https://mega.nz/file/xaFhCIQQ#nVcm45mFTPWVth674iSrJVDXL7SrQjBRjuy22q1kQkE

Trained using 174 images on PonyDiffusion

Activation keyword : sagittal
strongly recommend including bisection, brain, exposed brain in the prompt to make the body more likely to be separated. Do not include them, or include them in negative prompt to make it look like it's not completely cut apart

While gathering the training images, I found some images that were definitely AI generated, so there is already a LoRA or another model out there, but unfortunately I was not able to find it.

The image's quality degrades pretty much because I couldn't find enough high-quality images.
Also the image might break during generation. In that case, alter the prompt a little bit or use different epoch.

I'm planning to take a look at previous training sets and find a way to improve it. I'm giving up and coronal plane cutting because of a lack of decent images, and honestly, it seems to have a very low demand.