Train an AI to master the classic Snake game using Q-Learning! This project combines reinforcement learning with interactive visualization, allowing you to observe how an AI agent learns to navigate, collect food, and avoid collisions through trial and error.
- Reinforcement Learning Agent using Q-Learning algorithm
- Model Management: Save/Load trained models with custom names
- Real-Time Hyperparameter Controls:
- Learning Rate (α)
- Discount Factor (γ)
- Exploration Rate (ε)
- Epsilon Decay
- Dual Modes:
- 🏋️ Training Mode: Fast-paced learning with adjustable speed
- 🎮 Play Mode: Watch trained AI perform at normal speed
- Modern UI with dark theme and neon aesthetics
- LocalStorage Persistence for models and settings
- Interactive Statistics and performance metrics
- Collision Detection and optimized pathfinding