一、獨顯
二、下載顯卡驅動
https://www.nvidia.cn/Download/index.aspx?lang=cn
三、安裝顯卡驅動
1、卸載原有的驅動
sudo apt-get remove --purge nvidia*
2、禁用nouveau
sudo gedit /etc/modprobe.d/blacklist.conf
在文本最后添加以下內容
blacklist nouveau
options nouveau modeset=0
3、保存退出，執行命令使其生效
sudo update-initramfs -u
4、重啟電腦
5、執行命令沒有任何輸出則說明禁用成功
lsmod | grep nouveau
6、將安裝包放到Home里面
7、停掉linux圖形界面的服務
sudo service lightdm stop
8、按ctrl+alt+f1切換到純控制台界面
9、進入Home
cd ~
10、給驅動賦可執行權限
sudo chmod a+x NVIDIA-Linux-xxx.run
11、安裝
sudo ./NVIDIA-Linux-xxx.run -no-opengl-files
(提示安裝基本上都是accept, yes, 當提示你nvidia-xconfig時，如果有雙顯卡就選擇不安裝，如果單顯卡就選擇安裝)
12、檢驗是否安裝成功
nvidia-smi
13、啟用圖形界面
sudo service lightdm start
14、按ctrl+alt+f7切換回圖形界面
三、安裝cuda
1、下載cuda
https://developer.nvidia.com/cuda-downloads
2、查看nouveau是否禁用成功
lsmod | grep nouveau
3、給文件賦予可執行權限
sudo chmod a+x cuda_xxxx.run
4、安裝cuda
sudo ./cuda_xxx.run --no-opengl-libs
安裝步驟
Do you accept the previously read EULA?
accept/decline/quit: accept
Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 361.62?
(y)es/(n)o/(q)uit: n
Install the CUDA 8.0 Toolkit?
(y)es/(n)o/(q)uit: y
Enter Toolkit Location
[ default is /usr/local/cuda-8.0 ]:(直接回車)
Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: y
Install the CUDA 8.0 Samples?
(y)es/(n)o/(q)uit: y
Enter CUDA Samples Location
[ default is /home/tang]:(直接回車)5、配置環境變量
export PATH=$PATH:/usr/local/cuda_xxx/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda_xxx/lib64
6、使配置生效
source ~/.bashrc
7、驗證CUDA Toolkit
nvcc -V
四、配置cudnn
1、下載cudnn
https://developer.nvidia.com/rdp/form/cudnn-download-survey
2、解壓cudnn
cp cudnn_xxx-linux-x64.solitairetheme8 cudnn_xxx-linux-x64.tgz
tar -xvf cudnn_xxx-linux-x64.tgz
3、將cudnn復制到對應的文件夾內
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*
五、卸載顯卡驅動
sudo apt-get remove --purge nvidia*
六、卸載cuda
1、執行卸載腳本
sudo /usr/local/cuda_xxx/bin/uninstall_cuda_xxx.pl
2、刪除殘留的文件夾
sudo rm -rf /usr/local/cuda_xxx/
參考文獻：
https://github.com/karentwan/document/blob/master/deep%20learning%20install.md