linux下安装tensorflow-gpu

安装英伟达驱动

1、驱动下载地址： https://www.nvidia.cn/Download/index.aspx?lang=cn
2、卸载原有驱动
Ctrl+Alt+F1

sudo apt-get remove nvidia-*

sudo apt-get autoremove

sudo nvidia-uninstall

重启电脑

shutdown -r now

3、驱动安装
Ctrl+Alt+F1

sudo service lightdm stop

sudo sh ./filename.run -no-x-check -no-nouveau-check -no-opengl-files

sudo service lightdm restart

如果遇到如下问题：

ERROR: Unable to load the kernel module 'nvidia.ko'.  This happens most
   frequently when this kernel module was built against the wrong or
   improperly configured kernel sources, with a version of gcc that differs
   from the one used to build the target kernel, or if a driver such as
   rivafb/nvidiafb is present and prevents the NVIDIA kernel module from
   obtaining ownership of the NVIDIA graphics device(s), or NVIDIA GPU
   installed in this system is not supported by this NVIDIA Linux graphics
   driver release.

解决方法：更新Ubuntu16.04源

cd /etc/apt/

sudo cp sources.list sources.list.bak

sudo gedit sources.list

在sources.list中添加如下代码：

deb http://mirrors.ustc.edu.cn/ubuntu/ xenial main restricted universe multiverse 
deb http://mirrors.ustc.edu.cn/ubuntu/ xenial-security main restricted universe multiverse 
deb http://mirrors.ustc.edu.cn/ubuntu/ xenial-updates main restricted universe multiverse 
deb http://mirrors.ustc.edu.cn/ubuntu/ xenial-proposed main restricted universe multiverse 
deb http://mirrors.ustc.edu.cn/ubuntu/ xenial-backports main restricted universe multiverse 
deb-src http://mirrors.ustc.edu.cn/ubuntu/ xenial main restricted universe multiverse 
deb-src http://mirrors.ustc.edu.cn/ubuntu/ xenial-security main restricted universe multiverse 
deb-src http://mirrors.ustc.edu.cn/ubuntu/ xenial-updates main restricted universe multiverse 
deb-src http://mirrors.ustc.edu.cn/ubuntu/ xenial-proposed main restricted universe multiverse 
deb-src http://mirrors.ustc.edu.cn/ubuntu/ xenial-backports main restricted universe multiverse 

最后更新源和更新已安装的包后重新安装即可：

sudo apt-get update 

sudo apt-get upgrade 

安装Anaconda

1、Anaconda下载地址： https://www.anaconda.com/distribution/#linux
2、安装

bash Anaconda3-5.1.0-Linux-x86_64.sh

3、配置环境变量

sudo gedit /etc/profile

末尾添加：

export PATH=/root/anaconda2/bin:$PATH

保存并应用

source /etc/profile 

重启终端后更新

conda update -n base conda

安装Tensorflow-gpu和Keras

创建环境并指定python版本

conda create -n tf python=2.7

激活环境

conda activate tf  或  source activate tf

安装

conda install tensorflow-gpu==1.6.0 keras

安装其他库

conda install tensorflow-gpu keras

conda install matplotlib

conda install scikit-learn

conda install graphviz pydot

退出环境

conda deactivate  或  source deactivate

测试

运行 test.py 如下：

import tensorflow as tf
 
with tf.device('/cpu:0'):
    a = tf.constant([1.0,2.0,3.0],shape=[3],name='a')
    b = tf.constant([1.0,2.0,3.0],shape=[3],name='b')
with tf.device('/gpu:1'):
    c = a+b
    
sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True,log_device_placement=True))
#sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
sess.run(tf.global_variables_initializer())
print(sess.run(c))

运行结果如下，则环境正常：

参考

https://blog.csdn.net/gsww404/article/details/78328897
https://blog.csdn.net/weixin_39954229/article/details/79961172