Yes, I am training this Llama model with Nostr repos, and it also requires additional fine-tuning. I am working on a mechanism through which everyone can self host them for free on Colab or Pi.