-
-
Notifications
You must be signed in to change notification settings - Fork 246
Open
Description
I have been trying to do vid2vid with a batch of images and masks using a workflow similar to this one:
https://www.reddit.com/r/StableDiffusion/comments/18dhjjm/animatediff_based_inpainting/
And I wanted to use a proper SD1.5 inpainting model as it produces better results on the individual frames over using a non-inpainting model + inpaint controlnet.
However, I have been getting weird artefacts caused by the animate diff model:
Metadata
Metadata
Assignees
Labels
No labels