Sorry, this is an image-text to image-text model, but there doesn't seem to be such an option. What's your suggestion?
@Cxxs I have decided to add a task called any-to-any to Hub, so I think for now we can close this PR.
any-to-any
· Sign up or log in to comment
