This model is a Flux Kontext LoRA trained with the AI-toolkit.

It was trained on 70 image pairs, with around 80% of them featuring Asian subjects. However, since Kontext LoRA learns concepts rather than specific faces, it works well with various ethnicities.

I recommend using a LoRA strength of 1 for best results.
Please include the following trigger word in your positive prompt for optimal output:

Remove makeup of this person