Original character LoRA for Krea 2. One woman, not a style pack and not a celebrity likeness.

Trigger: zolacrml

Put the trigger first. Then describe the shot, the clothes, the place, and the light in plain English.

Who she is

Zola is a 23-year-old original character. Extremely tall, willowy, mixed heritage. Warm golden-caramel skin. Long sculpted oval face, high cheekbones, dark almond eyes, full glossy Cupid-bow lips. Small-to-medium teardrop chest. No tattoos.

Default hair is a honey-caramel to light-auburn chin-length wavy bob with blunt wispy baby bangs. Name that cut when you want it. Name a different cut when you do not. Hair follows the prompt.

She is an adult OC. Do not prompt her as a teen or as a real person.

What it does

Holds the face across close-ups, three-quarter portraits, and full-body shots when the trigger is present.

Keeps the long, narrow build if you give the prompt room to say so. Medium and full-body frames need an explicit body line. A face-only prompt will not invent the height for you.

Accepts wardrobe changes. Dresses, coats, jeans, knitwear, and swimwear all worked in testing. Describe the garment. Do not assume a default outfit. The black slip showed up often in training, so name a different garment if you do not want silk.

Follows natural-language captions. Krea 2 does not want Danbooru piles. Write sentences.

Works on Krea 2 Turbo for fast previews and on Krea 2 Raw when you want the slower pass.

What it does not do

It is not a universal beauty LoRA. Without zolacrml you should get ordinary Krea 2 output, not a soft Zola.

It does not lock a single art style. The set is photoreal editorial. Illustration, anime, and heavy stylization are on you and on whatever other LoRA you stack. No promises there.

It will not keep two different faces in one frame. This is a single-character adapter.

Backgrounds and lighting from the set can leak if you under-specify the scene. If you keep getting night flash or a pale plaster wall, say the room and the light out loud.

SFW and NSFW

It runs both. Portraits, street clothes, nude, and explicit scenes use the same file and the same trigger. No useful limit on heat. Prompt what you want.

She is written as 23. Keep her there.

How to run it

Trigger first, always.

LoRA strength: 0.8 to 1.0. Start at 1.0. Drop to 0.85 or 0.8 if clothes or lighting from the set start ignoring the prompt.

Krea 2 Turbo: 8 steps. Guidance 0 to 1. LoRA 1.0. This is the daily driver.

Krea 2 Raw: 28 to 52 steps. CFG around 3.5. Same LoRA range. Use Raw when Turbo smears the mouth or the hairline.

Sampler / scheduler: Euler or Euler Ancestral with Simple on Turbo. Match whatever already works in your Krea 2 workflow. Do not invent a special scheduler for this file.

Resolution: dataset was 1920p and higher, use whatever you like.

Prompt shape:

zolacrml, photograph, honey-caramel chin-length wavy bob with blunt wispy baby bangs, [shot size], [clothes], [place], [light], gaze aimed at the viewer

Drop the hair sentence when you want another cut, and write the new cut in full.

Krea 2 ignores attention weights like (word:1.5). Say the thing. Do not wrap it.

Negatives do little on Turbo. Steer with the positive prompt.

Example prompts

Portrait:

zolacrml, photograph, honey-caramel chin-length wavy bob with blunt wispy baby bangs, tight close-up of face and neck, neutral expression, soft window light, pale wall

Clothes change:

zolacrml, photograph, honey-caramel chin-length wavy bob with blunt wispy baby bangs, full body on a city sidewalk, black slip dress and heels, late afternoon light

Hair change:

zolacrml, photograph, long straight dark espresso hair, center part, tucked behind one ear, medium shot, cream silk blouse, evening lamp

Training

• Base: Krea 2 Raw

• Type: LoRA, rank 32, alpha 32

• Steps: 1750

• Captions: natural language, trigger first

• Flip augmentation: off