Announcing our improved experiment tracking tools in Azure Machine Learning

[Guest amipatel]

We are thrilled to announce the public preview of our enhanced experiment tracking tools in Azure Machine Learning, specifically designed to provide you with a more comprehensive understanding, management, and visualization of your training jobs. These improvements comprise a dynamic and customizable dashboard view, an upgraded jobs list view, and a wide range of features aimed at simplifying the organization and analysis of your experimentation results.

 

 

 

Key Features:

 

• New dynamic and flexible dashboard view
  
• Improved jobs list view with customizable columns and filters
  
• Easily compare metrics and images across training jobs
  
• Add markdown for notes and insights
  
• Create and save custom views for easy collaboration
  

 

 

 

Explore the New Dashboard View:

 

 

Our new dashboard view combines various tiles – including chart visualizations and markdown – to provide a comprehensive and customizable overview of your experimentation results. By default, the dashboard generates visualizations of your resource utilization, evaluation, and logged metrics for the selected jobs.

 

 

 

 

Customize Your Dashboard:

 

• Change job colors with the color picker
  
• Visualize and hide jobs as needed
  
• Add and edit charts for custom visualizations
  
• Compare metrics and images across training jobs
  
• Add markdown for notes and insights
  
• Create and save custom views for collaboration
  

 

 

 

 

Improved Jobs List View:

 

 

The jobs list view now allows you to easily track and organize information with features such as:

 

• Customizable job display names
  
• Select, pin, and reorder columns
  
• Filter jobs based on various criteria
  
• Perform batch actions on jobs
  

 

 

Get Started:

 

 

To try out these new features, simply enable the preview via the preview panel and navigate to the Jobs tab in the studio.

 

 

 

 

 

 

To learn more, read our documentation to:

 

• Visualize training results in studio (preview) - Azure Machine Learning | Microsoft Learn
  
• Organize & track training jobs (preview) - Azure Machine Learning | Microsoft Learn
  

 

 

 

Watch Microsoft Build 2023 sessions to get familiar with other Azure Machine Learning announcements:

 

• Breakout: Build and maintain your company Copilot with Azure Machine Learning and GPT-4
  
• Breakout: Practical deep-dive into machine learning techniques and MLOps
  
• Breakout: Building and using AI models responsibly
  

 

Continue reading...