环境搭建：linux + cuda8.0 + cudnn5.1 + tf1.2

1、安装Anaconda

a）在官网下载文件

b）安装命令：sh Anaconda3-5.3.0-Linux-x86_64.sh

      默认路径：/root/anaconda3

c）添加环境变量：export PATH=/root/anaconda3/bin:$PATH

 

2、安装cuda8.0

a）下载cuda：https://developer.nvidia.com/cuda-80-download-archive

b）安装命令：sh cuda_8.0.44_linux.run

c）安装选项

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: no
Install the CUDA 8.0 Toolkit?
(y)es/(n)o/(q)uit: yes
Enter Toolkit Location[ default is /usr/local/cuda-8.0 ]:
Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: yes
Install the CUDA 8.0 Samples?
(y)es/(n)o/(q)uit: no

3）下载cudann5.1（runtime版），解压后lib文件上传到root，将include拷贝入

   /usr/local/cuda/include

4）添加环境变量：

export PATH=/usr/local/cuda-8.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH=/root/cudnn5.1/lib64:$LD_LIBRARY_PATH
export CUDA_HOME=/usr/local/cuda-8.0

5）激活环境变量：source /root/.profile

 

3，tf12虚拟环境搭建

1）conda create -n tf12 python=3.6

     删除：conda remove -n tf12 --all

     克隆：conda create -n caf --clone caf_base

2）下载tf12，py36，x86_64的版本

3）安装：pip install tensorflow_gpu-1.2.0-cp36-cp36m-manylinux1_x86_64.whl

4）激活环境：source activate tf12

      退出：deactivate

5）查看已经安装的包：conda list

 

4、在pycharm中的使用

1）run->edit configuration

环境变量中添加：LD_LIBRARY_PATH /usr/local/cuda-8.0/lib64/;/home/name/cudnn5.1/lib64/