CivArchive
    Cellulite Ass - Flux2.Klein - v1.0
    NSFW
    Preview 121610685
    Preview 121610683
    Preview 121610694
    Preview 121610687
    Preview 121610684
    Preview 121610691
    Preview 121610692
    Preview 121610697
    Preview 121610690
    Preview 121610689
    Preview 121610696
    Preview 121610695
    Preview 121610693
    Preview 121610688
    Preview 121610686

    The idea of this lora is being realistc and not exagerated. So it's soft on point.


    Obviously works better on more fat and/or older women.

    Description

    FAQ

    Comments (11)

    StillbornFeb 19, 2026· 4 reactions
    CivitAI

    O herói que nós nem sabíamos que precisávamos.👏

    Skullkid
    Author
    Feb 19, 2026

    E o XL Girls sai hj a noite!

    CabaloniusFeb 19, 2026· 3 reactions
    CivitAI

    Hi, nice work, I was waiting for this one!! Can we have a qwen image/edit version too other than klein? It would be very nice since both models are used at the same time by the majority of the "local generation users" tnx in advance

    Skullkid
    Author
    Feb 19, 2026· 1 reaction

    I'll release XL girls soon too :)

    Skullkid
    Author
    Feb 19, 2026

    I'm not training for qwen yet, sorry. Maybe in the future...

    CabaloniusFeb 19, 2026

    @Skullkid would you like to share your dataset and captioning? I have failed more than once trying to train this concept on qwen, it always become body horror xD if I succeed I will of course share the result and give the lora and all the credit to you, let me know, I will not insist further.

    Still tanks again for your work!! And for the old lora that I miss so much!

    Skullkid
    Author
    Feb 19, 2026

    @Cabalonius unfortunately I don't share datasets. But I can share some previews of how do I crop the images to get this result.

    CabaloniusFeb 19, 2026

    @Skullkid Would be great!

    Skullkid
    Author
    Feb 19, 2026· 1 reaction

    @Cabalonius I will get it once I return to home.

    Skullkid
    Author
    Feb 20, 2026· 1 reaction
    CabaloniusFeb 20, 2026· 2 reactions

    @Skullkid WOW this is absolutely helpful! Now I see that my dataset is very wrong and from very wrong angles, I was trying to train more of the body with wider shots. I cannot thank you enough, keep up the good work!