I'm curious if there are any examples of fine-tuning the model, either using the transformers version or, ideally the galai library directly. I assume it's not too different from any other model but I'd still feel more confident doing so with some examples.
I'd also be curious to know if there are examples or attempts at training this model specifically to answer questions based on a person's understanding, i.e. either:
Using a dataset where the outputs are the model's own outputs adjusted to be understandable by that person or
Using a dataset of writing and conversation from that person to hopefully pick-up on speech patterns that make sense to them.
I've been toying around with this idea using scibert and a t5 trained on a similar corpus but this model seem superior to both.
I'm curious if there are any examples of fine-tuning the model, either using the transformers version or, ideally the galai library directly. I assume it's not too different from any other model but I'd still feel more confident doing so with some examples.
I'd also be curious to know if there are examples or attempts at training this model specifically to answer questions based on a person's understanding, i.e. either:
I've been toying around with this idea using scibert and a t5 trained on a similar corpus but this model seem superior to both.