Why Asset Groups Exist
When a project grows past a few hundred images, you need a way to slice the dataset without creating separate projects. Asset groups in Datature Nexus let you organize images into logical subsets: train/val/test splits, data from different collection sites, images captured under different conditions, or batches uploaded at different times.
What This Tutorial Covers
Datature demonstrates asset group creation and management:
- Creating asset groups to separate data subsets within a single project
- Moving assets between groups manually or with filters
- Using asset groups to control which data goes into training versus validation
- Tracking annotation progress per group
The walkthrough covers the full workflow in under three minutes.
Practical Applications
Asset groups are useful in several scenarios. A defect detection team might group images by production line. A wildlife monitoring project might group by camera trap location. A medical imaging team might group by scanner type or patient cohort. The key benefit: you keep all related data in one project while still controlling which subsets feed into each training run.
For more on dataset organization, read Introducing Asset Group Management for Managing Datasets and Visually Inspect Your Dataset and Annotations for Model Training.
