CivArchive
    Preview 66210785
    Preview 66211185
    Preview 66211186

    喵喵后宫才不是臭脚模型呢

    属于完整的 MiaoMiaoHarem Dogma 完整生态系统终于来了。👈详情请点击

    👑 MiaoMiao Harem Anima V1.0 | 二次元美学新纪元!基于 Anima 架构的华丽降临

    “不是把旧版本搬到新底模上,而是让喵喵后宫的审美灵魂,在 Anima 架构上完成一次真正的二次元觉醒!” “This is not a simple port to a new base—it's the true anime awakening of MiaoMiao Harem's aesthetic soul on top of the Anima architecture!”

    🚀 Update Highlights / 更新亮点

    MiaoMiao Harem Anima V1.0 正式登场!

    这一次,我们选择以 Anima 2B 架构为核心,重新塑造喵喵后宫的二次元表现力。 不同于偏向泛用或写实混合取向的底模,Anima 天生就是为 anime / illustration / artistic image 而生——它拥有更纯正的二次元语义理解、更稳定的角色风格表达,以及对标签体系与自然语言混合提示更强的兼容能力。

    基于这一点,MiaoMiao Harem Anima V1.0 不再只是“能画二次元”,而是向着更纯净的动漫审美、更鲜明的角色塑造、更统一的插画质感全面进化。 从角色立绘、单人美少女、双人互动,到轻剧情感场景、幻想服装设计、插画封面感构图——都能以更自然、更轻松、更高完成度的方式直出。

    这不是一次普通换底,而是一次真正面向二次元创作者的专属升级。

    提示词理解更友好 / Better Prompt Understanding:

    标签、自然语言、混合写法都能打

    Anima 原版就具备一个非常强的优势: 同时支持 Danbooru 风格标签、自然语言描述,以及二者混合输入。

    因此在这个版本中,你可以更自由地使用不同提示习惯:

    纯标签流:适合老手、控图党、精确指定元素

    自然语言流:适合描述氛围、构图、服装、剧情感

    标签 + 英文描述混合流:兼顾稳定性与创造力

    这意味着它既适合喜欢堆 tag 的用户,也适合只想用几句自然语言快速出图的用户。 对于角色名、作品名、服装设定、场景描述、画师风格倾向等内容,也能更自然地接收与还原。

    一句话总结: 更懂你想要什么,也更容易把“你脑子里的二次元画面”真的画出来。

    ⚡. 出图效率与可控性兼得 / Fast, Flexible, Effective

    虽然这是一个偏基础能力很强的动漫架构模型,但它并不意味着难用。 恰恰相反,MiaoMiao Harem Anima V1.0 保留了喵喵后宫系列一贯的低门槛、高回报特性:

    •简单 prompt 也能获得不错完成度

    •加质量标签后上限进一步提升

    •通过角色/服饰/姿势补充可快速细化画面

    •适配多种采样器,出图风格变化明显

    •可用于轻量控制,也可用于深度调校

    无论你是想“一抽出图”,还是想认真打磨构图与细节, 这个版本都能在效率审美稳定性之间取得非常优秀的平衡。

    非写实路线更加明确 / Illustration-Oriented by Design:不追求现实,只服务于幻想与艺术

    需要特别说明的是: MiaoMiao Harem Anima V1.0 是一个明确偏向动漫插画与非写实艺术表达的版本。

    它的强项不是现实摄影感,也不是高仿真写实人像, 而是:

    •动漫立绘•二次元角色插画•游戏卡面感视觉•轻小说封面感构图•梦幻幻想风•艺术化非写实场景

    如果你追求的是“像动画、像插画、像角色原画、像设定集封面”的视觉方向, 那么这个版本会比偏写实底模更加纯粹、更加直接、也更对味。

    MiaoMiao Harem Anima V1.0 不只是“二次元角色模型”, 更是一台可以在动漫、插画、艺术表现之间自由摆动的创作引擎。

    MiaoMiao Harem Anima V1.0,不是简单分支,而是喵喵后宫在二次元赛道上的全新主场!

    它继承了喵喵后宫系列的高审美表达,同时借助 Anima 架构获得了更纯正的动漫理解、更舒服的插画完成度,以及更友好的 prompt 兼容能力。 对于热爱角色创作、动漫插画、卡面海报、幻想美少女题材的创作者来说,这将会是一次非常值得体验的新进化。

    如果你想要的是更纯、更稳、更有“二次元味道”的喵喵后宫版本——那么这就是答案。

    立即体验,开启属于你的动漫幻想盛宴!欢迎分享返图,一起见证 Anima 架构下的全新蜕变!

    ⚠️ IMPORTANT: Usage Guide / 重要食用指南

    Model Output Parameters

    Scheduler Euler Normal Euler a Normal

    Steps 30

    CFG Scale 4.0-6.0 推荐5.5

    Model Recommended Resolution 512² ~ 1536²推荐常用尺寸768x1024 / 832x1216 / 896x1152 / 1024x1024 推荐 920x1536

    Positive Prompt (default )

    masterpiece, best quality, score_7, safe,very aesthetic, masterpiece, best quality, ultra-detailed,pale_skin,fair_skin, high contrast,huge filesize,{此处添加提示词 Add Prompt here}

    Negative Prompt

    worst quality,low quality,score_1,score_2,score_3,artist name,
    

    ⚡ 二次采样优化 / Secondary Sampling Enhancement:

    高清放大不糊不炸,细节质感全面觉醒!

    针对 Anima 模型 在高分辨率二次放大过程中容易出现的 模糊、发花、线条脏乱、细节崩坏 等问题,我在 WebUI Forge Neo 中反复实测,终于调出了一套更加均衡稳定的二次采样参数方案。 这套设置可以在尽量保留原图角色结构、神态、构图与插画气质的同时,显著提升整体清晰度与细节质感——放大后不容易糊,不容易花,线条更稳,画面更干净,成图完成度更高。无论是:•角色立绘•半身插画•卡面封面•壁纸级高分辨率输出•细节要求更高的成稿场景都能获得更舒服的放大效果。

    直接按照下图参数操作即可,基本就是当前 Anima 版本较为平衡、稳定、实用的二次采样方案。

    Sampler 推荐说明

    •er_sde默认推荐。风格中性、色块干净、线条利落,整体最稳。

    •euler_a线条更柔、更细,部分情况下会有轻微 2.5D 气质。可适当提高 CFG,适合柔和系角色图。

    •dpmpp_2m_sde_gpu创造性更强,变化更多,适合想要更丰富画面表现时使用。但某些 prompt 下可能会更“放飞”,建议搭配更明确的描述。

    Positive Prompt(推荐前缀)

    masterpiece, best quality, score_7, safe,

    如果你想保留原有识别,也可以继续加入:

    very aesthetic, ultra-detailed, high contrast

    Negative Prompt(推荐)

    worst quality, low quality, score_1, score_2, score_3, artist name

    如果你想沿用你原来更完整的负面,也可以融合成增强版,例如:

    worst quality, low quality, score_1, score_2, score_3, artist name, text, watermark, signature, jpeg artifacts, extra toes, bad feet, missing fingers, extra fingers, unfinished

    Prompt 使用建议

    •标签建议尽量小写

    •tag 中用空格,不要用下划线(除 score_7 这种特殊标签外)

    •可混用 标签 + 自然语言

    •想要更稳定角色表现时,建议写明:

    示例 Prompt

    Tag 风格

    masterpiece, best quality, score_7, safe, 1girl, solo, long hair, silver hair, blue eyes, school uniform, smile, looking at viewer, cherry blossoms, spring, soft lighting

    自然语言风格

    masterpiece, best quality, score_7, safe. An anime girl with long silver hair and bright blue eyes stands under cherry blossoms in spring. She wears a neat school uniform, smiles gently at the viewer, and the scene has soft lighting with a clean illustration style.

    混合风格

    masterpiece, best quality, score_7, safe, 1girl, solo, long silver hair, blue eyes, school uniform, cherry blossoms. A clean anime illustration with soft spring lighting, gentle expression, and a polished character-focused composition.
    
    

    有关如何使用 V-pred 模型的说明,请参阅 ←点击

    您可以通过加入我们的 Telegram 和 QQ 群组获取其他版本抢先体验资格,与我们一起见证这次重大的更新。感谢大家一直以来的支持和反馈!我将继续努力,为大家带来更多精彩的更新和功能。

    ☜ 请您赞助我 感谢支持 !Please sponsor me. Thank you for your support! 您的赞助将用于 GPU 费用与模型优化 All contributions go toward GPU costs and model improvements


    📥 Download Options / 模型下载说明

    ⚡ Sponsor with BUZZ / 赞助支持 (BUZZ下载)
    If you appreciate my work and wish to support the continuous development of the MiaoMiao series, you are welcome to download using BUZZ. Your support is the greatest motivation for my creative journey!
    如果您认可我的作品并愿意支持 MiaoMiao 系列的持续开发,欢迎使用 BUZZ 进行下载。您的支持是我持续迭代、不断创作的最大动力!

    🔗 Free Download (Non-BUZZ) / 免费获取 (免BUZZ渠道)
    I fully understand that BUZZ may not be available for everyone. To ensure creativity remains accessible, I have provided a non-BUZZ download link below:
    我深知 BUZZ 并不总是充裕,为了让创作不设限,我同样提供了免 BUZZ 下载渠道,您可以直接通过以下链接获取模型:👉 [downloadLink]

    ⬅️企鹅四周目⬅️Telegram


    ⚠️ IMPORTANT: Usage Guide / 重要食用指南

    Model Output Parameters

    Scheduler euler euler a

    Steps 30

    CFG Scale 3.8~6

    Model Recommended Resolution 768X1280 896x1152

    PVC Style(v1.2~v1.5)

    miaoka\(pvc style\), miaoka\(figure\),3d,

    Positive Prompt (default )

    {此处添加提示词 Add Prompt here} ,very aesthetic, masterpiece, best quality, ultra-detailed,miaoka, high contrast,huge filesize,

    Negative Prompt

    extra toes,lowres, (bad), text, error, fewer, extra, missing, worst quality, jpeg artifacts, low quality, watermark, unfinished, displeasing, oldest, early, signature, artistic error, username, bad feet,scan, [abstract], english text, shiny hair

    📥 获取方式 How to Access

    如果你不希望通过 BUZZ 获取下载权限,也完全没有问题。 欢迎加入我们的 Telegram 群组QQ 群组 与我联系,我将直接提供最新模型的下载链接,确保每位用户都能方便获取与使用。

    If you prefer not to use BUZZ for download access, no worries! Feel free to join our Telegram group or QQ group and contact me directly—I’ll provide the latest model download link to ensure easy access for everyone.

    这个模型如你们所见,由于我的个人兴趣,我更热衷于画脚。但是,如果你认为这是一个脚部特化模型,那就错了。这个模型的特点不在于画脚,而在于每个版本都在优化质感、语义理解和皮肤质感。我想创造一个能画出类似写实肌肤的动漫风格为主体,并尽量不影响其他风格的模型。这个模型可能很小众,没有多少人使用,但我会继续更新迭代以达到我的目标。如果你们喜欢,请帮我赞助我 BUZZ,帮忙点点赞、多评论、多返图就好了。

    As you can see, this model reflects my personal interest in drawing feet. However, if you think this is a foot-specialized model, you're mistaken. The model's strength lies not in drawing feet, but in optimizing texture, semantic understanding, and skin texture with each version. My goal is to create a model that can produce anime-style art with realistic skin textures while minimizing the impact on other styles. This model might be niche and not widely used, but I will continue to update and iterate to achieve my goal. If you like it, please support me on BUZZ by liking, commenting, and sharing more images.

    Model Usage Statement

    • 1. You are not allowed to host this model and its derivative versions (such as fusion model versions) on websites/applications that plan to earn income or donate. If you need to use this model and its derivative versions for commercial purposes (such as generative services, selling images, using images for public publication on short video platforms, etc.), please contact me through the website

    • 2. You are not allowed to directly sell images generated by this model, unless you have made sufficient manual modifications to the image so that it can be fully recognized as your personal work in a legal sense. If you violate this provision, all legal consequences resulting from it shall be borne by you personally, and I am not responsible.

    • 3. You cannot use this model to intentionally create or share illegal or harmful content for dissemination and output. Please abide by public order and good ethics, and

    Description

    MiaoMiao Harem 1.6G 是基于 MiaoMiao Harem Base 1.0 训练而成的版本。在延续之前版本质感的基础上,进一步优化了 Base 1.0 在肢体上的稳定性,并且完全兼容 Base 1.0 的 LoRA。稍后,我将展示 1.0 Base 版本中一些肢体容易崩坏的例子以及经过本次版本优化后得到的稳定性测试图。

    需要特别说明的是,该版本并不一定是最适合所有用户的最佳选择,请根据您的需求和喜好选择合适的版本使用。

    We are excited to introduce the MiaoMiao Harem 1.6G version update! To help you better understand the features of this update, please refer to the following details:

    MiaoMiao Harem 1.6G is trained based on MiaoMiao Harem Base 1.0. While maintaining the texture quality of the previous version, it further improves the stability of body structures from Base 1.0 and is fully compatible with Base 1.0 LoRA. Later, I will provide stability test images showing examples of body structures that were prone to issues in the 1.0 Base version, now improved in this updated version.

    Please note that this version may not necessarily be the best choice for everyone. Users are encouraged to select and use versions based on their individual needs and preferences.

    FAQ

    Comments (20)

    ThaiAI0Mar 29, 2025
    CivitAI

    Very good model like MiaoMiaoPixel

    ChameleonNeekoMar 30, 2025
    CivitAI

    Hello, I am new to AI generation;
    When trying to replicate the example pictures here, I get the same picture as you when I use CLIP Skip 1, but your generations use 6 instead, why could that be? Also, even when I do that, there are minor differences in the pictures, but I suppose that's due to the way AI works and FP precision/Different hardware/settings

    MIAOKA
    Author
    Mar 30, 2025· 1 reaction

    I can confirm that I used Clip Skip 2 when generating the image, but after uploading it to the website, the parameter changed to 6—which is likely a bug on the website. However, as far as I know, Clip Skip does not take effect in the SDXL and Illustrious series. No matter whether you set it to 1 or 6, the generated image remains unchanged.

    In SD1.5 and earlier versions, Clip Skip affects how the CLIP text encoder processes prompts by skipping certain transformer layers, which can influence image style and details. However, in SDXL and the Illustrious series, due to the use of OpenCLIP and an improved prompt processing mechanism, Clip Skip no longer takes effect. Regardless of the Clip Skip setting, the final generated image remains largely unchanged.

    If you want to verify this, you can compare Clip Skip 1 vs. Clip Skip 6 in SDXL and observe that there is almost no difference in the output.

    ChameleonNeekoMar 30, 2025

    @MIAOKA Hello, I tested what you mentioned, and for me:
    Clip skip 1 and 2: Roughly produces the same output (The output is exactly what I want)
    Clip skip 3 and 4: Change in camera angle and pose (The output is really similar to what I want)
    Clip skip 5 and higher: Image becomes a blob of flesh / very abstract failed generation (Output is a mess)
    This is different behavior than what you mentioned.
    I am using miaomiaoHarem_v16G.safetensors and sdxl_vae.safetensors.
    Also, is this model based on SDXL, Illustrious or something else entirely? The description says its based on MiaoMiao Harem Base 1.0, but I don't know what that's based on. Thank you for your answers.

    MIAOKA
    Author
    Mar 30, 2025· 1 reaction

    @ChamleonNeeko Then you can use clipskip2. By default, the 2-podel is based on noob AI

    T_eaMar 31, 2025
    CivitAI

    为何comfyui下的眼睛崩坏率比较高呢?wariza,seiza的脚有改善。不过有时候脚的大小不一,并排的双脚最明显。

    ShiinaApr 6, 2025

    猜测你给的cfg比较大, comfyui跟webui比要乘上0.7

    FarewellYhlApr 1, 2025
    CivitAI

    When I run hiresfix it has a high chance to create 2 navels, even when I set the denoise from 0.1-0.8.

    MIAOKA
    Author
    Apr 3, 2025

    The redrawing range should not exceed 0.3

    b1411651942370Apr 1, 2025
    CivitAI

    貌似1.6G版本经常会在背景生成小小的白色噪点...一开始以为是自己的问题,但是看了下图片区,好像也有这个问题

    MIAOKA
    Author
    Apr 3, 2025

    If you use my model to recommend parameters, this problem should not occur

    b1411651942370Apr 3, 2025

    @MIAOKA 谢谢,应该是因为我用了1024x1536的直出尺寸造成的,之前用illXL习惯了大尺寸

    moon____Apr 6, 2025· 1 reaction
    CivitAI

    丝袜和人物的质感都非常好,就此两点而言我目前找不到能出其右的模型。但头发是不是太硬了呢?一块一块的。还是说这也是有意为之?

    MIAOKA
    Author
    Apr 10, 2025

    如果你用过前几个版本的话你会发现前几个版本就是这样的风格头发质感就是这样的

    sol3raApr 13, 2025· 4 reactions
    CivitAI

    用什么调度类型(Schedule type)比较好

    AsdfghkknApr 18, 2025· 9 reactions
    CivitAI

    喵咔大佬(⁠つ⁠≧⁠▽⁠≦⁠)⁠つ我是AI初心者,我想请教一些问题...

    https://civitai.com/images/66369370 这张外国仁兄用MIAOMIAO后宫生成的布洛妮娅的光泽和色彩非常好看,于是我使用MIAOMIAO后宫然后使用和这张图片同样的提示词进行生成,但是生成结果不太好,然后我用网上的Metadata提取网站提取外国仁兄生成的图片的Metadata,发现似乎有使用“高像素修复”和由于不懂英语所以搞不懂很多参数,所以想询问一下喵咔大佬我要如何设置参数才能在civitai中生成出相同的图片?提前谢谢大佬制作和分享优秀模型!

    现在civitai无法使用MIAOMIAO后宫了...所以我是在Tensor生成的,如果喵咔大佬可以告诉我 布洛妮娅的图片里使用的Lora和是哪个“高像素修复”器的话我大概加上Lora和“高像素修复”就能生成出一样质量的图片。(我试图翻译Metadata,可是仍然看不明白使用的Lora和很多参数是哪里的参数,比如从一大堆英文里面翻译出来几个单词但是很多都没有翻译结果或是翻译成意义不明的中文,翻译得出来几个知道意思的中文比如去噪:0.25却不知道是哪里设置成0.25)

    MIAOKA
    Author
    Apr 18, 2025· 1 reaction

    他应该加了这个nyalia lora 你可以试试 我看到了这个lora的影子

    diupuriApr 21, 2025· 1 reaction

    可以在本地运行comfyui,然后把图片拖进去

    KazuhiraApr 19, 2025· 7 reactions
    CivitAI

    佬,想求一个例图的wan2.1的工作流可以吗,我自己搭的效果很差也不知道是什么设置出了问题

    MIAOKA
    Author
    Apr 19, 2025· 1 reaction

    我例图不是本地视频模型跑的

    Checkpoint
    Illustrious

    Details

    Downloads
    6,235
    Platform
    CivitAI
    Platform Status
    Available
    Created
    3/27/2025
    Updated
    5/22/2026
    Deleted
    -