VoiceAfrica
Domaine:
natural language processingType de record:
dataset
Créateur:
nai
Hôte:
Welcome to the VoiceAfrica dataset. VoiceAfrica is a Lanfrica–Meta collaboration (code-named VoiceAfrica 1) that set out to create authentic, conversational speech and expert-curated transcriptions for 11 under-represented African languages. The dataset contains ~125 hours of speech (about 10 hours per language) across 16,383 audio samples from 118 speakers in four countries.