CivArchive
    Comics Factory | Tome 1 - Opus IV
    NSFW
    Preview 98535575
    Preview 98535411
    Preview 98535120
    Preview 98535462
    Preview 98535408
    Preview 98535280
    Preview 98535349
    Preview 98535395
    Preview 98535440
    Preview 98535470
    Preview 98535492
    Preview 98535493
    Preview 98535738
    Preview 98535822
    Preview 98535943
    Preview 98536016
    Preview 98536171
    Preview 98536266

    NOTE :Base model : Z Image Turbo & Flux1.
    Showcase images were created with ComfyUI and are in PNG, include metadata & workflow.

    IMPORTANT NOTE: This model is a tribute to the artist and not an attempt to copy or reproduce his works.

    Description

    Model trained on civitai with 13 images.

    Recommended resources : Fluxmania

    Recommended inference : dpmpp_2m or heunpp2 / sgm_uniform / 20 steps / guidance : 3.5
    Weighting : around 0.5

    FAQ

    Comments (8)

    ReallyCleanTeethSep 6, 2025· 8 reactions
    CivitAI

    Ev Ganin makes AI stuff. So you're training AI on AI. That's how we get model collapse.

    Models should only be trained on non-generative AI output to prevent collapse. This is not helping.

    Adel_AI
    Author
    Sep 6, 2025· 5 reactions

    This theory is yours alone, I don't see how it's a problem. I've already created models from SDXL images and the results were good. Training flux on flux renderings can indeed create a sterile loop, but Ev Ganin uses MJ, which can bring a new graphic touch to the flux palette, without forgetting the contribution of the artist's creative universe.. So I don't agree with your theory.

    JigenSep 6, 2025· 1 reaction

    Bro living in 2022 understanding of generative AI

    CorruptedcupidJan 21, 2026· 2 reactions

    @Jigen train on slop, slop you get

    absolutely absurd; model output quality is determined ONLY by model input quality, the source is irrelevant.

    32BitshifterMar 19, 2026· 1 reaction

    This has always been wrong, from the beginning, full stop. Absolutely reams of ai models are trained on ai augmented or even generated inputs on purpose by leading labs.

    8914005Sep 13, 2025· 3 reactions
    CivitAI

    If you ever have a personal interest, could you create a LoRA in the style of another magnificent Russian artist, Sergey Kolesov aka Peleng? He also has a very interesting style, especially his older works. He masterfully works with form, indicating it with just a few strokes, and where needed, adds details — a kind of minimalism with an amazing result.
    I simply have no clue about the technical side of this. :(
    Especially his illustrations from the time he worked on Dishonored 2.
    https://www.artstation.com/artwork/3mOXE
    https://www.artstation.com/peleng

    Adel_AI
    Author
    Sep 13, 2025· 1 reaction

    Thanks for the links, I'll take a look

    LORA
    Flux.1 D

    Details

    Downloads
    948
    Platform
    CivitAI
    Platform Status
    Available
    Created
    9/5/2025
    Updated
    5/14/2026
    Deleted
    -
    Trigger Words:
    EvGan Style

    Files

    Ev_Ganin_IV.safetensors

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)

    Available On (1 platform)

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