Some question about low_lever low-level training

Regarding the wandb you provided, I saw that you used the feet height tracking reward. What is the difference between this and the feet height reward? Also, could you please open-source the relevant code?