CivArchive
    style-homx2_pudding-1(Anima) - v1.0
    NSFW
    Preview 146643121
    Preview 146643239
    Preview 146643275
    Preview 146643287
    Preview 146643424
    Preview 146643453
    Preview 146643487
    Preview 146643537
    Preview 146643576
    Preview 146643619
    Preview 146647409
    Preview 146647447
    Preview 146647475
    Preview 146647506
    Preview 146647543
    Preview 146647568
    Preview 146647587
    Preview 146647597
    Preview 146647630
    Preview 146647668

    This is a style LoRA for homx2_pudding.

    Trigger tag: @homx2 style v1

    Recommended tags: Style tag: @homx2 style v1

    Trained on high-quality anime illustrations, captures the core style features of homx2_pudding, including signature color tendencies, brushwork texture, lighting treatment, and composition habits. Works with various characters, poses, and backgrounds, allowing the style to be applied across different subjects.

    Recommended LoRA weight: 1.0.

    Works well with various outfits, poses, and backgrounds.

    It is recommended to use with anime-style base models, especially Anima-series base models.

    Avoid too high LoRA weight, which may cause oversaturation, stiff lines, and over-sharpened artifacts, reducing the natural feel of the style.

    Training notes:

    • Supports different angles, body poses, and various compositions.

    • Keep the signature color and brushwork features for best results.

    • For stable generation, use appropriate negative prompts.

    • If the base model is sensitive to the @ prefix, try removing @ and using only homx2 style v1 for comparison.

    Negative prompt suggestion:

    (worst quality, low quality, deformed, mutated, ugly, disfigured, blurry, bad anatomy, extra limbs, missing limbs, oversaturated, over-sharpened)

    这是一个 homx2_pudding 的画风 LoRA。

    触发标签: @homx2 style v1

    推荐标签: 画风标签:@homx2 style v1

    基于高质量动漫插画训练,捕捉 homx2_pudding 的核心画风特征,包括标志性的色彩倾向、笔触质感、光影处理与构图习惯。支持多种角色、姿势与背景,可将目标风格迁移到不同题材上。

    推荐 LoRA 权重:1.0。

    可以适配多种服装、姿势和背景。

    建议搭配动漫风格底模使用,尤其适合 Anima 系列底模。

    不要把 LoRA 权重拉得过高,过高可能导致画面过饱和、线条僵硬、细节过度锐化,反而削弱画风自然感。

    训练说明:

    • 支持不同角度、各类身体姿态与多种构图。

    • 想要最佳效果,请保留画风标志性的色彩与笔触特征。

    • 想要画面稳定,需要搭配合适的反向提示词。

    • 若底模对 @ 前缀敏感,可尝试去掉 @ 仅保留 homx2 style v1 做对比测试。

    反向提示词参考:

    (worst quality, low quality, deformed, mutated, ugly, disfigured, blurry, bad anatomy, extra limbs, missing limbs, oversaturated, over-sharpened)

    Description

    FAQ

    LORA
    Anima

    Details

    Downloads
    685
    Platform
    CivitAI
    Platform Status
    Available
    Created
    10/24/2026
    Updated
    11/7/2026
    Deleted
    -

    Files

    homx2_style_v1-000016.safetensors

    Mirrors

    homx2_style_v1.safetensors

    Mirrors

    CivitAI (1 mirrors)

    homx2_style_v1-000012.safetensors

    Mirrors

    homx2_style_v1-000014.safetensors

    Mirrors

    homx2_style_v1-000010.safetensors

    Mirrors

    homx2_style_v1-000018.safetensors

    Mirrors