GPU：12G+，建议16G以上，还是尽量勾搭，好像现在最大32G，目前个人性价比24G有时长出售。

内存：16G以上，建议32G，也是越大越好。

硬盘：最好使用固态硬盘，还是尽量够大

2、软件要求：

python：3.10

cuda：11.8

二、文件下载

github：GitHub - AUTOMATIC1111/stable-diffusion-webui: Stable Diffusion web UI

 整合好的版本：

链接: https://pan.baidu.com/s/1N0czN9ND_TeXpucKWYIrFg 提取码: t4de 

 选择文件：stable-diffusion-webui.zip

解压文件：

unzip stable-diffusion-webui.zip

三、开始安装stable-diffusion-webui

1、创建Python虚拟环境

我目前部署的是v1.9.3版本的Stable Diffusion，需要3.10.6版本的Python

//创建conda环境
conda create -n stable python=3.10.6
//激活conda环境
conda activate stable

2、安装python依赖包

pip install -r requirements.txt

四、启动服务

python launch.py --listen --xformers --enable-insecure-extension-access --theme dark --gradio-queue  --port=8080

五、会出现一些错误

问题1：libGL.so.1: cannot open shared object file: No such file or directory

sudo apt-get update
sudo apt-get install -y libgl1-mesa-glx

问题2： 

decoderF is not supported because:
        xFormers wasn't build with CUDA support
        attn_bias type is <class 'NoneType'>
        operator wasn't built - see python -m xformers.info for more info
    flshattF@0.0.0 is not supported because:
        xFormers wasn't build with CUDA support
        operator wasn't built - see python -m xformers.info for more info
    tritonflashattF is not supported because:
        xFormers wasn't build with CUDA support
        operator wasn't built - see python -m xformers.info for more info
        triton is not available
        Only work on pre-MLIR triton for now
    cutlassF is not supported because:
        xFormers wasn't build with CUDA support
        operator wasn't built - see python -m xformers.info for 

解决方式：

pip uninstall xformers
pip install xformers