This is my first attempt at training style lora. The dataset I used is the KAZUMA1-3 part of DanziEngine's work, so the characters are relatively simple, but the results are very successful. If you like this style, please support the original authorhttps://danziengine.fanbox.cc/