LLaVA-Mini: Efficient Image and Video Large Multimodal Models with One Vision Token Paper • 2501.03895 • Published 5 days ago • 42