该论文记录已去重归并

旧标识 oa-w4407806842 已合并到唯一正式记录。

前往 ChatVLA: Unified Multimodal Understanding and Robot Control with Vision-Language-Action Model