Magic Snake - Self-Aware DQN Snake Game

A reinforcement learning project where a snake learns to play the classic Snake game using Deep Q-Networks (DQN). The snake features self-awareness, allowing it to modify game parameters based on its performance.

Features

• Deep Q-Network agent with CNN layers
• Self-modifying game parameters
• Text-to-speech snake thoughts
• Voice selection menu
• Performance visualization

Requirements

See requirements.txt

Usage

Options

• --episodes N: Number of episodes to train (default: 1000)
• --grid-size N: Size of the grid (default: 20)
• --cell-size N: Size of each cell in pixels (default: 20)
• --fps N: Frames per second (default: 10)
• --load-model PATH: Path to model to load
• --no-gui: Disable GUI
• --no-speech: Disable speech
• --select-voice: Force voice selection

Project Structure

• magic_snake.py: Main script
• saved_models/: Directory for saved models
• snake_self_awareness.json: Self-awareness data