Ubuntu18.04从零配置tensorflow2.1 GPU计算环境

1. 更换Ubuntu 源
参考博文：https://www.cnblogs.com/gentlemanwuyu/p/11621636.html

阿里云源：
deb http://mirrors.aliyun.com/ubuntu/ bionic main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ bionic-security main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ bionic-updates main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ bionic-proposed main restricted universe multiverse
deb http://mirrors.aliyun.com/ubuntu/ bionic-backports main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ bionic main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ bionic-security main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ bionic-updates main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ bionic-proposed main restricted universe multiverse
deb-src http://mirrors.aliyun.com/ubuntu/ bionic-backports main restricted universe multiverse

清华源：
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-updates main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-backports main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-security main restricted universe multiverse
deb https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-proposed main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-updates main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-backports main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-security main restricted universe multiverse
deb-src https://mirrors.tuna.tsinghua.edu.cn/ubuntu/ bionic-proposed main restricted universe multiverse

中科大源：
deb https://mirrors.ustc.edu.cn/ubuntu/ bionic main restricted universe multiverse
deb https://mirrors.ustc.edu.cn/ubuntu/ bionic-updates main restricted universe multiverse
deb https://mirrors.ustc.edu.cn/ubuntu/ bionic-backports main restricted universe multiverse
deb https://mirrors.ustc.edu.cn/ubuntu/ bionic-security main restricted universe multiverse
deb https://mirrors.ustc.edu.cn/ubuntu/ bionic-proposed main restricted universe multiverse
deb-src https://mirrors.ustc.edu.cn/ubuntu/ bionic main restricted universe multiverse
deb-src https://mirrors.ustc.edu.cn/ubuntu/ bionic-updates main restricted universe multiverse
deb-src https://mirrors.ustc.edu.cn/ubuntu/ bionic-backports main restricted universe multiverse
deb-src https://mirrors.ustc.edu.cn/ubuntu/ bionic-security main restricted universe multiverse
deb-src https://mirrors.ustc.edu.cn/ubuntu/ bionic-proposed main restricted universe multiverse

#打开源地址文件
sudo gedit /etc/apt/sources.list
#添加上述国内地址，保存并更新
sudo apt-get update

2. 安装搜狗输入法
参考博文：https://www.cnblogs.com/lfri/p/10769144.html

#安装fctix插件，装完进入设置
sudo apt install fcitx
#修改sogou输入法安装文件权限后，安装
sudo dpkg -i sogoupinyin_2.3.2.07_amd64.deb
#安装出错，启动修复安装
sudo apt-get -f install
#再次安装sogou输入法，完成后设置，并重启
sudo dpkg -i sogoupinyin_2.3.2.07_amd64.deb

安装完fctix后进入设置——语言——选择fctix

安装完sogou后在系统桌面右上角有个键盘，右击进行配置，添加安装的sogou输入法并上调至第一顺位，重启即可。

3. 检查并安装软件环境
特别注意：tensorflow的版本和各种软件版本对应要求非常严格，比如tensorflow2.1，对应cuda10.1、cudnn7.6、python3.6-3.7、gcc ，且pip3的版本为20.0以上才能安装tensorflow2.1。
参考博文：https://blog.csdn.net/u014798590/article/details/84977444?utm_medium=distribute.pc_relevant.none-task-blog-title-1&spm=1001.2101.3001.4242

3.1 python3

#查看系统自带python3版本
python3 --version
#若不符合版本要求，安装其他版本
sudo apt-get install puthon3.7
#切换新老版本python3
sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.6（老版本python3） 1
sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.7（新版本python3） 2
sudo update-alternatives --config python3
#输入新版本对应的序号2，并回车，再次看看python3版本
#将python指向python3
sudo ln -s /usr/bin/python3 /usr/bin/python

3.2 pip3

#安装pip3，对应python3
sudo apt install python3-dev python3-pip python3-venv
#升级pip3
 sudo pip3 install --upgrade pip -i https://pypi.mirrors.ustc.edu.cn/simple/

3.3 gcc/g++
参考博文：https://blog.csdn.net/weixin_40543841/article/details/88350648

#查看gcc/g++版本
gcc -v
g++ -v
#安装新版本
sudo apt install gcc-7
sudo apt install g++-7
#配置新版本
sudo rm gcc
sudo ln -s gcc-7 gcc
sudo rm g++
sudo ln -s g++7 g++

3.4 camke

sudo apt install cmake

3.5 其他
缺什么装什么

pip3 install scikit-build -i https://pypi.mirrors.ustc.edu.cn/simple/
pip3 install matplotlib -i https://pypi.mirrors.ustc.edu.cn/simple/
pip3 install numpy -i https://pypi.mirrors.ustc.edu.cn/simple/
pip3 install scipy -i https://pypi.mirrors.ustc.edu.cn/simple/

4. cuda 安装
参考博文：https://www.cnblogs.com/booturbo/p/11834661.html

#下载对应deb文件进行安装，最好不要单独装显卡驱动，容易循环登录，最好在装系统时勾选安装第三方驱动和软件
sudo dpkg -i cuda-repo-ubuntu1804-10-1-local-10.1.105-418.39_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-10-1-local-10.1.105-418.39/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda
#安装完后配置环境变量
sudo gedit ~/.bashrc
source ~/.bashrc
#验证是否安装成功
nvcc -V

环境变量配置

验证安装

5. cudnn安装

#下载压缩文件版本，解压，在其中的"cuda”目录中打开终端，拷贝文件至系统中
sudo cp cuda/include/cudnn.h /usr/local/cuda/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
#修改权限
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
#在/usr/local/cuda-10.1/targets/x86_64-linux/lib目录下创建软连接
sudo ln -sf libcudnn.so.7.6.5 libcudnn.so.7
#验证安装
cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2

验证cudnn安装

若报软连接的错误，则按照上述创建软连接

6. tensorflow安装
6.1 简易安装

#修改pip下载等待时间为1000,-i表示用指定源，速度快，出错就多试几次。没有对应版本则检查所有的软件环境版本号
pip3 --default-timeout=1000 install -U tensorflow-gpu==2.1.0 -i https://pypi.mirrors.ustc.edu.cn/simple/
#验证安装
python3
import tensorflow
#tensorboard一般无需单独安装，若提示没有此命令，则单独安装
pip3 install tb-nightly -i https://pypi.mirrors.ustc.edu.cn/simple/
#一些情况下执行tensorboard安装命令后，会出现重复错误，表明系统中有多个tensorboard版本冲突。
#此时需要进到安装文件夹/home/wz/.local/lib/python3.7/site-packages/ 删掉一个tensorboard*.dist文件夹
#tensorboard使用
tensorboard --logdir=/日志目录
#终端运行tensorboard后，浏览器打开地址 http://127.0.0.1:6006
#若tensorflow无法使用GPU运算，再次检查对应的软件环境版本号，必须一一对应

6.2 编译安装
（1）java安装

sudo apt-get install software-properties-common
sudo add-apt-repository ppa:webupd8team/java
sudo apt-get update
sudo apt-get install openjdk-8-jdk

（2）bazel编译工具安装

echo "deb [arch=amd64] http://storage.googleapis.com/bazel-apt stable jdk1.8" | sudo tee /etc/apt/sources.list.d/bazel.list
curl https://bazel.build/bazel-release.pub.gpg | sudo apt-key add -
sudo apt-get update && sudo apt-get install bazel
sudo apt-get upgrade bazel 
sudo apt-get upgrade bazel

（3）tensorflow源码编译
参考博文：https://blog.csdn.net/HelloZEX/article/details/80739130

7. opencv安装
7.1 源码编译
参考博文：https://blog.csdn.net/kevineeo/article/details/83242516

#下载后解压源码，进入目录，创建release文件夹并进入
mkdir release
cd release
#用cmake编译
cmake -D CMAKE_BUILD_TYPE=Release -D CMAKE_INSTALL_PREFIX=/usr/local ..
#根据自己机器CPU选择-j
sudo make -j16
sudo make install
#配置环境变量
sudo gedit /etc/ld.so.conf.d/opencv.conf
#更新环境变量
sudo ldconfig
#配置bash
sudo gedit /etc/bash.bashrc
#更新bash
source /etc/bash.bashrc
sudo updatedb
#验证安装
pkg-config --modversion opencv

配置opencv.conf

配置bash

验证安装

7.2 简易安装

pip3 install opencv-python -i https://pypi.mirrors.ustc.edu.cn/simple/