3 ms·Show HN: VL-JEPA(Joint Embedding Predictive Architecture for Vision-Language) [video]3 points by JosefAlbers 9mo ago