Common Voice Scripted Speech 23.0 - Chinese (Hong Kong SAR China)

Locale: zh-HK

Size: 3.40 GB

Task: ASR

Format: MP3

License: CC-0

中文（香港） — Chinese (Hong Kong) (zh-HK)

This datasheet has been generated automatically, we would love to include more information, if you would like to help out, get in touch!

This datasheet is for version 23.0 of the the Mozilla Common Voice Scripted Speech dataset for Chinese (Hong Kong) (zh-HK). The dataset contains 123220 clips representing 143 hours of recorded speech (108 hours validated) from 3093 speakers.

Demographic information

The dataset includes the following distribution of age and gender.

Gender

Self-declared gender information, percentage refers to the number of clips annotated with this gender.

GenderPertentage
Undefined37.0%
Male Masculine41.0%
Female Feminine22.0%

Age

Self-declared age information, percentage refers to the number of clips annotated with this age band.

Age BandPercentage
Undefined38.0%
Twenties33.0%
Thirties11.0%
Teens2.0%
Fourties12.0%
Fifties3.0%
Sixties1.0%

Data splits for modelling

The official data splits for modelling this language are as follows. Of the validated clips, 20.92% are included in the splits.

SplitCount
Train8450
Test5604
Dev5604

Text corpus

The text corpus contains 20169 sentences, of which 20144 are validated, 25 are invalidated and 824 are reported.

Sample

There follows a randomly selected sample of five sentences from the corpus.

七
六
唔係
八
一

Text domains

DomainCount
Undefined122093
Automotive Transport26
Finance41
Service Retail255
General746
Healthcare82
History Law Government56
Language Fundamentals79
Media Entertainment47
Nature Environment46
News Current Affairs23
Technology Robotics35

Fields

Each row of a tsv file represents a single audio clip, and contains the following information:

  • client_id - hashed UUID of a given user

  • path - relative path of the audio file

  • text - supposed transcription of the audio

  • up_votes - number of people who said audio matches the text

  • down_votes - number of people who said audio does not match text

  • age - age of the speaker1

  • gender - gender of the speaker1

  • accent - accent of the speaker1

  • segment - if sentence belongs to a custom dataset segment, it will be listed here

Community links

Contribute

Licence

This dataset is released under the Creative Commons Zero (CC-0) licence. By downloading this data you agree to not determine the identity of speakers in the dataset.

Footnotes

  1. For a full list of age, gender, and accent options, see the demograpics spec. These will only be reported if the speaker opted in to provide that information. ↩ ↩2 ↩3