EfficientVLA: Training-Free Acceleration and Compression for Vision-Language-Action Models

Yang, Yantai, Wang, Yuhao, Wen, Zichen · arXiv · 2025