arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2110.03370cs.SDcs.CL
收藏

WenetSpeech:一个用于语音识别的10000小时以上多领域汉语语料库

WenetSpeech: A 10000+ Hours Multi-domain Mandarin Corpus for Speech Recognition

Northwestern Polytechnical University(西北工业大学);Mobvoi Inc.(出门问问公司);Beijing Shell Shell Technology Co., Ltd.(北京贝壳科技有限公司)

Binbin Zhang, Hang Lv, Pengcheng Guo, Qijie Shao, Chao Yang, Lei Xie, Xin Xu, Hui Bu, Xiaoyu Chen, Chenchen Zeng, Di Wu, Zhendong Peng

摘要:

在本文中,我们提出了WenetSpeech,一个多领域的普通话语料库,包括10000+小时的高质量标记的语音,2400+小时的弱标记的语音,约10000小时的未标记的语音,总22400+小时。我们从YouTube和播客中收集数据,其中涵盖了各种说话风格,场景,领域,主题和嘈杂条件。介绍了一种基于光学字符识别(OCR)的方法,用于在YouTube数据的相应视频字幕上生成音频/文本分割候选项,而高质量的ASR转录系统用于生成Podcast数据的音频/文本对候选项。然后,我们提出了一种新的端到端的标签错误检测方法,以进一步验证和过滤的候选人。我们还提供了三个手动标记的高质量测试集以及WenetSpeech用于评估- Dev用于训练中的交叉验证目的,Test_Net,从互联网上收集用于匹配测试,Test\_Meeting,从真实会议中记录,用于更具挑战性的不匹配测试。为三个流行的语音识别工具包,即Kaldi,ESPnet和WeNet提供了用WenetSpeech训练的基线系统,并且还提供了三个测试集上的识别结果作为基准。据我们所知,WenetSpeech是目前最大的开源汉语语音语料库,它有利于生产级语音识别的研究。

英文摘要:

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) based method is introduced to generate the audio/text segmentation candidates for the YouTube data on its corresponding video captions, 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-sourced Mandarin speech corpus with transcriptions, which benefits research on production-level speech recognition.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑