Speech Recognition Transformer Decoding Acceleration Method with Discarding Redundant Blocks

Transformer and its variants have become mainstream models in the field of speech recognition owing to their excellent contextual modeling capabilities. Although they can achieve good recognition results, the decoding speed is limited because the decoder uses all encoder features including redundant...

Fuld beskrivelse

Bibliografiske detaljer

Hovedforfatter:	Dechun ZHAO, Yang SHU, Ling LI, Huan CHEN, Zihao ZHANG
Format:	Artigo
Sprog:	Inglês
Udgivet:	Editorial Office of Computer Engineering 2023-10-01
Serier:	Jisuanji gongcheng
Fag:	speech recognition|transformer decoder|ctc model|feature compression|decoding acceleration
Online adgang:	https://www.ecice06.com/fileup/1000-3428/PDF/20231012.pdf
Tags:	Tilføj Tag Ingen Tags, Vær først til at tagge denne postø!