search

Sygil-Dev / sygil-webui

Stable Diffusion web UI
GNU Affero General Public License v3.0
7.86k stars 883 forks source link

How is the inpainting model supposed to be used? #1608

Open mrakgr opened 1 year ago

mrakgr commented 1 year ago

This is not a bug, but just as side question I am wondering about. In addition to the 1.5 model, I also got the inpainting one, but it does not seem like WebUI is making use of it.

ZeroCool940711 commented 1 year ago

Hi there, I leave this here for future reference for anyone also looking for this: We currently do not have a proper way to do inpainting in the streamlit version, the gradio version we have has a small image editor but this version is considered legacy and deprecated as we now have a lot more features on the streamlit version and it's a lot better overall. On our streamlit version we currently do not have an image editor which is why we do not have proper inpainting there yet, you can technically create the mask in an external editor and import it to the img2img tab but it doesn't do proper inpainting, I do want to implement a lot of stuff related to inpainting and masking like the infinite canvas and other stuff but until I can get a proper image editor working on the streamlit version then I can't to those things.

Sanster commented 1 year ago

For anyone who is looking for an inpainting tool, you may try this one Lama Cleaner. It's an image inpainting tool powered by SOTA AI Model. You can use it to remove any unwanted object, defect, or people from your pictures or erase and replace(powered by stable diffusion) anything on your pictures.

https://user-images.githubusercontent.com/3998421/196976498-ba1ad3ab-fa18-4c55-965f-5c6683141375.mp4

parlance-zz commented 1 year ago

www.g-diffuser.com is also a good tool for in-painting / out-painting. I hope to bring these features into Sygil in the near future.