Hosted on MSN

I stopped running the biggest local LLM that could fit, and a 2B model handles 90% of what I need

If you've got a regular consumer GPU and have been curious about running a local LLM, there's a good chance you've either ...