NVIDIA Releases 6 Million Multi-Lingual Reasoning Dataset

NVIDIA releases the 6 Million Multilingual Reasoning Dataset for Nemotron, translating English reasoning data into French, German, Italian, Japanese, and Spanish. The Nemotron Nano 2 9B uses a hybrid Transformer–Mamba architecture with a configurable thinking budget to balance accuracy, throughput, and cost for edge deployments, with weights released under an open license.

Source

HuggingFace Blog

Ingested

AUG 20, 2025 · 19:10

Editorial score

4.0 / 5