Yorùbá Pronunciation Resource & Speech Evaluation
Building a pronunciation resource and evaluation methods for Yorùbá voice AI. Language representation matters because pronunciation and tone help carry meaning, and people should be able to use voice technology in their own languages. Research and model-adaptation training are ongoing, alongside human listening evaluation, acoustic analysis and annotation-tool development.
Ongoing research · model adaptation & evaluation
The public repository contains code, research reports and metadata. The current resource is not a publicly released or population-representative training corpus. No participant audio or reviewer study portal is offered here as a public demo.