CUNY-CL/yoyodyne

Non-autoregressive sequence generation

Open

#262 geöffnet am 25. Okt. 2024

Auf GitHub ansehen
 (0 Kommentare) (0 Reaktionen) (1 zugewiesene Person)Python (21 Forks)auto 404
enhancementgood first issuenew architecture

Repository-Metriken

Stars
 (35 Stars)
PR-Merge-Metriken
 (PR-Metriken ausstehend)

Beschreibung

Right now, our dominant paradigm is autoregressive prediction. While this is SOTA ,it doesn't allow experimentation with other sequence-to-sequence tasks. Technically, we can define an arbitrary sequence length >>>> than any input or output and train a model to project an input to this length and then decode over it for non-autoregressive prediction. Example algorithms would be: Listen Attend Spell or CTC.

Contributor Guide