muse_ember` is a photorealistic SDXL character LoRA: one fictional adult woman, trained on 27 frames generated

for her, so a single trigger word carries her face, hair and body instead of a paragraph of description.

Trigger word: muse_ember — one token, typed exactly, near the front of the positive prompt.

Recommended weight: 0.8 (0.6 to 1.0). Clip skip: 1. Base: SDXL 1.0.

The look

An adult woman in her mid-20s. Long wavy copper-red hair worn loose with a centre part and no fringe. Soft muted

green eyes, hooded and almond-shaped and set level, pale eyelashes, light auburn eyebrows. Oval face, straight nose

with a slight upturn, full lips, softly defined jaw. Pale even ivory skin with natural texture, minimal makeup, no

freckles, no moles, no tattoos, no piercings, no jewellery. Slim and long-legged with a defined waist and a natural

full bust.

How she was trained

27 images: 7 whole figure (3 of them rear views, because a set without them teaches a LoRA that cannot render a

rear rung), 6 head and shoulders, 6 feet studies, 4 breast frames, 4 anatomy frames. 2000 steps, 10 checkpoints,

network dim/alpha 32, SDXL 1.0, learning rate 0.0005. Her frames were rendered on PornMaster Pro-SDXL-V8, which is

the family she is built to be used on. The live version is a mid-run checkpoint, not the final one: the 20-40 pass

window for a set this size is checkpoints 3 to 5.

How to use her

muse_ember, 1girl, solo, nude, full length standing figure front view, the whole figure from head to feet

inside the frame, bare feet on a plain grey studio floor, soft even studio light, photograph

Negative:

cropped feet, cut off at the ankles, feet out of frame, missing feet, no feet, cropped legs, zoomed in, close-up, clothing, swimsuit, bodysuit, one-piece, underwear, bra, worst quality, low quality, blurry, jpeg artifacts, plastic skin, airbrushed, waxy, doll, cartoon, anime, 3d render, cgi, illustration, watermark, text, signature, logo, extra limbs, extra arms, extra legs, extra fingers, fused fingers, missing fingers, malformed hands, malformed feet, extra toes, twisted limb, impossible pose, duplicate subject, multiple people, second person, two women, elongated torso, long torso, stretched torso, deformed face, asymmetric face, freckles, moles, blemishes, heavy makeup, dark eyeliner, eyeliner, mascara, eyeshadow, thick black lashes, dark eyebrows, kohl, smoky eye, false lashes, doll eyes, glossy iris

Swap the framing, the light and the setting for what you want. The trigger is what holds her; everything else is

yours.

## Honest limitations

- She is an identity and body LoRA, not a staging system. Framing, orientation and pose come from the canvas,

the staging and the control layer. Words alone do not command them.

- Hands land about two times in three on a full-length frame. Tight closes land about one time in five, so shoot

more cells than you need.

- Her build comes from this LoRA plus whatever body adapter you run; a body LoRA you stack also sets the framing.

- Same trigger in a two-character frame does not separate two women. She is one character, named once.

## Content

All training images and all output are of fictional adults, 18 and over, generated for this LoRA. No real person,

living or dead, is depicted or referenced, and no real person's likeness was used in training. Adult content: this

model is intended for adult use and is flagged as mature. Nothing here depicts a minor in any framing, coercion,

incapacitation or non-consensual framing, and any such use is outside the intended use of this model.

Tags: character, photorealistic, woman, nude, sdxl