Overview

Data annotation on Dataloop refers to the process of labeling datasets (images, video, audio, text, geo-spacial data, lidar, etc.) to train and validate machine learning models. The platform provides an intuitive interface that supports a variety of annotation types and workflows, enabling scalable and collaborative AI development.

Annotator - get started

1. Annotators are usually joined by an invitation to a project. If you're expecting an invitation, check with the project manager. Our customers also have the option to manage their workforce in Organizations, and from there assigned to different projects. In such cases, expect an invitation to the organization, but you'll need to be added to a project before you begin annotating.
2. The project you are currently working on is referred to as the active project. It is indicated as a breadcrumb at the header of the platform. It is simple if you have only one, otherwise, click My dashboard to see available projects and switch between them.
3. In your active project there's a list of My assignments where your work assignment with progress indication, number of items, and more. Double-click an assignment to start working on it.

Annotation work process

The following are the main concepts of annotation work in the Dataloop platform:

1. Your work in a project is contained in Assignments. A manager can create multiple assignments for you in the same project.

2. You work on files or items. The assignment can have all the files allocated to you in advance, or it can contain only a few, and more are added when those are completed.

3. You need to declare finishing work on a file by setting a status on it from the buttons in the bottom-right. The default ones for annotation work are:

   1. Complete - Annotation work is completed.
   2. Discard - The file is disqualified for annotation work.

4. You can browse between files in your assignment using the arrow keys at the bottom of the annotation studio. This can help in learning about edge cases or when waiting for information from your managers, reflecting new knowledge on uncompleted items. However, you are still expected to either Complete or Discard all the files.

Refer to the following pages for more information:

• Annotation studios
• Annotation tools