DreamVLA: A Vision-Language-Action Model Dreamed with Comprehensive World Knowledge

Benchmark Model Rank Results
robot-manipulation-on-calvinDreamVLA#1avg. sequence length (D to D): 4.44