Data

Roboflow provides a complete set of tools for managing your computer vision data -- from uploading images and annotations to organizing datasets and generating versioned snapshots for training.

Getting Started

• Create a Project -- Set up a new project in your workspace.
• Upload Data -- Add images, videos, and annotations to your project.
• Upload from Cloud Providers -- Import data from AWS S3, Azure Blob Storage, or Google Cloud Storage.
• Import from Universe -- Download or clone images from Roboflow Universe.
• Image Metadata -- Attach custom key-value pairs to images.

Managing Datasets

• Manage Datasets -- Overview of dataset management features.
• Batches -- Track groups of uploaded images through the annotation pipeline.
• Search -- Find images using text queries, semantic search, and filters.
• Manage Classes -- Lock, remap, or delete annotation classes.
• Tags -- Organize images with tags.
• Annotation Attributes -- Add structured attributes to annotations.
• Download an Image -- Download a single image from your dataset.
• Delete an Image -- Remove images from a project.

Dataset Versions

• Versions Overview -- Learn about point-in-time dataset snapshots.
• Create a Version -- Generate a new dataset version with preprocessing and augmentations.
• Preprocessing -- Prepare images for training with preprocessing steps.
• Augmentation -- Apply augmentations to improve model generalization.
• Delete a Version -- Remove a dataset version.

Analytics and Organization

• Dataset Analytics -- Assess and improve the quality of your dataset.
• Merge Projects -- Combine projects in your workspace.
• Rename a Project -- Change your project name.
• Delete a Project -- Permanently remove a project.
• Project Folders -- Organize projects into folders.
• Folder Permissions -- Control access to project folders.
• Make a Project Public -- Share your project on Roboflow Universe.
• Download a Dataset -- Export your dataset for local use.