CivArchive
    Avatar TV Style Lora The last airbender+Legend - V1
    Preview 1
    Preview 2
    Preview 3
    Preview 4
    Preview 5
    Preview 6
    Preview 7
    Preview 8
    Preview 9
    Preview 10
    Preview 11
    Preview 12
    Preview 13
    Preview 14
    Preview 15
    Preview 16
    Preview 17
    Preview 18
    Preview 19

    This Lora was trained on 890 still from Avatar the Last airbender and The legend of Korra, (Mainly the last Airbender).

    For the best results i would highly recommend Hassaku XL.

    It is aimed at trying to reproduce the Incredible Artstyle of the shows and possibly create new original characters with that artstyle, or convert established characters in avatar style.

    Because it was trained on such a large dataset(because I couldn't decide) I had to be very specific in the captions, and name some of the most important characters of the show. 

    Such as: Azula, Aang, Katara, Toph, Suki, Sokka, Zuko, Korra, Mai, Ty lee, Lin beifong, Suyin Beifong, Kuvira, Joo Dee, Master Piandao, Young Roku, young Sozin, Young Hama, Asami,  Yue, Amon, Young Amon, Kya, Fire lord Iroh, Senna.

    And probably some others i forgot.

    The trigger word is: In the Avatar Animation style,

    I will upload many pictures during the coming days and weeks, to illustrate the lora capabilities, and none of the prompts will be hidden. 

    So If you need to see how to get a specific character that was trained on the Lora you will have a prompt to copy and base your prompt on. 

    Because it’s a style Lora one of the most interesting thing to do, is to convert existing characters in the avatar animation style, or create your OC in that style

    Description

    FAQ

    LORA
    Illustrious

    Details

    Downloads
    3,940
    Platform
    TensorArt
    Platform Status
    Available
    Created
    2/7/2025
    Updated
    8/14/2025
    Deleted
    -
    Trigger Words:
    In the Avatar Animation style

    Files

    Avatar-Anime-Style2.safetensors

    Mirrors

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