Multimodal model

DreamActor-M1

DreamActor-M1 is a diffusion-transformer-based human image animation framework that turns a reference image into expressive, identity-preserving human motion video.