All Products
Search
Document Center

:PAI-Alink

Last Updated:Dec 20, 2023

The PAI-Alink component in E-MapReduce (EMR) refers to Alink, which is a general algorithm platform developed by the Machine Learning Platform for Artificial Intelligence team based on Flink or Blink. This topic describes how to configure and use Alink in the EMR console.

Prerequisites

  • An EMR Data Science cluster is created.

  • A Knox account is created.

  • Port 8443 is enabled.

Access Alink

  1. Log on to the Alibaba Cloud EMR console.

  2. In the top navigation bar, select the region where your cluster resides and select a resource group based on your business requirements.

  3. Click the Cluster Management tab.

  4. Find the cluster from which you want to access Alink, and click Details in the Actions column.

  5. Configure Flink-VVP.

    1. In the left-side navigation pane of the Cluster Overview page, click Connect Strings.

    2. Click the link of Flink-Vvp UI.

    3. On the logon page, enter the username and password of the created Knox account and click Sign in.

    4. In the left-side navigation pane of the page that appears, choose Administration > Deployment Targets.

    5. Click Add Deployment Target.

    6. In the Add Deployment Target dialog box, specify Deployment Target Name, set Yarn Queue to default, and then click OK.

      Add deployment target

  6. Configure Alink.

    1. Go back to the EMR console. In the left-side navigation pane of the Cluster Overview page, click Connect Strings.

    2. Click the link of PAI-Alink UI.

    3. In the left-side navigation pane of the page that appears, click Settings.

    4. Select the added deployment target from the Switch Running Deployment Target drop-down list and click Submit.

Use Alink

On the homepage of Alink, you can click Create from Template for a template to create an experiment based on the template.

Select the required widgets and drag them to the canvas. Use the Scaring Card Function template as an example. To read local HDFS data, you only need to enter the URL of the local HDFS data on the Read CSV widget.