Anaconda
First install anaconda in their own user directory
./Anaconda3-5.0.1-Linux-x86_64.sh
Add anaconda environment after the installation is complete
vim ~/.bashrc
export PATH=/home/zhangyt/anaconda3/bin:$PATH
source ~/.bashrc
Add Tsinghua mirror, execute the following command twice, the first place Tsinghua mirror
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
Creating pythonn3.6 interpreter
conda create -n py36 python=3.6
Add cuda and cudnn environment, the server is already installed and cudnn cuda, need to know where they are installed
vim ~/.bashrc
export PATH=/usr/local/cuda-10.1/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.1/lib64:$LD_LIBRARY_PATH
source ~/.bashrc
First upgrade about pip
pip install --upgrade pip
Start the installation tensorflow-gpu = 1.13.1
pip install tensorflow-gpu==1.13.1
Test TensorFlow (GPU) whether the installation is successful
source activate py36
Then type python
import tensorflow as tf
hello = tf.constant('hello world')
sess=tf.Session()
print(sess.run(hello))
Shown below, described tensorflow-gpu been installed successfully