CivArchive
    Clothing Tightness Slider (Wan2.2) - v1.0 HIGH
    NSFW

    This is a slider LoRA.  It makes clothing tighter or looser. I think this one turned out pretty good without too much color shifting.

    Positive values makes clothes looser.

    Negative values makes clothes tighter.

    High positive values seem to generate Asian people. Low negative values seem generate shorter people. I have no idea why but it can be avoided by specifying those characteristics in your prompt.

    All images in the previews have the same seed.   I added a comment to the positive prompt of each image with the strength used.

    The High noise samples represent the full scale of the slider. The Low noise samples are varied.

    Maximum Strength: ~1.75 The video quality degrade when going higher.

    Minimum Strength: ~ -1.25 The video quality degrade when going lower.

    The LoRA was trained with Ostris AI-Toolkit in concept slider mode: https://github.com/ostris/ai-toolkit

    Target class: person wearing clothes

    Positive training prompt: person wearing baggy, oversized, loose, roomy, Flowing, Voluminous, capacious, slack clothes

    Negative training prompt: person wearing tight, skintight, Form-fitting, Figure-hugging, clothes

    Anchor class: house

    Note:

    It should not have worked on the dog.

    Why did it work on the dog.

    I'm glad it worked on the dog.

    Description

    FAQ

    Comments (4)

    katsumendOct 4, 2025· 3 reactions
    CivitAI

    Woohoo!! Thanks for making this! This is really awesome, great visualization and thanks for examples.

    A pant, shirt, skirt shortness slider would be really awesome, but might be more challenging. Great for customizing clothing thanks so much!!

    qcgooner
    Author
    Oct 4, 2025

    I think it would be possible to make a separate LoRA for each type of garment. I have other projects I want to use my gpu time on for now but if someone wants to do it it should be as easy as replacing clothes with pants, shirts etc. in the prompts I used.

    azretion188Oct 5, 2025· 2 reactions
    CivitAI

    I didn't know i needed this, but now that you've made it i love it!

    imonaboat1Oct 7, 2025· 10 reactions
    CivitAI

    "It should not have worked on the dog.

    Why did it work on the dog.

    I'm glad it worked on the dog."

    Thank you for including this hilarious footnote.