Ω-QVLA: Robust Quantization for Vision-Language-Action Models via Composite Rotation and Per-step Scaling

返回详情 VLA / Vision-Language-Action 每日论文卡