Indroduction:
Lyla - red-skinned demon girl, original character of TheHelmetGuy. Loves big burgers.
Model used for previews: DivineEleganceMix [test]
Recommendations:
1.0 weight
lylaas "support" activation tagadultfor more "milf" versionshortstackfor round-faced shortstack one(nipples:1.2)in negatives to remove "covered nipples" effect
Suggested regular prompt:
red skin, black sclera, colored sclera, blue eyesfor skin and facewhite hair, long hair, blunt bangsfor default hairliterally anything you want next
Credits:
Trained on NAI Anime Full. Dataset of 65 pictures. 10 Epochs. Clip Skip - 2.
Really BIG thanks to:
TheHelmetGuy for beautiful character
Some good folks from Discord server
You, for reading, downloading and reviewing this LoRA
Description
Initial Release
FAQ
Comments (11)
If I remember correctly she doesnt just like big burgers.
Which opens up the question for other TheHelmetGuy OC's. Those with other big thinks Lyla likes.
This was just a little joke about one certain art of her.
yoooo would you ever make a lora for nekoarashi's character https://twitter.com/Arashi_Yagatake
id love to see her come to fruition
Hi, yeah, sure. I'll try later. If noone will be ahead of me. Working on my other LoRAs updates right now. I also like nekoarashi's character.
@TroubleDarkness I would also love to see a lora for nekorashi's characters but take it at your own pace Darkness, love your shit!
Hi, would start working on her soon.
@TroubleDarkness you a homie of homies <3
@Yabba Hi, actally i'm done with her, because of most images being from the same artist (nekoarashi, of course) it's overtrained in style. Was not uploading it until this day because of CivitAI lags, it's really a self-harm trying to do something on CivitAI while it's lagging. Would upload LoRA for her toaday.
@TroubleDarkness ahaha all good when isnt civit lagging or down T_T
So I had a question, I wanna train a model of a specific character, but the character has very little official art style, will the model of the character model get worse/change results if I train it with images of different art style from different artists? What's the Best thing I could do in this situation?
If you want to train character, you'll really want to use any accessible art on the internet. If you'll train character only with single artist style, of course, it would bake style as well. If your desired character available in different styles - this is your way to do LoRA.
Of course, if you don't want the same style as original authors style.
Than more different styles in dataset you have, the better result you'll get and less style baking you'll have.
If you're talking about mixing your character with just random art of different styles - well, it can help sometimes, but it's better not to rely on it.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.