V5 IS RETITLED;

The rebrand is due to the concept shifting in a drastic way, but still fulfilling the same original goals.

https://civitai.com/models/466894

V4 TLDR;

This fixes a lot of problems with v2, and v3 wasn't released due to being too unstable.

I'm confident this one has a little bit of something for everyone. Hope you all like it.

TAGS;

• colored <- colors the current

• <COLOR> <- blue, dark blue, red, green, dark green, aqua, lime, silver, white, purple, yellow, gold, black

• bulge <- creates a bulge where the and are both contained within a .

• cutout <- creates a hole for the

This solves many problems with v2 and v3, producing a much more solid output with the various tags. Good luck and show off your skill.

It's still a bit unstable after some testing, so I'll be releasing 2 other epochs and then working on version 4.5 soon enough.

V2 TLDR;

USE STRENGTH 0.5-0.7.

NEW TAGS; colored , <COLOR> <- blue, red, green, aqua, silver, white, purple, gold, sometimes black

This lora is my first released attempt at a strange .

The plan here was simple; fix the terrible . I did, to a degree of success. It has some limitations.

By default; they are terrible and only work once in a while without inpainting. Even with inpainting you can be lucky to get them working with many tags and no lora. I felt confident just finetuning the base tag, rather than making a new tag just for it.

Even with the current loras, the were very hit or miss, so I decided to make one that not only works well with other loras, but also works well with other pony models than the autism it was trained on.

This is version 2, as version 1 was needed to just fix the baseline output without having to use NovelAI. I may make a version 3 to fine tune and fix some poses, but for now a little tweaking can make anything work. It's missing some colors, but it seems to do well if you use the colored tag.

• 1boy, hair color, face expression, clothes,

  • size, colored , <COLOR> , on , used on

• 1girl, futanari, hair color, face expression, clothes,

• optional ->

  • cutout <- cuts a hole for a in a piece of clothing, sometimes too

  • in <- often fills a , makes it nice and full, my guess is the model is already trained to understand this.

The training data was created over a period of a few hours and there are multiple other loras that can both enhance the use of this, change the style, or even alter the output in a large way.