Задача с выбором из нескольких вариантов

AbstractIntermediate task fine-tuning has been shown to culminate in large transfer gains across many NLP tasks. With an abundance of candidate datasets as well as pre-trained language models, it has …...