Skip to content

适用于AMD显卡用户在Ubuntu中,一键安装stable-diffusion-webui。软件基于Docker运行,不依赖外部环境。

Notifications You must be signed in to change notification settings

k7212519/stable-diffusion-webui-AMD-onekey-deploy

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

35 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

image

stable-diffusion-webui-AMD-onekey-deploy

(停用!已有新版本,请移步新仓库!)适用于AMD显卡用户在Ubuntu中,一键安装stable-diffusion-webui。软件基于Docker运行,不依赖外部环境。
https://github.com/k7212519/ksd-launcher

介绍

1.该脚本适用于AMD rx5000,rx6000系列及其他专业显卡在Ubuntu20及以上版本系统中安装stable-diffusion-webui绘图软件。rx500rx400系列暂未适配,请期待后续更新, rx7000系列好像不支持ROCm,暂时未知,等这一代专业卡发布,应该就会一起官方适配了。
2.AMD显卡无法在Windows系统中使用ROCm深度学习环境,发挥全部性能。目前Windows中的转译效率低,在Linux中安装ROCm运行stable-diffusion,跑图能力将成倍提升!
3.依次两个脚本:
setup_docker.sh :一键安装最新Docker脚本
setup_sd.sh :一键拉取个人封装的 k7212519/stable-diffusion-webui 镜像并安装
4.运行时仅需打开终端Terminal,首先运行setup_docker.sh、完成后重启电脑,再运行setup_sd.sh即可,以下是运行命令:

./setup_docker.sh
./setup_sd.sh  

其他

1.运行脚本前,首先需要保证ubuntu系统中显卡驱动正确安装,并能在 设置-关于 中正确看到显卡型号,最好在终端输入rocm-smi能看到显卡相关信息。
2.如安装完系统无法识别显卡型号,请访问以下网站正确安装显卡驱动,并根据型号正确下载相关驱动文件:
https://www.amd.com/zh-hans/support/linux-drivers 下载DEB文件后,通过以下命令安装:

sudo apt install xxx.deb
amdgpu-install -y --usecase=graphics,hip,rocm --no-dkms

安装完成后,重新启动,在系统设置中看到显卡型号能正确识别后,再使用该脚本安装stable-diffusion-webui

stable-diffusion-webui-AMD-onekey-deploy

(Deactivate! There is a new version, please move to the new repository! For AMD graphics card users, install 'stable-diffusion-webui' in 'Ubuntu' with one click. The software runs on 'Docker' and does not depend on the external environment.
https://github.com/k7212519/ksd-launcher) For AMD graphics card users on Ubuntu, this script provides a one-click installation of the stable-diffusion-webui software. The software is based on Docker, which does not depend on external environments.

Introduction

1.This script is suitable for AMD rx5000, rx6000 series and other professional graphics cards to install stable-diffusion-webui graphics software on Ubuntu 20 and above systems. The rx500 and rx400 series are currently not compatible, please stay tuned for future updates. The rx7000 series seems to not support ROCm, it is currently unknown. When this generation of professional graphics cards is released, official adaptation should be available.
2.AMD graphics cards cannot use ROCm deep learning environment in Windows system to fully utilize their performance. Currently, the translation efficiency in Windows is low. Installing ROCm in Linux to run stable-diffusion can greatly improve the charting capability!
3.There are two scripts in order:
setup_docker.sh: One-click installation of the latest Docker script. setup_sd.sh: One-click pull of the personal packaging k7212519/stable-diffusion-webui image and installation.
4.When running, only need to open the terminal, first run setup_docker.sh, complete it and restart the computer, then run setup_sd.sh. The following are the commands:

./setup_docker.sh
./setup_sd.sh  

Notes

1.Before running the script, make sure that the graphics card driver is correctly installed in the Ubuntu system and the graphics card model can be seen correctly in Settings-About. It is best to enter rocm-smi in the terminal to see the graphics card related information.
2.If the graphics card model cannot be recognized by the system after installation, please visit the following website to correctly install the graphics card driver and download the related driver files according to the model:
https://www.amd.com/zh-hans/support/linux-drivers After downloading the DEB file, install it through the following command:

sudo apt install xxx.deb
amdgpu-install -y --usecase=graphics,hip,rocm --no-dkms

After the installation is completed, restart the system. After the graphics card model can be correctly recognized in the system settings, use this script to install stable-diffusion-webui.

About

适用于AMD显卡用户在Ubuntu中,一键安装stable-diffusion-webui。软件基于Docker运行,不依赖外部环境。

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages