research product . Model . 2021

Librispeech Masked Acoustic Model

Vyas, Apoorv; Madikeri, Srikanth; Bourlard, Hervé;
  • Published: 15 Feb 2021
  • Publisher: Idiap Research Institute
Abstract
We provide a Transformer based pretrained model trained on 1000 hours of Librispeech dataset using the Masked Acoustic Modeling objective.
Persistent Identifiers
Subjects
free text keywords: Self-supervised pretraining, masked acoustic modeling, ASR
Any information missing or wrong?Report an Issue