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

Open in new window