Feel free to experiment with the CUDA implementation in forks of this repo. However, I will not accept any pull requests due to time constraints and since this repo is not actively maintained. Instead, consider it as an educational resource alongside the published paper and YouTube video.