**Note: only outputs the mid-tf part, the example images are just showing what the input/output was to create the final image.
Input 2 images, will attempt to generate an image of image 1 mid-transformation into image 2. (Order doesn't really seem to matter outside of which background it prefers)
*Might take a couple tries to avoid body-horror
Each prompt must have the phrase:
"Make image 1 mid-transformation into image 2."
Other things can be added, like:
Make their clothes rip/morph
Give them a [emotion] expression.
Make them [action]. (raise their arms slightly, grab their stomach, look down at their body, etc.)
Add [effect] around their body (motion lines, magic particles, etc.)
(see the prompts I used for the images)