Currently it uses the Nous Hermes Llama 2 (7B). If you’re technical, you can customize the Dockerfile to run a different Llama model: https://github.com/getumbrel/llama-gpt/blob/c76225a6fc26a000fc07...
That is pretty cool but will wait for it to be fleshed out a bit more. Would be interesting to see how this eventually combines with home assistant and whisper. Can steal some good ideas from in regards to models and characters etc. https://github.com/oobabooga/text-generation-webui