Awesome Robotics Manipulation · full_paper

Token Expand-Merge: Training-Free Token Compression for Vision-Language-Action Models

arXiv · 2025-12-10 · Low-Level Learning-Based Action Modelling / Input Modelling / 2D Vision Language Action Models with Efficiency / Token Pruning

VLA数据筛选机器人学习
新标签打开 AlphaXiv新标签打开 Paper新标签打开 Code
Token Expand-Merge: Training-Free Token Compression for Vision-Language-Action Models figure
AlphaXiv 中文论文页面（可滚动查看）