Funny creative mix of some Pony and SDXL checkpoints.
Use DPM++ SDE or Euler A samplers.
Good luck, have fun!
P.S. I'll be glad to see your creations here.
Description
Recipe:
GoddessOfRealism https://civitai.com/models/212737?modelVersionId=1087831
PonyRealism https://civitai.com/models/372465?modelVersionId=914390
RealDream https://civitai.com/models/153568?modelVersionId=1163294
DreamShaper https://civitai.com/models/112902?modelVersionId=354657
RealVis https://civitai.com/models/139562?modelVersionId=798204
triple_sum(
0.5*triple_sum(
0.33*GoddessOfRealism,
0.33*PonyRealism,
0.33*RealDream
),
0.25*DreamShaper,
0.25*RealVis
)FAQ
Comments (3)
pretty nice creative gen even w/o any lora, keep it up
I checked the prompts of the example pictures and tried to recreate them and after doing this i have one question: How come the prompt and the resulting images often have absolutely nothing to do with each other? There is a prompt that seems to describe a marble statue and the picture shows a woman in a pink latex suit. Another prompt seem to describe a japanese high-tech cyborg and the picture shows some medieval witch in a orange latex outfit. WTF?
Thank you for your comment.
The images I generated are not examples, they are just free creativity. There are many images generated through large (more than 1000 words) dynamic prompts. Randomly selected fragments of these prompts often directly contradict each other and contain dozens of obviously impossible small details. Some of these details are visible in the generated images in bizarre combinations.
The accuracy of image generation by checkpoint is quite normal, you can check it, for example, on this picture: https://civitai.com/images/49847374
I hope I answered your question.