CivArchive
    Zanshou-kin(残照金) SDXL Hoko Manga Style - v1.5
    NSFW
    Preview 44902864
    Preview 44902757
    Preview 44902762
    Preview 44902766
    Preview 44902759
    Preview 44902756
    Preview 44902761
    Preview 44902771

    资源合集:闲话AI官网 talk2ai.cn

    简介:

    "残照金"这个名字是为了致敬日本漫画黄金时代的大师们,意指"黄金时代的余晖",旨在重现一些我个人喜爱的经典漫画家的独特笔触。

    本模型目前处于初始阶段(v0.9),主要聚焦于女性角色的半身特写绘制,也可以用于风格转换。

    特点:

    复刻经典:模仿黄金时代漫画家的绘画风格

    精细笔触:注重细节的线条艺术和高对比度阴影

    复古气息:呈现独特的复古漫画风格

    使用建议:

    大模型:Animagine XL V3.1

    推荐关键词:Hoko, 1girl, solo, retro manga style, detailed line art, high contrast shading, simple background

    步数:25-35步

    权重:0.65-0.8

    采样方法:DPM++ 2M Karras

    CFG:7

    单边小于1024的分辨率容易降低出图质量。

    注意事项:

    当前版本(v0.9)主要针对女性角色的半身特写进行了训练。

    未来展望:

    下一步计划持续优化和扩展模型功能,敬请期待后续更新。欢迎大家积极返图~

    Here's the English translation of the introduction:

    Introduction:
    The name "残照金" (Zanshou-kin) is a tribute to the masters of Japan's golden age of manga, meaning "the afterglow of the golden era." It aims to recreate the unique brushstrokes of some classic manga artists I personally admire.

    This model is currently in its initial stage (v0.9), primarily focusing on half-body close-ups of female characters. It can also be used for style transfer.

    Features:

    • Classic Recreation: Mimics the drawing styles of golden age manga artists

    • Refined Brushwork: Emphasizes detailed line art and high-contrast shading

    • Vintage Atmosphere: Presents a unique retro manga style

    Usage Recommendations:

    • Large Model: Animagine XL V3.1

    • Recommended Keywords: Hoko, 1girl, solo, retro manga style, detailed line art, high contrast shading, simple background

    • Steps: 25-35

    • Weight: 0.65-0.8

    • Sampling Method: DPM++ 2M Karras

    • CFG: 7

    • Resolutions with either dimension less than 1024 may reduce output quality.

    Note:
    The current version (v0.9) is primarily trained for half-body close-ups of female characters.

    Future Outlook:
    The next step is to continuously optimize and expand the model's functionality. Stay tuned for future updates. We welcome active feedback and image sharing from users.

    Description

    Version 1.5 Update:

    1. Overall Image Generation Improvements

    - Significantly enhanced image generation quality

    - Improved image stability with reduced artifacts and distortions

    - Enhanced prompt control capabilities

    2. Prompt-Controlled Style Options

    - Expanded from black and white only to dual modes: B&W and colored

    - Use "An original colored illustration of" to generate colored artwork

    - Use "A black and white manga of" to generate B&W manga style

    - Optimized B&W manga with enhanced screentone effects for more authentic manga look

    - Eliminated previous ghosting issues

    This update primarily enhances the model's practicality and reliability, making it more suitable for actual creative work. Users can now exercise more precise control over generation results through prompts, with notable improvements particularly in black and white manga style rendering.

    FAQ

    Comments (1)

    wlmsgApr 17, 2025
    CivitAI

    我说hoko是谁啊,原来是Hojo Tsukasa,怪不得看起来那么眼熟。

    LORA
    SDXL 1.0

    Details

    Downloads
    2,195
    Platform
    CivitAI
    Platform Status
    Available
    Created
    12/11/2024
    Updated
    5/13/2026
    Deleted
    -
    Trigger Words:
    An original colored illustration of
    A black and white manga of

    Files

    hoko_sdxl.safetensors

    Mirrors

    CivitAI (1 mirrors)
    TensorArt (1 mirrors)

    Available On (1 platform)

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