Ubuntu初始环境设置

系统升级

sudo apt update
sudo apt upgrade

安装python基础开发包

sudo apt install -y python-dev python-pip python-nose gcc g++ git gfortran vim

2.Anaconda安装

bash Anaconda2-4.3.1-Linux-x86_64.sh

然后

sudo gedit ~/.bashrc

在最后输入
export PATH="/home/wmy/anaconda2/bin:PATH" export LD_LIBRARY_PATH="/home/wmy/anaconda2/lib:LD_LIBRARY_PATH"

关闭，输入

source ~/.bashrc

重启终端

3.安装pycharm
解压

tar xfz pycharm-community-2017.1.tar.gz
将解压完的文件夹移动到/usr/local目录
sudo mv pycharm-community-2017.1 /usr/local
进入文件夹的bin文件夹内
cd pycharm/bin/
运行sh文件安装
./pycharm.sh
创建Pycharm桌面快捷方式：
gedit /usr/share/applications/studio.desktop
在这个文件里面下如下代码：
[Desktop Entry]
Type=Application
Name=Pycharm
GenericName=Pycharm3
Comment=Pycharm3:The Python IDE
Exec=sh /opt/pycharm-community-2016.3.1/bin/pycharm.sh
Icon=/opt/pycharm-community-2016.3.1/bin/pycharm.png
Terminal=pycharm
Categories=Pycharm;
桌面会有图标，右键锁定

4.下载CUDA8.0

sudo dpkg -i cuda-repo-ubuntu1604-8-0-local_8.0.44-1_amd64.deb
sudo apt update
sudo apt install cuda

将CUDA路径添加至环境变量 在终端输入：

sudo gedit /etc/bash.bashrc

在bash.bashrc文件中添加：

export CUDA_HOME=/usr/local/cuda-8.0
export PATH=/usr/local/cuda-8.0/bin{PATH:+:{PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-8.0/lib64{LD_LIBRARY_PATH:+:{LD_LIBRARY_PATH}}

之后

source gedit /etc/.bashrc即可 同样，在终端输入：

sudo gedit ~/.bashrc

在.bashrc中添加如上相同内容 （如果您使用的是zsh，在~/.zshrc添加即可）

测试 在终端输入：

nvcc -V

5.加速库cuDNN,先解压

sudo cp include/cudnn.h /usr/local/cuda-8.0/include/
sudo cp lib64/* /usr/local/cuda-8.0/lib64/

6.安装theano
conda install theano pygpu
配置theano文件 在终端中输入:

gedit ~/.theanorc

[global]
openmp=False
device = cuda0
floatX = float32
allow_input_downcast=True
[cuda]
root=/path/to/cuda/root
[lib]
cnmem = 0.8
[dnn]
library_path = /usr/local/cuda-8.0/lib64
include_path = /usr/local/cuda-8.0/include
[blas]
ldflags= -lopenblas
[nvcc]
fastmath = True

7.安装keras
pip install keras
修改默认keras后端 在终端中输入:

gedit ~/.keras/keras.json
{
"epsilon": 1e-07,
"floatx": "float32",
"image_data_format": "channels_first",
"backend": "theano"
}

在终端输入python

import theano
Using cuDNN version 5110 on context None
Mapped name None to device cuda0: GeForce GTX 970 (0000:01:00.0)

8.安装tensorflow1.1.0

pip install tensorflow_gpu-1.1.0rc0-cp27-none-linux_x86_64.whl

import tensorflow

全部搞定