Krea-2 V2:
V2 is focused on stability, and on keeping the womans body unaffected in size. Most, if not all other Breast size slider will make every woman chubby when you increase breast size, but this LoRA will keep the rest of her body lean and almost unaffected, in both directions.
It behaves different with and without clothes. And with different LoRA combinations etc. But in general i would stay within these values:
Range naked: -5 to 5
Range with clothes: -5 to 8
If you want hyper, or have problems with large breasts getting a forced bra, use V1 and V2 combined at half strength.
Krea-2 V1:
Recommended range: -4 to 5
ZIT
v1.1 is v1.0 merged 50/50 with a v2.0 candidate that i didnt think would live up to a v2.0 quality.
Pros of v1.1 is that is retains the original face, light and pose better. And therefor it can also be pushed to higher values.
I also changed the name of the LoRA to "Breast" instead of "Breasts", both here in civ and the filename.
Recommended range: -3.0 to 10.0
All my work is 100% free. The only thing that really matters to me is people leaving a review, good or bad. I will NEVER remove or block anyone for leaving a bad review, so don't be afraid to give me criticism.
I'm stating this because when a creator blocks someone for leaving a bad review, that person can no longer see any of the creator's models, essentially muting their critics. This is something I'm highly critical of, and I believe it's a fundamental quality problem on this site.
Description
FAQ
Comments (12)
Thank you very much! Krea lacks understanding to breasts sizes.
There is a V2 coming as well. In the making as we speak
krea-2 filter gets in the way in training in unexpected ways
But to get a clean and robust LoRA that work on all other checkpoints and with all other LoRAs, its nessesary to work around it.
Could probably use the turbo-adapter or something, but that is not a fix to the problem. The turbo-adapter just hides the problem for the LoRA creator and leave the users with horrible time when trying to combine multiple LoRAs.
It's likely the problem with almost mandatory usage of filter bypass LoRA with Krea 2, but for me beta is not isolated to just the breasts but the entire body build.
I sugest you use a dedicated checkpoint.
The NSFW LoRAs are much more prone to changes by other LoRAs.
What did you use for your positive prompt and negative prompt in your trainer? I've tried so hard to make this slider and I can't get it to impact the size of breasts at all. Other sliders, like emotion and stuff are just fine.
I help a lot of people with settings and generalt help training sliders. But my prompting is something i want to keep for myself.
But i can tell you how i go about finding out what prompt i should use. I make A vs B testes for all kind of different words and sentences. With different variations etc. And i use way more time finding good prompts then i use training.
For krea2 the prompting is a bit different as well since the text encoder is a LLM. So you should use sentences in all for prompts. Target should be "the size of her breastes" or something like that, instead of "breasts".
And if you cant get the effect you want from a normal image generation with that prompt, the slider will probably also fail.
@Loraholic Thank you. That is what I've been trying to do too. Just find the right words to make the change in a normal image gen and then work with that. I'll keep plodding along and thanks for all your work.
@namole5990210 I thing i used a modified version of ai-toolkit to train this when i think about it. So its probably not possible to make a similar LoRA as this without some tinkering in the codebase.
Details
Files
breast_size_krea2_loraholic.safetensors
Mirrors
BreastSliderKrea2.safetensors
breast_size_krea2_loraholic.safetensors
breast_size_krea2_loraholic.safetensors
breast_size_krea2_loraholic.safetensors
breast_size_krea2_loraholic.safetensors
131-breast_size_krea2_loraholic.safetensors
breast_size_krea2_loraholic.safetensors
breast_size_krea2_loraholic.safetensors
Breast Size Slider - Krea2 - -4 +5.safetensors




