Trained on 124 images. Previous version was trained on 89 images, this was trained on 124.
From what I can see, it generates more accurate-looking pictures of Ayumu Kasuga. Some of the pictures generated are starting to have that Kiyohiko Azuma art style.
Trained on 124 images. Previous version was trained on 89 images, this was trained on 124.
From what I can see, it generates more accurate-looking pictures of Ayumu Kasuga. Some of the pictures generated are starting to have that Kiyohiko Azuma art style.
Recommended you set the LoRA weight from 0.7-1.