DexGraspVLA
DexGraspVLA research team
A hierarchical vision-language planner and diffusion controller for dexterous grasping in cluttered scenes.
What you can explore
Study the released controller training, checkpoint and high-level planning integration.
Before you use it
The authors explicitly exclude hardware-related code because of intellectual-property constraints. A complete robot deployment cannot be reproduced from this repository alone. No root reuse license was identified.
Original contributors
Yifan Zhong, Xuchuan Huang, Ruochong Li and collaborators
License / access: Reuse terms unverified · hardware control is not released
Editorial reference · checked 2026-09-22