CivArchive
    Preview 2414289Preview 2415213Preview 2413486Preview 2413825Preview 2414231Preview 2414513Preview 2415237Preview 2413828Preview 2413826Preview 2413848Preview 2413847Preview 2413827Preview 2414654Preview 2413628Preview 2414813Preview 2414717Preview 2414719Preview 2414716Preview 2421534Preview 2414718

    v5.0BakedVAE_AutomaticFIX

    Blue dots/orbs gone

    V5.0

    added Better than words, PEAE-xl, Samaritan 3d Cartoon

    V4.0

    More focus on photorealism.

    Don't judge by examples. None of them are ordinary photo.

    V3.1

    • Better photorealism

    • Stable

    V3.0

    Quite a controversial version. It has many pros and cons. Gives the best results with authors and styles. It copes better with complex compositions. Improved drawing of small faces. Standard photorealism is not particularly good. Not as stable as previous versions. Needs improvements.

    Two new models have been added to the merging pool: Azure Wutong_XL and Endless coating v1.0

    V2.0

    • Much more artistic

    • Less distortion and noise

    • More beautiful faces and nudity

    • The size of the breast is less dependent on the constitution of the body

    This model specializes in turning your artistic and other fantasies into reality. Depending on the mood, these can be full body/half body/group portraits in a wide variety of poses and compositions and in a wide variety of styles from photorealism to anime or whatever you want. The model is completely uncensored, but adequate, and you won't see nudity unless you ask for it or put the personage in a scenario that is conducive to this.

    I am an Easy Diffusion user and for my previous merge I did separate examples on Automatic1111, but this time I'll not do this because they finally updated diffusers and now both of them give very similar results and the need to write separate prompts for each of them temporarily disappeared.

    • SDXL VAE !!!

    • no refiner

    • I prefer DDIM sampler, but you can use anything you like.

    • 7 GFG

    • 35 steps (75+ for complicated scenes)

    My negative:

    Monochromatic, blurry and fuzzy, chaos, compositional disorganisation, simple perspective, collage, distorted or deformed, (phantom limbs)+, bad anatomy, perfect skin+, cartoon rendering 3d doll, (additional limbs)+, (deformed or partially rendered face, deformed or partially rendered eyes, deformed or partially rendered hand gestures)+, broken or deformed geometry, cropped text watermark signature logo copyright artifacts noise, low resolution, partially rendered objects, deformed eyeballs

    List of models participated in this merge:

    • Sole Mio

    • Explicit Freedom - NSFW Waifu

    • SDXL InkPunkStyle

    • SDXL FaeTastic

    • WaifuGenieSDXL v1.0

    • Azure Wutong_XL (V3.0-V3.1)

    • Endless coating v1.0 (V3.0-V3.1)

    • Better than words (V5.0)

    • PEAE-xl (v5.0)

    • Samaritan 3d Cartoon (v.5.0)

    • personal non-shared checkpoints

    Description

    Checkpoint
    SDXL 1.0

    Details

    Downloads
    532
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    4/24/2025
    Updated
    4/24/2025
    Deleted
    4/24/2025

    Files

    lunaMia_v20.safetensors

    Mirrors

    CivitAI (1 mirrors)