🌸 Run large language models at home, BitTorrent-style. Fine-tuning and inference up to 10x faster than offloading - View it on GitHub