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

About the inpaint mode: fill, original, latent noise, latent nothing #5691

Closed
atlantistin opened this issue Nov 8, 2023 · 5 comments
Closed
Assignees

Comments

@atlantistin
Copy link

There is different mode to use inpainting in WebUI, and it can get the different result.
Could this be implemented in diffusers ???

@yiyixuxu
Copy link
Collaborator

yiyixuxu commented Nov 9, 2023

I think what we currently have is similar to "original". we are still exploring these options! and would definitely be interested in add if they are helpful

would you be interested in sharing more about your experience using these different inpaint mode? e.g. which mode you use most often and why, etc

thanks in advance!

@yiyixuxu yiyixuxu self-assigned this Nov 9, 2023
@atlantistin
Copy link
Author

Thank you very much!
I found that the parameter of strength can regulate it.
For example, if I need to create a new skin, I'll make it lower to make sure it's similar to the original skin.

@ka1tte
Copy link

ka1tte commented Mar 9, 2024

Can the mentioned function be used now?

@ka1tte
Copy link

ka1tte commented Mar 9, 2024

Thank you very much! I found that the parameter of strength can regulate it. For example, if I need to create a new skin, I'll make it lower to make sure it's similar to the original skin.

In fact, this is a different function from inpaint mode. We still need to control the initialization method of the masked area.

@dain5832
Copy link

dain5832 commented Mar 12, 2024

I guess #6072 mentions masked_content option.

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

No branches or pull requests

4 participants