CivArchive
    Preview 57389322
    Preview 57389375
    Preview 57389382
    Preview 57389409
    Preview 57389502
    Preview 57389523
    Preview 57389581
    Preview 57389587
    Preview 57389613
    Preview 57389749
    Preview 57389767
    Preview 57389791
    Preview 57389799
    Preview 57389827
    Preview 57389831
    Preview 57389854
    Preview 57389881
    Preview 57389887
    Preview 57389890
    Preview 57389917

    An attempt to replicate visual style of the Heroes of Might and Magic II game.

    Training dataset includes character portraits from the original game, "Price of Loyalty" expansion, several portraits from the Heroes 1, as well as fanart portraits, very close to the original style, made by me.

    Training data also includes all the creatures, primary and secondary skill icons, all artifact icons from the original game and expansion, as well a limited number of map objects and interface icons.

    Works best with image2image.

    Description

    Artifacts only style for FLUX.

    It does not require a trigger word.

    Model is optimized for 512x512, yet fully capable at 1024x1024.

    When generating in 1024x1024, bear in mind the LoRA is susceptible to general FLUX problems with blurry images on a certain prompts (namely: "background", background color, "artifact", even "HoMM2"). This is FLUX base model issue irregardless of an attached LoRA.

    FAQ

    LORA
    Flux.1 D

    Details

    Downloads
    69
    Platform
    CivitAI
    Platform Status
    Available
    Created
    2/12/2025
    Updated
    4/22/2026
    Deleted
    -

    Files

    HoMM2_artifacts_FLUX_v1.TA_trained.safetensors

    Available On (1 platform)

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