PonyXL is not necessarily optimized for realistic humans. Your mileage will vary across different style tags and LoRAs.
Stable Diffusion has an unfortunate tendency to make everyone very white unless explicitly told otherwise. I've found that adding (tanned:0.1) to your positive prompt shifts the skin tone just enough to resemble Stephanie's. If that overdoes it, try adding (light skin:0.5) to your negative prompt. You may need to adjust up or down depending on your model and any style LoRAs.
A LoRA weight of 1 mostly works, but I frequently had to shift up to 1.1 or 1.2. Some style LoRAs can have a tendency to overwhelm the likeness.
Because I use ComfyUI, not all the generation details are accurately captured in the image details. Particularly, LoRAs used and ADetailer usage are not included in the metadata.
I'm openly soliciting feedback on this one! Please let me know how you think it could be improved. I've included the training data as well, if you feel like checking that for improvements I could make!
First pass at training. 8000 total steps.
1. Modèle partagé uniquement à l'apprentissage et au partage. Droits d'auteur et interprétation finale réservés à l'auteur original.
2. Auteur souhaitant revendiquer le modèle : Contactez officiellement SeaArt AI pour l'authentification. Nous protégeons les droits de chaque auteur. Cliquer pour revendiquer
