Have you verified the AMP training results?

[soungsill]

I tested the AMP implementation using SKRL in Isaac Lab.

After training on the example "dance" motion data, the resulting behavior does not match what I expected based on the original Isaac Gym results. I don't think this is an issue with the data itself, but rather a discrepancy in the training results due to a possible problem in the code.

Have you verified that the training results for the "dance" motion work correctly?
If so, could you help me understand what might be wrong with the code I’m using?