research
∙
04/10/2022
Breaking Character: Are Subwords Good Enough for MRLs After All?
Large pretrained language models (PLMs) typically tokenize the input str...
research
∙
09/23/2021