Fetching the paper…

HumanOmni: A Large Vision-Speech Language Model for Human-Centric Video Understanding · Around