IRL-VLA: Training an Vision-Language-Action Policy via Reward World Model

Re-assign community