Use cases for tagging data

Tags enable you to quickly return to specific data items you found relevant. This can help you:

1. Data curation: You can select a subset of data relevant for a certain task such as training a model, and call it "representative" or use it to classify the slice, such as "train/test/validation"
2. Data cleaning: You can select the data you want to remove, and call it "low quality", "duplicates" or "outliers"
3. Workflow management: Highlight priority items with tags like "urgent" or "critical.", or assign states like "for review", "approved" and "rejected" when you review mislabeled items for example
4. Experiment Tracking: Tag datasets used in experiments (e.g., "Experiment-123") to track inputs and outputs

Managing Tags of multiple items

1. Start by opening a Dataset and exploring the data it contains
2. Select the Clusters, Images and Object you wish to tag

3. The selected items are added to "Selected Items" menu at the top right

4. Open the "Select items" menu and click on the "Manage Tags" button to associate and existing User tag to the items displayed in the list

4. Assign one or more User Tags to the selected items

Listing/removing Tags

1. Any tags assigned to the Selected Items can be viewed in the Assigned Tags Menu

2. Selecting a tag would allow you to remove it from any items that it's attributed to - even if the tag isn't applied to all images own a selected cluster, for example.

Managing Tags of a specific Image

User tags can also be applied directly to an image via the Image Details Page:

Clicking on "Assign User Tags" on the right pane will open a menu where you can either create a new tag or apply an existing one:

Any tags assigned to an image can be viewed in the right pane:

To remove a tag directly from Image View, click the X next to the tag name.