Delete a Model

Delete trained models you no longer need from Datature Vi to keep your workspace organized and free up storage. Deletion permanently removes the model along with all its weights, adapters, configurations, and metadata.

1

Open your training project

Go to Training in the sidebar and click the project containing the model you want to download.

Next →

You should see

The model no longer appears in the Models list. The training run history remains accessible in the Runs tab.

What gets deleted

Deleting a model permanently removes:

• Model weights (full model parameters)
• LoRA adapter files (if the model was trained with LoRA fine-tuning)
• Model configuration and architecture settings
• Training metadata and hyperparameters used during the run
• The model key (checkpoint identifier)

The following are not affected by deletion:

• Training run history: metrics and logs remain viewable in the Runs tab
• The dataset used for training
• Evaluation results and logs
• Workflow configurations
• All other models in the project

Impact on active systems

After deletion, any application using that model ID will fail. SDK calls to ViModel(run_id=...) return a ModelNotFoundError. API requests return HTTP 404. Any production deployments or automated workflows referencing the deleted model stop working immediately.

Update all services to use a replacement model before deleting.

Next steps