CivArchive
    GRIT-EYE_v4 - v4.0
    NSFW
    Preview 126963246
    Preview 126963251
    Preview 126963256
    Preview 126963252
    Preview 126963247
    Preview 126963259
    Preview 126963262
    Preview 126963249
    Preview 126963258
    Preview 126963260
    Preview 126963264
    Preview 126963248
    Preview 126963261
    Preview 126963253
    Preview 126963263
    Preview 126963250
    Preview 126963257
    Preview 126963254
    Preview 126963255

    Same "style" as v2 but a bit more sharper.

    ( I am very new to this... I still can't get the faces to look good when characters are small in relation to the entire image...

    Not too much free time to dial up more... Or learn more on how to finetune the model for better outputs)

    ... There are some persky tiny little creatures of faces that keep kreeping up on some images.

    ... Some times faces or text shows on clothes. I can't get rid of them as well.

    The suggested negative prompt helps a lot in both cases.

    A high-contrast, "indie-manga" style specialized in heavy, textured ink-work and hauntingly oversized vertical pupils. It blends the gritty, sketch-heavy lines of traditional underground comics with a dark, matte, gouache-painted atmosphere. Designed for characters that look both vulnerable and supernatural against deep, void-like backgrounds.

    SETTINGS:

    Steps: 8

    Cfg: 1.4

    Sampler: LCM

    Scheduler: Normal

    1st pass works well with 1 megapixel resolution.

    2nd pass up to 2.4 megapixels

    All sample images were upscaled 2.5 times using settings:

    Steps: 12

    Cfg: 1.1

    Sampler: LCM

    Scheduler: Normal

    Denoise: 0.4

    Suggested "helper-positive-prompt":

    highly detailed linework, textured shading, gritty inking, rough sketch lines, matte colors, textured cel shading, (indie comic style:1.3), traditional art style, gouache texture, dark atmosphere, depth of field, masterpiece, dramatic lighting, best quality, incredibly absurdres, masterpiece, very aesthetic

    Suggested negative prompt:

    (multiple subjects:1.5), (extra characters:1.5), (group:1.3), (crowd:1.3), (spirit:1.2), (ghost:1.2), (creature:1.2), (fairy:1.2), (chibi:1.2), (tiny girl:1.2), (miniature:1.2), (floating objects:1.2), (print on clothes:1.3), (graphic tee:1.3), (logo:1.3), (character on clothes:1.3), (patterned clothes:1.2), (text:1.2), (watermark:1.2), (signature:1.2), (split view:1.3), (comic:1.3), (2koma:1.3), (4koma:1.3), (grid:1.3), (multiple views:1.3), lowres, (bad), worst quality, low quality, jpeg artifacts, fewer, extra, missing, unfinished, displeasing

    Description

    Tried my best to make images sharper.

    Great results with upscaling using settings on description.

    FAQ

    Checkpoint
    Illustrious

    Details

    Downloads
    176
    Platform
    CivitAI
    Platform Status
    Available
    Created
    4/10/2026
    Updated
    5/1/2026
    Deleted
    -

    Files

    gritEYEV4_v40.safetensors

    Mirrors

    CivitAI (1 mirrors)
    Other Platforms (TensorArt, SeaArt, etc.) (2 mirrors)

    Available On (1 platform)

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