Team Ai
Datasetpublic

bashkorttele/broadcast-speech

Bashkir Broadcast Speech — Radio and Television 53.3 hours of speech in 293 recordings in the Bashkir language, from television and radio programmes produced by two public broadcasters of the Republic of Bashkortostan. Audio only — no transcripts in this release — which makes the set suitable for self-supervised speech pretraining for a low-resource Turkic language. 🌐 Languages of this card: English · Башҡортса · Русский Part of the Bashkorttele dataset series — preservation… See the full description on the dataset page: https://huggingface.co/datasets/bashkorttele/broadcast-speech.

sourceHugging Facecdla-permissive-2.0updated 1mo agoView on Hugging Face
2likes278downloads

bashkorttele/broadcast-speech · main · files are served by the source, never re-hosted here