Today StabilityAI released the weights for Stable Diffusion 3 Medium, the all new 2Billion parameter model. Now we should see people start to make some high powered fine tunes and LoRA's, but in the mean time you can try out this new state-of-the-art model over at pixeldojo.ai. https://pixeldojo.ai/stable-diffusion-3
On top of this, I also launched more updates today as well:
My Images now has multidiffusion upscales and consistent characters as options to save and filter. For multidiffusion upscales that you make public, people will be able to use the image slider to view the before and after. https://pixeldojo.ai/community-gallery/1718138877925-fqo343ov6
For consistent characters, I thought about just saving a single image when you click on it, but instead i'm saving the source image and all 4 of the generated images so you can scroll through and view them. Here is what that looks like if you make it public: https://pixeldojo.ai/community-gallery/1718199754041-tmzrefd1k
Enjoy!