I really like how this came out, I didn’t originally want the fire but the AI kept adding it with the Tanuki keyword so I leaned into it. It took a while to get though, it kept trying to make the Tanuki a horse and getting the fairy to ride it without it turning lewd or just ignoring the prompt was a pain. Anyone have any tips on that? anyways, heres a few more i liked.
Model and Prompt
width: 600
height: 800
model: nightmareshaper
count: 4
guidance: 7.0
steps: 40
sampler: ddim
seed: 1015036689
upscale: 1
lora1_strength: 0.7
lora2_strength: 0.7
prompt: a (beautiful powerful pixar-style fairy)+ (Traveling astride an) (armored blazing tanuki) action shot, close up, HD, Sharp Focus, DSLR, high detail, natural lighting, shadow play, High Dynamic Range, Light Falloff, realistic, detailed face, (realistic detailed hair), realistic skin, Artstation, Masterpiece,
negative_prompt: ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft, censored, anime, cartoon, fake, plastic skin, doll-like, fake hair, plastic, toy-like, not real, drawn look, large ears, filters, animated look, painting, cell shaded, digital art, pixar, cgi, porcelain, statues, monochrome, black and white, grayscale, colorless, greyscale, black and white, black-and-white, without color, not in color, tintless, drab, fat, pregnant, removed, omitted, multi-color eyes, malformed, misshapen, partner out of frame,
Awesome! Glad it worked for you. The choice of words for sure has a huge impact on the output. Sometimes I’ll just put my prompt into GPT and tell it to rewrite everything including parentheses and weights if I’m having issues getting the wrong result over and over.