← back to paper
arxiv: 2604.24764 · 3 revisions
World-R1: Reinforcing 3D Constraints for Text-to-Video Generation