Deep Q-learning for playing Flappy Bird

python source code for training an agent to play flappy bird. It could be seen as a very basic example of Reinforcement Learning's application.

Deep-Q-learning-for-playing-Flappy-Bird

How to use

With my code, you can:

  • Train your model from scratch by running python train.py
  • Test your trained model by running python test.py

Trained models

You could find my trained model at trained_models/flappy_bird

Requirements

  • python 3.6
  • pygame
  • cv2
  • pytorch
  • numpy

GitHub