This morning, the news broke that Larian Studios, developer of Baldur's Gate 3 and the upcoming, just-announced Divinity, is apparently using generative AI behind the scenes. The backlash has been swift, and now Larian founder and game director Swen Vincke is responding to clarify his remarks.
The Firefly image generator is a diffusion model, and the Firefly video generator is a diffusion transformer. LLMs aren’t involved in either process - rather the models learn image-text relationships from meta tags. I believe there are some ChatGPT integrations with Reader and Acrobat, but that’s unrelated to Firefly.
As I understand it, CLIP (and other text encoders in diffusion models) aren’t trained like LLMs, exactly. They’re trained on image/text pairing, which ya get from the metadata creators upload with their photos in Adobe Stock. Open AI trained CLIP with alt text on scraped images, but I assume Adobe would want to train their own text encoder on the more extensive tags on the stock images its already using.
All that said, Adobe hasn’t published their entire architecture. And there were some reports during the training of Firefly 1 back in '22 that they weren’t filtering out AI-generated images in the training set. At the time, those made up ~5% of the full stock library. Currently, AI images make up about half of Adobe Stock, though filtering them out seems to work well. We don’t know if they were included in later versions of Firefly. There’s an incentive for Adobe to filter them out, since AI trained on AI tends to lose its tails (the ability to handle edge cases well), and that would be pretty devastating for something like generative fill.
I figure we want to encourage companies to do better, whatever that looks like. For a monopolistic giant like Adobe, they seem to have at least done better. And at some point, they have to rely on the artists uploading stock photos to be honest. Not just about AI, but about release forms, photo shoot working conditions, local laws being followed while shooting, etc. They do have some incentive to be honest, since Adobe pays them, but I don’t doubt there are issues there too.
Adobe’s image generator (Firefly) is trained only on images from Adobe Stock.
Does it only use that or doesn’t it also use an LLM to?
The Firefly image generator is a diffusion model, and the Firefly video generator is a diffusion transformer. LLMs aren’t involved in either process - rather the models learn image-text relationships from meta tags. I believe there are some ChatGPT integrations with Reader and Acrobat, but that’s unrelated to Firefly.
Surprising, I would expect it’d rely at some point on something like CLIP in order to be prompted.
As I understand it, CLIP (and other text encoders in diffusion models) aren’t trained like LLMs, exactly. They’re trained on image/text pairing, which ya get from the metadata creators upload with their photos in Adobe Stock. Open AI trained CLIP with alt text on scraped images, but I assume Adobe would want to train their own text encoder on the more extensive tags on the stock images its already using.
All that said, Adobe hasn’t published their entire architecture. And there were some reports during the training of Firefly 1 back in '22 that they weren’t filtering out AI-generated images in the training set. At the time, those made up ~5% of the full stock library. Currently, AI images make up about half of Adobe Stock, though filtering them out seems to work well. We don’t know if they were included in later versions of Firefly. There’s an incentive for Adobe to filter them out, since AI trained on AI tends to lose its tails (the ability to handle edge cases well), and that would be pretty devastating for something like generative fill.
I figure we want to encourage companies to do better, whatever that looks like. For a monopolistic giant like Adobe, they seem to have at least done better. And at some point, they have to rely on the artists uploading stock photos to be honest. Not just about AI, but about release forms, photo shoot working conditions, local laws being followed while shooting, etc. They do have some incentive to be honest, since Adobe pays them, but I don’t doubt there are issues there too.