!!! Please do not use this LORA for commercial purposes !!!
This character comes from a Chinese game called 斩魂
With the Z-Image_Turbo published, this's a very good base model to generates image for scene composition and context.
My current workflow is to use ZIT generate the basic character compositions and environments, and then use Illustrious with LORA to partial redraws. The image generated in this way will have more tension and impact.
Tips for Qwen LORA:
The LORA will only be effective base model under qwen_image_fp8_e4m3fn or qwen_image_bf16 ！！！
Other quantitative models may be ineffective，such as svdq nuchaku models.
It also can be mixed with other styles LORA and 8-step LORA for use.
Qwen performs very well in natural language.Both Chinese and English can accurately express the composition of the picture. If you don't know Chinese, most of translation site is enough to help you translate.
Qwen 's prompt is very sensitive , remember If you don't want it to appear in your composition, then don't add it to your prompt.
For example, if you only want cowboy shot , do not add the lower body and legs foots prompt. Otherwise it will become fullbody shot.