Manage pipelines

To access a list of available pipelines (artifacts), select AutoML > Pipelines. Control Tower updates the list every time a new pipeline is deployed to your environment.

On the page, users can view the following information about existing pipelines:

Column	Description
Code	Pipeline identifier specified in its source code description
Title	Pipeline title specified in its source code description
Version	Pipeline version specified in its source code description
Imported	Date and time (in format dd.mm.yyyy, hh:mm) when the pipeline was added to your environment
Type	Problem type the pipeline can solve (for instance, information extraction)
ML SDK version	ML SDK version used to build the pipeline

View pipeline details

To view the detailed data on a particular pipeline, click its Code in the list or open the Actions menu and select View.

The pipeline details page not only contains pipeline data but also lists models associated with the pipeline in the Models section:

Clicking the model ID in the Models section brings you to the model's details page.

Filter and search

You can filter the list of displayed pipelines based on any or all the three criteria:

• Pipeline code from the source code description
• Pipeline type (for instance, information extraction)
• ML SDK version used to build the pipeline

To apply the criteria, click one or more filter buttons at the top of the pipeline list and select required items or find them using the search box. Filters turn blue as they become active, and the items in the pipeline list change accordingly.

To turn off an active filter, click it again. To reset them all, click Clear all.

You can also search for a particular pipeline using the Search by Code and Title box in the top-right corner.

Train model

You can train models from the Pipelines page as described below:

1. Open the Actions menu for a selected pipeline and choose Train model.

   

   Alternatively, you can click Train model from the pipeline's details page.

   

2. In the window that follows, the Pipeline and Version fields are prepopulated for you with those associated with the pipeline you selected in the list. Fill in the rest of the fields as follows:

   • In the Version drop-down box, choose another pipeline version, if needed.

   • In the Model ID field, specify a unique model ID so that you and other users can quickly identify the new model among others.

   • In the Training set drop-down list, select a dataset for training the model and choose its version.

   

3. Optionally, in the Advanced parameters JSON editor, set advanced model training and execution parameters. By default, the editor displays the following parameters:

   • executionProfile defines if evaluation results and metadata for them should be preserved (DEBUG mode). Possible values are PRODUCTION (default) and DEBUG.

   • enableFeatureImportanceExplanation prints top N features, each with detailed info and score. Possible values are TRUE (default) and FALSE.

   • extendedCrossValidation is a Boolean flag defining whether to apply extended cross-validation or not. Possible values are TRUE (default) and FALSE.

   For a full list of model training and execution parameters you can set in the section, read the Perform advanced model tuning in AutoML UI topic.

4. Click Next and set up the Testing step parameters: choose a test method and a testing set if appropriate.

5. Click Next and configure labels, post-processors, and decision rules if appropriate and available for the pipeline type you are using.

6. To launch the model training, click the Start button.

Export

You can export a pipeline as a bundle (zip archive) in one of the two ways:

• On the pipeline list, find the one you want to export. Open its Actions menu and select Export pipeline.

  

• On the pipeline's details page, open the Actions menu and select Export.

  

The resulting zip archive contains the selected pipeline (artifact).

Import

To import a pipeline, in the top-right corner of the Pipelines page, click the plus (+) button to open the menu and select Import.

info

An imported bundle can include only pipelines (artifacts) and models. When the bundle contains any other assets, the import fails.

The import status is visible above the pipeline and model lists. The following statuses are possible:

• Uploading: the bundle is being uploaded to Control Tower. At the stage, halt the procedure by clicking the Stop (cross-shaped) button.

  

• Importing: assets are being deployed to your environment. At the stage, you can't stop the procedure.

  

• Success: the status appears once the assets from the bundle are moved to the target environment. It disappears automatically after about 15 seconds, or you can click the close (cross) button. To view the bundle details, click View.

  

• Failed: the status appears for each error. Users need to close the messages manually. The assets imported before the error (if any) appear in the model and pipeline lists accordingly.

  

Delete

You can delete a pipeline in one of the two ways:

• On the Pipelines page, choose a pipeline in the list, open the Actions menu, and select Delete.

  

• On the pipeline details page, open the Actions menu, and click Delete.