Training Performance Decline due to Time Interval Mismatch

[XinyuSong123]

Hi, I am currently applying the motion dataset of the Unitree G1 in IsaacLab, but I have noticed that the training performance is not ideal. I suspect that one of the reasons might be the difference in time intervals between the IsaacLab simulation and the motion dataset. Below are the time intervals I am using for my training:

[INFO]: Base environment:
Environment device : cuda:0
Environment seed : None
Physics step-size : 0.005
Rendering step-size : 0.02
Environment step-size : 0.02

Could you kindly provide any insights on the time intervals used in the motion dataset? Thank you!