IR3D-Bench: Evaluating Vision-Language Model Scene Understanding as Agentic Inverse Rendering
返回详情
VLA / Vision-Language-Action 每日论文卡