This topic describes the FAQ about models and data.
How do I configure automated model service update for each pipeline run?
Does Machine Learning Designer support incremental training?
What do I do if I am prompted that no models can be used for deployment when I choose Model Deployment > EAS (Online Model Service)?
On the details page of a pipeline in Machine Learning Designer, add the component that you want to use to the pipeline. On the Fields Setting tab of the component, select Whether To Generate PMML. On the Pipeline Attributes tab of the pipeline, configure the Data Storage parameter. Then, rerun the component.
How do I configure automated model service update for each pipeline run?
Add the Update EAS Service component to the pipeline. On the Parameters Settings tab, enter the name of the Elastic Algorithm Service (EAS) service for which you want to configure automated update. This way, when you run this component, the system retrieves the path of the Object Storage Service (OSS) model generated by its upstream component and updates the EAS service.
How do I upload on-premises data for training?
Method 1: Use the Read CSV File component to upload the CSV file. This method is applicable if the CSV file is less than or equal to 1 GB in size.
Method 2: Upload the CSV file to an OSS bucket and use the Read CSV File component to read the file. This method is applicable if the CSV file is greater than 1 GB.
Method 3: Log on to the DataWorks console, find the workspace that you want to manage, and then click DataStudio in the Actions column. On the DataStudio page, create a table to upload the data. For more information, see Create tables and upload data.
How do I deal with the issue that some text is displayed as blob in Machine Learning Designer or Studio?
Symptom
When you right-click a component on the canvas and select View data, some text is displayed as blob.
Solution
Characters that cannot be transcoded are displayed as blob. This does not affect data processing and retrieval of downstream components.
Does Machine Learning Designer support incremental training?
Some algorithms, such as deep learning and EasyRec algorithms, in Machine Learning Designer support incremental training. For more information about the algorithms, submit a ticket for consultation.