Model Description:
This is a LoRA model for Red Rose from the game Agent17. It was trained on the Krea2 base model to accurately capture her semi-realistic art style, distinctive facial features, and specific outfit.

What does this model do?
It generates high-quality images of Red Rose, maintaining her key characteristics: elegant white hair tied back, large purple eyes, big black hoop earrings, and her signature dark green high-neck halter top. It is perfect for generating modern, stylish, or slightly mature illustrations.

Recommended Usage:

• Base Model: Krea2

• Weight / Strength: 1.0 - 1.5 (Recommended: 1.2)

• Trigger Words: rose

• Prompt Example: rose, a mature woman around 30 years old, mature feminine appeal, full bust, (extremely high, straight and upright nose bridge: 1.8), narrow nose bridge, small and delicate nose tip, extremely narrow nostrils, a pair of deep purple eyes, thick and long eyelashes; (lip color is bright rose-pink with a watery sheen: 1.6); hair is silver-white long hair, **coiled at the lower back of the head below the occiput into a gathered, slightly flat bun, the bun's volume is restrained and its top does not exceed the highest point of the head** (low chignon at the back of the head, NOT a tall updo), a ponytail falls naturally from behind the bun with soft curls, soft thin strands and wispy hair gently hang down on both sides of the forehead, overall hair texture is fine and silky; **on the earlobes she wears a pair of huge thin hoop earrings, the hoop itself extremely thin, black-green in color, hanging down to around the jawline**; adult body proportions

  wearing a dark green high-neck sleeveless tight-fitting top, the fabric has a smooth matte texture, with natural clothing folds and shadows.

  The body faces the camera directly, torso facing directly forward (torso facing directly forward). Head upright, gaze directed straight at the camera. Facial expression is (calm, cold and alluring with a slightly provocative gaze: 1.6), lips slightly pressed together, the overall demeanor appears extremely confident and full of a commanding older-sister aura.

  The scene is a solid color or slightly gradient dark brown background, without any extra environmental elements, the image is completely focused on the character's face.

  Close-up, head-and-shoulders framing, camera positioned at eye level with a full frontal view, (using a medium-telephoto lens, compressing perspective proportions, maintaining the character's natural head-to-body ratio: 1.5), the light source is strong side lighting from the upper left of the frame, casting clear and sharp geometric shadows on the right side of the face, the hair strands and the high collar, the image presents an extremely high-precision 2D illustration texture.

  **(High-quality anime 3D model style (like VAM or Unreal Engine 5 render): 1.9), excellent 3D CG texture, extremely smooth combination of cel-shading and realistic coloring (stylized cel-shading with PBR materials), (absolutely no real-person photography: 2.0) (NOT photorealistic, NOT live-action, NOT real human, NO skin pores, NO realistic physical details)**, medium saturation, medium contrast, overall color tone soft and pastel, **(with extremely subtle film grain and soft focus, preventing over-sharpening (subtle film grain, slight soft focus, NOT ultra-sharp 8k): 1.5)**, 4k resolution, masterpiece, best quality.

What is this model good at?

• Accurately recreating the 2.5D semi-realistic style specific to the game.

• Capturing her elegant, cool, and slightly mysterious expressions.

• Generating high-quality, stylish portrait shots.

Limitations:

• May require negative prompts to avoid generating purely 2D anime styles or overly realistic photos, keeping the semi-realistic game aesthetic.

• The specific green reflection on the earrings might not always be perfectly replicated at lower weights.