GraphPad: Inference-Time 3D Scene Graph Updates for Embodied Question Answering
返回详情
VLA / Vision-Language-Action 每日论文卡