CivArchive
    MeichiShuffle - Lust - v1.0
    NSFW
    Preview 40756830
    Preview 40754721
    Preview 40754743
    Preview 40754758
    Preview 40754757
    Preview 40754795
    Preview 40754808
    Preview 40754958
    Preview 40754984
    Preview 40754988
    Preview 40754638
    Preview 40756351
    Preview 40756377
    Preview 40755121
    Preview 40755212
    Preview 40755257
    Preview 40755265
    Preview 40755442
    Preview 40755641
    Preview 40755685

    Ok.

    So after a many try and attempt ( and a promess), i finally do it

    I wanted to do something, sharing like many other before me.

    A big thank to Citron Legacy https://civarchive.com/user/CitronLegacy, without him, his lora and his support, i wouldnt be able to go that far ( Also a crazy lora maker)

    A big thank to Snibbler https://civarchive.com/user/Snibbler for making my hype goes even higher due to his skill of prompting and great idea (NGL this guy is a beast)

    Marlosart https://civarchive.com/user/Marlosart for the first and my favorite lora of Chun-Li ( its was my first lora too), Deskgrenade https://civarchive.com/user/DeskGrenade (incredible lora) and MFCG https://civarchive.com/user/mfcg for his all mighty model and style (ickpot and MFCG PDXL)

    And Phinjo https://civarchive.com/user/phinjo for his being a maestro when it come to prompt ( and making the best comic model here)

    The main goal of this merge is to work with realistic, 3D feature and also be proeficient with 2D, usual stuff. I wanted to get a model who can "magnifiying" your prompt

    What i retain from the series of test:

    • Body generate well without to many errors ( depends of the prompt too for the outcome)

    • Background are really well made with a many details ( more than expected )

    • When using Lora with a specific style, you can be surprise with good result by playing with your prompt (play with the weight of the lora too)

      (exemple: rapscallion lora style was increasing the 3D shading, even if the style itself is more prone for 2D, sketch work)

    • Seem to be quite competent to respect the prompt (Mature theme included)

      • I use face detailer ( its not mandatory but work without it too) and Highres Latent 1.6 for image with that size - 976x1280 so i get 1560x2048, and for for 1280x1280, i was putting 1.4 for highres latent and get 1792x1792 after the highres. Im using a RTX 2080 ti. This setting could be change depend of your hardware.

      • I dont think i baked a Vae but i was using the one from MFCG PDXL when generating https://civarchive.com/api/download/models/353166?type=VAE&format=SafeTensor

    Now i leave it to you, see what you will accomplish and reveal some aspect i didnt notice.

    I will host a little contest, hoping to see how you will handle this model and maybe push MeichiShuffle further beyond.

    Thank you for listening until now

    Djoji

    PS: This is the main line when i using for my creation, its not a monolitic base, so feel free to experiment and change (use 3d and realistic in the prompt for addind these effect) :

    score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, colorful, masterpiece, 4k, highly detailed, detailed face, detailed hands, detailed background

    Description

    Until further change of two base of the merge, it will still be V1

    FAQ

    Checkpoint
    Pony

    Details

    Downloads
    270
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    11/18/2024
    Updated
    4/22/2026
    Deleted
    9/23/2025

    Files

    meichishuffleLust_v10.safetensors

    Mirrors

    Other Platforms (TensorArt, SeaArt, etc.) (1 mirrors)

    Available On (1 platform)

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