/dqn-lunar-lander

A Deep Q-Network reinforcement learning model trained to safely land a lunar lander in the Farama Gymnasium

Primary LanguagePython

Gymnasium Lunar Lander

This is a Pytorch-based implementation of the DQN for training Farama's Gymnasium's Lunar Lander

Run Instructions

  • The anaconda environment can be installed using the environment.yml file:
    conda env create --file environment.yml

  • Run the main.py file:
    python main.py

Results

Here's a video of the Lunar Lander in action:

Lunar Lander

Plot for the scores achieved while training:

Training Plot