Question

An architecture firm is building a lightweight vision app in Microsoft Foundry. Staff will upload photos of building floor plans, and the app must return a written paragraph describing the layout, room arrangement, and any labeled measurements visible in the image. Which type of model deployed in Foundry should the app use to accomplish this?

Accepted Answer

A multimodal model that accepts both image and text input and returns a text description

Answer

A text-only generative model prompted with a description of the floor plan

Answer

An image-generation model that creates new floor plan images from a text prompt

Answer

An embedding model that converts the floor plan into a numeric vector