The AI image generator Flux by Black Forest Labs founded by a group of disgruntled Stable Diffusion researchers has released a “raw mode” in its 1.1 version.
This raw mode is not like Adobe Camera Raw or similar RAW apps, but it still is photography-related because it is for “creators seeking authenticity”.
“Our new raw mode captures the genuine feel of candid photography. Toggle this feature to generate images with a less synthetic, more natural aesthetic,” Black Forest Labs writes on its website. “Compared to other text-to-image models, raw mode significantly increases diversity in human subjects and enhances the realism of nature photography.”
Black Forest Labs shared a grid of images (see below) from its new raw mode and most of them could pass as real photos when just glancing at small versions; aside from the top hat-wearing octopus taking a bath.
Flux Ultra Mode
Flux’s 1.1 version also sees an ultra mode introduced which lets users generate images at four times the resolution of the usual spec, without “sacrificing prompt adherence.”
“Unlike many high-resolution models that experience significant slowdowns at higher resolutions, our performance benchmarks show sustained fast generation times — over 2.5x faster than comparable high-resolution offerings.”
Flux, the AI Image Generator Born Out of Stable Diffusion
Since Flux broke cover this past summer, it has performed well in the Artificial Analysis Text to Image Model Leaderboard where Flux 1.1 currently sits second behind Recraft AI.
In March 2024, news broke that the three former Stability AI engineers — Robin Rombach, Andreas Blattmann, and Dominik Lorenz — had left the company. Shortly after, CEO Emad Mostaque resigned .
The three German researchers created Stable Diffusion while at university and it was only after Stable Diffusion was published did Stability AI became involved.
“Stability, as far as I know, did not even know about this thing when we created it,” Björn Ommer, the professor who supervised the researchers, has said on the record. “They jumped on this wagon only later on.”
What About Flux’s Training Data?
Black Forest Labs has not said what training data was used to make Flux but it is likely it used a huge unauthorized image scrape of the internet.
It is safe to assume at this point that unless a generative AI company says otherwise, it’s likely it got the training data from the open web without asking permission from copyright holders.
The practice remains controversial and could be decided in the courts. The researchers’ old company Stable Diffusion is facing a huge lawsuit brought by Getty Images.