hr0nix/omega

The YAML file of PPO algorithm training

Closed this issue · 3 comments

Could you please provide the YAML file of PPO algorithm training, millions of thanks!

Hello!

The implementation of PPO seems to be broken by a recent refactoring. I will fix it and provide an example config for PPO soon.

Keep in mind that in this repo I've been primarily focused on polishing MuZero implementation, PPO has been used mostly for the purpose of testing things before doing anything complicated. I've never tested it on anything harder than random_room_5x5.

I've fixed the implementation of PPO and added a PPO config for room_5x5 environment in 9d8c3c9

Thanks so much for your great work, it is really helpful.