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