An open source alternative to AI image generators like Fixed Diffusion Middle of the road and DALL-Eupdated as version 3.5. The new model attempts to correct some of the wrongs (which may be an understatement) of the popular Stable Diffusion 3 Medium. Stability AI says the 3.5 model follows directions better than other image generators and rivals larger models in output quality. In addition, it is more tailored for different styles, skin tones and features.
The new model is presented in three variants. The Stabile Diffusion 3.5 Large is the most powerful of the trio, with the highest quality of the bunch as well as the industry’s fastest adoption. Stability AI says the model is suitable for professional use at 1 MP resolution.
Meanwhile, the Stable Diffusion 3.5 Large Turbo is a “distilled” version of the larger model, focusing more on efficiency than maximum quality. Stability AI says the Turbo variant still produces “high-quality images with exceptionally fast matching” in just four steps.
Finally, Stable Diffusion 3.5 Medium (2.5 billion settings) is designed to run on consumer hardware, balancing quality with simplicity. With greater customization ease, the model can create images at 0.25 and 2 megapixel resolution. However, unlike the first two models available now, the Stable Diffusion 3.5 Medium doesn’t arrive until October 29th.
A new trio goes after the underdogs Stable Diffusion 3 Medium in June. The company acknowledged that the release “did not fully meet our standards or the expectations of our communities.” hilarious grotesque body horror in response to demands that do not want such a thing. It’s likely no coincidence that Stability AI repeatedly touted its exceptional operational compliance in today’s announcement.
Although Stability AI only mentioned it briefly in the announcement blog post, the 3.5 series has new filters to better reflect human diversity. The company describes the human results of the new models as “representative of the world, not a single type of human with different skin tones and features, without the need for extensive negotiation.”
Let’s hope it’s sophisticated enough to account for subtleties and historical sensitivities, unlike Google’s debacle earlier this year. Unsolicited Gemini has produced collections of highly inaccurate historical “pictures”. ethnically diverse Nazis and the founding fathers of the United States. The backlash was so strong that Google hasn’t reunited generations of people until six months later.
