no code implementations • NAACL (SIGMORPHON) 2022 • Patrick Cormac English, John D. Kelleher, Julie Carson-Berndsen
In recent years large transformer model architectures have become available which provide a novel means of generating high-quality vector representations of speech audio.