Reinforcement Learning in Snake Game 🐍

This project is an implementation of a Reinforcement Learning (RL) agent to play the classic Snake game. The RL agent should learn to navigate, avoid collisions, and collect food by maximizing rewards in a grid world.

Features

• Customizable RL Agent: Train the agent using a neural network for decision-making.
• Interactive Game Environment: Playable GUI and non-GUI versions of the Snake game.
• Live Progress Visualization: Monitor training progress with real-time plots.

Installation

Step 1: Clone the Repository

git clone https://github.com/Alireza2317/Snake_RL
cd rl-snake-game

Step 2: Create virtual environment (optional but recommended)

python -m venv env
source env/bin/activate  # On Windows: env\Scripts\activate

Step 3: Install dependencies

pip install -r requirements.txt

Step 4: Run and train the agent, and watch it play

python app.py

The code will first train the agent with the default configs and then play the game visually. Feel free to change the configs dictionary to change the outcome.

Project structure

.
├── app.py                 # Main entry point for training and playing the game
├── snake.py               # Snake game logic and environment
├── agent.py               # Reinforcement Learning agent implementation

Happy coding :)