ESPnet: End-to-end speech processing toolkit

Shinji Watanabe; Takaaki Hori; Shigeki Karita; Tomoki Hayashi; Jiro Nishitoba; Yuya Unno; Nelson Enrique Yalta Soplin; Jahn Heymann; Matthew Wiesner; Nanxin Chen; Adithya Renduchintala; Tsubasa Ochiai

ESPnet: End-to-end speech processing toolkit

Shinji Watanabe, Takaaki Hori, Shigeki Karita, Tomoki Hayashi, Jiro Nishitoba, Yuya Unno, Nelson Enrique Yalta Soplin, Jahn Heymann, Matthew Wiesner, Nanxin Chen, Adithya Renduchintala, Tsubasa Ochiai

Waseda University Senior High School

Research output: Contribution to journal › Article › peer-review

Abstract

This paper introduces a new open source platform for end-to-end speech processing named ESPnet. ESPnet mainly focuses on end-to-end automatic speech recognition (ASR), and adopts widely-used dynamic neural network toolkits, Chainer and PyTorch, as a main deep learning engine. ESPnet also follows the Kaldi ASR toolkit style for data processing, feature extraction/format, and recipes to provide a complete setup for speech recognition and other speech processing experiments. This paper explains a major architecture of this software platform, several important functionalities, which differentiate ESPnet from other open source ASR toolkits, and experimental results with major ASR benchmarks.

Original language	English
Journal	Unknown Journal
Publication status	Published - 2018 Mar 30

Keywords

Dynamical neural network
End-to-end
Kaldi
Open source software
Speech recognition

ASJC Scopus subject areas

General

Fingerprint Dive into the research topics of 'ESPnet: End-to-end speech processing toolkit'. Together they form a unique fingerprint.

Cite this

@article{38bea485914a4320ac1151b75accb00d,

title = "ESPnet: End-to-end speech processing toolkit",

abstract = "This paper introduces a new open source platform for end-to-end speech processing named ESPnet. ESPnet mainly focuses on end-to-end automatic speech recognition (ASR), and adopts widely-used dynamic neural network toolkits, Chainer and PyTorch, as a main deep learning engine. ESPnet also follows the Kaldi ASR toolkit style for data processing, feature extraction/format, and recipes to provide a complete setup for speech recognition and other speech processing experiments. This paper explains a major architecture of this software platform, several important functionalities, which differentiate ESPnet from other open source ASR toolkits, and experimental results with major ASR benchmarks.",

keywords = "Dynamical neural network, End-to-end, Kaldi, Open source software, Speech recognition",

author = "Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Soplin, {Nelson Enrique Yalta} and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai",

year = "2018",

month = mar,

day = "30",

language = "English",

journal = "Nuclear Physics A",

issn = "0375-9474",

publisher = "Elsevier",

}

TY - JOUR

T1 - ESPnet

T2 - End-to-end speech processing toolkit

AU - Watanabe, Shinji

AU - Hori, Takaaki

AU - Karita, Shigeki

AU - Hayashi, Tomoki

AU - Nishitoba, Jiro

AU - Unno, Yuya

AU - Soplin, Nelson Enrique Yalta

AU - Heymann, Jahn

AU - Wiesner, Matthew

AU - Chen, Nanxin

AU - Renduchintala, Adithya

AU - Ochiai, Tsubasa

PY - 2018/3/30

Y1 - 2018/3/30

N2 - This paper introduces a new open source platform for end-to-end speech processing named ESPnet. ESPnet mainly focuses on end-to-end automatic speech recognition (ASR), and adopts widely-used dynamic neural network toolkits, Chainer and PyTorch, as a main deep learning engine. ESPnet also follows the Kaldi ASR toolkit style for data processing, feature extraction/format, and recipes to provide a complete setup for speech recognition and other speech processing experiments. This paper explains a major architecture of this software platform, several important functionalities, which differentiate ESPnet from other open source ASR toolkits, and experimental results with major ASR benchmarks.

AB - This paper introduces a new open source platform for end-to-end speech processing named ESPnet. ESPnet mainly focuses on end-to-end automatic speech recognition (ASR), and adopts widely-used dynamic neural network toolkits, Chainer and PyTorch, as a main deep learning engine. ESPnet also follows the Kaldi ASR toolkit style for data processing, feature extraction/format, and recipes to provide a complete setup for speech recognition and other speech processing experiments. This paper explains a major architecture of this software platform, several important functionalities, which differentiate ESPnet from other open source ASR toolkits, and experimental results with major ASR benchmarks.

KW - Dynamical neural network

KW - End-to-end

KW - Kaldi

KW - Open source software

KW - Speech recognition

UR - http://www.scopus.com/inward/record.url?scp=85095012494&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85095012494&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:85095012494

JO - Nuclear Physics A

JF - Nuclear Physics A

SN - 0375-9474

ER -