#install torch in a venv python

[linneasmith]

How to Install PyTorch

PyTorch is a framework for the Python programming language, designed for machine learning. It includes a set of tools for working with models, commonly used in natural language processing, computer vision, and other similar fields. You can install it on your server manually using this guide.

PyTorch Installation on Linux¶

This instruction is suitable for the following operating systems: Ubuntu 22.04, and verified for Python versions: Python 3.10.

Install Python:sudo apt install python3.10 In Ubuntu 22.04, this version is installed by default, so we do not recommend installing a newer version.

Create a virtual environment for Python:python3 -m venv venv

Activate the virtual environment:source venv/bin/activate After successful activation, the prompt will include the name of the virtual environment in parentheses:(venv) user@49069:~$ NoteYou can create as many virtual environments as you like and install different libraries (including simultaneously, but sometimes this may cause conflicts).

Install PyTorch libraries in the virtual environment:pip3 install torch torchvision torchaudio

Verify PyTorch installation:To do this, run a Python console command python and then run the following program:import torch x = torch.rand(5, 3) print(x) If the installation is successful, you will receive the following output:(venv) user@49069:~$ python Python 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import torch >>> x = torch.rand(5, 3) >>> print(x) tensor([[0.80, 0.04, 0.6], [0.32, 0.59, 0.7], [0.8, 0.70, 0.25], [0.40, 0.9, 0.9], [0.8, 0.15, 0.5]])

Verify if PyTorch libraries use CUDA:To do this, run the following program in a Python console:import torch torch.cuda.is_available() If PyTorch can work in GPU mode, the output will be: >>> import torch >>> torch.cuda.is_available() True

[kuoxen]

Stable Diffusion Service Deployment —— SD WebUI on CentOS7

框架依赖安装

1. sudo yum update 并重启

更新ssl，为了安装Python3.10.6

whereis openssl | xargs rm -frv

yum install perl-ExtUtils-CBuilder perl-ExtUtils-MakeMaker perl-CPAN -y

<!-- 进入perl命令行内执行 -->

perl -MCPAN -e shellinstall IPC/Cmd.pm

<!-- 下载SSL version>=1.1.1k 2021-3-25 -->

wget https://www.openssl.org/source/openssl-1.1.1q.tar.gz

tar -zxvf openssl-1.1.1q.tar.gz

cd openssl-1.1.1q

./config --prefix=/usr/local/openssl --openssldir=/usr/local/openssl shared zlib

make && make install

echo "/usr/local/openssl/lib" >> /etc/ld.so.conf 

ldconfig -v

ln -s /usr/local/openssl/bin/openssl /usr/bin/opensslopenssl

安装Python3.10.6，为了能运行特定版本torch/CUDA

whereis python3 | xargs rm -frv

whereis pip3 | xargs rm -frv

yum install -y zlib-devel bzip2-devel openssl-devel ncurses-devel sqlite-devel readline-devel tk-devel gcc make python3-devel libevent-devel python-gevent libffi-devel 

wget https://www.python.org/ftp/python/3.10.6/Python-3.10.6.tgz

tar -zxvf Python-3.10.6.tgz

cd Python-3.10.6

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

make && make install

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

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

python3 --version

4. 安装Git(>=1.8.5)，为了能执行大文件下载

yum install epel-release

yum remove git

yum install https://repo.ius.io/ius-release-el7.rpm

yum search git2 //下行的具体版本会变，可通过search确定

yum install git236

yum install git-lfs

安装显卡驱动

yum install pciutils

wget https://us.download.nvidia.com/tesla/460.106.00/NVIDIA-Linux-x86_64-460.106.00.run

yum install -y kernel-devel-$(uname -r) kernel-headers-$(uname -r)

sh NVIDIA-Linux-x86_64-460.106.00.run

nvidia-smi

框架安装

wget https://raw.githubusercontent.com/AUTOMATIC1111/stable-diffusion-webui/master/webui.sh

<!-- 修改webui.sh脚本，将默认路径从 /home/$(whoami) 改为自定义路径 -->

<!-- 改个清华镜像加速-->

/......../stable-diffusion-webui/venv/bin/python3 -m pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple sh webui.sh

<!-- 如果安装出现问题，可以先自行git clone 到local，再如下从本地安装 -->

/......../stable-diffusion-webui/venv/bin/python3 -m pip install git+file:///data/sd/GFPGAN 

/......../stable-diffusion-webui/venv/bin/python3 -m pip install git+file:///data/sd/CLIP

/......../stable-diffusion-webui/venv/bin/python3 -m pip install git+file:///data/sd/open_clip

<!-- 如果安装出现zlib版本问题 -->

从 https://github.com/DiffusionHub/DiffusionHub 下载zlib-1.2.9

tar -xvf ~/Downloads/zlib-1.2.9.tar.gz

cd zlib-1.2.9

sudo -s

./configure; make; make install

make install完会生成zlib-1.2.9，将/lib下的libz.so和libz.so.1 软链接到这个zlib-1.2.9，如 ln -s -f /lib/libz.so.1.2.9/lib libz.so.1

模型安装

Stable Diffusion模型

2.0以下模型可以直接把ckpt模型文件放在WebUI目录内models/Stable-diffusion 路径下

2.0及以上模型需要额外的参数描述文件yaml，可以从 https://github.com/Stability-AI/stablediffusion/tree/main/configs/stable-diffusion 下载相应yaml，分辨率为512x512的用v2-inference.yaml，768x768的用v2-inference-v.yaml* NovelAI leaked model

服务发布

WebUI: sh webui.sh --listen

API: sh webui.sh --nowebui --listen

指定使用的显卡，例如