CivArchive
    q7's MILF style LoRA - q7-sexy clothes-0401-v2
    NSFW
    Preview 126058784
    Preview 126058781
    Preview 126058785
    Preview 126058780
    Preview 126058783
    Preview 126058782
    Preview 126058786
    Preview 126170114
    Preview 126170117
    Preview 126170113
    Preview 126170115

    Description

    Regarding the combined use of multiple LoRAs

    Since my LoRAs have some fixed datasets as anchor points during training, errors may occur when multiple LoRAs are used together with their weights all set to 1.0. Therefore, I suggest keeping the total weight of LoRAs for the Q7 style not exceeding 1.4. For example: <lora:q7-apple_ass-0402-v1:0.7>, <lora:q7-sexy-0401-v2:0.7>. Of course, the closer the total is to 1.0, the more stable it will be. You are free to adjust the weights based on your preference.

    About lora q7-sexy_clothes-0401-v2

    This LoRA was trained using several popular fashion styles from the Chinese internet, with detailed prompt annotations included in the images. You can check the prompts to see exactly what clothing items are featured (there are too many to list). Additionally, I've intensified the training specifically for the MILF aesthetic—you may notice frequent appearances of thicc thighs, voluptuous hips, and large breasts. Download with caution if that's not your preference.

    Although I trained this LoRA primarily with a focus on clothing, that doesn't mean I neglected attractive facial features, great body proportions, or other details.

    About lora q7-huge_ass-0402-v1

    This LoRA also uses source material collected from the internet. Any content involving portrait rights has been reviewed and handled to ensure that no legal risks or related issues arise. If you have questions about where the faces come from, please do not ask. (To be honest, the faces actually come from a 3D character I created in HS2, then repeatedly redrawn into a realistic style to form the dataset. I have also been struggling with how to further improve the dataset quality.)

    This LoRA is very easy to use. Simply add the trigger words huge ass, thick thigh


    Side note

    My skills have improved once again through this training session, but unfortunately, my inspiration has run dry once more. Quite depressing, really. Feel free to leave me comments about what you'd like to see!


    For best results

    These images were trained and generated using the ILL-V2 base model. Generally speaking, you shouldn't encounter issues with newer ILL models—the results should only improve.


    Regarding outfits

    You don't need to strictly adhere to specific style combinations; feel free to experiment with mixing and matching.


    Prompt Ideas for Clothing Combinations

    Each line below represents a complete outfit combination. I recommend reviewing or translating them so you can mix and match flexibly.
    If you simply want to use the facial features, remember to add the prompt tag "milf".

    thick thighs,white top,turtleneck,sleeveless,jean shorts,midriff,
    latex halter dress,shiny clothes,
    floral print dress,strap dress,cleavage,
    off-shoulder dress,bare shoulders dress,
    red lace top,thick thighs,leopard print skirt,black pantyhose,long sleeves, cleavage,
    sleeveless,turtleneck,deep blue top,thick thighs,black latex skirt,black pantyhose,
    sheer shoulders,black shirt,long sleeves,see-through cleavage,

    Description

    2026年4月1日训练完成,训练参数代码 q7-sexy-0401-v2

    FAQ

    LORA
    Illustrious

    Details

    Downloads
    120
    Platform
    CivitAI
    Platform Status
    Available
    Created
    4/1/2026
    Updated
    4/27/2026
    Deleted
    -

    Files

    q7-sexy-0401-v2.safetensors

    Mirrors

    CivitAI (1 mirrors)