The YARN (Yet Another Realistic NSFW) Collection includes diverse styles for use with Krea 2. While these Loras can be used one at a time, we recommend mixing different Loras (strength ~0.2 - 1.5) to combine their strengths.
Unless otherwise specified, pictures in the linked post are generated with strength 1.0, and only one Lora, which differs greatly from recommended use.
Here is a quick description of the style intended with each lora, starting with the latest releases.
Slayed
How many licks does it take to make a woman come? Find out with these glamour-chic women only studio photographs.
Zishy_v2
A new take on candid photographs of the cutest amateurs models, in a better Lora trained from scratch.
Deeper
What happens when the director thinks they're making absolute cinema.
ALSScan
Retro solo and BG shoots with shaved amateur models.
FH18
A simple story that repeats endlessly: Hot girl comes to the massage parlor, gets undressed and lies down on the table. She's in good hands, and the scented oils help the tension leave her body. But now she's so relaxed she wants more.
Hegre
Nothing is superfluous in this austere look on erotism.
ShowyBeauty
Who ordered the most beautiful women on the internet?
BlackedRaw
Same vibe as Blacked, but at night with a blinding flashlight in the performer's face.
GFRevenge
It's not about getting even, it's about remembering the good times that are past.
Throated
These women have the pornstar look, and do not care about their mascara getting smudged during messy blowjobs. Pictures in the linked post generated at strength 1.5.
Phil Flash
Take a trip back to the 2000s with these playful cuties. Short skirts, ultra tight low waisted denim, tiny plain thongs are all on the menu.
Tiny4K
These tight vertically challenged women take on girthy cocks for your viewing pleasure.
Ersties
These German lesbians are fully deconstructed and ready to take you into their world.
BrattySis
Thansksgiving, Christmas, Easter ... Every holiday is an excuse for unforgettable family memories. But coming back home to your step sister can wake some tensions. "Help, step bro! I'm stuck in the dryer."
AbbyWinters
Discover the best of the bush with these Australian amateur lesbians.
InTheCrack
Heavy makeup and shiny bubble butts for those who don't fear looking down the barrel.
Vixen
Glamour and provocative, these models pose around the villa or at the beach with "that" look, before giving themselves to their audience.
Blacked
Bold and beautiful women face shiny BCCs in front of studio lights with a polished professional aesthetics. Starts with high end lingerie or sporty underwear with the signature belt, end often ends up with a cum shower.
FTV
Beautiful young woman lift up their dresses in public, undress in front of the camera for the first time, and will make you blush the next time you're in a room with a cucumber. Side effects include upskirts, pussy spreading, and deep insertion of sextoys.
Tushy
The sun shines on the glistening bodies of these beautiful woman, but not where they like to take their cocks. In this primarily anal focused experience, the ultra-rich bang indoors and outdoors, near the pool or in the sand.
ThisYearsModel
Fall in love with these young mischievous women. Simple, warm and beautiful.
SexArt
The prettiest women know they are the prettiest. Mostly Eastern European woman wanting to make you feel good.
MetArt
These nude artistic pictures have soft edges for a dreamy look on the most flawless woman.
Zishy
Candid photographs of beautiful amateur girls next doors.
Description
These German lesbians are fully deconstructed and ready to take you into their world.
FAQ
Comments (14)
The BLACKED model is so good!
Reverse it, then I am in :)
I think you should merge these models all together in a single checkpoint. I'm getting really good results when doing that.
This would certainly be a good idea, which I'll investigate once I'm done training the few next LORAs I have in line. I'm concerned though with the loss of flexibility (you can't change the weights anymore obviously). What YARN Lora and weights are you using?
FTV, Blacked, BrattySis
They are all very good LoRAs for fixing the artificial faces
This should be helpful to merge models:
https://huggingface.co/docs/peft/developer_guides/model_merging
I strongly suggest using SVD since it seems to give better results than linear merging.
@WhyNaNWhenNyan Thanks a lot. Like I said, I'll investigate, but don't expect anything before a month at the very least. For everyone reading this, please do share your favourite combination (with weights). I'm convinced combination of these Loras is the right way to go. New Lora dropping tomorrow or the day after if training goes according to plan.
@TimmyHodor Wouldn't there be a caption overfit and overlap problem by doing this, reducing variation? As stated by the K2 devs themselves, its better to train several loras instead of training a single one. And even if several concepts in one are attempted, it better requires some good curation or a good distribution of concepts throughout the dataset in relation to its size, especially when it comes to a multitude of faces.
And unlike e.g. Qwen 2512 or ZIT/ZIB, I think K2 has a harder time not to generalise a wildly diverse dataset that is more true to the base than your dataset.
@JouzaDouza7135 I don't understand your point, can you please exaplain further? "Wouldn't there be a caption overfit and overlap problem by doing this, reducing variation?" ==> What is "doing this" ?
@TimmyHodor well if you would merge the loras it would certainly bring down the e.g. variation of facial diversity or distinctiveness of each lora since captions would be shared in a merge, requiring recaptioning or the use subtriggers to somehow retain the variation
@JouzaDouza7135 Unless I'm missing something, it would be exactly the same as using all the loras together at once. There is no captions involved.
@TimmyHodor true but imagine using multiple loras of yours and using a prompt with facial features x,y and z. if two or more loras are captioned and trained on exactly the same tokens of x, y and z, it will average out
@JouzaDouza7135 Isn't that the point? In any case, I personally have had a lot of very good results combining these Loras with various strengths. So have other users (who contacted me in comments or DMs).