PyTorch深度学习实战-1-Python3环境与PyTorch安装

Posted 2022-11-03 esc_ai

千里之行，始于足下。

本文是《PyTorch深度学习实战》系列的第一篇文章，介绍如何在一个纯净的Linux环境(Centos 7)下安装python 3以及PyTorch。

一、Python 3环境

1.1 下载Python 3安装包

下载地址:
https://www.python.org/downloads/release

wget命令获取3.7.5源码:

wget https://www.python.org/ftp/python/3.7.5/Python-3.7.5.tar.xz

解压缩:

tar  -xvJf  Python-3.7.5.tar.xz

1.2 安装Python 3

先装一下基础必须的依赖:

yum -y groupinstall "Development tools"


yum -y install zlib zlib-devel bzip2-devel gcc g++ openssl-devel ncurses-devel sqlite-devel readline-devel tk-devel gdbm-devel db4-devel libpcap-devel xz-devel gcc libffi-devel python-devel openssl-devel zlib* vim -y

编译：

mkdir /usr/local/python3 

cd Python-3.7.5

./configure --prefix=/usr/local/python3

make && make install

生成软链:

ln -s /usr/local/python3/bin/python3 /usr/bin/python3 
ln -s /usr/local/python3/bin/pip3 /usr/bin/pip3

更新pip:

pip3 install --upgrade pip 

pip3 install --upgrade setuptools

测试:

[root@node5 Python-3.7.5]# python3 -V
Python 3.7.5

至此，Python3环境安装成功。

二、PyTorch环境搭建

2.1 安装PyTorch

pip3 install torch==1.8.1 torchvision==0.9.1 torchaudio==0.8.1

如果镜像拉不下来，配置pip的镜像源:

mkdir ~/.pip
cd ~/.pip
touch pip.conf
vim pip.conf

写入以下内容:

[global]
index-url = http://mirrors.aliyun.com/pypi/simple/
[install]
trusted-host=mirrors.aliyun.com

2.2 测试

import torch
import time
print(torch.__version__)
print(torch.cuda.is_available())

a = torch.randn(10000, 1000)
b = torch.randn(1000,2000)

t0 = time.time()
c  = torch.matmul(a, b)

t1  = time.time()

print(a.device, t1 - t0, c.norm(2))

运行结果:

1.8.1+cu102
False
cpu 0.41905760765075684 tensor(140310.5156)

至此，PyThon3 和 PyTorch环境搭建完成。

PS: PyTorch深度学习交流群: 531360538, 欢迎加入！