Can this model be used for infilling?

#2
by JoaoLages - opened

From what I read in the paper (https://arxiv.org/pdf/2305.07922.pdf), fill-in-the-middle wasn't a task. I'd like to ask why or if it also works like that. FIM is what is missing for encoder-decoder models to just surpass decoder-only

Sign up or log in to comment