Another trashy LoRA! Yay!

DISCLAIMER!

Those who don't like experimentation better skip this LoRA and find something more stable and pretty.

So, well, I was going through my old LoRAs and came across a Tenebriflora. LoRA, that I really liked despite its AI-sloppiness. So I figured why not make another one!

Based on it and a few others of my LoRAs, I generated a dataset with a greater emphasis on maturity, curvy features and a bit of realism, while keeping the monochrome style and expressive eyes of a life-beaten dog.

However, this time I tried to make not a stand-alone style, but a style that can blend with other LoRAs and artist tags. It didn't work out xD Still, I think there will be a use for it, that's why I uploaded it here.

I'm still testing it, so I can't say exactly how to use it, however...

RECOMMENDATIONS:

- Put the weight at 1, and then reduce it to what you need, if it doesn't work in your workflow. I usually use 1. Sometimes 0.5-0.9.

- Trigger:

tmaix

Since the whole dataset was monochrome, I recommend to put the following tags:

monochrome, black background, simple background

Additionally you can put:

black eyes

- The quality improves after upscaling. In case of artifacts use Inpainting or Adetailer.

- By default more mature characters with curvy body features should be generated. But based on my first tests it can be controlled by reducing the weight and adding the necessary description to the prompt.

- Although LoRA has a style all its own, I recommend mixing it with other LoRAs and Artist Tags. As usual. You knew what you were signing up for >;D

Tested only on WAI-14, I don't know if it will work on other models. Meh.

I plan to upload examples of how this LoRA works. You will be able to view them here.

There's a very high chance I'll delete this model if it turns out to be very unstable. We'll see.