Snake AI with Reinforcement Learning

This project involves training a reinforcement learning agent to play the Snake game. The agent uses a Deep Q-Network (DQN) to learn optimal strategies and improve its performance over time.

Project Demonstration

During Training

During Training

Project Structure

• game.py: Contains the implementation of the Snake game, including the game logic and rendering.
• helper.py: Includes helper functions, such as plotting results.
• model.py: Defines the neural network architecture for the Q-Network and the QTrainer class for training.
• agent.py: Implements the reinforcement learning agent, including methods for training, saving, and loading the model.
• arial.ttf: A font file used for rendering text in the game.

Installation

1. Clone the Repository:

   git clone https://github.com/afrahshah/snake-rl.git
   cd snake-ai

2. Install dependencies

   pip install torch numpy pygame matplotlib
   

Usage

Training the Agent

To start training the agent, run the following command:

ls -l

python agent.py

The training script will:

Initialize the game and the agent. Train the agent through multiple games. Save the model state after training.

Happy coding!