[ICML 2024] 3D-VLA: A 3D Vision-Language-Action Generative World Model - View it on GitHub