ubuntu16.04LTS安装cuda9.0、cudnn7和tensorflow1.5

1.安装GeForce驱动程序

All Settings Software & Update Additional Drivers 

2.安装cuda8.0

下载cuda安装包 .runfile
sudo sh .runfile
由于之前设置好了显卡驱动，因此在Install NVIDIA Accelerated Graphics Drivers for Linux-x86_64时选择No
安装完成后，键入nvcc -V查询cuda版本号，如果nvcc还没装，键入sudo apt install nvidia-cuda-toolkit

3.安装cuDNN

下载cuDNN安装包，解压，键入：
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.配置环境变量

sudo gedit ~/.bash_profile
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64"
export CUDA_HOME=/usr/local/cuda
source ~/.bash_profile

5.安装pip

sudo apt-get install python-pip python-dev

6.安装tensorflow

sudo pip install tensorflow