The Koumankan4Dyula corpus consists of approximately 15 hours i.e. 10,929 recordings of Dioula language audio along with the corresponding texts and their translations into French and English.
Dioula is a low-resource language spoken by over 16.4 million people in several West African countries.