I have two questions and look forward to your answers:
1, there is only one training command bash train.sh 1021_twist2 cuda:0 in TWIST2, does this mean that there is no need to train the teacher model first and then the student model like TWIST?
2, Why is TAR_MOTION_STEPS_FUTURE = [0] set by default in g1_mimic_future_config.py instead of TAR_MOTION_STEPS_FUTURE = [1,2,3,4,5]? What is the effect of the different settings?