CivArchive
    道明寺ここあ Cocoa Domyoji | Virtual YouTuber, Vtuber | Animagine V3.1 - v2
    NSFW

    V4:

    I have completely changed the source of the dataset, thoroughly eliminating the 3D art style.

    This version makes it easier for the characters to change clothes, and it can generate images with strong contrasts of light and shadow.

    This should be the final version, unless I find something that can be further improved. Thanks and enjoy!

    Trigger word:

    Cocoa-oufit1 for the school uniform

    Cocoa-outfit2 for Riot music outfit

    Cocoa-outfit3 for her casual wear

    データセットのソースを完全に変更し、3Dのアートスタイルを徹底的に排除しました。このバージョンではキャラクターの着替えがより容易になり、また、強い光と影のコントラストを持つ画像を生成できます。

    これが最終版になるはずですが、さらに改善できる点が見つかればそれを適用します。ありがとうございます、お楽しみください!

    トリガーワード:

    「Cocoa-outfit1」は学生服、

    「Cocoa-outfit2」はRiot音楽の衣装、

    「Cocoa-outfit3」は彼女のカジュアルウェアです。


    V3:

    Add around 100 images,now Cocoa has 3 outfits.

    Trigger word: Cocoa-oufit1 for the school uniform

    Cocoa-outfit2 for Riot music outfit

    Cocoa-outfit3 for her casual wear

    For her casual wear,you can use "hat/black hat" to make her wear her hat.That will also work for her jacket I assume. Will test later.

    Thank you and enjoy!

    約100枚の画像を追加しました。これでここあは3種類の衣装を持っています。

    トリガーワード:

    「Cocoa-outfit1」は学生服、

    「Cocoa-outfit2」はRiot音楽の衣装、

    「Cocoa-outfit3」は彼女のカジュアルウェアです。

    彼女のカジュアルウェアには、「hat/black hat」という言葉を使うと、帽子をかぶることができます。これは彼女のジャケットにも使えると思いますが、後でテストします。

    ありがとうございます、お楽しみください!!

    ............................................................................................................................................................................................

    V2:

    Add more images,training for a lot more steps.

    Now it could generate detailed background, and the character will have more complex postures.

    This verson actually includes Cocoa's two outfits.

    For the Unlimited one please add promopts:"sweater,open jacket"

    For the Riot Music one please add promopts",blue coat,black skirt,blue necktie"

    Thanks!


    V1:

    This is the lora of Vtuber Cocoa Domyoji, trained on Animagine V3.1

    Trigger word: cocoa

    Recommended weights: 0.8-1.0

    Might work for other SDXL models as well.

    Thanks for using it!

    Description

    Add more images,training for a lot more steps.

    Now it could generate detailed background, and the character will have more complex postures.

    FAQ

    LORA
    SDXL 1.0

    Details

    Downloads
    115
    Platform
    CivitAI
    Platform Status
    Available
    Created
    5/24/2024
    Updated
    5/21/2026
    Deleted
    -
    Trigger Words:
    cocoa

    Files

    Cocoa_V2.safetensors

    Mirrors

    CivitAI (1 mirrors)

    Available On (1 platform)

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