One of the most powerful workflows in Nucleus is the ability to directly submit datasets to a Scale annotation project. This enables you to immediately fix any dataset or model issues that you have discovered by using Nucleus. Any annotations completed in Scale annotation projects are automatically synchronized with Nucleus, so you’ll see the new annotations updated in real-time.
Pre-reqs: Dataset, (slice created)
- Create a slice or navigate to an existing slice, see here
- Click on “Send to Annotation” in the top right corner
- Select the target project and the type of batch
- Click “Create tasks”
If you send data for annotation that has already been labeled by Scale, once the new annotations are ready, they will replace your existing annotations in Nucleus if the target project is of the same task type as your previous Scale annotations. This is useful if you want to send annotations to a project with an improved taxonomy for example. If the target project is of a different task type from your previous annotations, any new labels will simply be added to your existing data once they are available.
Updated over 1 year ago