DeepFlapPyBird

DeepFlapPyBird is a project that trains an agent using Deep Q-Learning to play flappybird game. Despite many other existed versions, this project does not use the game frames images as input and it is implemented inside the game and it will get the observations with the help of a custom gym environment.

This project is a modified clone of FlaPyBird repository.

Installation (as tested on MacOS)

1. Create a python virtual environment to install required packages (optional):

$ python3 -m venv venv
$ source venv/bin/activate

2. Install PyGame (2.x.x recommended):

$ pip install pygame

3. Install OpenAI Gym:

$ pip install gym

4. Install PyTorch as it's documentation.

5. Install NumPy:

$ pip install numpy

ScreenShot