CivArchive
    DunhuangMix Dunhuang large model - v1.1
    Preview 1
    Preview 2
    Preview 3
    Preview 4

    Description

    Connect to model customization and one-on-one teaching

    QQ Group: 895633778

    Telegram: https://t.me/+soU-Ws67aBlkY2I1

    For

    business cooperation, please email: [email protected]

    WeChat: 13353036335

    Dunhuang big model, using a large number of Dunhuang materials for training and multiple screening It came from fusion.

    The following are recommended parameters: va:840000

    sampling

    : DPM++ 2M SDE

    Karras steps

    Number: 20-30

    Hi-Res Fix (Hires. fix): HD Fix must

    be enabled!!

    HD iteration steps: 15-20 Magnification algorithm: 8x_nmkd-superscale_150000_g or R-ESRGaN

    4x+ redraw amplitude: 0.25-0.4

    Magnification factor: 2 or more cfg:

    7 (free to try)

    Special

    thanks: Moyu and Majic. The large model of the fusion includes MajicV6

    , Moyou's full version V1010


    FAQ

    Details

    Downloads
    381
    Platform
    ShakkerAI
    Platform Status
    Available
    Created
    7/14/2023
    Updated
    9/6/2024
    Deleted
    -
    Trigger Words:
    dunhuang
    dunhuang_face
    dunhuang_background

    Files

    DunhuangMixV1.1.ckpt

    Mirrors

    CivitAI (1 mirrors)
    ShakkerAI (1 mirrors)