A

Afro Xlmr Small

Developed by Davlan
A compact variant of XLM-R optimized for 17 African languages, enhanced through vocabulary reduction and multilingual adaptive training
Downloads 33
Release Time : 4/12/2022

Model Overview

This model is a lightweight version based on XLM-R-base, specifically optimized for African languages. Through vocabulary reduction and masked language modeling adaptation training, it delivers excellent performance in African language NLP tasks

Model Features

African Language Optimization
Specifically optimized for 17 major African languages, significantly improving NLP task performance for these languages
Lightweight Design
Reduced vocabulary from 250k to 70k, substantially decreasing model size while maintaining performance
Cross-Lingual Transfer Capability
Supports knowledge transfer between multiple African languages, enhancing performance for low-resource languages

Model Capabilities

African language text understanding
Named entity recognition
Text classification
Cross-lingual transfer learning

Use Cases

Natural Language Processing
African Language Named Entity Recognition
Identifying named entities in African languages on the MasakhaNER dataset
Achieved F-scores exceeding 90 in languages like Hausa, outperforming base models
African Language Text Classification
Topic classification or sentiment analysis for African language texts
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase