AfriMM/AFRICaption

Domain:

natural language processing

Record type:

dataset

Creator:

Afr

Host:

AfriMMD is a multilingual dataset created to enhance linguistic diversity in AI, focusing on African languages. This is a proof-of-concept experiment on the use of multimodal datasets to represent African languages in AI. The dataset contains translations of the captions in the widely-used Flickr8k dataset into 20 African languages. The goal is to address the underrepresentation of African languages

Visit

huggingface.co

Licenses

apache-2.0

AfriMM-ML-Collective/AfriMMD

All files including scripts for translations and final translations of Flickr8k captions in 24 Afric

AFRICAPTION: Establishing a New Paradigm for Image Captioning in African Languages

Multimodal AI research has overwhelmingly focused on high-resource languages, hindering the democrat