注意：CUDA自带显卡驱动，不用另装
1.进入tensorflow安装界面https://www.tensorflow.org/install/install_linux查看要求(需要v*n,不是必须）
2.安装CUDA
进入cuda documentation http://docs.nvidia.com/cuda/cuda-installation-guide-linux/#axzz4VZnqTJ2A
点击右上角older https://developer.nvidia.com/cuda-toolkit-archive
选择想要安装的版本，设置系统（这里是9.0的安装命令,官方有错）
wget https://developer.nvidia.com/compute/cuda/9.0/Prod/local_installers/cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64-deb
sudo dpkg -i cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64-deb
sudo apt-key add /var/cuda-repo-9-0-local/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda
3.安装CUDNN
在https://developer.nvidia.com/rdp/cudnn-download下载CUDNN（需要NVIDIA账号）
tar -xzvf cudnn-9.0-linux-x64-v7.solitairetheme8
sudo cp cuda/include/cudnn.h /usr/local/cuda/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h
/usr/local/cuda/lib64/libcudnn*
4.设置环境变量
在~/.bashrc 的最后添加
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
5.安装tensorflow-gpu
sudo pip install tensorflow-gpu