NVIDIA Introduces Nemotron-CC: A Massive Dataset for LLM Pretraining

NVIDIA has introduced the discharge of Nemotron-CC, a groundbreaking 6.3-trillion-token English language dataset designed to advance the pretraining of enormous language fashions (LLMs). This dataset, derived from Widespread Crawl, goals to raise the accuracy and effectivity of LLMs via revolutionary knowledge curation methods, together with the usage of 1.9 trillion tokens of synthetically generated knowledge, in response to NVIDIA.

Enhancing LLM Pretraining

NVIDIA’s initiative addresses a crucial want in LLM coaching, the place the standard of pretraining datasets performs a pivotal position. Whereas current fashions like Meta’s Llama collection have been primarily based on datasets comprising as much as 15 trillion tokens, the precise composition of those datasets stays largely undisclosed. Nemotron-CC seeks to fill this hole by offering the broader group with a high-quality dataset able to supporting each brief and lengthy token horizon coaching.

Conventional datasets usually sacrifice as much as 90% of information to enhance benchmark accuracies, limiting their utility for intensive coaching. Nemotron-CC, nevertheless, demonstrates learn how to remodel Widespread Crawl knowledge right into a superior dataset, surpassing even the Llama 3.1 8B mannequin via superior strategies reminiscent of classifier ensembling and artificial knowledge rephrasing.

Vital Outcomes

Nemotron-CC’s efficacy is evidenced by its efficiency in varied benchmarks. When coaching 8B parameter fashions for one trillion tokens, the high-quality subset Nemotron-CC-HQ outperforms main datasets like DCLM, growing MMLU scores by 5.6 factors. Moreover, the whole 6.3-trillion-token dataset matches DCLM on MMLU whereas providing 4 instances extra distinctive actual tokens. This permits efficient coaching over lengthy token horizons, with Nemotron-CC-trained fashions surpassing Llama 3.1 8B in a number of metrics, together with a 5-point improve in MMLU and a 3.1-point rise in ARC-Problem scores.

Modern Information Curation Methods

The event of Nemotron-CC concerned a number of key insights. By ensembling totally different model-based classifiers, NVIDIA was in a position to choose a broader array of high-quality tokens. Moreover, rephrasing methods decreased noise and errors, yielding various and beneficial knowledge variants. The choice to disable conventional heuristic filters additional boosted the dataset’s high quality with out compromising accuracy.

NVIDIA utilized its NeMo Curator device to extract and refine knowledge from Widespread Crawl, making use of filters for language, deduplication, and high quality classification. This course of was complemented by artificial knowledge technology, contributing roughly two trillion tokens to the dataset.

Future Prospects

Nemotron-CC is positioned as an important useful resource for pretraining state-of-the-art LLMs over various token horizons. NVIDIA plans to increase its choices by releasing extra specialised datasets, together with these targeted on particular domains like arithmetic, to additional improve LLM capabilities.

Picture supply: Shutterstock