Sweep

Hands-On: Try quick start to grasp a concept of SavviHub

Sweep in SavviHub

In Savvihub, a sweep is a term for hyperparameter optimization or tuning of a machine learning model. Sweep tries a group of experiments with different hyperparameters to optimize the model performance. The sweep results consist of monitoring, logs, and metadata which you can find at the corresponding tabs on each Sweep page.

Sweep in Web Console

You can find a list of sweeps on each project page. If you want to examine the sweep detail page, click the name of the experiment.

'freckle-shirt' sweep is running and 'maze-mouse' sweep is terminated

Sweep in SavviHub Client

Checkout savvihub sweep to learn how to manage the sweep in SavviHub Client.