After you create an experiment, you can manage the experiment and trials of the experiment, including operations such as cloning, modifying, stopping, and deleting. This topic describes how to manage experiments.
Modify an experiment
When an experiment is in the RUNNING state, you can modify the experiment parameters to adjust the experiment. After you modify the parameters, you may obtain the optimal hyperparameter combination in a shorter period of time, or obtain a better combination. The result varies based on the parameter that you modify. Perform the following steps:
Go to the AutoML page.
Log on to the PAI console.
In the left-side navigation pane, click Workspaces. On the Workspaces page, click the name of the workspace that you want to manage.
In the left-side navigation pane, choose .
Use one of the following methods to modify an experiment:
Method 1: If the experiment is in the RUNNING state, click Edit in the Actions column. On the Update Experiment page, modify the parameters and click Submit.
Method 2: Click the name of an experiment to go to the Experiment Details page. If the experiment is in the RUNNING state, click Edit in the upper-right corner. On the Update Experiment page that appears, modify the parameters and click Submit.
Rerun a trial
If the experiment is in the RUNNING state and the trial is in the FAILED state, you can rerun the trial. Perform the following steps:
Rerunning a trial is equivalent to creating a new trial. This operation starts a Deep Learning Containers (DLC) job or MaxCompute task and consumes your cloud computing resources.
On the AutoML page, click the name of the experiment that you want to view to go to the Experiment Details page.
On the Trials tab, find the trial that is in the FAILED state and click Rerun in the Actions column.
In the message that appears, click OK.
Stop an experiment
You can click Stop in the Actions column of an experiment that is in the RUNNING state on the AutoML page to stop the experiment and save resources. The system stops all trials that are in the RUNNING state in the experiment, and then stops the experiment.
Stop a trial
You can stop trials that are in the RUNNING state to save resources. Stopping a trial does not affect the overall status of the experiment.
On the AutoML page, click the name of the experiment that you want to view to go to the Experiment Details page.
On the Trials tab, find the trial that is in the RUNNING state and click Stop in the Actions column.
Clone an experiment
After you create an experiment, you can create an identical experiment by cloning it, or modify the experiment parameters based on the existing experiment to create a new experiment. On the AutoML page, use one of the following methods to clone an experiment:
Method 1: Click Clone in the Actions column of the experiment that you want to clone. On the Create Experiment page, confirm the parameters and click Submit.
Method 2: Click the name of the experiment that you want to clone to go to the Experiment Details page. In the upper-right corner of the page that appears, click Clone. In the Create Experiment dialog box, confirm the parameters and click Submit.
Delete an experiment
Use one of the following methods to delete an experiment that is not in the RUNNING state:
After you delete an experiment, you cannot view the experiment information or find the experiment data. Proceed with caution.
Method 1: On the AutoML page, find the experiment that you want to delete and click Delete in the Actions column. In the Note message, click OK.
Method 2: On the AutoML page, click the name of the experiment that you want to delete to go to the Experiment Details page. In the upper-right corner of the page, click Delete. In the Note message, click OK.
Search for an experiment
Search for an experiment by using the experiment ID or experiment name as a filter.
Search for an experiment by using the creator as a filter. Enter the ID of your Alibaba Cloud account in the Creator field.
Search for an experiment by using the status as a filter.