基于腾讯云GPU服务器的Stable Diffusion搭建

本项目基于腾讯云GPU服务器搭建，系统是选centos和cuda，cudnn配置。

• 下载和安装conda，一直选择回车，遇到选择填yes。

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh

./Miniconda3-latest-Linux-x86_64.sh

• 进入虚拟环境，看到（base）开头就算成功了

source .bashrc

• 拉取stable-diffusion-webui项目

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

cd stable-diffusion-webui/

• 安装依赖，如果下载速度太慢或者报错建议用豆瓣源，因为我后面遇到报错缺少cv依赖，所以这里要装个opencv-python-headless.

pip install -r requirements_versions.txt  -i https://pypi.douban.com/simple/

pip install opencv-python-headless -i https://pypi.douban.com/simple/

• 进入模型目录并下载模型，修改名字。后续下载的其他模型也是放在models/Stable-diffusion/下。

cd models/Stable-diffusion/

wget "https://cdn-lfs.huggingface.co/repos/22/ff/22ff773f5e71888a532282975a988033dcb7f20330c418f5e17da3e101e4b51e/e4bb82907cada4a51ea97953b488e68435a02833ac17894e1ca7677447eabc77?response-content-disposition=attachment%3B+filename*%3DUTF-8%27%27EimisAnimeModel_2-0.safetensors%3B+filename%3D%22EimisAnimeModel_2-0.safetensors%22%3B&Expires=1684738668&Policy=eyJTdGF0ZW1lbnQiOlt7IlJlc291cmNlIjoiaHR0cHM6Ly9jZG4tbGZzLmh1Z2dpbmdmYWNlLmNvL3JlcG9zLzIyL2ZmLzIyZmY3NzNmNWU3MTg4OGE1MzIyODI5NzVhOTg4MDMzZGNiN2YyMDMzMGM0MThmNWUxN2RhM2UxMDFlNGI1MWUvZTRiYjgyOTA3Y2FkYTRhNTFlYTk3OTUzYjQ4OGU2ODQzNWEwMjgzM2FjMTc4OTRlMWNhNzY3NzQ0N2VhYmM3Nz9yZXNwb25zZS1jb250ZW50LWRpc3Bvc2l0aW9uPSoiLCJDb25kaXRpb24iOnsiRGF0ZUxlc3NUaGFuIjp7IkFXUzpFcG9jaFRpbWUiOjE2ODQ3Mzg2Njh9fX1dfQ__&Signature=BHsCQuw1kXWXQMsV1ZzmVS-AqSIXZr4DfQ0%7Em1ZS-qXIEjJltYegDMJah5HXELnZ0AnTxPlrO5lGJRUAwz0Ob34chCoGsvlbPctvm6uDhb%7Ehcu8uU6XyXYKMDAQmU4G9pGC6nM6RhMQB5SNzImAYicBqT-Qr4vU0Y2-TO7z9Ec5ckWPCVS8LOGzQD3GgCCM1133VbuPRKEMw1yq-ugvUqqBELhdEsH5sAHHkcEchc7X0pBmxuygrmrnhPI8fwr2JcPukhc9LjMSUaTcBILduavqLiQETPktnWCDW8NiE7JcRdDFdhBAcLYdOn8X4HS%7E283zmCNfj6-JFokMofTjkbQ__&Key-Pair-Id=KVTP0A1DKRTAX" -O "EimisAnimeModel_2-0.safetensors"

• 回到项目目录下，安装汉化拓展，把localizations文件夹内的Chinese-All.json和Chinese-English.json复制到stable-diffusion-webui\localizations目录下

cd ../../

git clone https://github.com/VinsonLaro/stable-diffusion-webui-chinese

cp stable-diffusion-webui-chinese/localizations/chinese-all-0512.json   localizations/chinese-all-0512.json

cp stable-diffusion-webui-chinese/localizations/chinese-english-0512.json   localizations/chinese-english-0512.json

• 启动项目，因为腾讯云GPU服务器端口是全部打开的，外网可以直接访问公网 IP:5432即可，按照下面路径配置，然后点击Reload UI即完成汉化。

python3 launch.py --no-half --port 5432 --listen --xformers

• 现在到了成品展示阶段了，来看看效果

• 配置如下：

  Positive:1girl, solo, ahoge, bangs, blush, bridal gauntlets, capelet, closed mouth, crossed bangs, white long dress, final fantasy, winged capelet, yellow hair, hair band, hair between eyes, hair ornament, highres, jewelry, looking at viewer, extra short hair, beautiful detailed background, solo, upper body, shoulder wing, white gold theme, indoor, royal palace, glowing light, wind, flowers

  Negative:lowres, bad anatomy, error body, error hair, error arm, error hands, bad hands, error fingers, bad fingers, missing fingers, error legs, bad legs, multiple legs, missing legs, error lighting, error shadow, error reflection, error, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, blurry, flat, lowres, text, error, cropped, worst quality, low quality, jpeg artifacts, ugly

  Steps: 30, Sampler: DPM++ 2S a Karras, CFG scale: 7, Seed: 2762179779