Skip to content

Dataset · nvidia

Nemotron-CC-v2 for Pre-training

nvidia/Nemotron-CC-v2

Very large multilingual NVIDIA corpus for LLM pre-training featuring math, code, and Q&A.

What's inside

A text dataset for generative model pre-training: high-quality math and code complemented by multilingual Q&A. Volume ranges from 1 to 10 billion units per metadata. The card links the corpus to support for the NVIDIA Nemotron Nano 2 family, including 9B-v2 and Base.

What you can do with it

  • Pre-training or continued pre-training of LLMs
  • Fine-tuning Nemotron Nano 2-compatible models
  • Evaluating the impact of multilingual Q&A on pre-training
Tasks
Text generation
Size
1B–10B rows
Downloads this month
83.2K
Likes
155
Access
on request on Hugging Face

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("nvidia/Nemotron-CC-v2")