abalakrishna123 / recovery-rl

Implementation of Recovery RL: Safe Reinforcement Learning With Learned Recovery Zones.
https://sites.google.com/berkeley.edu/recovery-rl/
MIT License
51 stars 18 forks source link

How to implement navigation2 experiment #4

Closed ChenWeiHa closed 1 year ago

ChenWeiHa commented 1 year ago

I want to implement navigation2 experiment under MF method, However, I commented out the parts of train_MB_recover, pretrain_critic_recovery and pretrain_task_critic in the experi. py program with #, but they are still running. Does it mean that my experiment has not been pre-trained? How to implement the complete process (i.e. MF online data, pre-training, experiment)?