Mastering Multilingual GenAI - Open-Weights for Indic Languages

About

This session focuses on developing multilingual Generative AI inspired by the Aya project. You will see a demonstration of how to fine-tune an open-weights aya-23-8B model on a new language using QLORA. The session will emphasize the importance of multilingual capabilities in generative AI, their challenges and opportunities, and practical steps for evaluation and safety considerations.

Key Takeaways:

  • Importance of Multilingual AI: Learn why integrating multiple languages into generative AI models is crucial for creating versatile and inclusive applications.
  • Challenges and Opportunities: Understand the key challenges and opportunities that come with fine-tuning AI models for multiple languages.
  • Evaluation and Safety: Discover the criteria for evaluating multilingual models and considerations for ensuring their safety and user preferences.
  • Practical Fine-Tuning: Gain hands-on knowledge on how to fine-tune an open-weights model on a new language using QLORA.
  • Project Insights: Learn from the Aya project, a long-term open-science initiative, and how it has shaped the approach to multilingual AI.

Speaker

video thumbnail
Book Tickets
Stay informed about DHS 2025

Download agenda

We use cookies essential for this site to function well. Please click to help us improve its usefulness with additional cookies. Learn about our use of cookies in our Privacy Policy & Cookies Policy.

Show details