This Python repository contains a testing environment and implementation of the NEAT algorithm for the classic game of Snake.
The Snake game is played on a 10x10 map. Each time the snake eats an apple, its length increases and the fitness score is incremented by 1. The goal is to achieve a fitness score of 20.
To install the required dependencies, run the following commandpip install -r requirements.txt
There are several options available for running the Snake game:
- Test Environment: To test the environment, execute the following command:
python main.py test - Train Snake: To train the snake using the NEAT algorithm, use the following command:
python main.py train - Save and Create Winner Model: To save and create a winner model after training, run the following command:
python main.py save - Run Trained Model: To run a trained model, execute the following command:
python main.py run - Run Master Model: To run the master model, execute the following command:
python main.py run_master
Note: The last two commands are the same, but provided separately for clarity.
- Daniel Chang - Snake env logic and network render reference.
Feel free to contribute to this project by submitting pull requests or suggesting improvements. Happy coding!