Designer uses pipelines to build and debug models. Start by planning and creating a pipeline, then organize the various components according to your processing and scheduling logic. Machine Learning Designer offers multiple methods for pipeline creation, including demos to show you how to create a pipeline by using a template and how to create a custom pipeline.
Before you go
Some algorithm components in Designer depends on MaxCompute resources. Therefore, you need to go to MaxCompute buy page, activate and purchase the required MaxCompute resources. Then, associate the resources with your workspace. For instructions, see Manage workspace - Configure Computing Resource.
Use a preset template
Designer offers a variety of built-in templates tailored to different frameworks and the diverse needs of industry scenarios. You can use a preset template to create a pipeline, adjust the components or their configurations, and swiftly construct and deploy a model that aligns with your requirements. This demo shows how to create a pipeline using the heart disease prediction template, see Demo for creating a pipeline by using a template.
Custom pipeline
This demo walks you through the creation of a blank pipeline for developing a binary classification model for heart disease prediction. Starting from scratch, you will engage in data preprocessing, model prediction, and evaluation, culminating in a fully visualized model building and deployment process. For more information, see Custom pipelines.