!!! Please do not use this model for commercial purposes !!!
Version 5:
Based on the version4, the overall art style incorporates more semi-realistic, while improving the texture of character faces. The environment is optimized to be cleaner and more minimalist.
This Version also has its own SDXL VAE built in.
Version 4:
This version is primarily optimized for facial features. The native face model leans more toward traditional Chinese ancient style , while also removing some of the oily effects. The overall art style is more 2.5D, with character faces leaning more toward anime/2D style. Meanwhile, the scenes tend more toward a gorgeous, sparkling effect. If you prefer a simpler, flatter environment, you can add corresponding descriptions to the negative prompt.
This model has its own SDXL VAE built in.
Version 1~3:
***Please use SDXL VAE or Pony VAE, when you using this model(Version1~3) ,It will help you to solve the white spots and white-ish issue
https://civitai.com/models/296576/sdxl-vae
This is an original Checkpoint which combine many LORA I uploaded before.
The style is more like semi-realistic.
You can use the prompt attached by pictures.
I retrained the model, and the third version will be cleaner and more concise, with facial features more inclined towards Eastern aesthetics
If the image resolution less than 1536x1536 may have blurry and fuzzy result.
High-Res repainting will solve this problem.
Good pictures always need to be finely crafted
If you have some color or overexposure issue when you generate the armored character. You can put these high contrast,yellow light,overexposure, in your negative prompt.