Anime to Photoreal v1.1 (Qwen-Image-Edit-2511 LoRA)

The purpose of this LoRA is to make it easy to convert anime characters to a photorealistic style while preserving their identity. Every image edit model I've used for this in the past has always made an anime character look too human to the point that their identity is lost.

v1.1 improvements:
- Can convert more kinds of art styles
- Slightly better skin quality and detail in clothing

How to prompt:

Use the following prompt: Kigurumi of the anime character, realistic details


Examples:

The following generated images were the best of 2 tries each:

Input Generated
Input Generated
Input Generated
Input Generated
Input Generated

Tips:

  • The output dimensions should be close to the input image dimensions.
  • Ideally the input image should be ~1.5 megapixels.
  • Ideally use an output size of ~1.5 megapixels.
  • Sometimes the model will change the skin tone or other colors from the input, but you can fix this by adding more to the prompt after the required phrase, such as "Pale skin.", "Dark skin.", "Blue themed.", etc.
  • Some anime styles might work better than others
  • Try making the background of the input image a solid color if you're having issues.
  • You can get better face quality if you crop in closer to the face in the input image.

Download model

Download them in the Files & versions tab.

Downloads last month
37
Inference Providers NEW

Model tree for Hyperccino/Qwen-Edit-2511-Anime-to-Photoreal-v1.1

Adapter
(87)
this model