Enquiry of the root_joint choice for calculating root_mean_loss

[delaprada]

Thanks for the amazing work! I have a question regarding the choice of the root joint for calculating the root mean loss. Since the model is trained to predict the offset between the center_right joint and the center_left joint, would setting the root joint to 0 during evaluation (in eval.py) lead to a wrong value for the root_mean_loss on wrist joint whose prediction is based on center_joint?

Looking forward to your reply!