Anya Chen

Trigger Word: brunette-102

Description

Anya Chen is a polished, photorealistic character LoRA built around a beautiful 30-year-old Chinese woman with dark brown eyes, a softly defined oval face, and shoulder-length dark brown layered hair. Her look blends natural elegance with a calm, confident presence that works equally well in everyday lifestyle scenes, professional portraits, fashion editorials, swimwear, and refined glamour photography.

This LoRA was trained with a carefully curated dataset emphasizing facial consistency, natural expressions, multiple viewing angles, varied eye directions, changing hairstyles, indoor and outdoor lighting, and a broad range of wardrobe and environmental conditions. The result is a flexible identity model that maintains Anya’s recognizable facial structure while adapting well to different poses, settings, and visual styles.

Backstory

Anya Chen grew up in a creative, cosmopolitan environment and developed an appreciation for design, travel, photography, and modern architecture. She has a composed, observant personality and a style that moves easily between understated everyday clothing and sophisticated evening fashion.

She is equally at home reading beside a quiet window, working in a contemporary office, exploring a new city, or attending an elegant evening event. Her appeal comes from her natural confidence and refined simplicity rather than exaggerated styling, making her especially well suited to realistic lifestyle and editorial imagery.

Recommended Settings

LoRA weight: 0.75–0.9
Recommended starting weight: 0.8
CFG: 4.5–6

For the strongest identity consistency, include both the trigger and her name near the beginning of the prompt:

brunette-102, Anya Chen

Example:

score_9, score_8_up, source_photo, photorealistic, <lora:brunette-102:0.8>, brunette-102, Anya Chen, 30-year-old Chinese woman, dark brown eyes, shoulder-length dark brown layered hair, realistic skin texture, elegant waist-up portrait, soft window light, looking at viewer

LoRA Tips

• Use 0.8 as the default weight for most images.

• Lower the weight to 0.7–0.75 when using strong concept, hairstyle, fashion, or pose LoRAs.

• Increase toward 0.85–0.9 when maximum facial fidelity is more important than flexibility.

• Include dark brown eyes and shoulder-length dark brown layered hair when you want her standard appearance.

• Her hair can be changed successfully, but very high weights may pull it back toward the trained shoulder-length style.

• She performs particularly well in waist-up portraits, beauty photography, lifestyle scenes, professional fashion editorials, resort imagery, and cinematic evening settings.

• Use clear lighting descriptions such as soft window light, overcast daylight, golden-hour rim light, or large diffused studio light.

• Avoid stacking too many facial-enhancement LoRAs at high weights, as they may weaken her identity or alter her eyes.

• For epoch or model comparisons, keep the seed, sampler, resolution, steps, CFG, and prompt unchanged.

Suggested Negative Prompt

worst quality, low quality, low resolution, blurry, soft focus, out of focus, cartoon, anime, illustration, painting, CGI, 3d render, plastic skin, waxy skin, over-smoothed skin, distorted face, facial asymmetry, malformed eyes, crossed eyes, lazy eye, uneven pupils, bad teeth, extra teeth, distorted mouth, warped jawline, bad anatomy, malformed hands, extra fingers, missing fingers, fused fingers, extra limbs, duplicate person, multiple people, child, teenager, elderly, age drift, masculine features, blonde hair, red hair, gray hair, bangs covering eyes, harsh shadows, overexposed, underexposed, text, logo, watermark, signature