jeremiah-wa / neat_flappy_bird Goto Github PK

Basic A.I. that plays Flappy birds using the NEAT-python module

I used this project by Tech with Tom as a practical implementation of some of the reinforced machine learning techniques I've been learning. The NEAT-python module is a powerful tool for ML and I found it very ergonomic from a data science point of view. It allowed me to tweak hypo-parameters with no change to the code via the neat_config.txt file. This gave me the freedom to concentrate on experimentation. The built-in StdOutReporter function was very useful for displaying the results of each generation and the subsequent winning genome. I also found the Pygame library to be great for building quick and easy GUIs. Definitely one I'll use in future!

• install neat and pygame libraries using pip install -r requirements.txt

• configure algorithm by editing neat_config.txt file

• run experiment using the command py main.py