Fetching the paper…

Video-LLaVA: Learning United Visual Representation by Alignment Before Projection · Around