#6 · Primary category: Data Quality & Cleaning
Curator
Scalable data pre processing and curation toolkit for LLMs
Project last updated:08/28/26
GitHub Stars
1.7K
Forks
320
Contributors
70
License
Apache-2.0
Why we included this project
Preparing data for LLM pretraining or fine-tuning usually means writing the same cleanup scripts again and again, from scrubbing noisy web crawls to deduplicating documents. NeMo Curator packages that work as repeatable pipelines that run equally well on a laptop and on a multi-node Ray cluster, so you can test locally and scale up without rewriting anything. The GPU acceleration is the practical payoff, turning jobs that would crawl on a single machine into distributed runs. It also covers image, video, and audio curation, which keeps multimodal and speech work in the same codebase instead of forcing a separate toolchain.
Articles for this project
No articles for this project yet.
To suggest a topic or contribute an article, contact us.
Related projects in this category
cleanlab
Cleanlab's open-source library is the standard data-centric AI package for data quality and machine learning with messy, real-world data and labels.
DataFlow
Easy Data Preparation with latest LLMs-based Operators and Pipelines.
data-juicer
Data processing for and with foundation models! 🍎 🍋 🌽 ➡️ ➡️🍸 🍹 🍷
qsv
Blazing-fast Data-Wrangling toolkit
MNBVC
MNBVC is a massive Chinese corpus benchmarked against ChatGPT's 40T data, covering mainstream and niche cultures, with diverse text forms including news, essays, novels, and more.