60p1c/l3dy-agata-lora

l3dy Agata - Wan 2.2 LoRA

This is a LoRA trained on the appearance of l3dy Agata. The model was trained using the Wan-2.2 architecture, providing high-fidelity facial features and natural skin textures.

Model Description

• Character: l3dy Agata
• Base Model: Wan-2.2 (14B)
• Training Data: 30 high-quality portraits in various poses, lighting, and expressions.
• Trigger Word: l3dy Agata

Usage Instructions

1. Dual LoRA Setup

Since this model was trained on Wan-2.2, it produces two files. For the best results, it is recommended to use both:

• High Noise LoRA: Handles structure and composition.
• Low Noise LoRA: Handles fine details and skin texture.

2. Recommended Settings

• Trigger Word: Always include l3dy Agata at the beginning of your prompt.
• LoRA Weight: 0.7 - 0.9 (Start with 0.8 for both).
• CFG Scale: 5.0 - 6.5
• Sampling Steps: 25 - 40

Prompt Example

l3dy Agata, close-up portrait, looking at camera, soft cinematic lighting, high quality photorealistic, detailed eyes, messy hair.

Ethical Lead

This model is intended for creative and artistic use only. Do not use this model to generate NSFW or harmful content.