Go for it! Smaller domain specific models are a natural path. The more transparent you are about how you are preparing the data and training the models the better. Help educate people on how this stuff works.

Hi nostr… this message will be encoded in the nostr-LLM one day.