WENETSPEECH: A 10000+ Hours Multi-Domain Mandarin Corpus for Speech Recognition (2021-10-07T00:00:00.000000Z)

TL;DR

WenetSpeech is the current largest open-source Mandarin speech corpus with transcriptions, which benefits research on production-level speech recognition, and a novel end-to-end label error detection approach is proposed.

Abstract

In this paper, we present WenetSpeech, a multi-domain Mandarin corpus consisting of 10000+ hours high-quality labeled speech, 2400+ hours weakly labeled speech, and about 10000 hours unlabeled speech, with 22400+ hours in total. We collect the data from YouTube and Podcast, which covers a variety of speaking styles, scenarios, domains, topics and noisy conditions. An optical character recognition (OCR) method is introduced to generate the audio/text segmentation candidates for the YouTube data on the corresponding video subtitles, while a high-quality ASR transcription system is used to generate audio/text pair candidates for the Podcast data. Then we propose a novel end-to-end label error detection approach to further validate and filter the candidates. We also provide three manually labelled high-quality test sets along with WenetSpeech for evaluation – Dev for cross-validation purpose in training, Test_Net, collected from Internet for matched test, and Test_Meeting, recorded from real meetings for more challenging mismatched test. Baseline systems trained with WenetSpeech are provided for three popular speech recognition toolkits, namely Kaldi, ESPnet, and WeNet, and recognition results on the three test sets are also provided as benchmarks. To the best of our knowledge, WenetSpeech is the current largest open-source Mandarin speech corpus with transcriptions, which benefits research on production-level speech recognition.

Authors

References43 items

The People's Speech: A Large-Scale Diverse English Speech Recognition Dataset for Commercial Usage

Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-EndSpeech Recognition

GigaSpeech: An Evolving, Multi-domain ASR Corpus with 10, 000 Hours of Transcribed Audio

Hubert: How Much Can a Bad Teacher Benefit ASR Pre-Training?

Unsupervised Speech Recognition

WeNet: Production Oriented Streaming and Non-Streaming End-to-End Speech Recognition Toolkit

Cascade RNN-Transducer: Syllable Based Streaming On-Device Mandarin Speech Recognition with a Syllable-To-Character Converter

Recent Developments on Espnet Toolkit Boosted By Conformer

MLS: A Large-Scale Multilingual Dataset for Speech Research

The ASRU 2019 Mandarin-English Code-Switching Speech Recognition Challenge: Open Datasets, Tracks, Methods and Results

Unsupervised Cross-lingual Representation Learning for Speech Recognition

wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations

Conformer: Convolution-augmented Transformer for Speech Recognition

SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

fairseq: A Fast, Extensible Toolkit for Sequence Modeling

Semi-Orthogonal Low-Rank Matrix Factorization for Deep Neural Networks

AISHELL-2: Transforming Mandarin ASR Research Into Industrial Scale

A Time-Restricted Self-Attention Layer for ASR

Speech-Transformer: A No-Recurrence Sequence-to-Sequence Model for Speech Recognition

ESPnet: End-to-End Speech Processing Toolkit

AISHELL-1: An open-source Mandarin speech corpus and a speech recognition baseline

Attention is All you Need

Joint CTC-attention based end-to-end speech recognition using multi-task learning

Detecting Text in Natural Image with Connectionist Text Proposal Network

Purely Sequence-Trained Neural Networks for ASR Based on Lattice-Free MMI

Deep Speech 2 : End-to-End Speech Recognition in English and Mandarin

Listen, attend and spell: A neural network for large vocabulary conversational speech recognition

EESEN: End-to-end speech recognition using deep RNN models and WFST-based decoding

An End-to-End Trainable Neural Network for Image-Based Sequence Recognition and Its Application to Scene Text Recognition

Attention-Based Models for Speech Recognition

Librispeech: An ASR corpus based on public domain audio books

Speech recognition with deep recurrent neural networks

Sequence Transduction with Recurrent Neural Networks

Deep Neural Networks for Acoustic Modeling in Speech Recognition

Context-Dependent Pre-Trained Deep Neural Networks for Large-Vocabulary Speech Recognition

An overview of the SPHINX speech recognition system

Speech recognition

WeNet: Production First and Production Ready End-to-End Speech Recognition Toolkit

Joint CTC/attention decoding for end-to-end speech recognition

Jieba chinese word segmentation tool

The Kaldi Speech Recognition Toolkit

Connectionist Temporal Classiﬁcation: Labelling Unsegmented Sequence Data with Recurrent Neural Networks

Subtitle Change Detection : compute the structural similarity (SSIM) of subtitle area frame by frame until a change is detected at frame t + n . Then