Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
dn6 
posted an update Aug 30
Post
2550
Sharing for anyone using Diffusers from_single_file loading and affected by the Runway SD 1.5 issue.

If you have runwayml/stable-diffusion-v1-5 saved locally in your HF cache then loading single file checkpoints in the following way should still work.

from diffusers import StableDiffusionPipeline

pipe = StableDiffusionPipeline.from_single_file("<url or path to single file checkpoint>")


If you do not have the model repo saved in your cache, then automatically inferring the pipeline config will not work since the reference repo runwayml/stable-diffusion-v1-5 doesn't exist anymore.

You can use an alternative SD1.5 repo id to still configure your pipeline.

from diffusers import StableDiffusionPipeline

pipe = StableDiffusionPipeline.from_single_file("<url or path to single file checkpoint>", config="Lykon/DreamShaper")


We're working on resolving the issue ASAP.

@sayakpaul The fix should be very easy, but it is a fatal bug for from_single_file!
Even FluxPipeline should go read the repo for CLIP...😓

P.S.

In addition, there is something buggy with the behavior of Diffusers or Transformers related torch.float8_e4m3fn when loading the model. (I don't know which one it is because both are used...)
At first I thought it was a common torch version problem, but it seems to be different.
https://huggingface.co/Kijai/flux-fp8/discussions/4#66ce79dae0c8028b89a02c8d

·
In this post