The aim of this project is to master the game of Super Mario Bros using a Deep Reinforcement Learning approach.
Therefore a player agent with a Deep Q Learning Model was implemented.
pip install -r requirements.txt
Adjust hyper parameters as needed in the "config.py" file.
python train.py
Observe trianing loss using tensorboard:
tensorboard --logdir="logs/"
python play.py
Inspired by Deep Q Learning with Tensorflow and Space Invaders from Thomas SIMONINI:
https://www.youtube.com/watch?v=gCJyVX98KJ4
Using the Super Mario Bros gym from Kautenja:
https://github.com/Kautenja/gym-super-mario-bros