2048 Game with AI Self-Learning

A Python implementation of the popular 2048 game featuring an AI agent that learns to play through self-training using TensorFlow.

Features

• Classic 2048 game implementation
• AI agent powered by TensorFlow
• Self-learning capabilities through reinforcement learning
• GPU acceleration support
• Real-time visualization of game states
• Performance logging and monitoring
• Training progress visualization using Matplotlib

Requirements

• Python 3.7+
• TensorFlow 2.x
• NumPy
• Pygame
• Matplotlib

Installation

1. Clone the repository:

git clone https://github.com/scar17off/ai-2048.git
cd 2048-ai-self-learning

2. Install dependencies:

pip install tensorflow numpy pygame matplotlib

Project Structure

• self_learn.py: Main self-learning AI implementation
• game_logic.py: 2048 game mechanics
• train_model.py: Model training orchestration
• ai_play.py: AI gameplay execution
• colors.py: Game interface color definitions

Usage

1. Generate training games:

python generate_training_games.py

Or download pre-generated games (dataset) in releases.

2. Train the AI model:

python train_model.py

Or download pre-trained model in releases.

3. Watch the AI play:

python ai_play.py

Technical Details

Running on GPU

The project automatically configures GPU settings for optimal performance. It enables memory growth and selects the primary GPU for processing, as shown in self_learn.py.

To run the project on GPU:

1. Install TensorFlow GPU version:

pip install tensorflow-gpu

2. Install CUDA (https://developer.nvidia.com/cuda-downloads) version 12.0 or later.

3. Run the project as usual.

Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Original 2048 game by Gabriele Cirulli
• TensorFlow team for the excellent machine learning framework