Machine Learning Platform for AI:Manage workspaces

Limits

Only the administrator or owner of a workspace is allowed to manage the settings of the workspace.

View the details of a workspace

1. Log on to the Machine Learning Platform for AI console.

2. In the left-side navigation pane, click Workspaces. On the Workspaces page, click the name of the workspace that you want to manage.

3. On the right side of the Workspace Details page, view the details of the workspace.

   

Modify the description of the workspace

1. In the Details section of the Workspace Details page, click Modify next to Workspace Name.

2. In the Workspace Basic Configuration panel, modify the Description and click OK.

   Note

   You can modify the description of an existing workspace, but not the name of the workspace.

   

Manage computing resources of the workspace

• View the associated computing resources of the workspace

  1. In the Details section of the Workspace page, click Resources next to Computing Resources.

  2. In the Associated Resource Groups section of the Workspace Resource Configuration panel, you can view the associated resource groups, such as MaxCompute Resources, General Computing Resources, and Fully Managed Flink Resources. You can also view the detailed information about resource groups, such as Resource Quantity and Billing Method.

  Note

  If no resource groups are associated with the workspace, an error may occur when you perform operations in Machine Learning Designer or on the Create Job page.

• Associate computing resources with the workspace

  In the Associated Resource Groups section of the Workspace Resource Configuration panel, follow the instructions shown in the following figure to associate resource groups that are created by the Alibaba Cloud account with the workspace. In this example, a MaxCompute resource group is associated.

  Important

  You cannot disassociate an associated resource group. Contact your account manager to disassociate a resource group.

  When you associate MaxCompute resources, pay-as-you-go GPU-accelerated computing is enabled by default. You can modify the settings based on your business requirements.

• Use GPU resources

  You can use GPU resources in scenarios in which you want to build models for audio, video, or text analysis training in Machine Learning Designer.

  1. In the Associated Resource Groups section of the Workspace Resource Configuration panel, click the MaxCompute Resources tab.

  2. Find the resource group for which you want to use GPU resources and click Modify in the Actions column.

  3. In the Resource Settings dialog box, set the GPU parameter to Pay-as-you-go and click OK.

Set the temporary storage paths for the workspace

1. In the Details section of the Workspace Details page, click Settings next to Storage Settings.

2. In the Storage Settings panel, set the Workspace Default Storage parameter and click OK.

   

   Important

   To develop algorithms in the workspace, make sure that you have the read and write permissions on the specified Object Storage Service (OSS) storage paths.

Specify the retention period of temporary tables in the workspace

Temporary tables are generated in the workspace when you run components in Machine Learning Designer based on MaxCompute resource groups. To specify the retention period of temporary tables in the workspace, perform the following steps:

1. In the Details section of the Workspace Details page, click Settings next to Delete temporary table.

2. In the Delete temporary table dialog box, use one of the following methods to delete temporary tables:

   • Specify the retention period of temporary tables

     Set the Retention Period (Day) parameter and click OK. The default retention period of temporary tables is 28 days. After the retention period of a temporary table expires, the system automatically deletes the temporary table.

   • Delete all temporary tables that are generated before the specified date in the workspace

     Set the Delete Temporary Tables Before Following Date parameter and click Delete. The system automatically deletes all MaxCompute temporary tables that have a name prefix pai_temp_ and are generated before the specified date.

Manage members of the workspace