Generate datasets to check the quality of your image annotations in a grid view. This feature works with images that were uploaded to SuperAnnotate or were linked from external storages.

Generate dataset

To generate a project’s first dataset:

1. In Data, click Explore.
2. In the popup, type a dataset version name (mandatory) and a version description (optional).
3. Click Generate.

🚧

You can generate datasets only for Vector Projects.

If a project already has a dataset:

1. In Data, click Explore.
2. Click New Dataset.
3. Type a dataset version name (mandatory) and a version description (optional).
4. Click Generate.

When the dataset generation completes successfully, click the dataset version to go to Explore. Explore features all the project’s images.

📘

The popup shows the following information regarding each dataset version: dataset name, dataset description (if available), number of images, and creation date and time.

❗️

• You can't generate more than 50 datasets per project.
• You can't generate multiple datasets at once per project.

Publish dataset

You can make a dataset public. Anyone with the link of a published dataset can see the dataset and run a query.

To publish a dataset:

1. In the top right corner of Explore, click Publish.
2. In the popup, click Publish.
3. Click Copy link to copy the dataset’s link and share it.

📘

• When a dataset is published, a view count is displayed in the top bar. The view count is available in the editor and viewer modes.
• The view count shows the number of users who have opened the public dataset’s link.

🚧

• The URL expires 2 years after being generated.
• When the dataset is updated, the viewers will also be able to see the changes.

Rename dataset version

To rename a dataset version:

1. Click Explore.
2. Hover over the version you want to rename.
3. Click the edit button.
4. Type the dataset version’s new name.
5. Click Save.

Viewer mode

• Viewers can see the view count in the top bar.
• Viewers can't go to the editor when they click on an image.
• Viewers can't see the following information: image name, image status, copy button to copy the image name, QA name, and Annotator name.
• Viewers can't run a query with the following fields:
  • Instances (createdAt, createdBy, updatedAt, updatedBy)
  • Metadata (lastAction, annotatorEmail, qaEmail, status, pinned)
  • Comments
  • Folders

Unpublish dataset

To unpublish a dataset:

1. In the top right corner of Explore, click Published.
2. Click Stop publishing.

🚧

When a dataset is unpublished then published again, the previous URL will be invalid. So, you'll have to share the new URL.

Dataset statuses

A dataset can have 3 statuses: In progress, Complete, and Failed.

Dataset permissions

As a Team Owner, you can give a Team Admin permission to generate a dataset. Learn more.

📘

Only Team Owners and Team Admins can generate and explore datasets.

Dataset limitations

• You can't generate more than one dataset. Clicking Explore after generating a dataset will overwrite the current dataset.
• You need at least 8 images in the project to be able to generate a dataset.

Delete dataset version

To delete a dataset version:

1. In Data, click Explore.
2. Hover over the version you want to delete.
3. Click the delete button.
4. In the popup, click Delete.