Use pipeline inputs to retrain models in the designer

In this how-to article, you learn how to use Azure Machine Learning designer to retrain a machine learning model using pipeline inputs. You can use published pipelines to automate your workflow and set inputs to train your model on new data. Pipeline inputs let you reuse existing pipelines for different jobs.

In this article, you learn how to:

Prerequisites

• An Azure Machine Learning workspace
• Complete part 1 of this how-to series, Transform data in the designer

This article also assumes that you have some knowledge of building pipelines in the designer. For a guided introduction, complete the tutorial.

Sample pipeline

The pipeline used in this article is an altered version of the Income prediction sample pipeline. The pipeline uses the Import Data component instead of the sample dataset to show you how to train models using your own data.

If you're using the pipeline that you built in part one of this series, add the other components in the following example.

Create a pipeline input

Pipeline inputs are used to build versatile pipelines that can be resubmitted later with varying values. You can create pipeline inputs to dynamically set variables at runtime. Some common scenarios involve updating datasets or hyper-parameters for retraining.

Pipeline inputs can be added to data source or component parameters in a pipeline. When the pipeline is resubmitted, the values of these inputs can be specified.

For this example, you change the training data path from a fixed value to an input, so that you can retrain your model on different data. You can also add other component parameters as pipeline inputs according to your use case.

1. Double-click the Import Data component.

   Note

   This example uses the Import Data component to access data in a registered datastore. However, you can follow similar steps if you use alternative data access patterns.

2. In the component detail pane, select your data source.

3. Enter the path to your data. You can also select Browse path to browse your file tree.

4. Mouseover the Path field, and select the ellipses above the Path field that appear.

5. Select Add to pipeline input.

6. Provide an input name and a value.

   

7. Select Save.

   Note

   You can also detach a component parameter from pipeline input in the component detail pane, similar to adding pipeline inputs.

   You can inspect and edit your pipeline inputs by selecting the Settings gear icon next to the title of your pipeline draft.

   • After detaching, you can delete the pipeline input in the Settings pane.
   • You can also add a pipeline input in the Settings pane, and then apply it on some component parameter.

8. Submit the pipeline job.

Publish a training pipeline

Publish a pipeline to a pipeline endpoint to easily reuse your pipelines in the future. A pipeline endpoint creates a REST endpoint to invoke pipeline in the future. In this example, your pipeline endpoint lets you reuse your pipeline to retrain a model on different data.

1. Select Publish above the designer canvas.

2. Select or create a pipeline endpoint.

   Note

   You can publish multiple pipelines to a single endpoint. Each pipeline in a given endpoint is given a version number, which you can specify when you call the pipeline endpoint.

3. Select Publish.

Retrain your model

Now that you have a published training pipeline, you can use it to retrain your model on new data. You can submit jobs from a pipeline endpoint from the studio workspace or programmatically.

Submit jobs by using the studio portal

Use the following steps to submit a parameterized pipeline endpoint job from the studio portal:

1. Select Endpoints on the sidebar menu.

2. Select the Pipeline endpoints tab. Then, select your pipeline endpoint.

3. Select the Published pipelines tab. Then, select the pipeline version that you want to run.

4. Select Submit.

5. In the setup dialog box, you can specify the parameter values for the job. For this example, update the data path to train your model using a non-US dataset.

Submit jobs by using code

You can find the REST endpoint of a published pipeline in the overview panel. By calling the endpoint, you can retrain the published pipeline.

To make a REST call, you need an OAuth 2.0 bearer-type authentication header. For information about setting up authentication to your workspace and making a parameterized REST call, see Use REST to manage resources.

Related content

• Tutorial: Train a no-code regression model using designer

• Publish machine learning pipelines