This LoRA model has been fine-tuned to produce artwork in a style I like, with a focus on delicate touches and soft focus.
Since it was trained on a dataset with many crying expressions, the output often shows characters slightly crying even without specifying it.