The new Classes tab contains all classes created across all datasets in your V7 team.
This tab is visible to Users, Admins, and Team Owners, and can be filtered by class type, and sorted by class name or date added.
Classes can be created from the team-wide Classes page, or from within a dataset. Check out the section below to see how classes created from the Classes page can be added to an individual dataset.
Each class can be added with up to 5 thumbnails imported from your local device.
To accommodate class management at a team-level, we identified any classes created before the release of team-wide classes that shared identical names in different datasets, and merged them. All sub-annotation settings (Attributes, Instance IDs, Text) between classes have been combined in classes that share the same name.
The Classes tab in each dataset is broken down into two sections - the classes that have been added to the dataset itself, and the team-wide classes in the Add more classes section.
Classes must be added to a dataset in order to appear in the Select class dropdown in the workview:
Team members with Admin, and User permissions can also add classes directly from the Select class dropdown above.
Classes added in the Select class dropdown will be automatically included in the available classes for the dataset from which they are added, and will be available in the team-wide Classes page so that they can be quickly added to additional datasets.
Classes can be removed from a specific dataset, or deleted permanently from the team-wide Classes page.
In a dataset's Classes page, clicking Remove will retain all existing annotations made with a class, but remove the class from the Select class dropdown, as well as the classes filter in the data page.
Removed classes can easily be re-added from the Add more classes section of a dataset's Classes page.
In the team-wide Classes page, select one or more classes to delete them. As with any deletion, be careful: Deleting classes will remove all annotations using that class across all datasets within your team. This is irreversible.
Updated about 2 months ago