Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Tuning-Free Personalized Image Generation

ai.meta.com

Post reply on HN
Print view View original

41–49 of 49 posts

Re: Tuning-Free Personalized Image Generation

#41

Post by bastawhiz »

LarsDu88 wrote: post #5
fxtentacle wrote: post #4

I know exactly why Facebook / Meta are researching this. Just imagine the possibilities for advertisers: Instead of telling someone how happy they would be if only they bought your expensive car, let's just spam them with AI pictures of themselves sitting in said expensive car, ideally next to some very attractive other people that match their dating preferences. Facebook has all the data they need to create very ple…

This is a much bigger thing than the llama3.1 release. Llama 3.1 doesn't really help Meta's bottom line. But content creation and ads are Meta's killer app. By having a model that doesn't require finetuning, they just changed the whole game.

> Llama 3.1 doesn't really help Meta's bottom line

Not directly. But most genai needs text models. And Meta definitely doesn't want OpenAI or Google or someone else controlling the state of the art. Zuck is essentially preventing anyone from getting too big by preventing the only options for "quite good" being behind someone's metered API.

Re: Tuning-Free Personalized Image Generation

#42

Post by bastawhiz »

smokel wrote: post #13

Photographic images generated by these systems tend to look like the graffiti portraits you see on fairground attractions. I've done a lot of photorealistic drawings, and the trick to make something look real, is to get the tones exactly right. Misjudge a tone a bit, and the result looks like a mediocre drawing or a painting. In other words, the gradient of skin tones is off, which is ironic, I guess. I assume that t…

I have a feeling this is partly due to training data being selected by an "aesthetic score". If weirdly airbrushed skin consistently has a high aesthetic score, that's what the model gets trained on.

Re: Tuning-Free Personalized Image Generation

#43

Post by bastawhiz »

GaggiX wrote: post #15
smokel wrote: post #13

Photographic images generated by these systems tend to look like the graffiti portraits you see on fairground attractions. I've done a lot of photorealistic drawings, and the trick to make something look real, is to get the tones exactly right. Misjudge a tone a bit, and the result looks like a mediocre drawing or a painting. In other words, the gradient of skin tones is off, which is ironic, I guess. I assume that t…

The fact that these models rely heavily on classifier-free guidance has a strong impact on the tones of the image.

Exactly. No humans are involved in most of the process.

Re: Tuning-Free Personalized Image Generation

#44

Post by 01HNNWZ0MV43FF »

MasterScrat wrote: post #34
strongpigeon wrote: post #22

Earlier quoted context omitted.

This is quite thought provoking. I can totally see ads for, say Disney World, where they put you in the picture instead of an actor. I mean, the whole goal of these ads is already to have you imagine yourself there. Putting you in the picture makes it that much easier.

We sell text-to-image model finetuning (aka "Dreambooth") as a service and yes, this is one of the use cases. Recently a travel agency used our platform to generate images of people in the destinations they were advertising.

Ew

Re: Tuning-Free Personalized Image Generation

#45

Post by LarsDu88 »

ripdog wrote: post #40
LarsDu88 wrote: post #32

Earlier quoted context omitted.

Hmmm, that'd be an interesting startup idea! How do ad blockers work exactly?

They're entirely manual. A whole bunch of volunteers write filter rules to block known ads. There's a big github where people can post issues about ads they've found, and volunteers will write filter rules to block them. See https://github.com/easylist/easylist/issues

This does seem like something ai could automate

Re: Tuning-Free Personalized Image Generation

#46

Post by pyinstallwoes »

baq wrote: post #19
LarsDu88 wrote: post #5

Earlier quoted context omitted.

This is a much bigger thing than the llama3.1 release. Llama 3.1 doesn't really help Meta's bottom line. But content creation and ads are Meta's killer app. By having a model that doesn't require finetuning, they just changed the whole game.

Where do I sign up for my personalized AI content filter bot which can reliably detect ads and remove them from my browser?

Digital immune system
Post reply on HN