This style was based on the style I personally got from V1.5 - Unrefined. Needs more images for this to be any great at a lot of things.
But this was starting with around a dataset of 35 images. In version two it's 55 more added on with my base images I use for most of my LoRAs, its 105 images total.