This is my project for the 24th MMU Hackerspace Hackathon. The goal is to train a neural network to play snake game using RL (Reinforcement Learning) with Deep Q Learning in PyTorch.
This snake game is playing on a 8x8 grid, because I didn't have much time to train for a 20x20 board. Feel free to change self.N parameter in the environment.py and retrain it.
If you'd like, you can run main.py for a snake game simulation with the trained agent.