CivArchive
    GrayColor - General Model - v1.2
    NSFW
    Preview 123335838
    Preview 123335847
    Preview 123335892
    Preview 123441086
    Preview 123336404
    Preview 123336462
    Preview 123336486
    Preview 123336639
    Preview 123441345
    Preview 123441649
    Preview 123441797

    This model is based on ntrmix4.0 and is under active development.
    If the outputs are found to be overly similar to the original model,
    the author may choose to take the model down.

    Feedback and comparisons are highly appreciated to help improve originality.

    v1.1 Adjusted to enable general-purpose image output. (Fine details can occasionally collapse)

    1.2 made it easier to correct areas where the hands were prone to falling apart (it may not be perfect yet.

    MySetUp:

    webui: https://github.com/Zuntan03/EasyReforge

    clipskip:2

    VAE: sdxl_vae.safetensor
    CFGScale:4

    SamplingStep:20

    SamplingMethod:

    • dpm2m++ schedule

    • type: beta

    HighlesFix:(if need)

    • Latent(nearest-extract)

    • HighresStep:16step DenoisingStrength:0.6

    https://civarchive.com/models/971952/stabilizer-ilnaick?modelVersionId=2055853

    Description

    FAQ

    Comments (4)

    DjojiMar 6, 2026· 1 reaction
    CivitAI

    With what thought you build this one?

    What was your goal and what image set you use for build it?

    GrayColor
    Author
    Mar 7, 2026· 1 reaction

    My goal was versatility.

    My goal is to make it so that when you specify a style, the illustration will be displayed in a form close to the original artist's.

    At the moment, the poses tend to get a little out of whack, so my plan to gradually improve this.

    There may be some styles that don't look very similar, but I'd like to modify them to make them more versatile.

    1.2 is less likely to break down, but 1.1 may be better at resembling the original style.

    scottstitsMar 13, 2026· 2 reactions
    CivitAI

    Do you have a workflow handy? I would KILL for the quality in these images.

    GrayColor
    Author
    Mar 14, 2026

    Thanks! I'm really glad you like the images. They were generated using the WebUI, so there isn't a specific workflow available. If you match the provided settings such as Sampling Steps, Sampling Method, CFG Scale, and VAE, you should be able to achieve very similar results. Hope that helps!

    Checkpoint
    Illustrious

    Details

    Downloads
    512
    Platform
    CivitAI
    Platform Status
    Available
    Created
    3/6/2026
    Updated
    5/25/2026
    Deleted
    -

    Files

    graycolorGeneral_v12.safetensors

    Mirrors