intronhealth/afri-names

Afri-names: Read Speech Dataset of Numbers and African Named Entities This work is licensed under aCreative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Overview Afri-names is a curated African-accented read speech dataset comprising 6,307 single-speaker audio samples, totaling 8.92 hours of speech data. Each sample is densely populated with numbers or African named entities or voice commands (with African named entities), making it… See the full description on the dataset page: https://huggingface.co/datasets/intronhealth/afri-names.