[Policy Evaluation] Action sequence timing mismatch and disorder during dance playback

[zzHezzo]

When I use the following command line for policy evaluation, the robot dance in Issacsim has a different timing compared to the original dance. It is split into small segments of action sequences, and the order is messed up. Will this affect the subsequent deployment operations?
Policy Evaluation
Play the trained policy by the following command:
python scripts/rsl_rl/play.py --task=Tracking-Flat-G1-v0 --num_envs=2 --wandb_path={wandb-run-path}