制作16.04 系统安装U盘（使用Rufus 3.4制作Ubuntu 18.04.1 U盘启动盘的方法）
https://ywnz.com/linuxjc/3978.html
1 安装16.04
https://blog.csdn.net/zhengchaooo/article/details/79500209
安装过程中如果出现CD-ROM读取不到 可以将U盘拔出再插上 然后再重复一次读取CD-ROM的操作（奇怪的现象）
2 16.04 安装的是python 是2.7 和 3.5
3 替换python版本 为python3.6
操作：
参考步骤https://www.cnblogs.com/gaowengang/p/7736672.html
http://www.cnblogs.com/yqpy/p/9116590.html
1、打开终端
首先创建安装目录，
sudo mkdir /usr/local/python3
2、然后下载安装包，解压，并且进入
wget --no-check-certificate https://www.python.org/ftp/python/3.6.5/Python-3.6.5.tgz
tar -xzvf Python-3.6.5.tgz
cd Python-3.6.5
3、接着编译安装
sudo ./configure --prefix=/usr/local/python3
sudo make
sudo make install
如果第一步编译的时候出现了报错，则试试
sudo apt-get install build-essential
4、创建python3的链接
sudo ln -s /usr/local/python3/bin/python3 /usr/bin/python3
问题1
ln: failed to create symbolic link '/usr/bin/python3': File exists
说明已经有链接链到 /usr/bin/python上了，删除即可，使用命令
rm -rf /usr/bin/python3
问题2
zipimport.ZipImportError: can't decompress data; zlib not available
(1)先安装系统相应的依赖库文件
sudo apt-get install zlibc zlib1g-dev
然后重新编译安装
https://zhoujianghai.iteye.com/blog/1521993
python3.6.3 安装:
tar -zxvf Python-3.6.3.tgz
cd Python-3.6.3
./configure
make
make altinstall
`make altinstall` , 报错:
zipimport.ZipImportError: can't decompress data; zlib not available
解决:
cd Modules/zlib
./configure
make install
cd -
make altinstall
https://blog.csdn.net/jaket5219999/article/details/78583604
问题3
sudo apt-get install python3-dev libffi-dev libssl-dev
wget https://www.python.org/ftp/python/3.6.0/Python-3.6.0.tgz
tar xvf Python-3.6.0.tgz
cd Python-3.6.0
./configure --enable-optimizations --prefix=/usr/local/python3
make -j8
sudo make altinstall
python3.6
直接执行这两个命令即可：
sudo update-alternatives --install /usr/bin/python python /usr/bin/python2 100
sudo update-alternatives --install /usr/bin/python python /usr/bin/python3 150
Ubuntu下pip3的安装、升级、卸载
sudo apt-get install python3-pip
sudo pip3 install --upgrade pip
sudo apt-get remove python3-pi
http://www.lxins.com/2017/08/03/ubuntu%E4%B8%8Bpip-is-configured-with-locations-that-require-tlsssl-however-the-ssl-module-in-python-is-not-available/
安装
pip3 install tensorflow-gpu==1.10 -i http://pypi.douban.com/simple/ --trusted-host pypi.douban.com --user
4 Linux 装好后需要相互拷贝文件 可以在windows 上新建共享文件夹 然后在ubuntu 上挂载
挂载windows 共享文件夹 mount -t cifs //192.168.1.111/Share ./windows -o username=***,password=*****,vers=2.0
5 CUDA 9.0
https://developer.nvidia.com/cuda-90-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=runfilelocal
https://blog.csdn.net/wanzhen4330/article/details/81699769
https://developer.download.nvidia.cn/compute/cuda/9.0/Prod/docs/sidebar/CUDA_Quick_Start_Guide.pdf
6 安装CUDnn
https://developer.nvidia.com/rdp/cudnn-download
https://medium.com/beesightsoft/install-tensorflow-gpu-pytorch-on-ubuntu-18-04-dell-t1700-3e66bcab07a6
# Install
sudo dpkg -i libcudnn7_7.4.2.24-1+cuda9.0_amd64.deb
sudo dpkg -i libcudnn7-dev_7.4.2.24-1+cuda9.0_amd64.deb
sudo dpkg -i libcudnn7-doc_7.4.2.24-1+cuda9.0_amd64.deb
7 pip安装whl文件
pip install some.whl
8 运行实例
https://blog.csdn.net/qq_29893385/article/details/82084108
问题
ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory
https://blog.csdn.net/qq_34374211/article/details/81018320
sudo vim ~/.bashrc
在末尾加入：
export PATH=/usr/local/cuda/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
export CUDA_HOME=/usr/local/cuda
source ~/.bashrc
发现此时import tensorflow还是报一样的错误，所以进行步骤二
步骤二
检查 /usr/local/cuda-9.0/lib64 下是否有 libcublas.so.9.0
如果有，终端输入:
sudo ldconfig /usr/local/cuda-9.0/lib64
