Multilingual African Image Dataset with 13 Languages

Afri-Aya is a community-curated multilingual image dataset covering 13 major African languages with AI-powered categorization. It was created by the Cohere Labs Regional Africa community during the six-week Expedition Aya open-build challenge to include low-resource languages in vision-language tasks.

Use Cases

• Train vision-language models on image-text pairs across 13 African languages like Yoruba (yo) and Igbo (ig).
• Fine-tune models for visual question answering tasks using culturally relevant images and questions.
• Build multilingual image captioning systems for languages such as Hausa (ha) and Zulu (zu).
• Analyze AI-powered categorization labels for image content across diverse African cultural contexts.

Strengths

• Covers 13 major African languages, including Yoruba (yo), Igbo (ig), Hausa (ha), and Zulu (zu).
• Created through a structured, community-driven effort during the six-week Expedition Aya challenge.
• Includes AI-powered categorization for the image data.

Limitations

• Specific dataset size, row count, and file formats are unknown.
• The scope and balance of images per language are not detailed, which may affect model training.
• As a community-curated project, data consistency and labeling accuracy may vary.

Provenance

Source

CohereLabsCommunity on Hugging Face.

Collection Method

Community-curated during the Expedition Aya open-build challenge hosted by Cohere Labs.

Freshness

Last updated on 2025-09-17.

Geography

Africa, with a focus on 13 major African languages.