VLA ModelsarXiv preprint2026年1月

Show-Harness: Just a VLM Agent Can Play Robots

Yanzhe Chen, Zechen Bai, Zhijun Cao, Wenzheng Zeng, Kevin Qinghong Lin, Yiqi Lin, Guoqiang Liang, Kevin Yuchen Ma, Qiming Huang, Mike Zheng Shou et al.Show Lab, National University of Singapore

摘要

We introduce Show-Harness, demonstrating that a Vision-Language Model agent can directly control robots through demonstration-based interaction. Our approach eliminates the need for specialized robot training by harnessing the general capabilities of VLMs for perception, planning, and action execution.

VLM agentrobot controlharnessplaymanipulation

技术细节

仿真平台
测试机器人
模型骨架

VLM Agent

编码器

GPT-4V / SigLIP

解码器

Demonstration-based Controller

相关公司(1)

京ICP备2026064258号-1