CyberHarem/le_malin_azurlane

--- license: mit task_categories: - text-to-image tags: - art - not-for-all-audiences size_categories: - n<1K --- # Dataset of le_malin/ル・マラン/恶毒 (Azur Lane) This is the dataset of le_malin/ル・マラン/恶毒 (Azur Lane), containing 500 images and their tags. The core tags of this character are `long_hair, blue_eyes, white_hair, very_long_hair, bangs, breasts, small_breasts, animal_ears, rabbit_ears, braid, twin_braids, fake_animal_ears, hair_bun, double_bun, animal_ear_fluff`, which are pruned in this dataset. Images are crawled from many sites (e.g. danbooru, pixiv, zerochan ...), the auto-crawling system is powered by [DeepGHS Team](https://github.com/deepghs)([huggingface organization](https://huggingface.co/deepghs)). ## List of Packages | Name | Images | Size | Download | Type | Description | |:-----------------|---------:|:-----------|:-------------------------------------------------------------------------------------------------------------------|:-----------|:---------------------------------------------------------------------| | raw | 500 | 933.07 MiB | [Download](https://huggingface.co/datasets/CyberHarem/le_malin_azurlane/resolve/main/dataset-raw.zip) | Waifuc-Raw | Raw data with meta information (min edge aligned to 1400 if larger). | | 800 | 500 | 419.25 MiB | [Download](https://huggingface.co/datasets/CyberHarem/le_malin_azurlane/resolve/main/dataset-800.zip) | IMG+TXT | dataset with the shorter side not exceeding 800 pixels. | | stage3-p480-800 | 1378 | 988.88 MiB | [Download](https://huggingface.co/datasets/CyberHarem/le_malin_azurlane/resolve/main/dataset-stage3-p480-800.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. | | 1200 | 500 | 767.52 MiB | [Download](https://huggingface.co/datasets/CyberHarem/le_malin_azurlane/resolve/main/dataset-1200.zip) | IMG+TXT | dataset with the shorter side not exceeding 1200 pixels. | | stage3-p480-1200 | 1378 | 1.52 GiB | [Download](https://huggingface.co/datasets/CyberHarem/le_malin_azurlane/resolve/main/dataset-stage3-p480-1200.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. | ### Load Raw Dataset with Waifuc We provide raw dataset (including tagged images) for [waifuc](https://deepghs.github.io/waifuc/main/tutorials/installation/index.html) loading. If you need this, just run the following code ```python import os import zipfile from huggingface_hub import hf_hub_download from waifuc.source import LocalSource # download raw archive file zip_file = hf_hub_download( repo_id='CyberHarem/le_malin_azurlane', repo_type='dataset', filename='dataset-raw.zip', ) # extract files to your directory dataset_dir = 'dataset_dir' os.makedirs(dataset_dir, exist_ok=True) with zipfile.ZipFile(zip_file, 'r') as zf: zf.extractall(dataset_dir) # load the dataset with waifuc source = LocalSource(dataset_dir) for item in source: print(item.image, item.meta['filename'], item.meta['tags']) ``` ## List of Clusters List of tag clustering result, maybe some outfits can be mined here. ### Raw Text Version | # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | Tags | |----:|----------:|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | 0 | 7 |  |  |  |  |  | +_+, 1girl, choker, criss-cross_halter, solo, white_dress, white_rose, bare_shoulders, blush, hair_between_eyes, looking_at_viewer, sleeveless_dress, collarbone, sidelocks, bare_arms, parted_lips, barefoot, grey_hair, hand_up, on_back | | 1 | 17 |  |  |  |  |  | +_+, 1girl, solo, white_pantyhose, sidelocks, white_dress, hair_between_eyes, looking_at_viewer, long_sleeves, symbol-shaped_pupils, blush, navel, detached_sleeves, parted_lips, sitting, tutu | | 2 | 13 |  |  |  |  |  | +_+, 1girl, solo, white_dress, white_pantyhose, hair_between_eyes, long_sleeves, sidelocks, holding_sword, looking_at_viewer, thighband_pantyhose, closed_mouth, rapier, simple_background, underboob_cutout, blush, white_background, cape, hair_bow, full_body, pocket_watch, ribbon | | 3 | 5 |  |  |  |  |  | 1girl, looking_at_viewer, navel, nipples, nude, simple_background, solo, +_+, hair_between_eyes, blush, collarbone, hair_ribbon, loli, open_mouth, white_background, anus, clitoris, closed_mouth, spread_pussy, standing, symbol-shaped_pupils, uncensored | | 4 | 18 |  |  |  |  |  | +_+, 1girl, bare_shoulders, solo, looking_at_viewer, white_bikini, symbol-shaped_pupils, sidelocks, barefoot, blush, bare_legs, headphones_around_neck, hammock, arm_garter, on_back, hair_between_eyes, navel, ass, feet, thighs, purple_flower, stomach | | 5 | 9 |  |  |  |  |  | 1girl, looking_at_viewer, official_alternate_costume, playboy_bunny, solo, white_leotard, white_pantyhose, wrist_cuffs, blush, simple_background, white_background, choker, sitting, +_+, closed_mouth, thigh_strap, twintails, hairband, rabbit_tail | | 6 | 10 |  |  |  |  |  | 1girl, ass, looking_at_viewer, official_alternate_costume, playboy_bunny, rabbit_tail, solo, white_leotard, white_pantyhose, from_behind, hairband, looking_back, +_+, simple_background, thigh_strap, white_background, wrist_cuffs, blush, parted_lips, sideboob, thighs | | 7 | 7 |  |  |  |  |  | 1girl, covered_navel, looking_at_viewer, official_alternate_costume, playboy_bunny, simple_background, solo, white_leotard, white_pantyhose, blush, thigh_strap, white_background, wrist_cuffs, choker, closed_mouth, highleg_leotard, +_+, thighs, twintails | | 8 | 5 |  |  |  |  |  | 1girl, looking_at_viewer, official_alternate_costume, playboy_bunny, rabbit_tail, solo, white_leotard, white_pantyhose, +_+, ass, from_behind, looking_back, no_shoes, wrist_cuffs, couch, blush, full_body, low_twin_braids, sideboob, thigh_strap, wariza, white_hairband | | 9 | 16 |  |  |  |  |  | 1girl, hetero, solo_focus, penis, sex, 1boy, blush, censored, nipples, +_+, pussy, vaginal, open_mouth, pantyhose, official_alternate_costume, straddling, loli, looking_at_viewer, playboy_bunny, simple_background, spread_legs, clothed_female_nude_male, girl_on_top, white_leotard, crotchless, dark-skinned_male, interracial, navel, wrist_cuffs | ### Table Version | # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | +_+ | 1girl | choker | criss-cross_halter | solo | white_dress | white_rose | bare_shoulders | blush | hair_between_eyes | looking_at_viewer | sleeveless_dress | collarbone | sidelocks | bare_arms | parted_lips | barefoot | grey_hair | hand_up | on_back | white_pantyhose | long_sleeves | symbol-shaped_pupils | navel | detached_sleeves | sitting | tutu | holding_sword | thighband_pantyhose | closed_mouth | rapier | simple_background | underboob_cutout | white_background | cape | hair_bow | full_body | pocket_watch | ribbon | nipples | nude | hair_ribbon | loli | open_mouth | anus | clitoris | spread_pussy | standing | uncensored | white_bikini | bare_legs | headphones_around_neck | hammock | arm_garter | ass | feet | thighs | purple_flower | stomach | official_alternate_costume | playboy_bunny | white_leotard | wrist_cuffs | thigh_strap | twintails | hairband | rabbit_tail | from_behind | looking_back | sideboob | covered_navel | highleg_leotard | no_shoes | couch | low_twin_braids | wariza | white_hairband | hetero | solo_focus | penis | sex | 1boy | censored | pussy | vaginal | pantyhose | straddling | spread_legs | clothed_female_nude_male | girl_on_top | crotchless | dark-skinned_male | interracial | |----:|----------:|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:------|:--------|:---------|:---------------------|:-------|:--------------|:-------------|:-----------------|:--------|:--------------------|:--------------------|:-------------------|:-------------|:------------|:------------|:--------------|:-----------|:------------|:----------|:----------|:------------------|:---------------|:-----------------------|:--------|:-------------------|:----------|:-------|:----------------|:----------------------|:---------------|:---------|:--------------------|:-------------------|:-------------------|:-------|:-----------|:------------|:---------------|:---------|:----------|:-------|:--------------|:-------|:-------------|:-------|:-----------|:---------------|:-----------|:-------------|:---------------|:------------|:-------------------------|:----------|:-------------|:------|:-------|:---------|:----------------|:----------|:-----------------------------|:----------------|:----------------|:--------------|:--------------|:------------|:-----------|:--------------|:--------------|:---------------|:-----------|:----------------|:------------------|:-----------|:--------|:------------------|:---------|:-----------------|:---------|:-------------|:--------|:------|:-------|:-----------|:--------|:----------|:------------|:-------------|:--------------|:---------------------------|:--------------|:-------------|:--------------------|:--------------| | 0 | 7 |  |  |  |  |  | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 1 | 17 |  |  |  |  |  | X | X | | | X | X | | | X | X | X | | | X | | X | | | | | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 2 | 13 |  |  |  |  |  | X | X | | | X | X | | | X | X | X | | | X | | | | | | | X | X | | | | | | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 3 | 5 |  |  |  |  |  | X | X | | | X | | | | X | X | X | | X | | | | | | | | | | X | X | | | | | | X | | X | | X | | | | | | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4 | 18 |  |  |  |  |  | X | X | | | X | | | X | X | X | X | | | X | | | X | | | X | | | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 5 | 9 |  |  |  |  |  | X | X | X | | X | | | | X | | X | | | | | | | | | | X | | | | | X | | | | X | | X | | X | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | 6 | 10 |  |  |  |  |  | X | X | | | X | | | | X | | X | | | | | X | | | | | X | | | | | | | | | | | X | | X | | | | | | | | | | | | | | | | | | | | | X | | X | | | X | X | X | X | X | | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | 7 | 7 |  |  |  |  |  | X | X | X | | X | | | | X | | X | | | | | | | | | | X | | | | | | | | | X | | X | | X | | | | | | | | | | | | | | | | | | | | | | | X | | | X | X | X | X | X | X | | | | | | X | X | | | | | | | | | | | | | | | | | | | | | | | 8 | 5 |  |  |  |  |  | X | X | | | X | | | | X | | X | | | | | | | | | | X | | | | | | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | X | | | | | X | X | X | X | X | | | X | X | X | X | | | X | X | X | X | X | | | | | | | | | | | | | | | | | | 9 | 16 |  |  |  |  |  | X | X | | | | | | | X | | X | | | | | | | | | | | | | X | | | | | | | | X | | | | | | | | X | | | X | X | | | | | | | | | | | | | | | | X | X | X | X | | | | | | | | | | | | | | | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X |

原始信息汇总

数据集概述

基本信息

• 名称: Dataset of le_malin/ル・マラン/恶毒 (Azur Lane)
• 包含内容: 500张图像及其标签
• 核心标签: long_hair, blue_eyes, white_hair, very_long_hair, bangs, breasts, small_breasts, animal_ears, rabbit_ears, braid, twin_braids, fake_animal_ears, hair_bun, double_bun, animal_ear_fluff
• 许可证: MIT
• 任务类别: text-to-image
• 标签: art, not-for-all-audiences
• 大小类别: n<1K

数据集包列表

名称	图像数量	大小	下载链接	类型	描述
raw	500	933.07 MiB	下载	Waifuc-Raw	包含元信息的原始数据（最小边对齐至1400像素，如果更大）。
800	500	419.25 MiB	下载	IMG+TXT	短边不超过800像素的数据集。
stage3-p480-800	1378	988.88 MiB	下载	IMG+TXT	三阶段裁剪数据集，区域不小于480x480像素。
1200	500	767.52 MiB	下载	IMG+TXT	短边不超过1200像素的数据集。
stage3-p480-1200	1378	1.52 GiB	下载	IMG+TXT	三阶段裁剪数据集，区域不小于480x480像素。

加载原始数据集

• 使用工具: waifuc
• 加载方法: 通过Python代码下载并解压数据集，然后使用waifuc加载。

集群列表

• 描述: 列出了标签聚类结果，可能包含可挖掘的服装信息。
• 示例: 展示了不同集群的样本图像及其标签。

搜集汇总

数据集介绍

构建方式

在二次元角色数据集构建领域，CyberHarem/le_malin_azurlane 数据集聚焦于游戏《碧蓝航线》中的角色“恶毒”（ル・マラン）。该数据集通过自动化爬取系统从 Danbooru、Pixiv、Zerochan 等多个知名图像站点收集原始素材，爬取引擎由 DeepGHS 团队提供技术支持。原始图像经过筛选，最终保留 500 张高质量图片，并附有对应的标签信息。数据集以多种分辨率版本发布，包括原始分辨率、短边不超过 800 像素及 1200 像素的版本，以及经过三阶段裁剪处理、确保图像区域不小于 480×480 像素的增强版本，以满足不同训练需求。

特点

该数据集的核心特色在于其精细的标签体系与多版本结构。标签系统涵盖了角色的核心外观特征，如长发、蓝眼、白发、兔耳、双马尾等，并经过剪枝优化以突出关键属性。数据集提供了 Raw、800、1200 及 stage3-p480 等多种压缩与裁剪版本，其中 stage3 版本通过三阶段裁剪技术生成了 1378 张子图，有效增加了数据多样性。此外，数据集还提供了基于标签聚类的分组结果，将图像按服饰与场景主题归类，便于研究者挖掘不同装扮风格下的角色特征。

使用方法

使用者可通过 Hugging Face Hub 直接下载压缩包，或利用 waifuc 库加载原始数据集。加载时需先通过 hf_hub_download 函数下载 dataset-raw.zip 文件，解压至本地目录后，使用 waifuc 的 LocalSource 接口即可逐项读取图像及其元数据与标签。对于需要特定分辨率的场景，可直接选用 800、1200 或 stage3-p480 等预处理版本，无需额外处理。该数据集适用于文本到图像生成模型的训练，以及角色识别、风格迁移等二次元图像理解任务。

背景与挑战

背景概述

在生成式人工智能与动漫文化交汇的浪潮中，高质量、标签精细的角色数据集成为推动文本到图像（Text-to-Image）模型发展的关键基石。由DeepGHS团队于近年创建的CyberHarem/le_malin_azurlane数据集，聚焦于《碧蓝航线》中人气角色“恶毒”（Le Malin），通过自动化爬取系统从Danbooru、Pixiv、Zerochan等多源平台搜集了500张图像及其元标签。该数据集不仅服务于角色定制化生成的研究需求，还通过多尺度裁剪（如stage3-p480-800版本）与标签聚类分析（如服装、姿态分组），为模型在动漫角色精细特征（如兔耳、双马尾、特定服饰）的识别与生成提供了标准化训练资源，显著推动了二次元垂直领域的数据集建设与开源生态发展。

当前挑战

该数据集面临的首要挑战在于领域问题的复杂性：动漫角色图像生成需兼顾角色核心特征（如白发蓝瞳）的保真度与风格多样性，而现有模型在区分相似标签（如“动物耳”与“假动物耳”）时易出现语义混淆。构建过程中，爬取策略需应对多平台反爬机制与版权肖像权合规风险，标签自动提取依赖社区标注的准确性，存在噪声与遗漏（如部分图像含敏感内容）。此外，数据量仅500张，难以覆盖角色全部姿态与场景，聚类结果中“playboy_bunny”等特定主题样本稀疏，易导致模型过拟合，亟需通过数据增强或半监督学习扩展样本空间。

常用场景

经典使用场景

该数据集聚焦于《碧蓝航线》角色“恶毒”（le malin），包含500张高质量图像及其标签，是动漫风格角色图像生成与风格迁移研究的理想资源。通过提供原始、多分辨率裁剪及标签聚类等多种版本，它支持从文本到图像生成模型的微调训练，尤其适用于学习特定角色外貌（如白发、蓝眼、兔耳等核心特征）与不同着装形态的映射关系。研究者可借助这些数据探索角色一致性生成、局部属性编辑及多视角渲染等经典任务。

衍生相关工作

该数据集衍生出的经典工作包括基于角色标签聚类的服饰分类与姿态分析模型，以及结合多尺度裁剪策略的图像质量增强方法。研究团队利用其提供的原始与裁剪版本，开发了动漫角色局部属性编辑算法与跨数据集风格迁移框架。此外，该数据集的自动爬取与标注流程（由DeepGHS团队支持）为大规模动漫图像数据集的构建提供了可复现的技术范式，催生了更多针对特定IP角色的细粒度数据集，促进了二次元生成领域的标准化研究。

数据集最近研究

最新研究方向

在二次元角色生成与个性化内容创作领域，基于特定动漫游戏角色的精细化图像数据集构建正成为推动文本到图像模型发展的关键。CyberHarem/le_malin_azurlane 数据集聚焦于《碧蓝航线》中人气角色“恶毒”，通过从 Danbooru、Pixiv 等多源平台自动爬取并整理 500 张高质量图像及其标签，为角色定制化生成提供了结构化数据支撑。该数据集不仅包含原始图像与多种尺寸的裁剪版本，还创新性地引入了标签聚类分析，可挖掘角色在不同服饰、姿势下的视觉特征模式。这一前沿方向紧密关联当前扩散模型在可控生成、角色一致性保持方面的研究热点，其意义在于为动漫风格文本到图像生成任务提供标准化、可复现的基准数据，助力模型在特定角色外观、属性及场景下的精准生成与风格迁移，推动二次元内容创作自动化与个性化水平的提升。