A Wan video-to-video (V2V) workflow that uses a reference image for the first frame to give you more control over the result.
Description
FAQ
Comments (11)
So this is basically FlowEdit?
Why don't you call it FlowEdit anywhere?
I'm encountering an issue with WanVideoImageClipEncode: shape '[1, 4, 4, 54, 96]' is invalid for input of size 98496. Could you clarify which text encoder you're using?
Damn I have the same exact issue, did you ever find a solution?
Me too~
Did you ever resolve this? Can't seem to get the workflow to work personally.
Did you try to change frame_load_cap in Load Video (Upload). I know sounds crazy but it works for me. For example 48 fps gives me error but 49 fps works fine
Works really well thanks for sharing.
This is cool! Nice one! <3
How about using depth lora motion guidance with the first frame?