gripper's value in observation data

[IrvingF7]

Hi! Thanks for open-sourcing this work!

So in the LIBERO dataset and simulation environment, just eyeballing it, it seems like for the gripper values in the action vector, 1→close, -1→open and they are all binarized

Could you tell me how to interpret the gripper values in the state/proprio vector? From the looks of #26 It seems to be some kind of delta in gripper action, but I am not sure how to interpret it.

Thanks for your time!