README.md

metadata

configs: []
license: other
task_categories:
  - automatic-speech-recognition
language:
  - tg

WS2-audio — Unlabeled Speech Segments

Audio-only dataset of VAD-segmented speech clips. No transcripts. See AUDIO_DATASET_SCHEMA.md in the WorldSpeech repo for the full spec.