*** description: >- NeMo Curator is an open-source, scalable data curation platform for curating large datasets across text, image, video, and audio modalities to improve AI model training categories: * documentation * home tags: * data-curation * multimodal * scalable * gpu-accelerated * distributed personas: * Data Scientists * Machine Learning Engineers * Cluster Administrators * DevOps Professionals difficulty: beginner content\_type: index modality: universal *** # NeMo Curator Documentation Welcome to the NeMo Curator documentation. ## Introduction to Curator Learn about the Curator, how it works at a high-level, and the key features. Overview of NeMo Curator and its capabilities. target-users how-it-works Discover the main features of NeMo Curator for data curation. features capabilities deployments Explore the core concepts for each modality in NeMo Curator. data-loading data-processing data-generation ## Quickstarts Install and run NeMo Curator for specific modalities. Set up and run text curation workflows. Set up and run image curation workflows. Set up and run video curation workflows. Set up and run audio curation workflows. ## Data Curation Workflows ### Workflow Modalities Explore how you can use NeMo Curator across different content modalities. Curate and prepare high-quality text datasets for LLM training. filtering formatting deduplication Curate image-text datasets with embedding, classification, and deduplication. embedding classification semantic-deduplication Curate and process videos with GPU-accelerated pipelines and sharding. video-splitting video-sharding gpu-accelerated Transcribe, filter, and curate speech and audio datasets with ASR models. asr transcription quality-filtering ## Tutorial Highlights Check out tutorials to get a quick start on using the NeMo Curator library. Learn the basics of text data processing with NeMo Curator. beginner text-processing data-preparation Learn the basics of image data processing with NeMo Curator. beginner image-processing data-curation Learn the basics of video pipeline construction and execution. video-splitting video-sharding custom-pipelines Learn the basics of speech data processing with NeMo Curator. beginner asr-inference quality-assessment ***