CyberHarem/hiei_kantaicollection

---
license: mit
task_categories:
- text-to-image
tags:
- art
- not-for-all-audiences
size_categories:
- n<1K
---

# Dataset of hiei/比叡/比叡 (Kantai Collection)

This is the dataset of hiei/比叡/比叡 (Kantai Collection), containing 500 images and their tags.

The core tags of this character are `short_hair, brown_hair, hairband, blue_eyes, headgear, flipped_hair, breasts`, 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 | 528.09 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hiei_kantaicollection/resolve/main/dataset-raw.zip) | Waifuc-Raw | Raw data with meta information (min edge aligned to 1400 if larger). |
| 800 | 500 | 324.61 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hiei_kantaicollection/resolve/main/dataset-800.zip) | IMG+TXT | dataset with the shorter side not exceeding 800 pixels. |
| stage3-p480-800 | 1137 | 665.05 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hiei_kantaicollection/resolve/main/dataset-stage3-p480-800.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. |
| 1200 | 500 | 472.34 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hiei_kantaicollection/resolve/main/dataset-1200.zip) | IMG+TXT | dataset with the shorter side not exceeding 1200 pixels. |
| stage3-p480-1200 | 1137 | 904.79 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hiei_kantaicollection/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/hiei_kantaicollection',
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 | 14 |  |  |  |  |  | 1girl, bare_shoulders, detached_sleeves, japanese_clothes, nontraditional_miko, thighhighs, pantyhose, smile, solo, open_mouth, thigh_boots, turret, cannon, one_eye_closed, plaid_skirt, ribbon-trimmed_sleeves, ;d |
| 1 | 6 |  |  |  |  |  | 1girl, detached_sleeves, green_skirt, japanese_clothes, nontraditional_miko, plaid_skirt, ribbon-trimmed_sleeves, simple_background, smile, solo, white_background, cowboy_shot, looking_at_viewer, wide_sleeves, blush, clenched_hands, grey_eyes, one-hour_drawing_challenge |
| 2 | 6 |  |  |  |  |  | 1girl, detached_sleeves, japanese_clothes, looking_at_viewer, nontraditional_miko, plaid_skirt, ribbon-trimmed_sleeves, solo, thigh_boots, thighhighs, green_skirt, simple_background, smile, white_background, blush, cowboy_shot, grey_eyes, open_mouth |
| 3 | 16 |  |  |  |  |  | 1girl, detached_sleeves, japanese_clothes, ribbon-trimmed_sleeves, solo, smile, upper_body, looking_at_viewer, nontraditional_miko, simple_background, white_background |
| 4 | 7 |  |  |  |  |  | 2girls, bare_shoulders, detached_sleeves, japanese_clothes, nontraditional_miko, smile, blush, open_mouth, skirt, long_hair |
| 5 | 5 |  |  |  |  |  | 1girl, detached_sleeves, hetero, nipples, open_mouth, penis, 1boy, blush, medium_breasts, nontraditional_miko, solo_focus, japanese_clothes, mosaic_censoring, pussy, ribbon-trimmed_sleeves, sex, thighhighs, vaginal, bare_shoulders, handjob, large_breasts, open_clothes, spread_legs, sweat, tears |
| 6 | 13 |  |  |  |  |  | looking_at_viewer, 1girl, blue_sky, cloud, cowboy_shot, day, smile, solo, outdoors, white_bikini, large_breasts, navel, open_mouth, side-tie_bikini_bottom |
| 7 | 21 |  |  |  |  |  | playboy_bunny, rabbit_ears, strapless_leotard, 1girl, detached_collar, solo, looking_at_viewer, fake_animal_ears, wrist_cuffs, simple_background, black_pantyhose, cowboy_shot, large_breasts, white_background, alternate_costume, cleavage, medium_breasts, bowtie, white_leotard, rabbit_tail, thighband_pantyhose, green_necktie, smile |

### Table Version

| # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | 1girl | bare_shoulders | detached_sleeves | japanese_clothes | nontraditional_miko | thighhighs | pantyhose | smile | solo | open_mouth | thigh_boots | turret | cannon | one_eye_closed | plaid_skirt | ribbon-trimmed_sleeves | ;d | green_skirt | simple_background | white_background | cowboy_shot | looking_at_viewer | wide_sleeves | blush | clenched_hands | grey_eyes | one-hour_drawing_challenge | upper_body | 2girls | skirt | long_hair | hetero | nipples | penis | 1boy | medium_breasts | solo_focus | mosaic_censoring | pussy | sex | vaginal | handjob | large_breasts | open_clothes | spread_legs | sweat | tears | blue_sky | cloud | day | outdoors | white_bikini | navel | side-tie_bikini_bottom | playboy_bunny | rabbit_ears | strapless_leotard | detached_collar | fake_animal_ears | wrist_cuffs | black_pantyhose | alternate_costume | cleavage | bowtie | white_leotard | rabbit_tail | thighband_pantyhose | green_necktie |
|----:|----------:|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------|:-----------------|:-------------------|:-------------------|:----------------------|:-------------|:------------|:--------|:-------|:-------------|:--------------|:---------|:---------|:-----------------|:--------------|:-------------------------|:-----|:--------------|:--------------------|:-------------------|:--------------|:--------------------|:---------------|:--------|:-----------------|:------------|:-----------------------------|:-------------|:---------|:--------|:------------|:---------|:----------|:--------|:-------|:-----------------|:-------------|:-------------------|:--------|:------|:----------|:----------|:----------------|:---------------|:--------------|:--------|:--------|:-----------|:--------|:------|:-----------|:---------------|:--------|:-------------------------|:----------------|:--------------|:--------------------|:------------------|:-------------------|:--------------|:------------------|:--------------------|:-----------|:---------|:----------------|:--------------|:----------------------|:----------------|
| 0 | 14 |  |  |  |  |  | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 1 | 6 |  |  |  |  |  | X | | X | X | X | | | X | X | | | | | | X | X | | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 2 | 6 |  |  |  |  |  | X | | X | X | X | X | | X | X | X | X | | | | X | X | | X | X | X | X | X | | X | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 3 | 16 |  |  |  |  |  | X | | X | X | X | | | X | X | | | | | | | X | | | X | X | | X | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 4 | 7 |  |  |  |  |  | | X | X | X | X | | | X | | X | | | | | | | | | | | | | | X | | | | | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 5 | 5 |  |  |  |  |  | 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 | | | | | | | | | | | | | | | | | | | | | |
| 6 | 13 |  |  |  |  |  | X | | | | | | | X | X | X | | | | | | | | | | | X | X | | | | | | | | | | | | | | | | | | | | | X | | | | | X | X | X | X | X | X | X | | | | | | | | | | | | | | |
| 7 | 21 |  |  |  |  |  | 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

数据集基本信息

• 名称: Dataset of hiei/比叡/比叡 (Kantai Collection)
• 许可证: MIT
• 任务类别: text-to-image
• 标签: art, not-for-all-audiences
• 大小类别: n<1K

数据集内容

• 图像数量: 500张
• 核心标签: short_hair, brown_hair, hairband, blue_eyes, headgear, flipped_hair, breasts

数据包列表

名称	图像数量	大小	类型	描述
raw	500	528.09 MiB	Waifuc-Raw	原始数据，包含元信息（最小边对齐到1400像素，如果更大）
800	500	324.61 MiB	IMG+TXT	短边不超过800像素的数据集
stage3-p480-800	1137	665.05 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素
1200	500	472.34 MiB	IMG+TXT	短边不超过1200像素的数据集
stage3-p480-1200	1137	904.79 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素

标签聚类结果

原始文本版本

#	样本数量	图像示例	标签
0	14		1girl, bare_shoulders, detached_sleeves, japanese_clothes, nontraditional_miko, thighhighs, pantyhose, smile, solo, open_mouth, thigh_boots, turret, cannon, one_eye_closed, plaid_skirt, ribbon-trimmed_sleeves, ;d
1	6		1girl, detached_sleeves, green_skirt, japanese_clothes, nontraditional_miko, plaid_skirt, ribbon-trimmed_sleeves, simple_background, smile, solo, white_background, cowboy_shot, looking_at_viewer, wide_sleeves, blush, clenched_hands, grey_eyes, one-hour_drawing_challenge
2	6		1girl, detached_sleeves, japanese_clothes, looking_at_viewer, nontraditional_miko, plaid_skirt, ribbon-trimmed_sleeves, solo, thigh_boots, thighhighs, green_skirt, simple_background, smile, white_background, blush, cowboy_shot, grey_eyes, open_mouth
3	16		1girl, detached_sleeves, japanese_clothes, ribbon-trimmed_sleeves, solo, smile, upper_body, looking_at_viewer, nontraditional_miko, simple_background, white_background
4	7		2girls, bare_shoulders, detached_sleeves, japanese_clothes, nontraditional_miko, smile, blush, open_mouth, skirt, long_hair
5	5		1girl, detached_sleeves, hetero, nipples, open_mouth, penis, 1boy, blush, medium_breasts, nontraditional_miko, solo_focus, japanese_clothes, mosaic_censoring, pussy, sex, thighhighs, vaginal, bare_shoulders, handjob, large_breasts, open_clothes, spread_legs, sweat, tears
6	13		looking_at_viewer, 1girl, blue_sky, cloud, cowboy_shot, day, smile, solo, outdoors, white_bikini, large_breasts, navel, open_mouth, side-tie_bikini_bottom
7	21		playboy_bunny, rabbit_ears, strapless_leotard, 1girl, detached_collar, solo, looking_at_viewer, fake_animal_ears, wrist_cuffs, simple_background, black_pantyhose, cowboy_shot, large_breasts, white_background, alternate_costume, cleavage, medium_breasts, bowtie, white_leotard, rabbit_tail, thighband_pantyhose, green_necktie, smile

表格版本

#	样本数量	图像示例	1girl	bare_shoulders	detached_sleeves	japanese_clothes	nontraditional_miko	thighhighs	pantyhose	smile	solo	open_mouth	thigh_boots	turret	cannon	one_eye_closed	plaid_skirt	ribbon-trimmed_sleeves	;d	green_skirt	simple_background	white_background	cowboy_shot	looking_at_viewer	wide_sleeves	blush	clenched_hands	grey_eyes	one-hour_drawing_challenge	upper_body	2girls	skirt	long_hair	hetero	nipples	penis	1boy	medium_breasts	solo_focus	mosaic_censoring	pussy	sex	vaginal	handjob	large_breasts	open_clothes	spread_legs	sweat	tears	blue_sky	cloud	day	outdoors	white_bikini	navel	side-tie_bikini_bottom	playboy_bunny	rabbit_ears	strapless_leotard	detached_collar	fake_animal_ears	wrist_cuffs	black_pantyhose	alternate_costume	cleavage	bowtie	white_leotard	rabbit_tail	thighband_pantyhose	green_necktie
0	14		X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X	X
1	6		X		X	X	X			X	X						X	X		X	X	X	X	X	X	X	X	X	X
2	6		X		X	X	X	X		X	X	X	X				X	X		X	X	X	X	X		X		X
3	16		X		X	X	X			X	X							X			X	X		X						X
4	7			X	X	X	X			X		X														X					X	X	X
5	5		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
6	13		X							X	X	X											X	X																					X					X	X	X	X	X	X	X
7	21		X							X	X												X	X																					X					X	X	X	X	X	X	X

搜集汇总

数据集介绍

构建方式

在二次元文化与人工智能的交叉领域中，高质量的角色图像数据集是驱动文本到图像生成模型发展的关键资源。CyberHarem/hiei_kantaicollection 数据集专注于收录《舰队Collection》中的角色比叡（hiei），其构建过程严谨而系统。数据集的原始图像经由自动化爬取系统从Danbooru、Pixiv、Zerochan等多个知名图站获取，该系统由DeepGHS团队开发并维护。为确保数据质量与一致性，原始图像经过预处理，将最小边长对齐至1400像素。最终数据集包含500张图像及其对应的标签信息，并剔除了角色的核心标签（如短发、棕发、蓝眼等），以提升数据集的聚焦度与可用性。

特点

该数据集展现出鲜明的结构化与多层次特点，以满足不同应用场景的需求。除了提供包含元信息的原始数据包外，还衍生出多种规格的压缩版本，例如短边不超过800像素或1200像素的标准化图像集。尤为突出的是，数据集引入了创新的三阶段裁剪版本（stage3-p480-800与stage3-p480-1200），确保裁剪后的图像区域不小于480x480像素，从而保留了关键视觉元素的完整性。此外，数据集还提供了详尽的标签聚类结果，将图像按视觉主题（如传统巫女装、泳装、兔女郎装等）进行分组，为风格迁移、服装识别等下游任务提供了丰富的语义线索。

使用方法

为便于研究者高效利用，该数据集提供了灵活的使用路径。对于偏好原始数据处理的用户，可通过Waifuc库加载原始数据包，利用其提供的Python接口便捷地访问图像、文件名及标签信息。加载过程首先通过Hugging Face Hub下载压缩包，解压至指定目录后，即可使用LocalSource进行迭代读取。对于需要直接投入模型训练的场景，用户可直接下载IMG+TXT格式的打包文件，这些文件已包含对齐的图像与文本标签，无需额外预处理，大幅降低了数据加载与格式转换的复杂度。

背景与挑战

背景概述

在二次元文化日益繁荣的背景下，角色图像数据集成为文本到图像生成模型训练的重要基石。CyberHarem/hiei_kantaicollection数据集由DeepGHS团队于近年创建，聚焦于《舰队Collection》中的角色“比叡”，汇集了500张涵盖多种姿态与场景的高质量图像。该数据集旨在为动漫风格图像生成提供精细化的角色特征标注，如短发、棕发、蓝眼等核心标签，并通过自动爬取系统从Danbooru、Pixiv等站点采集数据。其发布不仅丰富了二次元角色数据资源，也为相关领域的研究者提供了标准化训练素材，推动了角色定制化生成技术的发展。

当前挑战

该数据集面临的核心挑战在于领域问题的复杂性与构建过程的精细度。一方面，角色图像生成需解决多姿态、多服装变体下的特征一致性难题，例如原始数据中比叡的巫女服、泳装、兔女郎装等不同造型，要求模型具备强泛化能力以准确还原角色身份。另一方面，构建过程中自动爬取系统面临图像质量参差不齐、标签噪声显著等问题，尤其是涉及敏感内容（如聚类5中的成人向图像）时，需通过裁剪、过滤等预处理手段确保数据纯净性。此外，数据规模较小（仅500张原始图像）限制了模型训练效果，而多分辨率版本（如800px、1200px）的扩充虽部分缓解了这一问题，但仍需更高效的数据增强策略来克服样本不足的瓶颈。

常用场景

经典使用场景

CyberHarem/hiei_kantaicollection数据集专为文本到图像生成任务而设计，其核心应用在于基于角色“比叡”（出自《舰队Collection》）的视觉特征进行可控图像合成。该数据集包含500张经过精细标注的图像，并提供了多种分辨率版本（如800px、1200px）以及三阶段裁剪版本，以适应不同粒度的生成需求。通过引入短发、棕发、发带、蓝眼、头饰等核心标签，研究者能够利用该数据集训练或微调扩散模型，实现对特定动漫角色的一致性与风格化生成。此外，数据集的聚类结果（如服饰分类）为探索角色在不同场景下的外观变体提供了结构化参考，从而推动角色保持与属性编辑等任务的深入研究。

实际应用

在实际应用层面，该数据集被广泛用于动漫角色IP的自动化视觉内容生产，例如为游戏《舰队Collection》的衍生创作提供角色立绘的快速生成与变体设计。通过结合waifuc等工具，开发者能够便捷地加载原始数据并构建自定义的生成管线，从而在社交媒体、粉丝艺术平台或商业宣传中实现角色形象的批量定制。此外，数据集的聚类结果（如泳装、巫女服、兔女郎等服饰类别）可直接用于角色换装系统的训练，推动虚拟偶像或互动娱乐中的实时外观切换功能。这些应用不仅降低了人工设计成本，还提升了内容生成的多样性与响应速度。

衍生相关工作

该数据集衍生了一系列经典工作，主要集中在动漫图像生成与特征解耦领域。基于其标注结构，研究者提出了角色身份保持的扩散模型微调策略，如使用DreamBooth或LoRA技术对核心标签进行低秩适应，从而实现特定风格下的高保真生成。此外，数据集的聚类信息被用于训练服饰分类器与姿态估计网络，推动了如“角色换装”（Outfit Transfer）和“背景替换”（Background Removal）等下游任务的发展。在更广泛的范畴内，CyberHarem系列数据集（如hiei_kantaicollection）为多角色联合生成与属性编辑提供了基准，催生了基于对比学习的角色特征分离方法，这些工作共同构成了动漫AIGC领域从数据构建到模型优化的完整技术链条。