EfficientVLA: Training-Free Acceleration and Compression for Vision-Language-Action Models Paper • 2506.10100 • Published 29 days ago • 10