This is Jeremy Barnel's Sexyfur character Sasha, trained on images drawn by (Elton Portihol) Eltonpot/XXXPot
Use this with Pony Diffusion V6 or with Incursio's Meme Diffusion (SDXL, Pony), Autism mix
Keep the LoRa weight to either 6-7 for some slightly accurate results
Trigger Words:Sasha, rabbit, lagomorph, leporid, anthro, bedroom eyes, breasts, Spanish green eyes, cheek tuft, chest tuft, clitoral hood, clitoris, labia, ruddy pink nipples, ruddy pink pussy, vagina, vulva, rabbit girl, shimmering blush rabbit nose, rabbit ears, ear tuft, rabbit tail, eyelashes, eyelids, facial tuft, right pastel pink eyeshadow, left sandrift eyeshadow, half-closed eyes, lips, long hair, narrowed eyes, cape honey fur, grandis fur, makara fur markings, two-tone body, two-tone fur, two-tone tail, pastel pink lips, true v hair, tuft, fur tuft, wide hips,
Sampler: DPM2 Karras or Euler a
Steps: 30
CFG Scale: 7
Description
FAQ
Comments (14)
Pardon my ignorance but would it be a better qualifier for "Elton Portilho Style" in the description? It's a great model!
I'll be honest I've been trying to figuring out properly source these models. Mainly because the model I trained it on is on character not style. I was considering to do it as you suggested, but should I also label it to the one who own/created the character?Which in Sasha case would I guess is JB. Kinda like Mojita was (I think) created by scappo, or like Cassie was created by Bonk.
But anyway I am planning on renaming the models for Pony XL, to make it easier to identify them for users and make it less confusing. So I'll just label them as trained images by EltonPot and character owned/created by JB.
I'm planning on releasing SD 2.5. now that civitai allowed multi submission training. I just really wish they made training a 4 steps instead of 3 cause in step 2 is where I keep getting issues and having to restart from scratch. Even making sure the images are there by clicking "next" to save even after auto tagging and going "back" to edit tags. Just think they should just do some sort of "save as" function.
@ScorchingFlames
I think the best way would to be sighting the creator and then the style. In this case it would be Jeremy Bernal's Sexyfur Sasha in Elton Portihol's Style. The formula would be:
"(Character Creator) (Homesite) (Character) (Applicable Artist style)"
That way you cover your bases, give the proper credit, and acknowledge the "Originators".
While some artists will be angry at the use of AI it's only a matter of time before they start using it to speed up their output and present themselves as the curators of the IP and custodians of its direction enabling them to monetize their own style directly.
I have to say, I wonder which sexyfur character you do next, there’s some like SurfBunny, Kaylee, Cheery Berry and Lontra for example
It's going to be a bit before I get another one out. I have plans on reworking models prior to Beth, (even Dawn) since every lora I made has something new I've been doing to the training. Since then, those models have been optimized more efficiently. So I want to do the same with the past models. I might do either Anita version 2 or Miami version 2, after that I'll do a new model.
Plus I'm also busy chipping away at Elden Ring's Shadow of the Erdtree DLC. So yeah maybe a week or two finishing that. But I'll do the rework models since I still have the datasets. Also including working my job as well.
@ScorchingFlames Sounds cool, but hey, wait until you see Kaylee the Arctic Fox, one of those sexyfur types bro, she's cute.
Great Start, but this Model still need Work.
The Art look more like the Work of the Co Artist Elton Portilho, not like the Art of Jeremy Bernal, the Creator of Sasha and the other Bunny Girls like Orchid and Gunbunny.
For some reason it's claiming to be an SD1 Lora rather than an SDXL and so it's not working with Incursios
That's weird there doesn't seem to be an issue with the settings it still says it's PONY/XL and not an SD1. It still generates on my end it might have been something with the site.
@ScorchingFlames Well here's the attempt that says it's SD1
https://cdn.discordapp.com/attachments/1218110763835723817/1345522856821522643/Untitled.png?ex=67c4db4f&is=67c389cf&hm=e84ed654336e3d74b7dfd40f29794493419fd63dedf2fd56d50ae2f94a0766e8&
@SpongeO wait are you doing this locally or on another site generator? If you are I think I know what happened you probably downloaded the wrong LoRa file. I just noticed that for some reason I have two files, I don't know why I have two. But the larger file is the one you're going to want to use. I try to remove the file that's not meant to be there.
@ScorchingFlames Okay thank you it's working now! and Yes I am doing it locally.
plan creating sexyfur style?
Honestly I wasn't originally but if I can get all of my data sets back (I have redo like 6 of them) I might have a prototype eltonpot style somewhere next year maybe by Jan. I Want to at least include every species that Elton has drawn. It will make the LoRa a bit more refined. I'll be more active once I go on my vacation that I might get it done faster in Nov. And Dec.
You can use most of the LoRa I have post as a style you just don't put any of the character descriptions in the prompt. You put what you want in there. But I have seen some elton LoRa made by other users that are good. Though I think the Weights might be a little too high I feel 0.6 - 0.7 to be the balanced point.