Go2 Forward Walking Policy - How to Train?

[AlejandrooBC]

I am training a PPO forward walking policy for the Unitree Go2 quadruped using MuJoCo and StableBaselines3. I found this website: https://genesis-world.readthedocs.io/en/latest/user_guide/getting_started/locomotion.html#reward

That talks about training a stable forward walking policy on the Go2.

Is this the correct repository directory that contains the files used to train the forward walking policy: https://github.com/Genesis-Embodied-AI/Genesis/tree/main/examples/locomotion

One more question: Was this trained using curriculum learning? What is the approach to training the Go2 to walk stably here? In my own project, I am struggling to achieve a stable forward walking policy. The Go2 achieves either galloping, forward lunging, etc. Any help is greatly appreciated, thanks!