Compatibility of RDT2 with OpenArmX Pro Max Humanoid Robot？

[backrobot]

First of all, I want to commend you on the remarkable work you've done with RDT2! The claim of zero-shot cross-instance generalization is truly impressive, and it's exciting to see how your approach could potentially revolutionize robotic learning and task adaptation. We're really excited to explore the capabilities of this model in our own research.

I am working with a humanoid robot, the OpenArmX Pro Max（openarmx.com）, which is an enhanced version of the OpenArm configuration. Our laboratory has recently procured this robot, and we are eager to test RDT2 on it. Specifically, we are curious whether RDT2 is compatible with this robot and how it might handle tasks involving dual-arm humanoid configurations in terms of cross-instance generalization.

Could you provide any insights on whether RDT2 can be effectively applied to such a robot configuration?

[backrobot]

7 dofs per arm

[ethan-iai]

On my point of view, the problem mainly lies on designing the correct flange and installing the required end effector.