Delete a Run

Deleting a run in Datature Vi permanently removes the run along with everything it produced: training metrics, loss charts, logs, and the trained model. The source workflow and dataset are not affected. There is no recovery option once you confirm the deletion.

Before You Start

Deleting a run is irreversible.
Active runs (status: Running, Queued, or Starting) must be killed before they can be deleted.
Any trained models you want to keep already downloaded.

Open your training project

Go to Training in the sidebar and click the project containing the run you want to review.

You should see

Runs tab without the deleted run, workflow still visible in the Workflows tab

The run no longer appears in the runs list. The workflow that launched it remains available for future runs.

What gets deleted and what stays

Deleted immediately:

Trained model artifacts and checkpoints
Training metrics and loss charts
Training logs and error output
Run configuration snapshot
Resource usage records for that run

Not affected:

The source workflow (remains available for new runs)
The source dataset
Other runs in the project
Compute credit history (past consumption stays in usage records)

Do this with the Vi SDK

import vi

client = vi.Client(
    secret_key="your-secret-key",
    organization_id="your-organization-id"
)

client.runs.delete(run_id="your-run-id")

For more details, see the full SDK reference.

Next steps

Kill A Run

Stop an active run before deleting it.

Monitor A Run

Watch your next training session in real time to catch issues early.

Manage Workflows

Organize the training configurations in your project.