It may not be immediately clear why someone wants a LoRA for something that the model can produce with some acceptable cohesion, after excessive prompting, but I can explain it very simply. Token efficiency. I'm not sure why someone would spend 150 tokens on an item when 2 are enough with a LoRA so yea, while someone might be the worlds most awesome "Prompt Engineer" I figure taking a shortcut doesn't fall into some sort of moralistic category.
The training run was done with 6 images, of those only 1 was real.
Some people like the "real" look while I tend to enjoy things that I don't see often, these days, like 3D, airbrush and deep contrast pencil work.
This LoRA might accidentally lean towards real, or something that irritates some people enough to feel a need to critique the lighting, in order to feel good about themselves.
This can do at least two very different configurations which are a split leg pose and legs behind the head, the latter being more prominently trained and has more fun doohickeys without having to prompt them in.
Description
FAQ
Comments (7)
Didn't work for me but here are some prompts that are now working for me with no lora: "Her legs are flat on the ground, she is doing the splits, 180 degrees." or "Her ankles are shackled together behind her neck. This causes her legs to be pulled up to each side of her body." You can also say the shackles are invisible if you want to get rid of the bdsm aspect, lol.
The shackles idea to accomplish the position was a good idea. BDSM side effect would have been fun too. Was trying the legs behind head the other day and couldn't figure it out. Krea2 is pretty wild if you can think out of the box, it can probably do just about anything.
Realistic fantasy enactment... rest of prompt... along with realismslider fairly low (.2-.5) can get you some fairly unrealistic things but they look realistic... I think this is helping me some here even though it's realistic, it doesn't have training data on it.
Something something... "This action ~exposes/causes~... is also pretty powerful. "This action exposes the marbles that were inside of her vagina as they begin falling out of her." haha, it doesn't even have to be marbles, had me cracking up for a while. Rainbows is about where it stopped working but it tried, especially if you give her special powers, I can't quit laughing as I type this. I didn't try rainbows and unicorns for obvious reasons, though you may want to?
Krea2 can really figure out some wild things if you can figure out how to explain it. Sometimes if it's a really complicated scene with a bunch of things/people/events... preface the scene with a high level concept mixed lightly in with your quality/style header, then say "From right to left: personOne, personTwo standing, mashed potatoes, fork, whatever, personThree."
To the left:
personOne:
A 75 year old man with grey hair and dim blue eyes. is whatever act to personTwo. (brief description)
personTwo:
A 300 year old powerful wizard with glowing white eyes and an awesome wizard hat. Mention whatever personOne is doing somehow.
To the right:
personThree...
In the middle:
Detailed description of the events taking place. Use "he/his" or "she/her" sparingly if at all, use personOne is.. personThree has a.. and personThree's doing whatever with the thing he has...
It's kind of mind numbing, but kind of like this, personTwo holds a magic wand in his (probably fine) hand, personTwo is waves the magic wand back at personOne.
That structure is oddly only useful for fairly complex things and starts to fall apart with simpler things...
I should probably post this stuff somewhere more visible, I just started typing out some of my eureka moments and figured I'd keep going. Have fun, thanks for the effort! Like I said, I finally got the legs behind head pose from this concept.
There's quite a few great tutorials out there on how to get started with Krea 2, without a little bit of that knowledge it's unlikely any of my "adult" LoRA will work, for anyone.
Some people seem adamant about mentioning the accompaniments they use but I have all of that in every single prompt, in every single posted image, I do this routinely. I think maybe some don't expect that type of thoroughness so they don't bother to read the prompt, or even copy it I guess.
"didn't work for me", not "I can't get it to work." Cranking realism slider and the candid header kills so much of what you can do with Krea2.
Also, not accompaniments. Prompts that accomplish the same thing without using a lora.
I was just trying to be friendly.
@rando2048
Sorry, I don't quite grasp what you're trying to say. Maybe that's part of the problem you're experiencing, you're not fully capable of relaying an idea to the model, nothing wrong with that, I'm sure you'll improve with time. But, yea, you lost me but, you know, good luck over there.
good.
It's got potential, not sure what to make next.
Keep going !! Do it bester!!