CivArchive
    New Year and Christmas greeting cards in style of Vladimir Zarubin / Новогодние открытки в стиле Владимира Зарубина - v1.0
    Preview 5028700Preview 5028590Preview 5028604Preview 5028586Preview 5028621Preview 5028778Preview 5028608Preview 5028654Preview 5028653Preview 5028664

    This LoRA creates images in the style of Soviet and russian artist and animator Vladimir Zarubin, best known as a New Year greeting card artist.

    Works good on realistic and semirealistic models. Best results i got on epiCPhotoGasm zUniversal. Add "human, woman, man, child" to negative prompt to exclude elements of human anatomy from animals. Sometimes LoRA draw animals with three ears, therefore add "three ears" to negative prompt.

    Prompt examples see in attached images.

    Recomended parameters:

    Strength: 0.8-0.9. Sampler: DPM++ 2M SDE Exponential. Steps: 30-40. CFG Scale: 6-8. Aspect ratio 2:3, 3:2, 4:3, 3:4. Short side size 512-576 pix. Negative embeddings sometimes reduce quality of generation.

    Hires.fix:

    Upscaler: 4x-UltraSharp. Steps - half of your sampling steps. Ratio 1.3-1.5. Denoise 0.3-0.4. Than proceed with SD Upscale. Or Latent (nearest) upscaler with ratio 1.25 and denoise 0.5.

    To draw different animals on image use inpaint function (see example picture).

    Mask mode - Inpaint masked. Masked content - Latent noise. Inpaint area - Only masked. Sampler - DPM++ 2M SDE Exponential. Steps - 30-40. CFG - same as in text2image. Denoise 0.95.

    In prompt area - name of animal. For example - hare, dancing, paint by zarubin, <lora:zarubin:0.8>. Negative prompt - same as in text2image.

    Description

    LORA
    SD 1.5

    Details

    Downloads
    560
    Platform
    CivitAI
    Platform Status
    Available
    Created
    12/30/2023
    Updated
    9/27/2025
    Deleted
    -
    Trigger Words:
    zarubin
    hare
    bear
    squirrel
    santa claus
    christmas tree
    forest
    winter
    gift box
    gifts bag
    snowman
    hedgehog
    stump

    Files

    zarubin.safetensors

    Mirrors

    CivitAI (1 mirrors)