[chpatrick]

I remember seeing a post about someone running the full size DeepSeek model in a dual-Xeon server with a ton of RAM.