CivArchive
    [Wan2.2] Breast/Chest slider - v1.0 - High (Chest)
    NSFW
    Preview 101255771

    Wan2.2 Breast/Chest Slider LoRA

    V2

    Adjust breasts size effortlessly:

    • Negative (-2 to 0): Small (use prompt to help)

    • Positive (0 to 2): from big to huge, (bigger than 2.5 works but destroy the image colors and contrast)

    This version adds less fat out of breasts.

    V1

    Adjust chest size effortlessly:

    • Negative (-1 to 0): Slim down for petite, natural looks.

    • Positive (0 to 1): Enhance curves for a fuller bust.

    Note: May slightly alter composition or add body fat at higher strengths; avoid extreme settings for best results.

    Description

    • Negative (-1 to 0): Slim down for petite, natural looks.

    • Positive (0 to 1): Enhance curves for a fuller bust.

    Note: May slightly alter composition or add body fat at higher strengths; avoid extreme settings for best results.

    FAQ

    Comments (28)

    LuxariaSep 19, 2025· 7 reactions
    CivitAI

    Nice, first slider for Wan2.2

    strong134Sep 19, 2025· 6 reactions
    CivitAI

    We would need a waistline slider to counter the effect of large waistline when setting big breasts.

    alcaitiff
    Author
    Sep 19, 2025· 3 reactions

    Yes, we can do it

    alcaitiff
    Author
    Sep 21, 2025· 3 reactions

    The second version is better. But the Skinny/Fat is cooking.

    firemanbrakeneckSep 22, 2025

    @alcaitiff I'm seeing some colour corruption in the samples, is that due to instability at high weights?

    alcaitiff
    Author
    Sep 22, 2025· 1 reaction

    @firemanbrakeneck Yes, testing the extremes to show the problems

    firemanbrakeneckSep 22, 2025· 1 reaction

    @alcaitiff Respect.

    strong134Sep 23, 2025· 1 reaction

    @alcaitiff 

    Thanks for this.

    Unfortunately, I wasn't able to make it work with the breasts slider and the skin/fat slider. My goal is to find an easy way with wan to get a slender woman with much bigger natural breasts (not huge or too saggy). So far, no lora does this.

    firemanbrakeneckSep 29, 2025

    @strong134 I think all the models I've seen that weren't trained on booru tend to struggle with that concept. Best that can be obtained is probably by training a slider that does both simultaneously, or balancing the synergy of a cocktail of extreme proportion loras combined at low weight. Which civ now prohibits.

    strong134Sep 29, 2025

    @firemanbrakeneck 

    I know there is one for Flux called tiny waist line. It's pretty powerful. It's also altering the face, but since it's powerful, we just need a low strength, and it work perfectly.

    MemorableSep 19, 2025· 5 reactions
    CivitAI

    Legend! Excited to try this out.

    alcaitiff
    Author
    Sep 19, 2025

    Thanks

    alvamos354Sep 19, 2025· 3 reactions
    CivitAI

    I was surprised by the small size of only 37 MB. But this LoRA exceeded expectations - it works and performs quite well, although it doesn't enhance the appearance of nipples. Thank you!

    firemanbrakeneckSep 19, 2025· 3 reactions

    Wan has great inbuilt knowledge of physics. Most models compress to ~1/3 of the original size with ~95%+ retention. A model trained with a low rank should perform roughly as well as the standard 150 / 300mb in many cases. But almost every creator favours minute detail over size considerations, even though in a year 2.2 will probably be forgotten.

    yedeb63824620Sep 19, 2025

    @firemanbrakeneck 2.2 dead in 1 year? so what's already replacing it? xD

    firemanbrakeneckSep 20, 2025· 1 reaction

    @yedeb63824620 Not dead, forgotten. A few people are still making sd1.5 content but do people care about it, is it groundbreaking? For the most part, no.
    Hunyuan (~01/25) was succeeded by 2.1, despite framepack sustaining it a bit longer. 2.1 (~03/25) was succeeded by 2.2. I'd say the most likely succession (maybe wishful thinking) would be by wan 3, integrating the high-low positioning and details in a larger architecture. Or it could be some new contender out of the blue, who knows. Maybe they'll make an anime oriented video model that's much more responsive to detailed prompts and is less restricted by the laws of physics.

    firemanbrakeneckSep 20, 2025· 2 reactions

    To be clear, there are definitely still things that are more difficult to accomplish on newer models, and that's part of the problem - can't be sure whether you're ever going to want to use the older ones again, and by then finding the right loras might be a problem.

    Light6969Sep 21, 2025

    @yedeb63824620 2.5 supposedly coming in a week or two

    yedeb63824620Sep 21, 2025· 3 reactions

    @Light2020462 This is all going too fast for me. I didn't even start installing Wan2.2, just collecting what I think is necessary for it to work, and another version will be out before I can figure things out. Potato-pc doesn't help either

    firemanbrakeneckSep 22, 2025· 1 reaction

    @yedeb63824620 Man, I hear you. If it's any consolation, it'll probably take about a month, maybe 2, before the new version's magna opera loras kick in. Assuming it's actually good and not like sd2. More importantly, don't let the fomo get to you. Experiment in your own time, upgrade because you're dissatisfied with your own creations, or don't. Pretty much anything recent, the archive can get it later - just ask.

    blackhumantg735Sep 20, 2025· 4 reactions
    CivitAI

    I2v pls

    alcaitiff
    Author
    Sep 21, 2025

    why?

    idevstSep 21, 2025

    @alcaitiff i second an I2V request would be lovely to modify the breast size of a good imeage or charcter we made

    alcaitiff
    Author
    Sep 21, 2025

    @idevst I'm sorry but I2V will use your image how you will benefit from a lora?

    firemanbrakeneckSep 22, 2025

    @alcaitiff At the least, if the body is out of frame in the initial image, then it's entirely up to the vid model to decide. And it may stabilise gens even when the body begins in frame.

    Though it might be worth seeing if the t2v version suffices for this purpose, such was the case for 2.1.

    alcaitiff
    Author
    Sep 22, 2025· 1 reaction

    @firemanbrakeneck Yeah, you have a point.

    Lora_AddictSep 25, 2025· 1 reaction

    It works fine with I2V. Use face pics and prompt a zoom out, remove clothes with a prompt or just "inpaint" the chest area with a black or white square. Just be a little creative.

    LORA
    Wan Video 2.2 T2V-A14B

    Details

    Downloads
    1,595
    Platform
    CivitAI
    Platform Status
    Available
    Created
    9/19/2025
    Updated
    5/14/2026
    Deleted
    -

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.