DiffVLA++: Bridging Cognitive Reasoning and End-to-End Driving through Metric-Guided Alignment

返回详情 VLA / Vision-Language-Action 每日论文卡