Common Voice Scripted Speech 25.0 - Wakhi

Wakhi (Wuk̃hikwor) — Wakhi (wbl)

This datasheet is for cv-corpus-25.0-2026-03-09 of the Mozilla Common Voice Scripted Speech dataset for Wakhi [Wakhi (Wuk̃hikwor) - wbl]. The dataset contains 8192 clips representing 15.36 hours of recorded speech (12.12 hours validated) from 14 speakers, recorded from a text corpus of 5,607 sentences.

Language

Wakhi or Wakhani is indigenously termed as K̃hikwor (contraction of Wuk̃hikwor). It's an old eastern Iranian or Iranic language within the Pamiri branch. Though, diasporas also live in Russia and Turkey as well as in the European, American an Australian continents, the Wakhi (or K̃hikwor) language is spoken indigenously in Pakistan, China, Afghanistan, Tajikistan and Kirghizistan.

Demographic information

The dataset includes the following self-declared age and gender distributions. A coverage summary is shown below each table.

Gender

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.

Gender declared: 968 of 8,192 clips (11.8%), 0 of 14 speakers (0.0%)

Age

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.

Age declared: 6,451 of 8,192 clips (78.7%), 2 of 14 speakers (14.3%)

Data splits for modelling

Clip buckets

Training splits

Training split coverage: 4,780 of 6,464 validated clips (73.9%)

The dataset contains 6464 validated, 89 invalidated, and 1639 unresolved clips. The average clip duration is 6.753 seconds.

Text corpus

The corpus has sentences of Hunza Wakhi and based on extensive anthropological and linguistic fieldwork in Pakistan, China, Afghanistan and Tajikistan.

Validated sentences: 5,493

The corpus contains 5,607 sentences: 5,493 validated and 114 unvalidated (114 pending review, 0 rejected), with 1 reported for review.

Writing system

The script used is Roman Anglicized writing system, which is the approved script by Wakhi Tajik Cultural Association (WTCA), Pakistan, an Ishkoman Wakhi Welfare Organization (IWWO). Through this script, the literate Wakhi people easily and happily interact with each other across the borders on social media forums: it thus facilitates their creativities, thought expression in textual form and binds them together.

Symbol table

D̃d̃ Dh Ee Ẽẽ Ff Gg Gh g̃h Ii Jj J̃j̃ Kk Kh K̃h Ll Mm Nn Oo Pp Qq Rr Ss Sh S̃h Tt T̃t̃ Th Uu Ũũ Vv Ww Yy Zz Zh Z̃hZ̃z̃

Sample

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

1. Yan yi gheyr mulkiyi bimor vitk.

2. Sake Kitob Joyetu

3. Wozi k̃hat: “Wuz beh nomardi taw neh yundem.”

4. Johil zẽmanẽs̃h dowlatẽt sũratẽ dẽstan k̃hũ tat nanẽ lecrẽn

5. Yan, dẽ yi khun ney dẽ yi khun yow goten.

Sources

1. Texts (sentences) made out of my own brain (creation) during the assignment period.

2. Texts out of selected Wakhi poetries.

3. New Wakhi transcriptions (texts) of the interviews out of my extensive fieldwork in Pakistan, china, Afghanistan and Tajikistan

4. Wakhi publications from formal website: www.fazalamin.com

Text domains

General

Fields

Clips

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.tsv

The 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.tsv

The 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)

Get involved

Community links

• Common Voice translators on Pontoon

• Common Voice Communities

Discussions

• Common Voice on Matrix

• Common Voice on Discourse

• Common Voice on Discord

• Common Voice on Telegram

Contribute

• Speak

• Write

• Listen

• Review

Acknowledgements

Datasheet authors

• Mazdak Beg

• Ahmad Jami Sakhi

• Mr. Amanullah

• Fazal Amin Beg

Funding

This dataset was partially funded by the Open Multilingual Speech Fund managed by Mozilla Common Voice.

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 demographics spec. These will only be reported if the speaker opted in to provide that information. ↩ ↩2 ↩3 ↩4