Embedding to help make web UI and UX designs in Stable diffusion.
Description
First Version uses lots of images from midjourney and pinterest. Training data was not well curated as this was my first embedding test. Will improve original data set in v2
FAQ
Comments (9)
This is awesome. I had just discussed this idea with someone in the community earlier this week. So cool, thanks for making and sharing this.
How functional are the results? In UI design usability and functionality comes first, and artistic quality comes last...
The screenshots look like something you might want for presentation, but not like something people have to interact with on a daily basis.
This is a step before you take it to something like Figma, Use it to generate a rough idea then design based on that. I would not take any of the generated images directly to production.
This is great idea. May be you should more higher height, or you train based sections. “Header”, “hero section”, “footer” etc.
Thanks for the suggestions will try to add better descriptors to be able to control generations better in next version.
This is really good model, which I've been looking for! Totally awesome idea and quite nice replacement for midjourney, which I used to generate my ux/ui design promts! Hope for a successful training and model improvement in the future! I really want it!
waiting for the next update, to get this even better.
Thanks, great work!
Please do not forget to upgrade this embeding too :)
If others like me couldn't get the same results, the creator used this checkpoint: https://huggingface.co/stabilityai/stable-diffusion-2-1/resolve/main/v2-1_768-ema-pruned.safetensors