Does RDT2 require no joint positions, or transform between camera link and base link?

[OmMandhane]

Does RDT2 require no joint positions, or transform between camera link and base link. How does it output tcp position in base frame on new embodiement/ robots while running zero shot? it outputs direct end effector position and not delta postions right?

[LBG21]

1. RDT2 requires no joint positions. 2. RDT2 outputs in VIVE Tracker base frame, and the whole calibration process is to make sure the transform matrix of VIVE Tracker->Robot is properly set when deployed on new embodiments. 3. The model outputs delta EEF positions.