AIOPEN Service User Manual

Contents:

• Introduction
  • Structure of this Manual
  • High-Level Architecture
  • GitHub
  • Development Environment
  • Development Services UI
  • Global Catalogue
  • Marketplace
  • Exploitation
  • Becoming a User
• Developer Manual
  • On-Boarding
  • Authentication
  • Repository Management
  • Workspace
  • Development Services Landing Page
    • Configuring remote access to MLflow
    • Configuring GitHub Access to IDEs
    • Controlling Personal IDE
    • Using the STAC Validator
  • Training Data Management
    • Import Files
    • Delete Files
  • AI Models Development and Training
    • Accessing training data in the workspace datastore (mounted S3 bucket)
    • Training your Model - Simple cookbook
    • Inspecting the training sessions in MLflow
  • Sharing and Publishing
    • Trained Models
    • Training Data
• Exploitation Manual
  • User roles and capabilities
  • The AIOPEN Marketplace
  • User Authentication
  • Discovering trained models
  • Executing a model to obtain predictions
    • Selection and parameterisation
    • Inference Pipeline: Providing valid input data
    • Execution monitoring and reports
  • Running a model as an OGC processing service
    • Starting the Model as a Service
    • Authenticating and Accessing the Model Service
    • Calling the Model Service Remotely
    • Stopping the Model Service
    • Inspecting the Models running as Services
  • Discovering and Ordering Training Data
    • Discovering Training Data
    • Ordering Training Data
    • Using Ordered Datasets
  • Terms and Conditions Management
  • Resources Usage Tracking
• Using the STAC Specification
  • Purpose of the STAC Item descriptors
  • Trained Models Information
    • Required Information
    • Recommended Information
  • Training Data Information
    • Required Information
  • Publish & Unpublish status
  • Target catalogue collection
  • Reference to the resource assets
    • Trained Model assets
    • Training Data assets: label or feature
  • Resource versioning
  • Terms and Conditions (license)
  • Custom thumbnail or logo
  • Contact persons
  • Themes
  • Publication DOIs and Citations