Datasets | Mozilla Data Collective

Sociocultural dataset for the Kanuri community of Cameroon: identity, kinship, religion, economy, onomastics and a linguistic component.

A Kanuri Sociocultural Dataset

License: NOODL-1.0

Locale: knc

Task: OTH

Format: TSV, PDF

Size: 77.79 KB

Filtered Sindhi translation of the Stanford Alpaca instruction tuning dataset for Sindhi LLM instruction fine tuning.

Sindhi Alpaca

License: CC-BY-NC-SA-4.0

Locale: snd

Task: NLG

Format: Parquet

Size: 18.94 MB

Curated collection of 10,300 Sindhi personal names in Arabic script for NLP and language modeling tasks.

Sindhi Names by Proxima AI

License: CC-BY-NC-SA-4.0

Locale: snd

Task: NLG

Format: CSV

Size: 57.15 KB

English medical speech dataset of 11,563 audio-text pairs covering clinical terminology and medical education content.

English Medical Speech Dataset

License: CC-BY-NC-SA-4.0

Locale: en

Task: ASR

Format: Parquet

Size: 5.69 GB

Synthetic Urdu sentences audio dataset of 85,327 verse style audio-text pairs.

Synthetic Urdu 2

License: CC-BY-NC-SA-4.0

Locale: urd

Task: TTS

Format: Parquet

Size: 14.80 GB

Curated Urdu language legal question answer corpus covering Pakistani family law, religious ordinances, and criminal statutes.

Urdu Legal Instruct

License: CC-BY-NC-SA-4.0

Locale: urd

Task: NLG

Format: Parquet

Size: 132.33 KB

Urdu translation of Anthropic's HH-RLHF preference dataset for RLHF/reward-model research in Urdu.

Helpful and Harmless RLHF Urdu

License: CC-BY-NC-SA-4.0

Locale: urd

Task: NLP

Format: Parquet

Size: 10.37 MB

Urdu translation of the Financial PhraseBank sentiment dataset, covering financial news sentences labeled by polarity.

Urdu Financial Phrasebank

License: CC-BY-NC-SA-4.0

Locale: urd

Task: NLP

Format: Parquet

Size: 194.21 KB

Bahasa Bahana

BAHANA-Speech Corpus of English Learners from Indonesia

A corpus of read speech by learners of English from Indonesia.

License: CC-BY-SA-4.0

Locale: en

Task: CALL

Format: WEBM, WAV, TSV

Size: 1.14 GB

Taruen

Kumyk Proverbs and Sayings Text Corpus

A corpus of traditional Kumyk proverbs and sayings extracted from a 2015 dictionary, designed for NLP tasks and LLM cultural competence evaluations.

License: CC0-1.0

Locale: kum

Task: NLP

Format: TXT

Size: 86.48 KB

Synthetic Urdu conversational speech dataset of 7,963 audio text pairs featuring casual, humorous dialogue exchanges.

Syntheic Urdu Audios

License: CC-BY-NC-SA-4.0

Locale: urd

Task: TTS

Format: PARQUET

Size: 4.58 GB

Sociocultural dataset for the Bum community of Cameroon: identity, kinship, religion, economy, onomastics and a linguistic component.

A Bum Sociocultural Dataset

License: NOODL-1.0

Locale: bmv

Task: OTH

Format: TSV, PDF

Size: 84.33 KB

Sociocultural dataset for the Foulbe / Pular community of Cameroon: identity, kinship, religion, economy, onomastics and a linguistic component.

A Foulbe / Pular Sociocultural Dataset

License: NOODL-1.0

Locale: fub

Task: OTH

Format: TSV, PDF

Size: 76.62 KB

Sociocultural dataset for the Gbaya community of Cameroon: identity, kinship, religion, economy, onomastics and a linguistic component.

A Gbaya Sociocultural Dataset

License: NOODL-1.0

Locale: gya

Task: OTH

Format: TSV, PDF

Size: 83.68 KB

Safety aligned Urdu medical question answer dataset of 3,072 records for supervised fine-tuning of medical-domain LLMs.

Urdu Medical SFT

License: CC-BY-SA-4.0

Locale: urd

Task: NLG

Format: PARQUET

Size: 357.97 KB

This dataset consists of segmented Bamun (Shupamem) speech audio clips paired with text, read by a female native speaker, designed for TTS applications.

Bamun-TTS-Dataset (female voice)

License: NOODL-1.0

Locale: bax

Task: TTS

Format: WAV, TSV

Size: 976.53 MB

MDC DatasetsCompensated

🚧 Single Speaker Urdu TTS dataset (Female Speaker, South Punjab) 🚧

A high-quality single-speaker Urdu text-to-speech (TTS) dataset with 25 hours of paired speech recordings and transcriptions for speech synthesis.

License: MDC DLA-1.0

Locale: ur

Task: TTS

Format: WAV, TSV

Size: 6.02 GB

MDC DatasetsCompensated

🚧 Hazargi-English Parallel Text Corpus 🚧

A Hazargi-English parallel corpus derived from transcribed and translated video recordings

License: MDC DLA-1.0

Locale: haz, en

Task: MT

Format: TSV

Size: 434.04 KB

Multimodal dataset for Badja (Badjia), an Ewondo/Beti dialect of Cameroon: lexical datasheet, native-speaker audio, and audio-sentence mapping files.

Badjia-ALCAM-MultimodalDataset

License: NOODL-1.0

Locale: ewo

Task: NLP

Format: WAV, TSV

Size: 42.77 MB

Arabic Common Voice scripted speech 26.0 segment of Female speakers.

Common Voice Scripted Speech 26.0 - Arabic (Female)

License: CC0-1.0

Locale: ar

Task: ASR

Format: MP3, TSV

Size: 185.04 MB

Arabic Common Voice scripted speech 26.0 segment of Male speakers.

Common Voice Scripted Speech 26.0 - Arabic (Male)

License: CC0-1.0

Locale: ar

Task: ASR

Format: MP3, TSV

Size: 235.91 MB

Portuguese Common Voice scripted speech 26.0 segment of Brazilian Portuguese accents.

Common Voice Scripted Speech 26.0 - Brazilian Portuguese

License: CC0-1.0

Locale: pt-BR

Task: ASR

Format: MP3, TSV

Size: 122.18 MB

Chinese Common Voice scripted speech 26.0 segment of Wu Chinese accents.

Common Voice Scripted Speech 26.0 - Wu Chinese

License: CC0-1.0

Locale: zh-CN-shanghai

Task: ASR

Format: MP3, TSV

Size: 191.22 MB