### 🏰 Welcome to the 13th Century!
This LoRA transforms any concept into authentic, bizarre, and flat medieval manuscript marginalia. It specializes in ugly proportions, creepy human-like animal faces, zero shading, and dirty aged parchment paper textures.
Initially trained on medieval cats, it can stylize modern objects (skateboards, laptops, motorcycles) into hilarious historical adaptations!
Recommended Settings:
* Trigger word: mdvlcat_art style
* LoRA Weight: 0.8 to 1.0 (for SDXL) | 0.3 to 0.4 (for Flux)
* CFG Scale: 7.0 (SDXL) | 1.0 to 3.5 (Flux)
* Sampler: Euler a or DPM++ 2M
Pro Tips for best results:
1. Use ADetailer! If generating characters from a distance, use ADetailer to keep the creepy human-like faces sharp.
2. Force the flatness: Add flat 2d perspective, lack of shading, tempera and ink to your positive prompt.
3. Kill the 3D: Use this negative prompt: 3d render, realistic, shading, shadows, lighting, volume, fluffy, cute, modern art, masterpiece.
Have fun generating historical memes!
Description
🐱 The Medieval Weirdness has evolved! (Flux.1 D Release)
The mdvlcat_art style is finally available for Flux.1 Dev.
This version brings even more "uncanny" detail, better text rendering, and more authentic 14th-century manuscript textures than ever before.
⚠️ CRITICAL USAGE NOTE: "LESS IS MORE" This LoRA is extremely powerful and concentrated. Because of the training density, using standard weights will result in artifacts ("deep-fried" images).
* Recommended Strength: 0.25 – 0.3 (This is the sweet spot).
* Maximum Strength: 0.35 (Do not go higher unless you want a chaotic ink mess).
* Base Model: Works perfectly with Flux.1 Dev (including FP8 versions).
What's new in this version:
* Superior Human Faces: The "uncanny" human-like expressions on cats are now more consistent.
* Manuscript Elements: Better handling of gold leaf, ornate borders, and aged parchment textures.
* Drolleries & Hybrids: Improved ability to create "cat-snails" and other bizarre medieval hybrids. Trigger word: mdvlcat_art style
Pro Tip: If the image looks too "clean," try adding aged parchment, ink stains, or naive art to your prompt.