Experiments (KFP)
Last updated
Last updated
The Experiments (KFP) menu is used to group a number of experiments or pipelines that are run to achieve certain goals. Each experiment in KFP is a container that manages a number of pipeline runs related to one specific experiment or goal. This helps in organizing, comparing, and analyzing the results of the various pipelines run for the experiment. On the MLOps menu page, select the Experiments (KFP) menu.
In the Experiments menu (KFP), there are several submenus along with their explanations as follows.
Active, is designed to help you manage and monitor ongoing machine learning experiments in Portal Cloudeka Deka MLOps Service.
Refresh
This option allows you to update the menu view with the latest information regarding ongoing experiments. It is important to ensure that users are always seeing the most accurate and up-to-date data.
Archive
This feature is used to store completed or inactive experiment runs. Archiving runs helps keep the interface clean and organized while allowing you to retain data for future reference.
Clone Run
This option enables you to duplicate an existing experiment run. It is useful when you want to make variations of the same experiment without having to reconfigure all the parameters.
Compare Runs
This feature is used to compare the results of multiple experiment runs. By comparing runs, you can analyze the performance of different models and determine which one is the most effective.
Create Experiment
This option allows you to initiate a new experiment. You can define the necessary parameters and configurations required to run their machine learning experiments.
Archived, when experiments are archived, they are no longer active but can still be accessed for reference or analysis. This helps keep the workspace organized by separating ongoing experiments from those that have been completed or are no longer in use.
Refresh
This option allows you to update the menu view with the latest information regarding archiving experiments.
