Search for a command to run...
LLaVA-Mini: Efficient Image and Video Large Multimodal Models with One Vision Token