Compressor-VLA: Instruction-Guided Visual Token Compression for Efficient Robotic Manipulation

作者：Juntao Gao, Feiyang Ye, Jing Zhang, Wenjing Qian · 单位：School of Information Science and Technology, Beijing University of Technology, Beijing Key Laboratory of Computational Intelligence and Intelligent System, Beijing University of Technology · 会议/期刊：arXiv · 日期：2025-11-24 · 来源：Low-Level Learning-Based Action Modelling / Input Modelling / 2D Vision Language Action Models with Efficiency / Token Pruning