Dataset Enrichment Flow
Select models to enrich your dataset and
Overview
Enriching your dataset is a 2-step flow:
- Select Models
- Enrich Dataset
Pre-requisites
Verify dataset is in the READY state:
- Navigate to the dataset view and confirm that the dataset status is "READY". Alternatively, go to the dataset inventory, locate your dataset and verify its status.
On-premise deployment
- For pre-requisites for on-premise deployment please contact [email protected]
Model Selection
- Navigate to Enrich tab, select a model and click on "Generate Models Preview" button.
- Add additional models by clicking on Add Models (optional)
- Navigate between models to review the model output before enriching the entire dataset.
- To initiate the model run, click on "Enrich Dataset".
- All the images/frames in the dataset will be enriched by the models listed in the Models section.
- This operation may take some time, depending on the size of the dataset.
- Track enrichment progress
- You can track progress by navigating to the Enrich tab in the dataset view. Alternatively you can track progress by navigating to the Dataset Inventory view - Notifications.
- Notification will be displayed once process completes.
- Note: User can explore their dataset throughout the enrichment process.
-
Explore Enriched Dataset
- Select Exploration View - Images (default) or Objects.
- Expand the VL Enrichment section in the right pane to view the metadata added to the dataset.
-
Select metadata type to display on cluster card:
- Labels
- Captions
- User Tags (not generated by models but by Users. See User Tags section ).
- Select Exploration View - Images (default) or Objects.
Updated 3 months ago