One of the biggest pain points when handling large, machine learning datasets is being able to easily slice and manipulate labels in various ways. Machine Learning engineers and researchers often need to select a subset of the dataset (i.e. those that contain a particular set of object classes) and tag them to revisit them later or share with colleagues (i.e. "Used to train model v1").
A big benefit of using the Suite is exactly this — the Suite provides features like tag, filter and search to give users (both ML engineers and Labeling teams) the versatility to manipulate labels easily and boost their productivity.
A Label Tag is a customizable marker that can be used to leave reminders or other vital information to oneself or other users. For example, an engineer may choose to notify another engineer that the tagged image data has been used on an AI prototype and does not want the tagged image(s) to be re-used by another engineer who is viewing the same dataset. A manager may choose to mark a set of labels that they must be reviewed for quality assurance purposes.
1. To tag a label, select the labels you would like to tag, then select Edit Tags on the top right of your list.
2. Once you click on the Edit Tags button, you'll be shown a list of existing Label Tags, and whether they have been applied to the selected labels:
An empty checkbox means none of the selected labels have the particular tag.
A checked box means all of the selected labels have the particular tag.
A checkbox with a dash ([-]) means some of the selected labels have the particular tag.
You may either choose from the list of existing tags, or create a new tag by writing out the new tag and then pressing the Enter key.
The Superb AI Suite allows versatile filtering of your labels for maximum productivity and organization. Upon entering Label List, **users may sift through the data using any one of the following filters:
Label Tag Filter
Users can filter labels that have or do not have particular label tags.
Users can filter labels based on which annotations (such as object classes, categories, etc.) are present in the label.
Users can filter labels based on which status each label is in. Please refer to the following user manual for more information on label status: What is a Label?
Date Added Filter
Users can filter labels based on the date the corresponding raw data was added to the specific project being viewed.
Last Updated Filter
This filter is used to find labels that were updated on a chosen date or between a chosen date range. Any changes to labels, such as the annotations, status, label tags, and assignee are all regarded as an update.
Users can filter labels based on the name of the assigned users.
Open Issues Filter
This filter is used to view labels that have open issue threads (issue threads that have not been resolved yet). Users can create issue threads on any label to flag them for any reason (i.e. when an annotation was not done properly, a manager may create an issue thread). Please refer to the following related manuals for more information.
If raw data from multiple datasets have been uploaded to a project, users can use this filter to view the labels that are from a particular dataset.
Users can filter auto-labeled labels, which are marked with blue icons in the Pre-label column, by using the filter feature.
Auto-label Request Filter
This filter allows users to filter out labels by the status of the auto-labeling. Labels are marked with ‘is processing’ once requested, and the status will change to ‘has failed’ if the request is unsuccessful.
To search for labels, go to the search bar at the top right of the screen. There, you will find the Search bar where you can search for labels by the data key. Data key is simply the raw data name that each label is linked to. You can search for the full data key, or search using a partial name.
Any other questions? E-mail us at firstname.lastname@example.org