DISCLAIMER: THESE IMAGE(S) ARE GENERATED BY THE PROTEUS V2 STABLE DIFFUSION XL MODEL, THEY ARE NOT REAL NOR DO I OWN RIGHTS TO THE IMAGE I just thought this generation for "Celeste Spring Colab" was cool. For more information on this cool model visit the following link:
https://huggingface.co/dataautogpt3/ProteusV0.2
dataautogpt3/ProteusV0.2 · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
With that said and out of the way, I'm including reproducible steps to generate this image although I'm not sure this will work?
`prompt:Celeste Spring Colab inspired vector art, imaginative, masterpiece, 16:9, (best quality, HD, "~*~aesthetic~*~":1) negative_prompt:disfigured, ugly, cursed, missing body parts, oversized, undersized, bad ratio, low detail, imperfections, jpeg artifacts, compression artifacts, double moons, double noses, extra nose, detached foot, broken body parts, injury, bad posture, city, civilization, dress height:1088 width:1920 sampler:Euler a steps:80 styles:illustration n_iter:1 seed:3610834443`
A lower quality sample using the Euler sampler
Back at Euler A with 60 steps
20 steps has a different position
35 steps somehow has a landscape?