VL-JEPA: Joint Embedding Predictive Architecture for Vision-language Paper • 2512.10942 • Published Dec 11, 2025 • 59
