Trained on 100 copies of the same image.
One of 4 dataset images i used for my main Wakana LoRA.
https://pixai.art/artwork/1764386334511263209
Since the dataset image has usual AI mistakes, it kinda got reflected in the LoRA model as well.
This is the only image of the 4 Wakana images that has a side ponytail. Also her holding up her arms in this image influenced the main Wakana LoRA to generate most of the images like this.
