/kemono/ - kemono.party

Without more details or at least more than a single cropped image, I'm not sure how anyone can really give you a confident answer, but as a rule the best place to look for AI related tools like checkpoints or LoRAs is probably CivitAI (with PixAI coming in a semi-distant second). Where they "come from" to get there is a matter of people training those models themselves.

Also, that text is most likely just garbage spat out because the images in its training set contained signatures consistently enough that it thinks they are a part of what's being asked for by the prompt. It's an especially strong artifact with artists who have a very consistent size, placement, and style of signature, and also for things like watermarks or other icons that some artists put in the corners to advertise their patreon/substar/etc.

>>41093
He used novelai. It costs $25 a month. This particular style is usually a mix of cutesexyrobutts, blushyspicy and nyantcha, YD, etc. You'll see this style all over pixiv if turn off your AI filter. Some of the uploads there have unscrubbed metadata.

If you don't want to spend $25 to generate slop, just use stable diffusion (a1111 or forge) with the ponyXL checkpoint + the floox lora from civitai.

>>41233 TensorAI is pretty good, too.

SeaArt, though, depending on the generated image, will censor your output because prudish, dinosaur, kike overlords (whom, more than likely, are monopolistic pedofiles/hypocrites themselves) have final say. They also cucked AI training for TensorArt from using explicit images, speaking of.

@AnonAdmin I've a VPN. Cry harder, bitch nigger.

>>41093
Thats truck kun if im not mistaken

>>41233
I went on NovelAI and all I see is NAI Diffusion Anime V3 V2 and V1

Is that the style you are talking about?

And for stable diffusion (a1111 or forge) can you run that locally, is there a video guide on how to install it and run it locally.

And finally most important part: what are the types of prompts used. Because tbh making ai slop is way cheaper and from the looks of it easier if you know what to do.

>>41333
It's v3.

Just know that local requires a fairly beefy nvidia gpu. There are install instructions on their respective github pages, or you can watch a YT video about it.

Anything else you need to know you'll have to figure out yourself, it's very painful process overall and requires a lot of trial and error.

Just start with ponyxl/autismmix and adetailer.

You can learn more about loras and prompts by checking the stuff uploaded to civitai.

>>41345
I have an rtx 3060, and for yt vids you got any recommendations because github pages ion think that'll help me out