Since v1 was trained using only official illustrations, its art style may tend to be somewhat rigid.

It has been trained on Varar as well, but due to the limited amount of training data, it has actually no ability to reproduce that form.

v1は公式イラストだけで学習したため絵柄が固定されがちな可能性があります。

ヴァーラルも一応学習してますが、素材が少ないため再現性はほぼないです。