Trained on "kobold", "lizardfolk", and "dragonborn". Using Dreambooth, trained for 6000, 10000, or 14000 steps. I recommend using the 14000 step model with a CFG 4-8. You may need to use the models that were trained for fewer steps if you're having difficulty getting certain elements in the image (e.g. hats).
You can also use a higher CFG if attempting to generate inked images. E.g: CFG 9 and "photo octane 3d render" in the negative prompt:
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API:
The model has no library tag.