Didnt have time to test it out. But it turned out okay from some of my test. Trained it using civitai :)

Dont know which dataset i used though.. some of the images have white borders.. Also dont know how to caption correctly.

Captions that matter -

[digital anime-style illustration of a young woman with red hair, pink peach hairpin on the right side of her head, large brown eyes, wearing a green long-sleeve shirt under a blue sleeveless vest and a pink short pleated skirt,]

WARNING : the model is caption sensitive (not sure if its the model I used or the lora) so make sure your caption is written plainly as possible.