MEMO is a state-of-the-art open-weight model for audio-driven talking video generation.
Want to make some of these yourself?