[orthoxerox]

It's nice that Unsloth has already published the model on HF, but it requires a fork of llama.cpp to run at the moment.