Skip to content

This repo contains PPO implementation in PyTorch for LunarLander-v2

Notifications You must be signed in to change notification settings

junkwhinger/PPO_PyTorch

Repository files navigation

PPO Implementation in Pytorch for LunarLander-v2

Being fastinated by "IMPLEMENTATION MATTERS IN DEEP POLICY GRADIENTS: A CASE STUDY ON PPO AND TRPO", I wrote PPO code in PyTorch to see if the code-level optimizations work for LunarLander-v2. And they do! for some extent.

How to train

Find a config .yaml file in the config directory and run the following command. You can make your own .yaml file, but make sure they have all the necessary options.

$ python main.py --config PPO_M.yaml

How to play

$ python main.py --config PPO_M.yaml --eval

How to run Bayesian Optimization for hyperparameters

$ python search.py

Reference

About

This repo contains PPO implementation in PyTorch for LunarLander-v2

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published