Skip to main content
Version: Next

Overview

Annotation Tasks are workflows designed to allow your teams to quickly and efficiently apply labels or classes to your training datasets. With Annotation Tasks, you can:

  • Choose specific data points for annotation.
  • Assign annotators who are responsbile for labeling.
  • Track the progress of the task.
  • Review the applied labels or classes and make a selection to commit to Ground Truth.

Workflow overview​

The typical end-to-end flow of uploading, annotating, reviewing, and finalizing Ground Truth for your datasets are simple with Annotation Tasks:

  1. Import the dataset you want to label.
  2. Create an Annotation Task for the dataset.
  3. Select the data points to label.
  4. Select the questions that the annotators should answer.
  5. Assign annotators to the data points.
  6. Annotators go through their assigned data points and apply labels.
  7. Reviewers see the results, and select the final answers to commit as Ground Truth.