Archive runs

To prevent runs from appearing in comparison views and search results, you can archive them.

Archive runs in the web app

You can view all archived runs in the Archived runs custom view.

note

Deleting data from individual runs isn't supported.

From runs table

1. In the runs table, navigate to the Name column of the runs table.
2. Click next to the run name.
3. Select Archive run.

From homepage

1. Go to the homepage.
2. From the right-most column of the table, access the run menu.
3. Select Archive run.

Archive runs in bulk

To archive multiple runs at once:

1. In the runs table, at the left edge of the Name column, tick the checkboxes.
2. In the bottom toolbar, select → Archive.

note

Run bulk actions aren't available in reports.

Archive experiments

When the runs table is in Experiments mode, Neptune shows only experiment head runs.

When you archive an experiment head, other runs in the experiment aren't affected:

• Neptune doesn't archive any other runs that belong to the experiment.
• No other run in the experiment is promoted to head.

As a result, the experiment is hidden because there's no head run to show. Runs remain visible in Runs mode unless you archive them individually.

Unarchive a run

To unarchive a run:

1. In the Neptune app, access the Archived runs custom view.
2. Click next to the run name and select Restore run.

To restore runs in bulk:

1. At the left edge of the Name column, tick the checkboxes.
2. In the bottom toolbar, select → Restore.

Archive runs via API

To archive a run programmatically, use the log_configs() method to manage the sys/archived attribute:

Archive a run

from neptune_scale import Run


run = Run(...)
run.log_configs({"sys/archived": True})

Unarchive a run

run.log_configs({"sys/archived": False})

Related

For to how to access a previously created run, see Resume a run.