Skip to content

Files

Latest commit

alexbuckgitalexbuckgit
[BULK UPDATE] DocuTune - Apply event-tier1-build-2022 ms.custom tag t…
May 23, 2022
55762f7 · May 23, 2022

History

History
111 lines (68 loc) · 7.4 KB

how-to-create-component-pipelines-ui.md

File metadata and controls

111 lines (68 loc) · 7.4 KB
title titleSuffix description services ms.service ms.subservice ms.author author ms.date ms.topic ms.custom
Create and run component-based ML pipelines (UI)
Azure Machine Learning
Create and run machine learning pipelines using the Azure Machine Learning studio UI.
machine-learning
machine-learning
core
keli19
likebupt
05/10/2022
how-to
devplatv2, designer, event-tier1-build-2022

Create and run machine learning pipelines using components with the Azure Machine Learning studio (Preview)

[!INCLUDE cli v2]

In this article, you'll learn how to create and run machine learning pipelines by using the Azure Machine Learning studio and Components. You can create pipelines without using components, but components offer better amount of flexibility and reuse. Azure ML Pipelines may be defined in YAML and run from the CLI, authored in Python, or composed in Azure ML Studio Designer with a drag-and-drop UI. This document focuses on the AzureML studio designer UI.

[!INCLUDE preview disclaimer]

Prerequisites

Register component in your workspace

To build pipeline using components in UI, you need to register components to your workspace first. You can use CLI or SDK to register components to your workspace, so that you can share and reuse the component within the workspace. Registered components support automatic versioning so you can update the component but assure that pipelines that require an older version will continue to work.

In the example below take using CLI for example. If you want to learn more about how to build a component, see Create and run pipelines using components with CLI.

  1. From the cli/jobs/pipelines-with-components/basics directory of the azureml-examples repository, navigate to the 1b_e2e_registered_components subdirectory.

  2. Register the components to AzureML workspace using following commands. Learn more about ML components.

    az ml component create --file train.yml
az ml component create --file score.yml
az ml component create --file eval.yml

  3. After register component successfully, you can see your component in the studio UI.

:::image type="content" source="./media/how-to-create-component-pipelines-ui/component-page.png" alt-text="Screenshot showing registered component in component page." lightbox ="./media/how-to-create-component-pipelines-ui/component-page.png":::

Create pipeline using registered component

  1. Create a new pipeline in the designer.

    :::image type="content" source="./media/how-to-create-component-pipelines-ui/new-pipeline.png" alt-text="Screenshot showing creating new pipeline in designer homepage." lightbox ="./media/how-to-create-component-pipelines-ui/new-pipeline.png":::

  2. Set the default compute target of the pipeline.

    Select the Gear icon Screenshot of the gear icon that is in the UI. at the top right of the canvas to open the Settings pane. Select the default compute target for your pipeline.

    :::image type="content" source="./media/how-to-create-component-pipelines-ui/set-default-compute.png" alt-text="Screenshot showing setting default compute for the pipeline." lightbox ="./media/how-to-create-component-pipelines-ui/set-default-compute.png":::

    [!Important] Attached compute is not supported, use compute instances or clusters instead.

  3. In asset library, you can see Data assets and Components tabs. Switch to Components tab, you can see the components registered from previous section.

    :::image type="content" source="./media/how-to-create-component-pipelines-ui/asset-library.png" alt-text="Screenshot showing registered component in asset library." lightbox ="./media/how-to-create-component-pipelines-ui/asset-library.png":::

    Drag the components and drop on the canvas. By default it will use the default version of the component, and you can change to a specific version in the right pane of component if your component has multiple versions.

    :::image type="content" source="./media/how-to-create-component-pipelines-ui/change-component-version.png" alt-text="Screenshot showing changing version of component." lightbox ="./media/how-to-create-component-pipelines-ui/change-component-version.png":::

  4. Connect the upstream component output ports to the downstream component input ports.

  5. Select one component, you'll see a right pane where you can configure the component.

    For components with primitive type inputs like number, integer, string and boolean, you can change values of such inputs in the component detailed pane.

    You can also change the output settings and compute target where this component run in the right pane.

    :::image type="content" source="./media/how-to-create-component-pipelines-ui/component-parameter.png" alt-text="Screenshot showing component parameter settings." lightbox ="./media/how-to-create-component-pipelines-ui/component-parameter.png":::

Note

Currently registered components and the designer built-in components cannot be used together.

Submit pipeline

  1. Select submit, and fill in the required information for your pipeline job.

    :::image type="content" source="./media/how-to-create-component-pipelines-ui/submit-pipeline.png" alt-text="Screenshot of set up pipeline job with submit highlighted." lightbox ="./media/how-to-create-component-pipelines-ui/submit-pipeline.png":::

  2. After submit successfully, you'll see a job detail page link in the left page. Select Job detail to go to pipeline job detail page for checking status and debugging.

    :::image type="content" source="./media/how-to-create-component-pipelines-ui/submission-list.png" alt-text="Screenshot showing the submitted jobs list." lightbox ="./media/how-to-create-component-pipelines-ui/submission-list.png":::

    [!NOTE] The Submitted jobs list only contains pipeline jobs submitted during an active session. A page reload will clear out the content.

Next steps