Ubuntu 16.04下安装Tensorflow(GPU)
安装之前参考安装英伟达驱动的文章安装相关的显卡驱动。1.首先安装nvidia显卡驱动:
系统设置->软件更新->附加驱动->选择nvidia最新驱动(第一项)->应用更改
在ubuntu16.04中,更换驱动非常方便,去
系统设置->软件更新->附加驱动->切换到最新的NVIDIA驱动即可。应用更改->重启
2.下载CUDA8.0地址https://developer.nvidia.com/cuda-release-candidate-download(需要登陆)
请先确定显卡型号和是否支持GPU加速,查询网址:https://developer.nvidia.com/cuda-gpus
下载.run文件,进入文件目录,执行安装命令:
sudo ./cuda_8.0.61_375.26_linux.run (目录和文件名由你下载的文件进行更改)
根据提示输入y或回车等操作:此安装过程可选择不安装显卡驱动。
Do you accept the previously read EULA?
accept/decline/quit: accept
Install NVIDIA Accelerated Graphics Driver
for
Linux-x86_64
367.48
?
(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
]:
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
安装完毕后,再声明一下环境变量,并将其写入到 ~/.bashrc 的尾部:
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64"
export CUDA_HOME=/usr/local/cuda
3.测试是否安装成功(可选)
cd /usr/local/cuda/samples/1_Utilities/deviceQuery
sudo make
./deviceQuery
4.安装CuDNN
如果要使用gpu来对tensorflow进行加速,除了安装CUDA以外,cuDNN也是必须要安装的。跟cuda一样,去nvidia的官网下载cuDNN的安装包。不过这次没法直接下载,需要先注册,然后还要做个调查问卷什么的,稍微有点麻烦。我下的是cuDNN v5.1 Library for Linux这个版本。不要下cuDNN v5.1 Developer Library for Ubuntu16.04 Power8 (Deb)这个版本,因为是给powe8处理器用的,不是amd64.
下载地址:https://developer.nvidia.com/cudnn(需要登录)
下载完成后复制文件到cuda目录/usr/local/cuda/,解压下载文件:
tar xvzf cudnn-8.0-linux-x64-v5.1-ga.tgz ###(解压这个文件)
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn* (root用户可以忽略)
1.4 Tensorflow 0.11
tensorflow github上面提到 4 种安装方式,本教程使用 第四种 源码安装:
参看https://github.com/tensorflow/tensorflow/blob/master/README.md
https://github.com/tensorflow/tensorflow(下载地址)
说明:
(1)打开README.md页面,往下翻,直到下图这个位置:
Installation
See Installing TensorFlow for instructions on how to install our release binaries or how to build from source.
People who are a little more adventurous can also try our nightly binaries:
- Linux CPU-only: Python 2 (build history) / Python 3.4 (build history) / Python 3.5 (build history)
- Linux GPU: Python 2 (build history) / Python 3.4 (build history) / Python 3.5 (