wolverinn / stable-diffusion-multi-user Goto Github PK

stable diffusion multi-user django server code with multi-GPU load balancing

Home Page: https://image.stable-ai.tech/

License: GNU General Public License v3.0

Python 85.56% JavaScript 1.48% Jupyter Notebook 8.85% C++ 1.24% Cuda 2.12% Shell 0.05% HTML 0.31% Dockerfile 0.01% CSS 0.28% CMake 0.06% Roff 0.05%

aigc stable-diffusion stable-diffusion-api

stable-diffusion-multi-user's People

Contributors

Stargazers

Watchers

Forkers

aaronbesson harrison001 weihaigang 3dsworks2022 tmaaefu192 mouhao wangchencom titer18 wowworldpotato ukaserge baymikyou sanbuphy fengluo colin-viola tianbohuang ryanxia2016 qianyebuleng crl228 tulanglaoli eamanread liucanpeng marvion2 fuyunzhishang khk084 ac-km ferdiz-afk openainext bitaminkim linecode happyvee wcswcswcs xiaoqingwang mengkek hunkjun usesless icarried liu954326053 grdval1 ethan2018t1 letheliu zhigangjiang lldacing alphapro-pro randyluo85 simbaprince hawanlee ik4sumi choijiwung shuoshadow xiaosong light333333 erickech diegovivas medalcollector huangxihuang flyinsky222 lucameschiari render-ai bearla

stable-diffusion-multi-user's Issues

请问不用API的话要怎么解决多人模型切换问题和队列的问题呢？

公众号里似乎只讲了使用API的方法。

Do any one have docker image for this

I was trying to run but there are a lot of version issues as it's in Python so it's expected but I have very basic knowledge of Python it will be easy forks like just pulling the image or building to run it and then we can also use that in production.

请问能出教程吗

请问能出详细些的部署调试教程吗,愿意支付给您费用!

manage.py makemigrations报错

麻烦问下，当我执行setup.sh中的python manage.py makemigrations时，会出现如下错误，像是把webui的参数和manage.py想冲突

bash setup.sh env|venv 之后运行会提示缺少各种各样的库

基本都是SD里面用到的库，setup.sh env|venv 这个流程重复试了好多遍，每次都这样，然后只能手动pip install，但这样手动又会遇到很多报错，而且缺少的库实在是太多了，感觉全部手动pip install也不现实。

基本都是这个init_model.py报的错

Skipping unknown extra network: lora

lora:xxxxx:0.7 我按照这种格式添加到prompt中 ,控制台跳出 Skipping unknown extra network: lora lora
请问如何修复?

项目启动报错

您好！我按照所有流程安装完依赖之后启动项目，浏览器访问和接口请求都是返回500
运行环境：ubuntu20.04本地服务器，nvidia驱动版本525.125.06，cuda11.8
打开apache2 log查看错误日志如下

这个项目只有一个api？没有前端么？

如题

Can we run it with CPU only

So right now I only have an Ubuntu laptop that does not have any GPU so before using the Cloud GPU I want to test whether it is working or not. @wolverinn

apache报错

您好，我按照操作方法来做，到最后一步apache报错，错误日志为：
4463 2023] [wsgi:error] [pid 311707:tid 139956991420160] [client 10.11.140.104:64206] ModuleNotFoundError: No module named 'django'
[Wed Jun 07 09:37:22.628

我看了下，apache下面的配置文件貌似是对的，现在好像就是说这个虚拟环境不能用，请问应该如何解决，谢谢。

IP配置方面，是否支持域名+端口的形式，解决家庭网络动态DDNS的情况

如题

插件

请问可以支持插件嘛，或者后续可否支持插件呢

关于单机多gpu的应用

您好,请问如果使用该框架在单机上调用多个gpu可以实现吗?
ip_list = {
"1.1.1.1": {},
"2.2.2.2": {},
}
这个设置调用的逻辑是?

请问docker可以实现这样的功能吗,还是django效果更好?

在哪里能够获取更详细的api教程,因为在示例中并没有提到其他能用的扩展,例如脚本的使用?

使用问题

这个**是使用django做了一层外壳，使用django完成后续的接口调用操作吗？这样写的目的是什么，为什么不是在原有的基础上做出修改？谢谢

Need multi user code only

I have tried every method that you have shown in the readme file and I have tried all those methods and none of them work. Every code has some kind of bug or something.
So I'm running webui on my GPU server where I have my UI and I'm making API calls to generate images.
So how can installing multiple webui with API enabled and load balancer can redirect traffic of the API calls?
I just need those files that will help me to achieve this because your present code has a lot of bugs and missing files as I tried to run it.
@wolverinn please help me with this. Does anybody else have solved this issuer?