The stylized model are based on the 8chan and 4chan models made by two different anons
4chan model https://8chan.moe/abdl/res/16220.html#17641
8chan model https://8chan.moe/abdl/res/16220.html#21661
This is my attempt at making a model that can generate women wearing diapers for any diaper fetishist out there.
Since Stable Diffusion works a lot with broad associations and "diaper" being associated with a lot of toddler and baby apparel images, it is advised in most cases to put emphasis on the age of the character you want, this is less of a problem with the newer models but is still adviseable.
The models have not been trained on any sexual activity nor exposed genitals, so images of a sexual nature combined with a diaper CAN work, but also has a high probability of giving distorted genitals and warped results.
--- Update Log ---
24/3 - Updated with V2, trained with a curated and hand captioned set of 168 images, it can sort of... do pullups, it has been tagged with a lot of patterns but it doesn't really get them. The dreamboothed model out of the gate wasn't very good on anatomy and such on it's own, and it still faces some issues with crushed faces and whatnot, but I did some merges with other abdl models of mine and others and at least you can get a lot of doable generations with it while most of the uniqueness of the model is retained.
23/2 - Gotten back into training again now that most of the issues with D8's dreambooth seems to be gone and I can find the energy to do so. Did a test run with auto-generated prompts for the training images and got pretty good results with that, so did a 200 epoch run on 50 images I got done manually prompting for training today and merged that in with the Semi model and an overfit model of mine with a Reverse Smoothstep 4 blockmerge, keeping the alpha/clip of the original models, then merged those two models together 50/50 with the alpha/clip of the Semi model. Since this is just the spawn of a test run I call it 1.5 since it's intended to be much more versatile rather than just a side-grade/minor upgrade.
12/2 - Updated with two different levels of cartoon/anime styled models in a long attempted and trialed merge with the anime novelai 4chan model and the cartoony 8chan model. Both who are overfit in their own very different and respective ways that made it hard to get a good result I could be satisfied with.
30/1 - updated with broken clip values fixed, updated with new images to be representative of the new prompting understanding of the model, easier to work with now. Wish I knew about this issue before permanently deleting a bunch of my overfit training attempts.
Description
An earlier mixing point before 1.0.A, better in some ways, worse in others.
FAQ
Comments (23)
This was long awaited by the ABDL community. Thank you. I had made my own but yours looks better.
ok... but why? 😅
Just make 3 categories.
SFW NSFW Fetish
While I'm not into the ABDL thing, remember the quote: "First they came for the socialists, and I did not speak out-because I was not a socialist"
Ending with...
"Then they came for me-and there was no one left to speak for me."
Don't be that person. If you don't like something that CANNOT harm you, grows some stones and IGNORE IT.
Pretty excellent start! Doing a 50/50 merge with RealisticVision gives some great results.
This is well done. Might not be everybody's cup of tea, but well done.
I can't seem to find a good tutorial to train a certain piece of clothing (like this) into a model. Can anyone direct me to one?
Not really a fan of normal diapers, needs more pull ups/goodnites content
I absolutely love this model. great work, and it gets even better. Question tho: would it be possible to get the trained part of it as a lora?
(probably a noob question, i have no idea if thats even possible)
Hey, nice Model but care to remove 8chan link?
They aren't meant for public
I'd like to further work in this area. Any pointers on generating a model checkpoint? Or, is there any manual work that could help further this model, such as specific description writing or classification of images?
.ckpt plz
Does it help to have more training data with more of the subject's body in focus? I noticed that it seems easy to generate close-up views of the diapers with other limbs and head cropped out of image. In case more data helps, I see lots of possible contributing images on DeviantArt. Maybe tags existing there can serve as pre-existing metadata?
Nice! Now we need a pacifier lora to go with this.
I love you!
It looks like those who get turned on by diapers will thank you for this... damn that's bizarre haha
How many trigger words is too many trigger words? 🤣
i absolutely love the progress you make with this model. what base model did you use to train it on?
How can I make the diapers look wet or messy consistently?
a novice I am, and downloaded the Tensor I did. How to I begin to use the file for anything?
could you please make a LORA ?
I would just like to say thank you for making this.
v2 is awesome. Thank you!
For those who like anime models, note this mixes well with AOM3 (50/50).
Seems to also mix pretty well with photo models - for example, Deliberate v2 (50/50), or Life Like Diffusion (70% this, 30% LLD).
Is there a version of something like this but without the anime style? Just want IRL / Photorealistic results trained on real images. Seems everyone wants this stuff in anime or furry style, except me!
do you have any tips if you want to change things like the "print" on the diaper