Unlike many service providers and in-house labeling tools, you can start using Labelbox to create training data in a matter of minutes. To help you get familiar with our platform, we’ve created three short videos to introduce you to Labelbox and show you how our platform is set up.
Organizations & projects
Get a quick introduction to how organizations, projects, and roles work within Labelbox.
- Organizations: This is intended to be the account for your enterprise or organization. You can only be logged into one organization at a time.
- Projects: You can create multiple projects within an organization. We recommend making separate projects for labeling different data types (i.e. video vs image) and creating training data for different models.
- Roles: Assigning a role to each person with access to your organization's account can help secure your projects.
Adding & managing datasets
Learn about the different options for uploading data into Labelbox and managing data within Labelbox. To bring data into the platform, you can:
- Use Delegated Access to store your data in your preferred cloud storage and securely connect it to Labelbox
- Connect your cloud data via signed URLs
- Upload data directly to the Labelbox cloud
This video also covers the metadata that Labelbox creates for each asset uploaded, labeled, and iterated on within the platform.
Preparing your labeling workflow
You can set up your labeling workflow within minutes in Labelbox. The basic steps are:
- Create a project
- Select a dataset
- Choose a labeling interface
- Create or attach an ontology
- Set up optional quality management workflows
In this video, you'll also learn about the configurable labeling interfaces, ontology building and cloning options, and quality management features.