Lora for fusing two characters together at the midsection. Not as consistent as I'd like it to be when fusing two different characters together, but works very well making a conjoined character where both halves are the same.

The main tags are (no legs), (duo), and (conjoined). I recommend against using the (solo) tag for most cases.

Using (gradient body) is helpful for getting two characters with very different fur colors to conjoin properly.

Not very much influence on style so increasing strength should help with consistency if needed.

The most consistent pose is standing upright using the tag (hand stand) with the main two will almost always work as long as (solo) is not in the tags.

(Lying) is less consistent but can work with several different variations, I highly recommend using a wide aspect ratio when attempting to make this style of image.

(Taur) is included in the dataset and should try to make a taur with two sets of front legs facing away from each other with a torso on each end. This one is fairly inconsistent. Weirdly keeping (no legs) seems to better than removing it. Increasing the lora strength should improve consistency.

If only using one character for both halves, you can use the (solo) tag, but you still want (duo) in the prompt as well.

Aspect ratio is important! Square and landscape images will have different results from each other, portrait is best for (hand stand).

If a character has a large tail that is causing problems using (no tail) should help.

Does not work with humans as far as I can tell.