MAGREF: Masked Guidance for Any-Reference Video Generation with Subject Disentanglement

Open in new window