AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
A Captin for an Image and Its Corresponding Text
The original dali was, me 20 19 or so. And then the diffusion work, i would say, either i canntlate 20 20 or y 20 21. In both of these cases mark the same kind of data situation. You need hundreds of millions. Yet for diffusion, i think the early work, you could use much smaller data sets. Buti think we had some work here at adopen also kind of 20 20, 20 21. We were pushing these diffusion models to the state of the art on image net and a, yes.