[MacsHeadroom]

You can do even better!. You can run the second smallest one (better than GPT-3 175B) on 24GB of vram, ie LLaMA-13B. https://github.com/oobabooga/text-generation-webui/issues/14...