> This project provides a Docker-based inference engine for running Large Language Models (LLMs) on AMD GPUs.

First sentence of the README in the repo. Was it somehow unclear?