lllyasviel / lllyasviel/LayerDiffuse

SDXL Lightning Models Produce Awful Results

Open
#12 8 comments 1 reaction 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
No language data
Stars
2.2k
Forks
36
PR merge metrics
No merged PRs in 30d

Description

I tried LayerDiffusion using a ComfyUI workflow with an SDXL Lightning model and the results were really quite bad. When I switched to a standard SDXL model, the result was much better. Is this something you expect? Have you tried a lightning model yourself? Is there any way to improve the output for lightning checkpoints? I'm so addicted to their speed. 😂

Thank you for your work.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

The issue reports poor LayerDiffusion results in a ComfyUI workflow with an SDXL Lightning checkpoint, compared with a standard SDXL model. Start by reproducing both workflows and comparing their outputs; done would be a documented explanation of the compatibility issue or a confirmed way to improve Lightning results.

Written by the indexing model from the issue text.

Assessment

Domain
machine-learning
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
18/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.