUDE := $(ANACONDA_HOME)/include \$(ANACONDA_HOME)/include/python3.6m \$(ANACONDA_HOME)/lib/python3.6/site-packages/numpy/core/includeh. 使用 Python3# Uncomment to use Python 3 (default is Python 2)PYTHON_LIBRARIES := boost_python3 python3.6mPYTHON_INCLUDE := /usr/include/python3.6m \/usr/lib/python3.6/dist-packages/numpy/core/include \/home/senius/anaconda3/include/python3.6mi. 修改 PYTHON_LIB 路径# We need to be able to find libpythonX.X.so or .dylib.#PYTHON_LIB := /usr/libPYTHON_LIB := $(ANACONDA_HOME)/lib \$(ANACONDA_HOME)/pkgs/python-3.6.3-h0ef2715_3/lib以上内容一定要仔细查看自己相应的目录一一对应，若没有相关目录一定是少安装了某些依赖，这一步配置好了后面就不会有什么错误

修改 Makefile 文件把下面第一行代码改为第二行代码

LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_hl hdf5LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial

编辑 /usr/local/cuda/include/host_config.h ，将其中的第115行注释掉

//#error -- unsupported GNU version! gcc versions later than 5 are not supported!

编译 caffe

mkdir buildcd buildcmake ..sudo make allsudo make installsudo make runtest

所有的 RUN 都为 OK 的话就编译成功了

编译 python 接口

cd .. #到caffe-master目录sudo make pycaffe -j4

若出现 cannot find -lboost_python3 错误（参考/u012675539/article/details/51351553），首先检查是否有文件存在ls /usr/lib/x86_64-linux-gnu/libboost_python-py35.so，若存在则建立软链接

sudo ln -s libboost_python-py35.so libboost_python3.sosudo ln -s libboost_python-py35.a libboost_python3.asudo ln -s libboost_python-py35.so.1.58.0 libboost_python3.so.1.58.0然后 sudo gedit /etc/ld.so.conf添加 /usr/lib/x86_64-linux-gnusudo ldconfig

然后重新编译

sudo make clean sudo make pycaffe -j4

添加环境变量

sudo gedit ~/.bashrc最后一行添加export PYTHONPATH=/home/senius/Software/caffe-master/python:$PYTHONPATHsource ~/.bashrc

测试 python 接口

进入终端import caffe不报错说明安装成功