11.5 Pipelines

Model Pipeline enable you to create and publish models based on the workspaces created from datasources. The published models are then deployed in production to be consumed by other services and applications. Modelers create models by using model templates available as part of ML4AML or by developing them independently in Compliance Studio Notebooks. After building and evaluating, multiple models, the best model or champion model can be selected. The champion model can then be deployed for scoring. In this document the champion model can also be referred to as the scoring model.

• Prerequisites
  
• Access Model Pipelines
  
• Manage Models
  
• Pipeline Designer
  
• Parameter Set
  
• Import/Export Parameter Sets using Utility
  
• Scheduler Configuration UI
  
• Publish Model Draft
  
• Scope Detail
  
• Model Governance
  
• Import a Model Data into a New Model
  
• Git Functionalities
  
• Execute Models using Scheduler Service
  
• Export/Import/Retire Models via Utility
  
• View Models
  
• Edit Models
  
• Delete Objectives and Draft Models
  
• Programmatic Dynamic Forms and Fixed Dynamic Forms
  
• Session Management in Model Pipeline