Libri-Light: A Benchmark for ASR with Limited or No Supervision

Kahn Jacob
Kahn Jacob
Rivière Morgane
Rivière Morgane
Zheng Weiyi
Zheng Weiyi
Kharitonov Evgeny
Kharitonov Evgeny
Mazaré Pierre-Emmanuel
Mazaré Pierre-Emmanuel
Karadayi Julien
Karadayi Julien
Liptchinsky Vitaliy
Liptchinsky Vitaliy
Fuegen Christian
Fuegen Christian
Likhomanenko Tatiana
Likhomanenko Tatiana
Synnaeve Gabriel
Synnaeve Gabriel

ICASSP, pp. 7669-7673, 2019.

Cited by: 41|Views70
EI

Abstract:

We introduce a new collection of spoken English audio suitable for training speech recognition systems under limited or no supervision. It is derived from open-source audio books from the LibriVox project. It contains over 60K hours of audio, which is, to our knowledge, the largest freely-available corpus of speech. The audio has been seg...More

Code:

Data:

Your rating :
0

 

Tags
Comments