Manage Datasets

Datature Vi gives you tools to keep your datasets organized and healthy throughout the model development lifecycle. From the dataset Settings page, you can rename or delete a dataset. From the Explorer, you can download your data, view annotation statistics, and remove individual or bulk assets.

This section covers all dataset maintenance operations.

By the end of this guide

Rename, delete, download, and maintain your datasets throughout the model development lifecycle.

Rename A Dataset

Update the display name without affecting the dataset ID, annotations, or any active training runs.

Delete A Dataset

Permanently remove a dataset, all its assets, and all associated annotations from your workspace.

Dataset Overview

Check asset counts, annotation distribution, and class balance before you start a training run.

Manage Assets

Delete individual images or use bulk selection to remove multiple assets at once.

Download Data

Export a full dataset with assets and annotations, or download annotation files only in Vi JSONL or TFRecord format.

Next steps

Rename A Dataset

Give your dataset a more descriptive name for better organization.

Download Data

Export your dataset before making large-scale changes.

Dataset Overview

Check annotation statistics to prepare for training.