VLM

From Seeing to Doing: Bridging Reasoning and Decision for Robotic Manipulation