LoRA Library
Expand your creative possibilities with specialized LoRA models for unique styles and characters
What are LoRA Models?
LoRA (Low-Rank Adaptation) models are small add-on models that teach FLUX new styles, characters, or concepts without replacing the base model. They're much smaller than full models but add powerful creative capabilities.
Artistic Styles
Studio Ghibli, anime, watercolor, photorealism, and more
Characters & Concepts
Specific characters, clothing styles, or unique concepts
Small & Fast
Typically 20-100MB vs 3-12GB for full models
Browse by Category
How to Install LoRA Models
Download LoRA
Click the download link to get the .safetensors file from Hugging Face.
Open ImageMates
Go to Preferences → LoRA Library in ImageMates.
Add LoRA
Click "Add LoRA" and select your downloaded file.
Configure & Use
Add trigger words, set weight, and start generating!
LoRA Usage Tips
📝 Use Trigger Words
Each LoRA has specific trigger words (like "GHIBSKY style" or "anime"). Include these in your prompts for best results.
⚖️ Adjust Weight
Start with weight 0.8-1.0. Lower weights (0.3-0.7) for subtle effects, higher weights (1.0-1.5) for stronger style.
🔄 Combine LoRAs
You can use multiple LoRAs together, but keep total weight under 2.0 to avoid artifacts.
🎯 Model Compatibility
Use LoRAs trained for FLUX.1-dev with the FLUX Dev model for best compatibility.