AI Snake Player

Project Overview

Using reinforcement learning to teach a nueral network how to play snake!

Project Steps:

1. Create a snake game
2. Create an agent to control the game
3. Create a model to learn how to play the game
4. Add in a human controlled game that can feed information to the model to train based on human interactions

Code

File overview:

1. game.py = The game of snake
2. agent.py = Agent that can generate a move and feed it into the game to play snake
3. model.py = A neural network with 16 input nodes and 3 output nodes
4. game_human.py = A human controlled version of the snake game which can feed informaion into the model for training

Installation

To run the code locally you will need:

• Python 3.7+
• Python Packages:
  • Pygame
  • NumPy
  • PyTorch

Running the Code

To run the code, run the agent.py file to automatically train the model or run the game_human.py file to train the model with human inputs