📚 About LumoAfrica

This initiative is a product of Zambian Cyber Security Initiative Foundation (ZCSIF). The core objective is to provide a centralized and accessible translation tool for local Zambian languages—bridging the gap between indigenous dialects and English using modern AI models.

🎯 Project Aim

To develop a multilingual translation platform tailored for Zambia’s underrepresented languages, enabling real-time two-way translations and supporting dataset contributions from native speakers for future model training and improvement.

🌍 Currently Supported Languages

• 🇲🇼 Nyanja ↔️ English (via NLLB-200)
• 🇿🇲 Bemba ↔️ English (via NLLB-200)
• 🇰🇪 Swahili ↔️ English
• 🇿🇦 Zulu ↔️ English
• 🇿🇼 Shona ↔️ English
• 🇿🇦 Ndebele ↔️ English

🚧 Languages in Progress

• Lozi
• Luvale
• Tonga
• Kaonde
• Lunda

🧪 Technologies & Models Used

• NLLB-200 (Distilled 600M) for African language translation
• Google Translate API for low-coverage dialects

📁 External Resources & Documentation

• OPUS-MT by Helsinki-NLP
• Meta NLLB-200 Distilled
• Bemba Dataset by CSikasote
• Streamlit Docs

🚀 Project Status (May 2025)

The tool is in its beta phase, supporting multiple languages, with fully functional dataset creation modules, admin features for deletion/export, and ongoing training of additional dialects. Community-driven contributions are being collected to expand training resources for Lozi, Kaonde, Tonga, Lunda, and Luvale.

Empowering African voices through technology — one word at a time.