V0 contains read-aloud and prompt-based recordings from one adult Swahili speaker from the Democratic Republic of the Congo, two adult Yoruba speakers, and two Yoruba child participants aged 10–15. It is intended for development of the Ntina TaRL reading-assessment benchmark.

Development status

This is a public development V0 dataset. Contributor consent is confirmed for all speakers, and guardian consent is confirmed for both child… See the full description on the dataset page: https://huggingface.co/datasets/Ash11/ntina-voice-dataset-v0.