[d3nj4l]

I use ollama (https://ollama.ai/) which supports most of the big new local models you might've heard of: llama2, mistral vicuna etc. Since I have 16GB of RAM, I stick to the 7b models.