CivArchive
    Gacha Style | ガチャスタイルイラスト - EXV0
    Preview 1074145
    Preview 1074146
    Preview 1074148

    Gacha Style

    ガチャスタイルイラスト

    (EN/日本語/中文)

    |___________________________________________________________________________________________|

    Versions:

    EXV0 (Experimental Version 0) : Experimental Version

    V1 「>T|-|E: BeGinning<」: Works pretty well with a lot of checkpoints. Face is still a problem.

    EXV [FairY] : Trained on different database. Over 100 times smaller than EXV0. Please use nightskyyozora checkpoint for amazing results. Some checkpoints doesn't work AT ALL.

    ________________________________________________________________________________________

    About this LoRA:

    -->Actually, I wanted to make my first LoRA a gacha splash LoRA after seeing silvermoong's fantastic work. At that time I wasn't really familiar with the making of LoRAs etc so I failed miserably. This LoRA is basically a remake of the old ones.

    -->Only 5 images have been used for the training to counter the argument of so many, who are convinced that you need a couple hundred images to train a style LoRA.

    -->Please do not use this LoRA for commercial purposes because of copyright issues. I am working on a copyright free version.

    -->Does not work on 2.5d Checkpoints. (AOM, Anylora, ...)

    -->Please use a 1x1 formal when generating.

    ___________

    >Trained with >CouterfeitV3.0<

    ___________

    >How to use it:

    ___________

    Recommended checkpoints:

    NotSoXJBMix-1

    EndlessMiraiMix [Re:AL]

    Pastel-Mix [Stylized Anime Model]

    MeinaMixV10

    9527

    INFP

    jasminiquemix

    nightskyYOZORA

    betterMix_Anime (V2)

    clearVAE

    pastel-waifu-diffusionVAE

    ___________

    -->Prompts: (white background:1.5)

    ___________

    -->LoRA strength: 0.6

    -->Highres fix: Latent

    -->Sampler: DPM++ 2M Karras

    -->CFG scale: 7~11

    ________________________________________________________________________________________

    Prompt example/プロンプトの例/提示的例子:

    ___________

    Prompt:

    masterpiece,best quality, (white background:1.5), 1girl, far shot, full body, green circle, fairy, fantasy, magic, green leafs, dome,<lora:GachaGachaV1:0.6>, standing, clear face,

    ___________

    Negative:

    1girl,badhandv4, EasyNegative, verybadimagenegative_v1.3,illustration, 3d, sepia, painting, cartoons, sketch, (worst quality:2), (low quality:2), (normal quality:2), lowres, bad anatomy, normal quality, ((monochrome)), ((grayscale)),text,title, artifacts,

    ___________

    *Important keywords are marked >BOLD<.

    *重要なキーワードは >太字< で表記しています。

    *重要的关键词被标记为 >粗体字<

    ________________________________________________________________________________________

    ________________________________________________________________________________________

    このLoRAについて:

    ___________

    -->実は、silvermoongさんの素晴らしい作品を見て、初めてのLoRAをガチャのスタイルLoRAにしようと思ったんです。当時はLoRAの作り方などにあまり詳しくなかったので、大失敗してしまいました。今回のLoRAは、基本的に昔のものをリメイクしています。

    -->数百枚の画像でLoRAを作るという意見に対抗するため、5枚の画像で構成されています。

    -->著作権の問題から、このLoRAを商業目的で使用することはご遠慮ください。著作権フリーのバージョンを作成中です。

    -->2.5dチェックポイントでは動作しません。(AOM、Anylora、...)

    -->生成の際は1x1フォーマルでお願いします。

    ___________

    >CouterfeitV3.0<でトレーニングしました。

    ___________

    >おすすめのせってい:

    ___________

    おすすめのチェックポイント :

    NotSoXJBMix-1

    EndlessMiraiMix [Re:AL]

    Pastel-Mix [Stylized Anime Model]

    MeinaMixV10

    9527

    INFP

    jasminiquemix

    nightskyYOZORA

    betterMix_Anime (V2)

    clearVAE

    pastel-waifu-diffusionVAE

    ___________

    -->プロンプト:(white background:1.5)

    ___________

    -->LoRA strength: 0.6

    -->Highres fix: Latent

    -->サンプラー DPM++ 2M Karras

    -->CFGスケール:7〜11

    ________________________________________________________________________________________

    关于这个LoRA:

    ___________

    -->实际上,在看到silvermoong的精彩作品后,我想把我的第一个LoRA做成gacha splash LoRA。当时我对制作LoRA等还不是很熟悉,所以我失败了。这个LoRA基本上是旧版的翻版。

    -->由于版权问题,请不要将这个LoRA用于商业目的。我正在开发一个无版权的版本。

    -->在2.5d检查点上不工作。(AOM, Anylora, ...)

    -->生成时请使用1x1的形式。

    ___________

    >使用>CouterfeitV3.0<进行训练。

    ___________

    >建议:

    ___________

    推荐的检查点:

    NotSoXJBMix-1

    EndlessMiraiMix [Re:AL]

    Pastel-Mix [Stylized Anime Model]

    MeinaMixV10

    9527

    INFP

    jasminiquemix

    nightskyYOZORA

    betterMix_Anime (V2)

    clearVAE

    pastel-waifu-diffusionVAE

    ___________

    -->提示: (white background:1.5)

    ___________

    -->LoRA强度:0.6

    -->Highres修复: Latent

    -->采样器: DPM++ 2M Karras

    -->CFG规模:7~11
    _________________________________________________________________________________________

    Description

    Experimental Version

    FAQ

    Comments (1)

    zeppelinofthe
    Author
    Jun 8, 2023
    CivitAI

    Ah yes...forgot to mention that this Version is extremly bad at generating characters. Their faces look weird. Will fix that in later Versions.

    LORA
    SD 1.5

    Details

    Downloads
    842
    Platform
    CivitAI
    Platform Status
    Available
    Created
    6/8/2023
    Updated
    5/13/2026
    Deleted
    -
    Trigger Words:
    (white background:1.5)

    Files

    GachaSplashEXV0.safetensors

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)

    Available On (1 platform)

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