I've trained this on one of my characters to learn how to train a SDXL LoRA with my hardware. You can customize the character's look by changing the triggers.
ComfyUI workflow is embedded in the NoobAI example images. Generated with ChromaXL Sorbet.
Weight 0.8 - 1.0
Pairs well with style lora's.
You may need to add the following tags to negative prompt if you get a strong 3D look. (This is because a majority of the training data was from a 3D character model)
source_3d, 3dFeel free to generate whatever you want.
Enjoy!