Judith Lanster (Kuro no Kiseki)

V2 Released! Details Below.

This is my first ever Lora! I hope you like it!

Includes 2 outfits, her regular and her swimsuit. Be sure to check the trigger words.

V2 Details:

Tested and works great with AAM XL, JuggerMagine XL and GhostXL.

Here are the triggers for the two outfits:

KnKJudithLanster wearing SignatureOutfit, coral-pink hair, orange hair tips, turquoise eyes, one earring, open dress with belt, detached sleeves, bracelet, coral-colored shorts, black thigh-highs, aqua high-heels,

and

KnKJudithLanster wearing SwimsuitOutfit, coral-pink hair, orange hair tips, turquoise eyes, gold earrings, necklace, bracelet, thigh accessory, lace high heels,

Not usually needed but in some cases if the default outfit doesn't come through precisely you can add the basic trigger of the other one to the negative prompt. i.e. if you want the SignatureOutfit ,you can add "SwimsuitOutfit" in the negative, and vice-versa.

If you aren't using any of the included outfits I advise adding "detached sleeves" to the negative prompt, otherwise there is a high chance all clothes will have them. In some cases also adding "open dress" to the negative helps if the clothes you are trying to make are "open" in the front or back.

Haven't seen major issues from my testing, let me know if there are any.

Weight: Best between 0.9-1.0

Enjoy!

If you liked my work you can buy me a coffee:

Ko-Fi: https://ko-fi.com/z1gamer

or crypto: https://coindrop.to/z1gamer

V1 Details:

Recommended Settings:

For best results use:

"KnKJudithLanster wearing her SignatureOutfit"

or

"KnKJudithLanster wearing her SwimsuitOutfit"

Weight: Best between 0.9-1.0

Sampler: Depends heavily on the model used, but generally DPM++ SDE Karras and Euler are good starting options.

if you have some issues with her hair or eye colors, you can put:

"coral-pink hair with orange hair tips" and " cyan/turquoise eyes" (one of the colors :D), at about 0.6-0.8 weight and somewhere towards the end of your prompt. Usually helps.

I highly recommend using the AAM XL model, it had really good results from my testing.

Test other models as well if you like. Digital Dream, Juggermagine and YamersAnime also produced some good results for me.

I know it has some issues, some of which are due to the lower settings used in the training process (for example 768x768 instead of 1024x1024), due to not enough VRAM on my GPU :D.

Still, I hope you can make some nice images :). Feel free to share them here.

Enjoy!

P.S. I am Sorry there are no generational parameters on the images. I Am using ComfyUI and I thought CivitAI could now properly read the parameters from comfy images, so idk what the problem is. I will try to look into it when I get some time.

If you liked my work you can buy me a coffee:

Ko-Fi: https://ko-fi.com/z1gamer

or crypto: https://coindrop.to/z1gamer