SD seems to love big breasts. But not all people prefer them too. Even most of the NSFW Lora's for SDXL only produce big ones. This is a Lora for really small breasts.
SD 1.5 Version: https://civarchive.com/models/285031?modelVersionId=320676
Read the "About this version" to get the best results.
I have nothing against 5* ratings, but please refrain from doing so if this LoRA has obviously not been used...
I recommend a strength between 08 and 1.25.
I strongly recommend the "flat chested" versions. Other versions are experimental.
Hopefully it is clear that this LoRA may not be used to create underage or otherwise illegal content
Feel free to share the images you created with this LoRA if you like. I'm excited to see what others here have done with it.
There are not many tutorials for training NSFW stuff. So I am absolutely no expert either.
For this LoRA everything in the image was tagged except for the breasts. For the head I made it easy and it was either not on the pictures or I just tagged it with "head". Not much different than for SFW stuff.
Description
This version is based on a new, slightly smaller but better data set.
I recommend to use adetailer because faces can look a bit worse than in the previous versions, but the rest is better.
also add ``cropped, close up`` to negative prompt can help
FAQ
Comments (16)
just put "fat-tits" and "xl-tits" in negative....
checkpoints react on "tits" like crazy, so anything you prompt with that will 100% work.
I really prefer the output of the SD 1.5 over the SDXL Version. Have not being able to recreate my pictures and went back to using weighted prompts for SDXL sadly. But I really love your SD 1.5 Version.
interesting, im not as happy with the sd 1.5 version as with the SDXL.
@Explatord Hehe :) Guess tastes are different.
@FPVx yes haha. But glad that you like the sd1.5 version :)
yeah i prefer the sdxl more. Since sdxl can create better 1024x1024 better, which is just enough for upscaling. sd1.5 images are hard to upscale.
@yofoton174609 IT is not about the Size, just that - to me - the flat chests in SD 1.5 look more beautiful than with SDXL. I always upscale anyway.
@Explatord These here are perfect, and I really struggle to get anything alike in SDXL. The outcome in the SD 1.5 version was also realiable... the model I used here is DeliberateV5 or 6
For XL, I find that weighted prompts are sufficient most of the time. However, certain character loras (and even entire checkpoints) just refuse to go below a c-cup, and this lora will reliably force them to obey. It's extremely useful in those situations.
Best LoRa ever hehe, love it
I think you mean "breast LoRA ever"
I'm just learning LoRA's. I've never seen a model that starts out with a giant jpeg embedded in the beginning. Is this a different type or toolchain? Is there an easy way to parse the training caption keywords, because I'm not seeing them?
I'm Using OneTrainer (https://github.com/Nerogar/OneTrainer). Unlike kohya ss it does not store the training captions
This is fantastic! It really had the desired effect of smaller breasts that look natural. Not all woman have big boobs! Only issue I noticed is that the cropping of the image gets tighter with the higher number.
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.