Is it possible to train a model on a specific hit repository?
I'm working a lot on Ceph specifically. I have used ollama a year ago and concluded that the available models spat out more nonsense than anything else when asking stuff about Ceph in particular. It hallucinated well over 80% of the commands I asked it for. That's not helpful at all.
So my idea would be to "augment"/"train" any reasonable model that happens to be good at coding with the documentation of the Ceph git repository, which also contains its documentation.
Is such a thing possible at all with ollama? Or do I need extra tooling to do this? Eg. OpenWeb-UI?