Note: This only applies to Domino deployments that are version 5.5+ and have mlflow experiments enabled
If you trained models that produced Domino experiments as noted in our docs, there may come a time when you want to remove some of those experiments from your project.
Our engineering team is working on a UI solution to archive experiments, but in the meantime you can use the following method.
First, open a workspace with an environment that has mlflow installed.
Next, run the following command to in a workspace terminal to look up experiments:
mlflow experiments search
The output should consist of something like the following:
Please locate the experiment you would like to delete and note the "Experiment Id".
Next, type the following within the terminal:
mlflow experiments delete --experiment-id=##
replacing the ## with the proper number of your experiment.
From the photo above, if we wanted to delete the "file-previews-test" experiment our command would look like this:
mlflow experiments delete --experiment-id=13
The system should return a confirmation of the experiment deletion. Now you can go back to your Domino project and see the experiments are also absent within the "Experiments" portion of your Domino project as well.
Comments
0 comments
Please sign in to leave a comment.