Monolingual speech recognition models beat multilingual models ~30x biggerarxiv.org3 points·dbreunig··0 commentsOpen articleSaveView on HN