CivArchive

    Built for artists who love alluring, confident women, this model gives you both realism and fantasy charm. Ideal for portraits, fashion, and erotic artwork

    -- Illustrious Versions --

    There are two versions of it :

    Semi-realistic : This version is trained on a predominantly realistic flux images dataset with a relatively high learning rate, this version prioritizes appearance consistency and produces a stable, coherent look.
    The trade-off is reduced adaptability, meaning it does not blend as well with highly stylized checkpoint or a wide range of lora models.

    Stylized : This one is trained on a broader and more diverse stylized dataset with a lower learning rate.
    As a result, it offers higher adaptability and flexibility, allowing it to blend well with a wide range of illustration styles, checkpoints, and LoRA models.

    Same goes for the SDXL versions

    Description

    support me by adding posts

    FAQ

    LORA
    SDXL 1.0

    Details

    Downloads
    109
    Platform
    CivitAI
    Platform Status
    Available
    Created
    12/18/2025
    Updated
    6/28/2026
    Deleted
    -
    Trigger Words:
    R0XY

    Files

    SDXL_Roxy.safetensors

    Mirrors

    CivitAI (1 mirrors)