README.md

metadata

dataset_info:
  config_name: read
  features:
    - name: audio
      dtype: audio
    - name: text
      dtype: string
    - name: duration
      dtype: float64
    - name: file_id
      dtype: string
    - name: speaker_id
      dtype: string
  splits:
    - name: train
      num_bytes: 2398892289
      num_examples: 68493
  dataset_size: 2398892289
configs:
  - config_name: read
    data_files:
      - split: train
        path: read/train-*
task_categories:
  - automatic-speech-recognition
language:
  - tt
tags:
  - Tatar
  - ASR
  - Speech
  - Audio
  - Turkic
  - STT
pretty_name: Hackathon Tatar ASR
size_categories:
  - 10K<n<100K
license: cc-by-sa-4.0

Hackathon Tatar ASR

Dataset Summary

Hackathon Tatar ASR is a speech dataset distributed during the "Татар.Бу Хакатон" (Tatar.Bu Hackathon) held in Tatarstan in May 2024. This dataset likely consists of newly collected crowdsourced recordings created after the last release of TatSC (Tatar Speech Corpus), although some intersections with TatSC might be present. While TatSC contains 269.1 hours of transcribed speech with 271,914 utterances, this hackathon dataset comprises 89.99 hours of audio with 68,593 utterances.

Dataset Origin

The dataset was shared during the Tatar.Bu Hackathon 2024, which took place on May 17-19, 2024, at the IT Park in Kazan.

Dataset Structure

Parts:
The dataset contains a single set of recordings, primarily comprised of crowdsourced audio samples.

Data Fields:

• audio: The speech audio file in the Tatar language
• text: The transcription of the audio file in Tatar
• duration: Length of the audio in seconds
• file_id: Unique identifier for the audio file
• speaker_id: Identifier for the speaker in the recording

Data Processing

Audio Processing:
The audio samples range from very short utterances (0.21s) to longer segments (31.1s).

Technical Details

Dataset Type	Speech corpus for ASR
Language	tt, Tatar
Speech Style	Primarily crowdsourced
Content	Varied
Audio Parameters	16 kHz sampling rate
File Format	MP3, TXT (UTF-8)
Number of lines	68,593
Total duration	89.99 hours

Relation to Other Datasets

This dataset complements existing Tatar language resources such as the original TatSC (269.1 hours) and TatarTTS (~70 hours). It represents ongoing efforts to expand speech recognition capabilities for the Tatar language through community participation and hackathon initiatives.