Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Alternative inpainting algorithm #36

Closed
wants to merge 1 commit into from

Conversation

oobabooga
Copy link
Contributor

Seems to give better results.

Extracted from Sygil-Dev/sygil-webui#308

Seems to give better results
@oobabooga
Copy link
Contributor Author

I'm not sure if this is different at all, actually.

@AUTOMATIC1111
Copy link
Owner

I've tried that before and there wasn't any meaningful difference. Are you sure this improves anything?

@oobabooga
Copy link
Contributor Author

oobabooga commented Aug 31, 2022

I am not sure. The alternative implementation seemed to give perfectly blended in results every single time when I first tried it, and the implementation in this repository seemed to often give mismatched results.

But it might have been pure luck as suggested by @1blackbar and the fact that I used different images.

@oobabooga
Copy link
Contributor Author

I have posted a direct comparison of the two algorithms on #35

I will close this PR since the implementation here is severely bugged doesn't work (it generates colorless blur in the masked region every time). It's better to start from scratch

@oobabooga oobabooga closed this Sep 1, 2022
nne998 pushed a commit to fjteam/stable-diffusion-webui that referenced this pull request Sep 26, 2023
When the user does not give an additional image, the init_image is used as the default
cabelo pushed a commit to cabelo/stable-diffusion-webui that referenced this pull request Nov 29, 2023
Atry pushed a commit to Atry/stable-diffusion-webui that referenced this pull request Jul 9, 2024
Atry pushed a commit to Atry/stable-diffusion-webui that referenced this pull request Jul 9, 2024
…UTOMATIC1111#157

now XYZ-plot really works with upscaler, but assure "Keep input image size" is checked
inner API hooks are hijack/unhijack and hooked/unhooked at each run, it is stateless now and should work with other sampler hijacks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants