A simple neural network, learning a simple game, using reinforcement learning and epsilon greedy implementation.
Install Dependancies :
- random
- math
- numpy
- pygame
- pymunk
- matplotlib
- keras
Execution :
The most basic model is already provided in /src/models with the plotted graphs in src/data.
To train, run the learning.py from the /src directory.
To plot the results of training, run plot.py
To test the training, run play.py