Snake game reinforcement learning

Example of a snake game ML with Ray RLlib, PyGame, and Gymnasium.

Check out on this Medium article

Training

To train the network, check out the Ray_RLlib_PyGame_Snake_Gym_Training.ipynb notebook.

Environement

In order to run these files, you need a Python environment with few libraries installed.

Check the requirements.txt.

Conda environment for Ray with Pytorch, Nivida GPU:

conda create --name ray_torch python=3.9
conda activate ray_torch
conda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia
pip install pygame gymnasium opencv-python ray ray[rlib] ray[tune] dm-tree pandas scipy lz4

Inference

You can do inference on a PC by adding your network checkpoint to the ray_checkpoints and modifying the checkpoint_path inside the rayInferenceSnake.py file.

Run game

If you simply want to play the snake game with WASD, run playSnake.py. R for reset, T to kill the game.