Using vLLM to get an LLM running fast locally (live stream)

I got Llama 3.1 running on my local machine, and now I will try to speed up inference by using vLLM.