vnahata/AfriMCQA-qa-retrieval

Afri-MCQA visual QA retrieval (MTEB) Culturally grounded multiple choice questions about photographs, in 16 African languages. A query is a question with its photograph; the corpus holds every distinct answer option in that language, so a question is ranked against the whole option pool rather than only its own four. Built from Atnafu/Afri-MCQA at revision 8b8c53d, cc-by-nc-4.0, using the official dev split, the only one where the correct answer is labelled. Built by… See the full description on the dataset page: https://huggingface.co/datasets/vnahata/AfriMCQA-qa-retrieval.