Open-source, lightweight, and fast model server for running large language models (LLMs) on your own hardware.