A general purpose "detailer" sort of Lora in the same vein as Zoot Fluxxer XL, Zoot Detailer XL, etc. Trained on more than 500 images all generated specifically with Flux Pro Ultra and recaptioned with JoyCaption. Recommended inference strength is around 0.7 to 0.8. Load this in ComfyUI with the built-in "LoraLoaderModelOnly" node (since there's no CLIP component as Kolors only uses ChatGLM-3-8B for text encoding).