lllyasviel / lllyasviel/stable-diffusion-webui-forge

How to install xFormers ? xformers==0.0.30 requires torch==2.7.0pip

Open
#2,898 3 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
13k
Forks
1.7k
PR merge metrics
No merged PRs in 30d

Description

Hi There,
So I'm using Stability Matrix to run ForgeUI, and I wanted to install xFormers to boost performance and use ControlNet. I tried installing via pip, but hit this error:

xformers==0.0.30 requires torch==2.7.0pip
And updating Torch breaks compatibility with the other

= -----> Loop

Python 3.10
Torch 2.3.1+cu121

the Cuda I manually installed :1.24

Then I tried the "[Forge with CUDA 12.4 + Pytorch 2.4]<- Fastest, but MSVC may be broken, xformers may not work" package, but it seems it doesn't have xFormers either.

Does anyone have any workaround to get this working ?

Thanks in advance!

Edit : Overall, is all of this worth it? If I manage to get it working, will it be stable, or am I likely to run into crashes and issues? Would it be better to just stick with [CUDA 12.1 + PyTorch 2.3.1] instead? Also, regarding Stability Matrix — is there a way to import my manual installation into it?

Contributor guide

No contributing guide indexed for this repository

Research direction

The issue mentions Python 3.10, Torch 2.3.1+cu121, CUDA 12.4, xFormers, ForgeUI, ControlNet, and Stability Matrix, but names no repository file, test, or entry point. Determine whether this should become compatibility documentation and define a supported installation path and stability guidance before treating it as contribution work.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
machine-learning
Issue type
Documentation
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.