Nota : Pony being based on SDXL, my pony / SDXL models can work on SDXL / Pony checkpoints... or not, and in sometimes surprising ways. You have to try.
Sometimes, a Pony/SDXL lora can even work on SD 1.5, but of course isn't made for good results like that.
Here's my version of kitty katswell. Kitty is a character from an old cartoon called T.U.F.F. puppy, in the years 2010's (really ? I though it was 1990's).
Trigger : The trigger is kittykatswell.
As the choice of training images is poor, you will have better results by adding more prompts used by the training.
Change the word COLOR for the color you want, you silly goose ! Or just don't give color at all. Looks at my examples !
Honestly, look at the examples. Do you know the time it took me to do them on all those checkpoints ? I don't have a 600$ graphic card !
the word color is meant to be replaced by the color you want. Throw as much paint on the poor girl as you want to !
Prompting :
The points here are captioned (used in the training, meaning have strong effect when you use them in your prompt) :
COLOR hair, COLOR bangs : By default Kitty has black hair. I'm not sure color bangs has a big effect.
cat tail : For some reason, Kitty lack sometimes a tail.
COLOR body fur : By default, Kitty is beige
COLOR headband : By default, Kitty has a white or a pink one, but it rarely appears without this prompt.
COLOR eyes : By default, Kitty has green ones (putting perfect before this prompt will give better eyes).
Other prompt :
All images have also been captioned with anthro and furry. Will help on Anime checkpoints, maybe also on furries, you try. Particularly true on Pony.
Her cartoon uniform can be shown by :
COLOR pants, COLOR sweater, COLOR jacket, COLOR elbow gloves, COLOR high heels boots.
By default, Kitty has black or grey pants and jacket, white for the others.
belt and belt buckle are captioned, and making some color belt, color belt buckle will often do some difference between the belt and the buckle, and make the buckle more visible, but don't expect a strict respect of colors : blue belt, red belt buckle can make red belt, or a belt of the pants color (probably because the buckle is nearly always present with the belt in the training images, and have the same color than the pants).
Also, you will probably want to weight more the color of the gloves, since the model like to put them in white on most checkpoints (That's basically why I took the 16th epoch of training and not the 20th : More epoch, more the model become an*l on some things).
Note for Pony :
Remember : With Pony, it's better to start your prompt with :
Positive (normal) prompt:
score_9, score_8_up, score_7_up, score_6_up, source_furry BREAK
The BREAK can be ignored or not. Depend of your needs and your checkpoint, honestly.
Negative prompt :
score_4, score_3, score_2, score_1
That can also help.
SDXL version.
anthro, furry in the prompt can be necessary on some checkpoints.
Tips:
Basically, the more detail you add in your prompt, the better Kitty will be.
This is because the few number of images I could find for the training dataset.
clip skip can be 1 or 2. You should try.
Lora weight : between 0.7 and 1. depends of the checkpoint.
Trigger : The trigger is kittykatswell. You can put no weight on it, or give it 1.1 or 1.2, particularly if you put a lower weight to the lora.
Also :
Generation with DPM++ 2M / SDE and Automatic / Karras works well
20 - 25 sampling steps should be largely enough.
Upscaling upgrade the image a lot. I use R-ESRGAN 4x+ Anime6B, hires steps 0, Denoising strength 0.35, upscale by 2, and no refiner.