Loading datasets...
Task: ASR
Release Date: 3/23/2026
Format: MP3
Size: 21.38 GB
Share
A collection of read speech recordings in Chinese (China) (汉语（中国大陆）).
Restrictions/Special Constraints
None provided.
Forbidden Usage
It is forbidden to attempt to determine the identity of speakers in the Common Voice datasets. It is forbidden to re-host or re-share this dataset.
Intended Use
This dataset is intended to be used for training and evaluating automatic speech recognition (ASR) models. It may also be used for applications relating to computer-aided language learning (CALL) and language or heritage revitalisation.
zh-CN)This datasheet is for cv-corpus-25.0-2026-03-09 of the Mozilla Common Voice Scripted Speech dataset for Chinese (China) [汉语（中国大陆） - zh-CN]. The dataset contains 851641 clips representing 1073.77 hours of recorded speech (239.14 hours validated) from 7546 speakers, recorded from a text corpus of 60,052 sentences.
| Code | Accent | Clips | Speakers |
|---|---|---|---|
| 110000 | 出生地：11 北京市 | 7,174 (0.8%) | 128 (1.7%) |
| 360000 | 出生地：36 江西省 | 5,638 (0.7%) | 41 (0.5%) |
| 440000 | 出生地：44 广东省 | 4,621 (0.5%) | 94 (1.2%) |
| 230000 | 出生地：23 黑龙江省 | 3,971 (0.5%) | 45 (0.6%) |
| 320000 | 出生地：32 江苏省 | 3,913 (0.5%) | 95 (1.3%) |
| 370000 | 出生地：37 山东省 | 3,701 (0.4%) | 98 (1.3%) |
| 310000 | 出生地：31 上海市 | 3,526 (0.4%) | 59 (0.8%) |
| 330000 | 出生地：33 浙江省 | 2,917 (0.3%) | 87 (1.2%) |
| 210000 | 出生地：21 辽宁省 | 2,904 (0.3%) | 42 (0.6%) |
| 120000 | 出生地：12 天津市 | 2,889 (0.3%) | 19 (0.3%) |
| 510000 | 出生地：51 四川省 | 2,698 (0.3%) | 74 (1.0%) |
| 410000 | 出生地：41 河南省 | 2,275 (0.3%) | 68 (0.9%) |
| 130000 | 出生地：13 河北省 | 1,930 (0.2%) | 58 (0.8%) |
| 350000 | 出生地：35 福建省 | 1,810 (0.2%) | 36 (0.5%) |
| 420000 | 出生地：42 湖北省 | 1,792 (0.2%) | 54 (0.7%) |
| 450000 | 出生地：45 广西壮族自治区 | 1,685 (0.2%) | 24 (0.3%) |
| 340000 | 出生地：34 安徽省 | 1,411 (0.2%) | 42 (0.6%) |
| 500000 | 出生地：50 重庆市 | 1,404 (0.2%) | 21 (0.3%) |
| 140000 | 出生地：14 山西省 | 1,391 (0.2%) | 30 (0.4%) |
| 430000 | 出生地：43 湖南省 | 1,053 (0.1%) | 51 (0.7%) |
| 610000 | 出生地：61 陕西省 | 685 (0.1%) | 37 (0.5%) |
| 220000 | 出生地：22 吉林省 | 534 (0.1%) | 24 (0.3%) |
| 640000 | 出生地：64 宁夏回族自治区 | 414 (0.0%) | 6 (0.1%) |
| 650000 | 出生地：65 新疆维吾尔自治区 | 355 (0.0%) | 18 (0.2%) |
| 460000 | 出生地：46 海南省 | 331 (0.0%) | 2 (0.0%) |
| 150000 | 出生地：15 内蒙古自治区 | 277 (0.0%) | 17 (0.2%) |
| 530000 | 出生地：53 云南省 | 240 (0.0%) | 14 (0.2%) |
| 620000 | 出生地：62 甘肃省 | 182 (0.0%) | 16 (0.2%) |
| 520000 | 出生地：52 贵州省 | 164 (0.0%) | 13 (0.2%) |
| 810000 | 出生地：81 香港特别行政区 | 113 (0.0%) | 4 (0.1%) |
| 710000 | 出生地：71 台湾省 | 85 (0.0%) | 2 (0.0%) |
| 630000 | 出生地：63 青海省 | 5 (0.0%) | 1 (0.0%) |
| 540000 | 出生地：54 西藏自治区 | 5 (0.0%) | 1 (0.0%) |
| - | Other | 3,898 (0.5%) | 79 (1.0%) |
The dataset includes the following self-declared age and gender distributions. A coverage summary is shown below each table.
Self-declared gender information. The table shows clip and speaker counts with percentages. Speakers who did not declare a gender are listed as Unspecified. A dash (-) indicates zero.
| Code | Gender | Clips | Speakers |
|---|---|---|---|
| male_masculine | Male, masculine | 48,479 (5.7%) | 975 (12.9%) |
| female_feminine | Female, feminine | 12,282 (1.4%) | 255 (3.4%) |
| transgender | Transgender | - | - |
| non-binary | Non-binary | 238 (0.0%) | 2 (0.0%) |
| do_not_wish_to_say | Prefer not to say | 536 (0.1%) | 9 (0.1%) |
| - | Unspecified | 790,106 (92.8%) | 6,434 (85.3%) |
Gender declared: 61,535 of 851,641 clips (7.2%), 1,112 of 7,546 speakers (14.7%)
Self-declared age information. The table shows clip and speaker counts with percentages. Speakers who did not declare an age are listed as Unspecified. A dash (-) indicates zero.
| Code | Age | Clips | Speakers |
|---|---|---|---|
| teens | Teens | 11,550 (1.4%) | 230 (3.0%) |
| twenties | Twenties | 41,678 (4.9%) | 877 (11.6%) |
| thirties | Thirties | 11,719 (1.4%) | 192 (2.5%) |
| fourties | Fourties | 2,192 (0.3%) | 59 (0.8%) |
| fifties | Fifties | 165 (0.0%) | 10 (0.1%) |
| sixties | Sixties | 6 (0.0%) | 2 (0.0%) |
| seventies | Seventies | 5 (0.0%) | 1 (0.0%) |
| eighties | Eighties | - | - |
| nineties | Nineties | 30 (0.0%) | 2 (0.0%) |
| - | Unspecified | 784,296 (92.1%) | 6,318 (83.7%) |
Age declared: 67,345 of 851,641 clips (7.9%), 1,228 of 7,546 speakers (16.3%)
Clip buckets
| Bucket | Clips |
|---|---|
| Validated | 189,674 (22.3%) |
| Invalidated | 59,226 (7.0%) |
| Other | 602,741 (70.8%) |
Training splits
| Split | Clips |
|---|---|
| Train | 29,608 (15.6%) |
| Dev | 10,653 (5.6%) |
| Test | 10,653 (5.6%) |
Training split coverage: 50,914 of 189,674 validated clips (26.8%)
The dataset contains 189674 validated, 59226 invalidated, and 602741 unresolved clips. The average clip duration is 4.539 seconds.
Validated sentences: 59,143
| Category | Count |
|---|---|
| Unvalidated sentences | 909 |
| Pending sentences | 22 |
| Rejected sentences | 887 |
| Reported sentences | 1,145 |
The corpus contains 60,052 sentences: 59,143 validated and 909 unvalidated (22 pending review, 887 rejected), with 1,145 reported for review.
There follows a randomly selected sample of five sentences from the corpus.
这可能导致真正的社会不平等和不公正。
京沈公路过境。
平谷区长城列表旨在列出中国北京市平谷区的长城墙体及附属设施。
但高速铁路毋须自行驾车会较为舒适。
归入第五批全国重点文物保护单位直波碉楼。
| Source | Sentences |
|---|---|
| wiki | 54,638 (92.4%) |
| cn | 2,881 (4.9%) |
| Other | 1,623 (2.7%) |
| Code | Domain | Clips | Speakers |
|---|---|---|---|
| general | General | 890 (0.1%) | 153 (2.0%) |
| agriculture_food | Agriculture and Food | 55 (0.0%) | 36 (0.5%) |
| automotive_transport | Automotive and Transport | 75 (0.0%) | 54 (0.7%) |
| finance | Finance | 111 (0.0%) | 53 (0.7%) |
| service_retail | Service and Retail | 58 (0.0%) | 44 (0.6%) |
| healthcare | Healthcare | 139 (0.0%) | 71 (0.9%) |
| history_law_government | History, Law and Government | 394 (0.0%) | 111 (1.5%) |
| media_entertainment | Media and Entertainment | 1,672 (0.2%) | 156 (2.1%) |
| nature_environment | Nature and Environment | 63 (0.0%) | 42 (0.6%) |
| news_current_affairs | News and Current Affairs | 194 (0.0%) | 74 (1.0%) |
| technology_robotics | Technology and Robotics | 249 (0.0%) | 88 (1.2%) |
| language_fundamentals | Language Fundamentals | 102 (0.0%) | 61 (0.8%) |
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
accents - accents of the speaker1
variant - variant of the language1
segment - if sentence belongs to a custom dataset segment, it will be listed here
prompt_upvotes - number of upvotes the sentence prompt received
prompt_reports - number of reports the sentence prompt received
is_edited - whether the clip's transcription has been edited
validated_sentences.tsvThe validated_sentences.tsv file contains one row per validated sentence in the text corpus:
sentence_id - unique identifier for the sentence
sentence - the sentence text
variant - the variant of the language
sentence_domain - the domain(s) the sentence belongs to
source - the source the sentence was collected from
is_used - whether the sentence is still in circulation for recording
clips_count - number of clips recorded for this sentence
unvalidated_sentences.tsvThe unvalidated_sentences.tsv file contains one row per unvalidated sentence in the text corpus:
sentence_id - unique identifier for the sentence
sentence - the sentence text
variant - the variant of the language
sentence_domain - the domain(s) the sentence belongs to
source - the source the sentence was collected from
up_votes - number of upvotes the sentence received
down_votes - number of downvotes the sentence received
status - current status of the sentence (pending or rejected)
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.
For a full list of age, gender, and accent options, see the demographics spec. These will only be reported if the speaker opted in to provide that information. ↩ ↩2 ↩3 ↩4