High-Fidelity Character & Style LoRA This model captures the distinct art style of the Houkago Play series (originally by Alxyon) with a focus on logical consistency and sharp vector aesthetics.

Why use this version? Unlike other trainings, this version hits the perfect "sweet spot":

• Clean Character Separation: It correctly understands that Kareshi wears glasses and Kanojo does not. No more concept bleeding!

• True Monochrome: Delivers deep, crushed blacks and sharp outlines without unnecessary greyscale noise.

• Versatile Backgrounds: Capable of generating both the iconic "Spot Color Red" backgrounds and clean Manga-style white backgrounds.

Best for:

• Fanart of Kanojo & Kareshi with accurate interactions.

• Minimalist, high-contrast monochrome illustrations.

• Creating "stiff/nervous" vs. "fluid/dominant" dynamic poses.

2. Für das Feld "Notes" / "About this version" (Bild image_c7f354.png)

Hier kommen die technischen Anweisungen rein.

Text für die Notes Box:

🟢 Usage Guide

Trigger Word: Houkago

To spawn the specific characters, use their dedicated tags. The model is trained to recognize them individually:

1. Kanojo (The Girl)

• Trigger: hp_kanojo

• Essential Tags: 1girl, very long hair, hime cut, black hair.

• Pose/Vibe: standing, arms crossed, smirk, expressionless, looking down.

2. Kareshi (The Boy)

• Trigger: hp_kareshi

• Essential Tags: 1boy, short hair, glasses (Important!), gakuran, school uniform.

• Pose/Vibe: nervous, sweatdrop, shocked, clenched teeth.

🎨 Background Styles

• Classic Cover Look: spot color, red background, simple background.

• Manga Look: monochrome, greyscale, simple white background.

⚙️ Settings

• Strength: 1.0 is recommended (this version is stable). Lower to 0.8 only if you want to mix styles.

• Resolution: Trained on 1024x1024 (SDXL/Pony).