Confusing training results

[yihangyao]

Hi, thanks for the great work!

I am writing to ask for your advice. I used your task file ''imitation_learning'' for my RL training. I utilized the PPO algorithm and both go1 and a1 robots. However, the results are not good (the gif is attached). Do you have any idea why the result looks so funny? Thank you!

Best

[yihangyao]

Also, would you mind doing me a favor to check the curriculum settings? I encounter an error "AttributeError: 'LocomotionGymEnv' object has no attribute 'set_curriculum_scale'" while enabling the curriculum training.

Thanks for your help!