Trained for PonyXL with anime screencaps - thanks to https://fancaps.net/ - using kohya_ss. I would mostly recommend a weight of 0.85 to 1.
For specific traits/outfits use:
halterneck, black choker, armlet, arm strap, bare shoulders, black gloves, single glove, black dress, sleeveless dress, pleated dress, black belt, calflet, thigh strap, black footwear
bound by spell, glowing rings (may want to combine with: hula hoop)
half-closed eyes
polearm, holding polearm
Since this model is trained on screencaps I also deliberately train to replicate the anime style to get the character likeness as close as possible.
If you want to apply a different style you can lower the weight (around 0.85 should be good for most).
What is Pony? It's one of the best SDXL checkpoints for anime (and more). It intrinsically understands a lot of styles, most characters and most importantly it will generate what you ask for with a much higher probability.
Description
FAQ
Comments (8)
I can fix her
I'll help you fix her
No you can't she's already perfect
what tagger do you use?
DeepDanbooru (https://github.com/KichangKim/DeepDanbooru)
However I go through all images and adjust the tags - tagging seems to be the most important part of the training process if you care about precision.
Awesome! Female characters such as Fern are also pleased♪
Hey guys I'm new in SDXL and I'm trying the lora but it's not working, what can I be doing wrong? I tried with Animagine XL 3.0 and 3.1
also doesn't work by downloading on bactchlink as well
Hey, this LoRA only works with Pony XL V6 based checkpoints (check suggested resources). It doesn't work with Animagine.






