Product development /

Archive Experiment

December 2 2021/6 min read

We are excited to bring you yet another product update that simplifies the management of multiple experiments in your projects. This time, we introduce the archive experiment feature. Enjoy!

Introducing the Archive experiment feature 

Deep learning is an iterative process. Oftentimes you might need to run several experiments in different ways before finding the best performing model. With so many experiments running simultaneously, the Evaluation view can become cluttered, and it can become harder to find the best experiment. 

With that in mind, we’re introducing the possibility to archive experiments. That way, you can run and save multiple experiments while easily managing them, archiving the ones you’re not interested in or not using at the moment without permanently deleting them. 

How to archive an experiment

This is what your evaluation view looks like with multiple experiments running simultaneously.

To archive an experiment, and therefore remove it from the evaluation and modelling views while still keeping the experiment - not deleting it - simply select the experiment you wish to archive and click on the three dots of the experiment options menu on the top right hand corner of the experiment box.

You will then see a drop-down menu with options to manage your experiment. Select “Archive”.

After selecting “Archive” in the drop down menu, your experiment will now be archived. You will still be able to see it in the experiments list in a faded greyed-out tab, but it will not be shown in the experiment ranking list.

The archived experiment will not appear in the metrics table nor in the loss graph.

To restore the archived experiment to active status, select the greyed-out faded experiment tab, click on the same three dots Experiment options menu in the top right hand corner, and select “Unarchive” from the drop down menu. 

Your unarchived experiment will return to your evaluation view in both the metrics table and the loss graph and in the experiment ranking list.

There it is!

If you need any extra help, take a look at our Knowledge Center or send us a message and we’ll gladly assist you.

Happy clutter-free modeling!

  • Björn Treje

    Björn Treje

    Head of Technical Enablement

    Björn has a Master of Science in Electrical Engineering. He strives to put engineering into the business and business into the engineer. Secretly he hopes all projects involves helmets or reflex vests at some point.

02/ More on Product development