Duration: 0:00

1,577 views

1

In our fifth stream of the Python + AI series, we're going to explore the two main ways to get LLMs to output structured responses that adhere to a schema: function calling and structured outputs.

We'll start with function calling, which is the most well supported way to get structured responses, and discuss its drawbacks. Then we'll focus on the new structured outputs mode available in OpenAI models, which can be used with Pydantic models and even used in combination with function calling.

Our examples will demonstrate the many ways you can use structured responses, like entity extraction, classification, and agentic workflows.

📌 Follow-along live, thanks to GitHub Models (https://github.com/marketplace/models) and GitHub Codespaces.
If you'd like to follow along with the live examples, make sure you've got a GitHub account.

📌 You can also join a weekly office hours to ask any questions that don't get answered in the chat, in our AI Discord: https://aka.ms/aipython/oh

📌 This session is a part of a series. To learn more, click here: https://aka.ms/PythonAI/series

#MicrosoftReactor #learnconnectbuild

[eventID:25087]