Labeling instructions are essential for any large-scale labeling operation whether in-house or outsourced. It is one of the primary means for ML teams to offer the best practices with labeling their data. All good labeling instructions should include good and bad labeling examples. Often these are curated over time through various iterations. Among the most successful labeling operations, we often see ML teams keep labeling instructions up to date.
Get started with creating your own labeling instructions using the template below
Select a project.
Go to Settings.
Click Label editor. Then, click Edit.
Click on the Instructions tab.
Click Attach PDF and select a file to attach.
You can update labeling instructions anytime in the project
Simply attach the new labeling instructions within project settings. New instructions instantly are updated for labelers to view.
Updated 3 days ago