Multimodal model

DreamID-Omni

DreamID-Omni is a human-centric audio-video generation framework that supports reference-conditioned generation, identity-and-voice video editing, and audio-driven identity animation with lip-sync.