Please Read
Hello everyone, I wanted to take a minute to address something. As I began thinking of styles to create, one of the first ones that came to mind was one for Skuddbutt’s artworks. I have always had an appreciation for Skuddbutt, and when I tried my hand at animating many years ago I even downloaded some models that were posted by them!
However, I have always been reluctant to make this LoRA, and that’s because Skuddbutt has recently had some health issues that are affecting their personal life substantially. I have heard that this has been very debilitating for Skuddbutt; a full link to their public post is available here. As a result, I didn’t want to release a LoRA that might affect their profits and thus their ability to achieve treatment.
So, why am I making this LoRA now? As I reviewed Skuddbutt’s many releases, it became clear that most of their works remain as animations with only a few being still images. I feel as though this LoRA is less likely to gain traction as a result, and that it would also be less likely to affect Skuddbutt’s profits. Skuddbutt is a phenomenal animator and artist; I feel as though this LoRA is only a shoddy imitation of the artwork that they produce. Hopefully it will provide some degree of entertainment during Skuddbutt's indefinite hiatus.
Please consider supporting Skuddbutt, a link to their profiles are included below.
@skuddbutt on Twitter
Skuddbutt on SubscribeStar
About This Model
This LoRA was created from approximately 100 high quality images of Skuddbutt’s artworks that were posted online. The majority of the dataset is comprised of images of Gwen, Mavis, Nani, Helen, and Azula. As a result, some images might default to appearances resembling these characters unless otherwise specified in your positive prompt.
Possible Issues
It is rather well-known that IllustriousXL struggles with 3D styles (moreso photorealistic 3D). I recommend using a model that has a dataset that is not strictly anime as this LoRA might struggle to produce desired images under those circumstances. I have found that NAI models (or merges with NAI) tend to produce better results.
Description
FAQ
Comments (11)
Nice quality overall.
Thanks; I already made a V2 that I am testing. The main issue with getting the quality better is that images start to get that overbaked/overtrained look. I'm trying to find a good balance before I release a V2.
awesome
If you think it's possible, can you, one day, make also a lora of his Meru the Succubus series?
It's a great series that he sadly couldn't make more episodes due some issues.
Also thank you very much for the loras you share!
Also hope Skudd's health improves. Here to wish him only good things on his life.
Have a great day Twisted and also, wish lots of great days to Skudd's too!
I can absolutely make a LoRA of that as well. Just send me a DM about what you would like the LoRA to accomplish. If it's just Meru, then the LoRA shouldn't be too difficult to make. If you want peripheral/side characters it is all dependent on how many images I can get of them.
I will also say that there are a few other users on this site who might already have a version of that available for Illustrious, if you would like to try theirs! For example, TecnoAI has a great version available here: https://civitai.com/models/108077/meru-the-succubus-skuddbutt-version
For now, I have delayed a V2 because I'm not certain that there is a need for one. Nor am I confident that the V2 I have generated will satisfy anyones' needs. Does anyone have feedback regarding this?
NEED a good Faye Artemis LoRa bro
DM me details or put them here
@Oddities not sure if you’ve seen the video Skuddbutt made of her, but it was a futa shapeshifter character from LemonFont I think. Skuddbutt made the animation a while back so I’m not sure how much there is to work off, but lowkey this LoRa is more accurate to his style than I thought so maaaybe it’d be possible
@iHateTheMIC I can take a look and see how much content there is of the character. If I can't get enough unique angles/shots then it is going to be difficult.
Edit: Gave it a look, and there's two main problems. There are very few different views, which means that the model might struggle to put the character in different contexts. Additionally, the fact that the character shapeshifts means that it would be difficult for a model to understand them as a concept with so few images.
All of that being said, I might be able to make it work? Can't promise anything. I don't know when I will get it done.
@Oddities oh no pressure I know she’s kinda niche but it’s also nice to just get her name around yk