Fine-Tuning NVIDIA Nemotron to Know Local Rules
According to NVIDIA Developer. NVIDIA releases Nemotron-Personas datasets for France and Belgium to improve model adaptation for local markets.
- • NVIDIA built Nemotron-Personas-France with Pleias and Nemotron-Personas-Belgium with Pleias and KU Leuven.
- • Datasets provide synthetic residents grounded in regional demographics, geography, language, and culture.
- • The Nemotron repository supplies reusable steps for data generation, training, and evaluation.
- • Pleias applies the persona datasets in its Lamina spine-care assistant to expand coverage.
Introduction to Nemotron-Personas
NVIDIA built Nemotron-Personas-France with French AI startup Pleias, and Nemotron-Personas-Belgium with Pleias and Belgian university KU Leuven. Both are part of the open NVIDIA Nemotron data program, which includes locale-grounded persona datasets for 12 countries that developers can use to adapt open models market by market.
Workflow and Application
The Nemotron repository supplies reusable steps for data generation, training, and evaluation, plus agent skills that help coding agents assemble the workflow. Pleias applies the France and Belgium persona datasets in its Lamina spine-care assistant to expand coverage from limited real patient cases and test across diverse patient profiles.
- 1. developer.nvidia.comOpen source