Warning.
The use of this LoRA is prohibited for illustrations where the generated results appear to be, or would be, underage.
Whether the generated illustration is R18 or not is completely irrelevant. It is prohibited even for all-ages content. "It is troublesome to judge whether it is OK or not one by one! It would be easier to ban it uniformly."
Even if the actual age is 495, if the generated result looks like a minor, it is not permitted.
In the unlikely event that I discover a violation of this rule, there is a "possibility*1" that I will take action as I deem appropriate, taking into account platform, maliciousness, etc.
The "possibility*1" here is not a common sense possibility. The "possibility*1" is within the scope defined at the end of this document.
As a general rule, I will deal with violations that I myself find according to "possibility*1". In the case of violations against the rules of each platform, please deal with it according to the rules provided by the platform provider, as it will be the platform's problem.
Important things to prevent problems
Only for the above violations, even if you guys find them, please avoid pointing them out directly. It may cause trouble with users.
What? Did you get into trouble with the violator by pointing it out directly? I don't know, solve it by yourself. Even if you ask me for advice, I will throw your message in the trash.
Purpose
This LoRA is designed to generate illustrations dressed by mature women and to enjoy the gap between them! No other use is envisioned! Well, even a good-looking gentlemen will work. Maybe.
How to use
Trigger words: kindergarten_uniform
For "big breasts~huge breasts" adult women: 1.2~1.3 is recommended.
1x is recommended for flat chested adult females.
Optional word: school_hat, bag, school bag, [color]_skirt OR [color] skirt, pleated_skirt
Negative (Oprional): animal ears
Deprecated in most cases: white_background
possibility*1
The definition of "possibility*1" is "n" within the following range.
1 / R3 ≤ n ≤ 1
Note that R3 refers to the value stored in register 3, not R*3.
Theoretical machine executing pseudo-assembly code: it has infinite computation speed and registers (R1~R8) capable of storing infinite bits.
Well, it doesn't matter if it is finite; a register that can store G(64) bits should be sufficient.
The specifications of each instruction are as follows.
Instruction: LOAD Rn, x
Summary: Loads the value x into the register Rn.
Instruction: LOAD Rn, Rm
Summary: Loads the value of register Rm into register Rn.
Instruction: CMP Rn, x
Summary: Compares the value of register Rn with the value x.
Instruction: JZ label
Summary: If the results of the previous CMP instructions are equal, jump to label label.
Instruction: JMP label
Summary: Jump to label label.
Instruction: DEC Rn
Summary: Rn <- Rn - 1
Instruction: POW Rn, Rm
Summary: Rn <- Rn ^ Rm
Code
#############################
START
LOAD R1, 2
LOAD R2, 65536
LOAD R3, R1
LABEL loop
CMP R2, 0
JZ end
DEC R2
POW R3, R1
JMP loop
LABEL end
END
#############################
Description
Adult women with "huge breasts" can be difficult to control.
This epoch can be corrected more strongly, but is not recommended.
FAQ
Comments (14)
If you find someone violating any of the rules mentioned in the warning, please do not come to me to report it, as it will add disproportionately to the cost of dealing with it. I am checking to see if anyone is violating the rules, so there is no need to report it.
Or rather, it would take too long to notice if it was reported here in the first place.
pray tell us, how do you manage to check what people do on their privacy?
Also, learn some real math and some LaTeX formatting.
@theBrebs
Oh, you want to know how privacy is ensured?
As of this stage, we do not specifically record the subject of the checks in any electronic or physical medium. Since there is no recorded data, all people know is that it is not recorded. If they ask me to disclose it, all I can say is that I have no record of it.
There is an exception that my brain cells are recording, but GDPR and other privacy protections cannot intervene in the recording of the brain, and I myself cannot manipulate brain data.
I have no intention of learning about math and formatting.
I am not going to learn about math and formatting, because there is no benefit to me in learning. Of course, the viewer will benefit. If there is an objection, I may be able to enforce it as an order through the courts!
“good" students(baby) always come to teacher(parents) and ask teacher(adult) to punish the "bad” students (w)
What are the differences between these three versions?
There is not much difference in practice, but in this LoRA they are classified as follows
Lion: Good quality
AdaFactor20: So-so quality
AdaFactor30: Special (too large adult females (huge ) may not be controlled even with trigger:1.3. In such cases, this more learned version can be used for even stronger control).
If you are still in doubt, we recommend Lion, which is easy to handle. Most mainstream models are stable. Maybe...
If you are not satisfied, you can use other versions.
lmao that 'warning'
I know, right? And calls his creation "Kindergarten" uniform...er...wtf? Does he even know what that word means? Weirdest gaslighting by a white knight I've ever seen. Hilarious.
al least a good amount of people: We do not care 👍
She looks so tired! POOR LADY![doge]
Hmmm... some people may have downloaded "ver: Lion" with the name "LKKw_LionDiv61024Px.safetensors." "LKKw..." I have no problem using the name "LKKw..." as it is! (The hashes are identical, so needless to say, the licenses are equivalent.)
This name is very confusing as it is my identification name. Especially LKKw was too aptly typed ...
If it bothers you, please rename it to "Kindergarten_Uniform_Cosplay_Lion10.safetensors". You are of course free to continue using it as it is without changing the name.
This matter has been resolved.
It appears that the update now refers to metadata when LoRA is selected. I have renamed the file for everyone's clarity. It's a specification due to the update, so if you want to complain, tell the WEBUI😡 I don't expect the specification to reference metadata in the first place, and metadata is not unique for the same directory. File names are always guaranteed to be unique in the same directory.
If you are uncomfortable with this specification, please change it from "When adding to prompt, refer to lora by" in the settings.
What Base Model do I use? Is PDXL okay?
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.