Download
1 variant available
Created on Civitai
*First attempt at doing as single model, work very much in Progress*
Dreampersona is a cinematic portrait LoRA trained entirely on AI-generated images, designed to produce expressive, highly realistic female characters with strong facial consistency and dynamic lighting.
Ideal for close-ups, portraits, soft lighting, and realistic aesthetic workflows using SDXL.
⚠️ This model does not depict any real person — it is a stylized fictional identity.
Best results with:
– Trigger word: dreampersona
– Sampling method: DPM++ 2M Karras
– Best model for me has been : Jib Mix Realistic XL
🔧 Works well in styles ranging from cinematic realism to moody fantasy.
Trained using carefully curated Civitai-generated images with consistent lighting, angles, and expression diversity.