quantus17/rise

basemodel: stabilityai/stable-diffusion-xl-base-1.0 instanceprompt: frhn style tags:

• —text-to-image
• —diffusers
• —autotrain inference: true ---

Lora dreambooth testing with 6 fully blank purple images

I just wondered what would happen if I have a fine tuning with 6 exactly same image which are fully blank purple. For trigger word please use "frhn style" Here are the generated images with just prompt 'frhn style', it is getting sometimes an even uniformly colored image. I have some other generated images from 200 to 220 with prompt 'cat, frhn style'. It is interesting to see the generated images trying to converge an even colored canvas.