CivArchive
    Preview 81314090
    Preview 81317879
    Preview 81316527
    Preview 81313807
    Preview 81313672
    Preview 81313702
    Preview 81313722
    Preview 81313697
    Preview 81313817
    Preview 105831990
    Preview 105832525
    Preview 105832596
    Preview 105860454
    Preview 105860840
    Preview 105860844
    Preview 105860831
    Preview 105860836
    Preview 105860833
    Preview 105860830

    Great current CartoonPink style of artist Amy Matthews.

    For Anima

    Trigger

    @cartoonpink

    Place it near the start of the prompt, after the safety and count tags:

    nsfw, 1girl, @cartoonpink, solo, pink hair, twintails, looking at viewer, ...

    JabComix Middle 0.5

    large eyes - add to prompt if u need more cartoon style face

    Anima Turbo (recommended)

    For better results use

    • More detail tags

    • sampler: Euler A (euler_ancestral)

    • scheduler: Kullback–Leibler Optimal (kl_optimal) | Simple | Normal

    • weight: 0.75-1.0

    • resolution: 880×1200 or similar

    Anima Base

    For better results use

    • More detail tags

    • sampler: Euler A (euler_ancestral) / euler

    • scheduler: Kullback–Leibler Optimal (kl_optimal) | normal

    • steps: 30

    • CFG: 3-4

    • weight: 0.75-1.0

    • resolution: 880×1200 or similar

    Onlyfans 1.0

    Anima Turbo (recommended)

    For better results use

    • More detail tags

    • sampler: Euler A (euler_ancestral)

    • scheduler: Kullback–Leibler Optimal (kl_optimal)

    • weight: 0.75-1.0

    • resolution: 880×1200 or similar

    Onlyfans 0.2

    Anima Turbo (recommended)

    Sampler: euler

    Scheduler: normal

    Steps: 10 · CFG: 1.0

    LoRA strength: 1.0

    Resolution: 880×1200 or similar

    Anima Base (experimental)

    Sampler: uni_pc, er_sde,

    Scheduler: ddim_uniform (uni_pc), beta (er_sde),

    Steps: 28 - 35

    High CFG (5+) washes the style out noticeably. Start around CFG 3 and go lower

    if the texture disappears. These values are still being tested.

    Resolution: the training set is mostly portrait. Square 1024×1024 can produce

    duplicated figures — prefer 880×1200 or similar.

    For Pony

    Scores:

    score_9, score_8_up, score_7_up, score_6_up, score_4_up

    Negative prompt:

    3d

    If u has pastic or 3d result - use neg tag: 3d from 0.5 to 1.0

    Steps:

    25 - 30

    Cfg:

    5 - 6.5

    Sampler:

    Euler_a

    It u'll find another best settings - please share images with metadata

    Description

    First version

    FAQ

    LORA
    Pony

    Details

    Downloads
    171
    Platform
    CivitAI
    Platform Status
    Available
    Created
    6/9/2025
    Updated
    10/26/2026
    Deleted
    -

    Files

    MB_CartoonPink_v1.safetensors

    Mirrors

    Available On (1 platform)

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