VL-JEPA: Joint Embedding Predictive Architecture for Vision-Language(arxiv.org)3 points·by hbarka·6 เดือนที่ผ่านมา·0 commentsarxiv.orgVL-JEPA: Joint Embedding Predictive Architecture for Vision-Languagehttps://arxiv.org/abs/2512.109420 commentsPost comment—