AI and .NET: Exploring the AI samples repo and the model evaluation sample

This post has been republished via RSS; it originally appeared at: New blog articles in Microsoft Community Hub.

Hi, and welcome to the recap of our latest video on AI and .NET.
In this video, we talked about some of the exciting projects and samples that we have been working on to help you get started with AI using .NET. We also answered some of the questions that you asked us in the previous episode, such as:
  • How to work with local models
  • How to use semantic kernel
  • How to evaluate different models.
One of the main topics that we discussed was the AI samples repo, where we are adding artificial intelligence samples for .NET developers. You can find scenarios such as text summarization, chat, data generation, image generation, and more. You can also choose between using semantic kernel or the Azure Open AI SDK to interact with the models.
We also showed you how to switch between different models, such as GPT-4, Phi-3 and Llama, using the same code and API. We also shared some links to other resources, such as the Azure Developer .NET Day, where you can learn more about AI, GitHub Copilot, SQL Server, and other topics.
Another topic that we covered was the model evaluation sample, which is a work in progress that we are going to present at Build 2024.  (you can check the PR here). The idea behind this sample is to help you evaluate different models that you may want to use or fine-tune for your own applications. We showed you how to use a console application that uses semantic kernel to generate questions and answers, user stories, and other inputs, and then evaluate them using different metrics, such as coherence, groundedness, relevance, and length.
The full recording is here, 
Check the next sessions here: .NET Live

Leave a Reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.