This is the beginning of an all-in-one breast contest model. I intend to continue fleshing this out until it can generates every scenario you can think of for this concept. One big focus is splitting the tag asymmetrical docking into it's different variants. In case you don't know, asymmetrical docking consists of misaligned docking, height difference docking, size difference docking, and one breast docking. Each of these will be given their own tag so less prompts are needed to generate the desired result.

V1.0

Apparently I'm too dumb to realize how posting images works and thought I deleted this model for 2 days. V1 has a larger data set and fixes the forehead-to-forehead issues, as well as increases the overall consistency of tags.

V0.3

Views, one breast press, and sizes are now consistent. I've also changed the tagging to try an define the different types of asymmetrical docking, but not all work yet. Asymmetrical docking will generate images with breasts misaligned or on breast pressed more than the other. Size difference will generate breasts of different sizes, but also seems to generate height difference as well. Height difference currently doesn't work as intended.

Furthermore, this version can produce decent catfighting scenarios. Helper tags like holding hands, hug, wrestling, fighting, etc. can help narrow down the fighting specifics. Grabbing another's hair doesn't generate consistently yet.

Also, face-to-face and cheek-to-cheek are support now. Forehead-to-forehead seems to generate the same as forehead instead of what it should. That will be remedied in the next version.

V0.1

Currently, this model supports views from above (top down, from above), from the side (side view), and from below (bottom up, from below). I tried to include height difference (diff, asymmetrical docking), and one breast press, but I haven't found enough data to make these consistent. You can sometime generate these concepts, but don't count on it being consistent yet.

To do:
- Support lightning glare
- Make one breast press generate consistently
- Add fighting aspects (catfight, hair pulling, etc.)
- Support a wider array of breast sizes
- Support more views
- Improve facial expressions
- Better distinguish clothed/unclothed generation