Shawn Lewis, CTO of Weights & Biases

New Feature: Run Spam Filter

TLDR: We’ve updated the project page to hide runs less than 10 seconds long. You can configure this behavior with custom filters in the project settings︎ ⚙️.

Building a machine learning model comes with a lot of trial and error. Debugging often generates a bunch of short runs that are not useful to analyze, so we’re giving you a way to hide the noise in your project.

The Spam Filter lets you configure which runs are automatically hidden on your project page. By default we hide runs that are shorter than 10 seconds or tagged ‘hidden’.

Hover over the shield icon in the project page’s run table to expand the spam filters. From there you can remove them temporarily, or permanently modify the project-level spam filter by clicking the gear icon at the top right of the project page.

When you click “New Report”, the report captures the spam filters configured in the project run table. To edit the global report spam filter, click the shield icon at the top of the report page.

Happy filtering!

Partner Program

We are building our library of deep learning articles, and we're delighted to feature the work of community members. Contact Carey to learn about opportunities to share your research and insights.

Try our free tools for experiment tracking →