Multimodal models are great because they ground language in the world we actually live in. Vision language models (VLMs) can see images and answer questions in natural language. Video models understand how actions unfold and how cause and effect play out over time. Audio models can hear speech and sound. All of them bring AI closer to human-like perception.
What if you want to build an assistant that can see, hear, and speak? Based on what we have seen so far, you would need a vision model, an audio model, a speech synthesis model, and careful orchestration among them.
The next frontier is any-to-any models: unified architectures that handle understanding and generation across text, images, audio, and video within a single system (Figure 10-1).
To be truly any-to-any, a model must handle at least two input modalities and two output modalities. If that sounds ambitious, it is…and the engineering choices involved are fascinating—as you will see in this chapter.
Particularly, you will learn about the three architectural trends for any-to-any models, along with their pros and cons. You will also run inference and train those models to gain experience and judge these different approaches.
Why three families? That is a fair question. The challenge of unifying modalities creates a fundamental puzzle: how do you represent continuous signals (images, audio, video) in a way that a transformer can both understand and generate?
Figure 10-1. Overview of a unified ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month, and much more.
O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.
Julian F.
Head of Cybersecurity
I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.
Addison B.
Field Engineer
I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.
Amir M.
Data Platform Tech Lead
I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.
O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.
