April from Microsoft Developer demonstrates how to deploy and compare models in the Microsoft Foundry portal, guiding viewers through model benchmarking, deployment, interactive testing, and coding integrations.

Deploying and Comparing Models in Microsoft Foundry

Author: Microsoft Developer (April)

This session guides developers through the capabilities of the Microsoft Foundry portal for model management, including:

1. Introduction

  • Quick overview of the Microsoft Foundry portal and its focus on AI model workflows.

2. Exploring Models

  • Browse available models within the Foundry portal.
  • Understand model capabilities and intended use cases.

3. Comparing Model Benchmarks

  • Evaluate and contrast different models using built-in benchmarking tools.
  • Learn how to interpret and leverage benchmark results for selecting the right model.

4. Deploying a Model

  • Step-by-step demonstration on how to deploy a model to a production environment via the portal.
  • Access deployment status and configuration information post-deployment.

5. Model Playground and Chat

  • Use the interactive playground to chat with deployed models directly in the portal.
  • Test model responses and performance before integrating into applications.

6. Transition to Code-First Experience

  • Seamlessly shift from chatting and experimentation in Foundry to code development.
  • Use the web-based Visual Studio Code editor to interact with, modify, and manage deployed models via code.

Key Takeaways:

  • Learn how to compare model performance using Foundry.
  • Deploy models and view deployment status directly in the portal.
  • Test models interactively and shift smoothly into a code-centric workflow with Visual Studio Code for further development.

For more developer resources and in-depth AI tools guidance, see the linked Microsoft Foundry materials and join the community channels.