Datasets:
| 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. | |