Agile-VLA: Few-Shot Industrial Pose Rectification via Implicit Affordance Anchoring

Teng Yan, Zhengyang Pei, Chengyu Shi +8

Deploying Vision-Language-Action (VLA) models on resource-constrained edge platforms encounters a fundamental conflict between high-latency semantic inference and the high-frequenc…