Seamless: Meta's New Speech Models

Re:Seamless: Meta's New Speech Models by Jayakumark on Friday December 01, 2023 @11:11AM (#38488338)

How does this compare to whisper-large-v3 on STT?

[ Parent | on HN ]

Re:Seamless: Meta's New Speech Models by trovas on Friday December 01, 2023 @11:47AM (#38488860)

they evaluate it in the paper. M4Tv2 is significantly ahead (Whisper Large v3 - 16.9 BLEU vs. M4Tv2 26.6). These are averages over 81 directions X->english

[ Parent | on HN ]