Phonon-2: most accurate open speech recognition model in a 164 MB downloadfermionresearch.com5 points·handfuloflight··0 commentsOpen articleSaveView on HN