Skip to content

Datasets

Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.

Data type

For which models

Find a dataset for your goal

Answered by Gemini 4 · the chat is kept only in this tab

Datasetturiabu/Sagalee

Read speech dataset for the Oromo language (ASR and TTS)

An open speech corpus for automatic speech recognition and text-to-speech in the low-resource Ethiopian language Oromo.

  • Automatic speech recognition
  • Text → speech
  • Oromo
350 510K–100K rows
DatasetAlex-Song/MSR-86KTop pick

Multilingual speech recognition corpus (86,300 hours)

A large 15-language YouTube audio corpus for training and researching speech recognition systems.

  • Automatic speech recognition
  • Spanish
  • Korean
  • English
  • +9
344 60🔒 on request
Datasetseanghay/km-speech-corpus

Khmer speech corpus

Around 10.4 hours of audio and text pairs in Khmer for training ASR and TTS models.

  • Automatic speech recognition
  • Text → speech
  • Khmer
334 1010K–100K rows
DatasetPorameht/processed-voice-th-169k

Cleaned Thai Speech Dataset for ASR

169k Thai audio recordings with transcriptions for ASR fine-tuning.

  • Automatic speech recognition
  • Thai
318 18100K–1M rows
DatasetNandemoGHS/Japanese-Eroge-Voice

Japanese voice data from visual novels with transcriptions

A large corpus of Japanese audio recordings from games and automatically generated subtitles for TTS and ASR.

  • Text → speech
  • Automatic speech recognition
  • Japanese
317 38100K–1M rows
Datasetalvanlii/cantonese-youtubeTop pick

10k hours of Cantonese YouTube speech

Large-scale corpus of 10,000 hours of Cantonese YouTube audio with pseudo-transcriptions from Whisper and SenseVoice.

  • Automatic speech recognition
  • Audio classification
  • Chinese
  • Cantonese
296 501M–10M rows🔒 on request
Datasetshb777/gemini-flash-2.0-speechTop pick

Synthetic speech from Gemini Flash 2.0 (284 hours)

An audio dataset with paired texts and voiceovers in two voices from Gemini Flash 2.0 for training TTS and ASR.

  • Text → speech
  • Automatic speech recognition
  • English
277 6110K–100K rows
Datasetechonova-ai/arabic-dialect-asr-benchmark

Arabic and 8 dialects speech recognition benchmark

20 hours of annotated Arabic speech in Modern Standard Arabic and eight regional dialects for evaluating ASR systems.

  • Automatic speech recognition
  • Arabic
277 101K–10K rows🔒 on request
Datasetysdede/khanacademy-turkish

Turkish speech corpus from Khan Academy videos

78 hours of short audio clips with Turkish voiceovers and text for ASR and TTS training.

  • Automatic speech recognition
  • Text → audio
  • Turkish
270 3910K–100K rows
Datasetupb-nlp/echo

300 Hours of Romanian Speech by 300 Speakers

Crowdsourced Romanian audio corpus for training speech recognition systems.

  • Automatic speech recognition
  • Romanian
250 5100K–1M rows🔒 on request
Datasetdavidchan/anim400k

425k English and Japanese anime audio and video clips

A large dataset of synchronized video and audio clips for training automated dubbing systems.

  • Text → speech
  • Automatic speech recognition
  • Audio → audio
  • English
  • Japanese
211 43100K–1M rows🔒 on request
Datasetuvci/koumankan4dyula

Speech and translation for the Dyula language

A corpus of 15 hours of audio in West African Dyula with transcriptions and translations in French and English.

  • Automatic speech recognition
  • Translation
  • Bambara
194 1210K–100K rows🔒 on request
Datasetvhdm/persian-voice-v1

Cleaned Persian speech corpus based on Common Voice 17

Version of the Mozilla Common Voice 17 Persian section with corrected spelling errors for speech recognition tasks.

  • Automatic speech recognition
  • Persian
150 3210K–100K rows
Datasetplaylogue/playlogue-v1

Corpus of adult-child conversations

Audio data and texts of natural conversations between adults and preschool children.

  • Automatic speech recognition
  • Audio classification
  • Text classification
  • English
142 13🔒 on request
Datasetdubbing-ai/vaja-thai

Combined Thai speech corpus for TTS and ASR

Nearly 555 hours of cleaned Thai speech in WAV format for voice synthesis and recognition.

  • Text → speech
  • Automatic speech recognition
  • Thai
102 7100K–1M rows🔒 on request