CivArchive
    美杜莎/meidusha2.5D-SD1.5-lora模型 - v1.0
    NSFW
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined
    Preview undefined

    Based on majicmixRealistic_v7 as the base model

    基于majicmixRealistic_v7为底模

    Trigger words can be added or not OK

    Weight setting: 0.8 is the best

    Model selection: 2.5D model is the best, the author tested it with: zhmixRealistic_v40

    The effect of using multiple LORA at the same time has not been tested much, but the preliminary test will have an impact, you can also test it and give me feedback

    The model of majicmixRealistic will make the picture blue, I don’t know if it is a VAE problem, you can try it and tell me the reason.

    HD repair needs to be turned on

    ADetailer: Just turn on face repair, full body repair will make hands and feet deformed

    You can try other controllers, welcome to return pictures.

    This version is the first generation version, the data tends to be upper body, the whole body is drawn, the accuracy will not be high

    The procedure version will adjust the hands and feet, the whole body details, so stay tuned.

    基于majicmixRealistic_v7为底模

    触发词加不加都行

    权重设置:0.8为最佳

    出图模型选择:效果以2.5D模型最佳,作者测试用的是:zhmixRealistic_v40

    多重LORA同时使用的效果测试的不多,但初步测试会有影响,你们也可以测试一下给我反馈

    majicmixRealistic的模型出图,会让图片泛蓝,不知是不是VAE的问题,你们可以尝试一下,顺便告诉我原因,

    需要开启高清修复

    ADetailer:开启脸部修复即可,全身修复会让手脚畸形

    其他控制器可以多尝试,欢迎返图。

    此版为一代版本,数据上倾向于上半身,全身出图,精度会不高

    手续版本会调整手脚,全身细节,敬请期待。

    Description

    LORA
    SD 1.5

    Details

    Downloads
    74
    Platform
    CivitAI
    Platform Status
    Available
    Created
    3/8/2025
    Updated
    4/30/2026
    Deleted
    -
    Trigger Words:
    meidusha

    Files

    meidusha_0308_4.0-000005.safetensors

    Mirrors