This was one of my earliest loras, so I'm revisiting it with what I've learned in the interim. Compared to v3, it's larger. The training images are the same, but the way they're used is different.
There are 84 images with 2 repeats, and 64 images with 5 repeats. No regularization.
This was one of my earliest loras, so I'm revisiting it with what I've learned in the interim. Compared to v3, it's larger. The training images are the same, but the way they're used is different.
There are 84 images with 2 repeats, and 64 images with 5 repeats. No regularization.