The specific goal of this model is well rendered painted toenails, pretty toes, with a realistic foot shape.

(Sample generation config given at bottom, civitai doesn't parse my generator's metadata)

Accurate and aesthetic five-toes-per-foot for models wearing nylon stockings / pantyhose. Trained on ponydiffusion v6 xl with quality photographs of beautiful women's feet in black nylon.

All training images had some color of nail polish with a bias towards red, white, and black, but works well with other colors.

Typically, tagging "toes, nylon, black nylon, painted toenails" does all the heavy lifting.

Success with values between 0.3 and 1. around 0.75 is a good starting point.

Tags used:

toes

nylon

feet

black nylon

painted_toenails

red_painted_toenails

white_painted_toenails

stockings

thighhighs

close_up_nylon_toes

sitting

standing

lying

(assortment of views)

{
"prompt": "score_9, score_8, score_7, score_6_up, view from above, missionary, 1girl, 1boy, ???, nylon, toes, black nylon, stockings, red painted toenails, small perky breasts, petite body, ?????? dripping ???, tattoos, body paint, cinematic lighting, g0thicPXL, glowing, neon, extremely high detail, realistic artwork, day time, highly detailed skin, imperfect skin",
"negative_prompt": "score_6, score_5, score_4, (bad ???????), hairbun, pony, gaping, muscular, censored, furry, child, kid, chibi, 3d, (monochrome:1.4), (embedding:easynegative:1), watermark, text, flowers on wall, cracks on wall, bad hands, extra long finger, strange fingers, underwear, futa, ?????, bad feet",
"prompt_expansion": "",
"styles": "[]",
"performance": "Speed",
"steps": 40,
"resolution": "(1152, 896)",
"guidance_scale": 4,
"sharpness": 2,
"adm_guidance": "(1.5, 0.8, 0.3)",
"base_model": "ponyDiffusionV6XL_v6StartWithThisOne.safetensors",
"refiner_model": "None",
"refiner_switch": 0.7899,
"clip_skip": 2,
"sampler": "dpmpp_3m_sde_gpu",
"scheduler": "exponential",
"vae": "Default (model)",
"seed": "657556562576920735",
"lora_combined_1": "Expressive_H-000001.safetensors : 0.3",
"lora_combined_2": "Smooth Anime Style LoRA XL.safetensors : 0.29",
"lora_combined_3": "princess_xl_v1.safetensors : 0.46",
"lora_combined_4": "g0th1cPXL.safetensors : 0.48",
"lora_combined_5": "nylontoes_2.safetensors : 0.6",
"lora_combined_6": "Hyperrealistic_pony.safetensors : 0.81",
"metadata_scheme": false,
"version": "Fooocus v2.5.5"
}