This is my first SDXL Lora, and I trained it right here on CivitAI. He seems to work well with weight between .6-1 depending on the style.
P.S. It seems the metadata in the sample images I created in Fooocus-MRE does not carry over. I manually added it. I believe all of these were using dreamshaperXL10_alpha2Xl10 and refiner model: sd_xl_refiner_1.0_0.9vae, but if I am incorrect about that, please do not feed me to Lord Tatsuro. Incidentally, try Fooocus-MRE, it rules- so fast, even with my so-so 3070ti, and no nodes if you are like me and like a straight forward GUI.