CivArchive
    ai-孙燕姿-syz-so-vits-svc-4.0 - v50
    Preview 801742

    Some AI cover songs:here

    Extract the zip file,Rename configs and place them in the so-vits-svc-4.0\configs folder

    Put G_ Rename xxx.pth and place it in the so-vits-svc-4.0\logs\44k folder and load the model to use.

    This is the initial version of the model, and the model trained based on 800 songs will be updated in the future.

    The modal source and training code come from the Internet, and the model is only used for exchange of scientific research interests. If there is any infringement, please contact to delete, thank you.

    图源和训练代码来源于网络,模型仅作为科研兴趣交流使用,如有侵权请联系删除,谢谢!

    画像ソースとトレーニング コードはインターネットから取得され、モデルは科学研究の利益の交換にのみ使用されます.侵害がある場合は、連絡して削除してください。

    Legal & Risk

    It is prohibited to use this model for commercial purposes and any scenarios of illegal acts and purposes.

    本人原则上不支持使用此模型生成任何成人内容。用户在使用本模型时,请务必遵守当地法律法规,勿侵犯他人的名誉权、隐私权或肖像权。

    Description

    FAQ

    Comments (18)

    nbbxs22809May 16, 2023
    CivitAI

    这个配置文件一直显示不兼容呢,模型用不了

    mingsheMay 16, 2023

    这个不是SD的,这个是AI歌声的文件

    553737344342May 17, 2023

    报错:Given groups=1, weight of size [xxx, 256, xxx], expected input[xxx, 768, xxx] to have 256 channels, but got 768 channels instead

    答:v1分支的模型用了vec768的配置文件,如果上面报错的256的768位置反过来了那就是vec768的模型用了v1的配置文件 作者:羽毛布団 https://www.bilibili.com/read/cv22206231/ 出处:bilibili

    leeel2003Jun 14, 2023

    @553737344342 但要怎麼處理呢!?

    scoollerMay 16, 2023
    CivitAI

    she look like this IG cosplayer https://www.instagram.com/kaorucatcosplay/ it her?

    Jojo9999May 19, 2023

    She's Stefanie Sun, famous Singaporean singer.


    scoollerMay 19, 2023

    @Jojo99999999999999999999999999 it look like her

    astro_art891May 19, 2023
    CivitAI

    你训练的步数太少了,杂音太多。基本不能用来推理歌曲

    VonChenPlusMay 26, 2023
    CivitAI

    声音转出来有种沙哑的感觉,感觉还是差点意思

    usa503504399Jun 9, 2023
    CivitAI

    waiting for the model trained based on 800 songs, plz

    leeel2003Jun 14, 2023
    CivitAI

    报错:Given groups=1, weight of size [xxx, 256, xxx], expected input[xxx, 768, xxx] to have 256 channels, but got 768 channels instead

    lafezzi1900428Aug 1, 2023

    我也报这个错,似乎是因为配置文件的问题。请问你解决了吗?

    JisooLoverOct 15, 2023

    @lafezzi1900428 "gin_channels": 768,

            "ssl_dim": 768,

            "n_speakers": 1,

            "speech_encoder": "vec768l12",

            "speaker_embbeding": false

    JisooLoverOct 15, 2023

    @lafezzi1900428 修改配置文件config.json,找到"gin_channels": 256, 把这之后的改为以上配置。其实出现这个错误已经表示这个模型并不适用你

    zoeyzoeyJun 15, 2023
    CivitAI

    天黑黑~

    zoeyzoeyJun 15, 2023
    CivitAI

    大佬好厉害啊,但是我这听不到歌曲怎么办,能救我一下吗?

    micodeJul 3, 2023· 1 reaction
    CivitAI

    多謝分享,可惜很多雜音,不能用。

    sdjlfkhlJul 13, 2023
    CivitAI

    大佬,可以再训练一个完美的出来吗?

    Other
    Other

    Details

    Downloads
    1,618
    Platform
    CivitAI
    Platform Status
    Available
    Created
    5/15/2023
    Updated
    5/13/2026
    Deleted
    -

    Available On (1 platform)

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