CyberHarem/hayashimo_kantaicollection

--- license: mit task_categories: - text-to-image tags: - art - not-for-all-audiences size_categories: - n<1K --- # Dataset of hayashimo/早霜/早霜 (Kantai Collection) This is the dataset of hayashimo/早霜/早霜 (Kantai Collection), containing 500 images and their tags. The core tags of this character are `black_hair, long_hair, hair_over_one_eye, ribbon, very_long_hair, hair_ribbon, bow, brown_eyes`, 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 | 428.63 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hayashimo_kantaicollection/resolve/main/dataset-raw.zip) | Waifuc-Raw | Raw data with meta information (min edge aligned to 1400 if larger). | | 800 | 500 | 281.71 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hayashimo_kantaicollection/resolve/main/dataset-800.zip) | IMG+TXT | dataset with the shorter side not exceeding 800 pixels. | | stage3-p480-800 | 1071 | 558.06 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hayashimo_kantaicollection/resolve/main/dataset-stage3-p480-800.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. | | 1200 | 500 | 390.21 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hayashimo_kantaicollection/resolve/main/dataset-1200.zip) | IMG+TXT | dataset with the shorter side not exceeding 1200 pixels. | | stage3-p480-1200 | 1071 | 723.97 MiB | [Download](https://huggingface.co/datasets/CyberHarem/hayashimo_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/hayashimo_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 | 11 |  |  |  |  |  | 1girl, solo, blush, looking_at_viewer, simple_background, sitting, white_background, barefoot, medium_breasts, navel, nude, purple_eyes, smile, bikini, yellow_eyes | | 1 | 6 |  |  |  |  |  | 1girl, black_bra, black_panties, solo, looking_at_viewer, navel, simple_background, small_breasts, underwear_only, blush, white_background, cleavage | | 2 | 6 |  |  |  |  |  | 1girl, blush, looking_at_viewer, open_shirt, solo, black_panties, large_breasts, open_mouth, purple_eyes, grey_pantyhose, navel, nipples, no_bra, smile, white_shirt, dakimakura_(medium), full_body, on_back, skirt | | 3 | 19 |  |  |  |  |  | 1girl, bowtie, school_uniform, solo, white_shirt, long_sleeves, looking_at_viewer, purple_dress, halterneck, simple_background, white_background, white_ribbon, cowboy_shot | | 4 | 9 |  |  |  |  |  | 1girl, bowtie, grey_pantyhose, looking_at_viewer, school_uniform, solo, white_shirt, lace-up_boots, simple_background, long_sleeves, white_background, full_body, skirt, smile | | 5 | 5 |  |  |  |  |  | 1girl, bowtie, holding_umbrella, long_sleeves, looking_at_viewer, school_uniform, solo, white_shirt, rain, smile, sleeveless_dress, hydrangea, upper_body | | 6 | 14 |  |  |  |  |  | 1girl, enmaided, solo, black_dress, maid_headdress, looking_at_viewer, simple_background, maid_apron, smile, white_apron, long_sleeves, white_background, frilled_apron, puffy_sleeves, twitter_username, blush, one-hour_drawing_challenge, breasts, dated, pantyhose, purple_eyes | | 7 | 21 |  |  |  |  |  | playboy_bunny, 1girl, bowtie, detached_collar, fake_animal_ears, rabbit_ears, wrist_cuffs, solo, strapless_leotard, purple_leotard, looking_at_viewer, rabbit_tail, grey_pantyhose, simple_background, fishnet_pantyhose, adapted_costume, small_breasts, blush, cleavage, white_background | | 8 | 7 |  |  |  |  |  | 1boy, 1girl, blush, hetero, penis, solo_focus, fellatio, medium_breasts, open_mouth, closed_eyes, mosaic_censoring, nipples, shirt, bowtie, breasts_out, cum_on_tongue, facial, male_pubic_hair | ### Table Version | # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | 1girl | solo | blush | looking_at_viewer | simple_background | sitting | white_background | barefoot | medium_breasts | navel | nude | purple_eyes | smile | bikini | yellow_eyes | black_bra | black_panties | small_breasts | underwear_only | cleavage | open_shirt | large_breasts | open_mouth | grey_pantyhose | nipples | no_bra | white_shirt | dakimakura_(medium) | full_body | on_back | skirt | bowtie | school_uniform | long_sleeves | purple_dress | halterneck | white_ribbon | cowboy_shot | lace-up_boots | holding_umbrella | rain | sleeveless_dress | hydrangea | upper_body | enmaided | black_dress | maid_headdress | maid_apron | white_apron | frilled_apron | puffy_sleeves | twitter_username | one-hour_drawing_challenge | breasts | dated | pantyhose | playboy_bunny | detached_collar | fake_animal_ears | rabbit_ears | wrist_cuffs | strapless_leotard | purple_leotard | rabbit_tail | fishnet_pantyhose | adapted_costume | 1boy | hetero | penis | solo_focus | fellatio | closed_eyes | mosaic_censoring | shirt | breasts_out | cum_on_tongue | facial | male_pubic_hair | |----:|----------:|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------|:-------|:--------|:--------------------|:--------------------|:----------|:-------------------|:-----------|:-----------------|:--------|:-------|:--------------|:--------|:---------|:--------------|:------------|:----------------|:----------------|:-----------------|:-----------|:-------------|:----------------|:-------------|:-----------------|:----------|:---------|:--------------|:----------------------|:------------|:----------|:--------|:---------|:-----------------|:---------------|:---------------|:-------------|:---------------|:--------------|:----------------|:-------------------|:-------|:-------------------|:------------|:-------------|:-----------|:--------------|:-----------------|:-------------|:--------------|:----------------|:----------------|:-------------------|:-----------------------------|:----------|:--------|:------------|:----------------|:------------------|:-------------------|:--------------|:--------------|:--------------------|:-----------------|:--------------|:--------------------|:------------------|:-------|:---------|:--------|:-------------|:-----------|:--------------|:-------------------|:--------|:--------------|:----------------|:---------|:------------------| | 0 | 11 |  |  |  |  |  | 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 | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 2 | 6 |  |  |  |  |  | X | X | X | X | | | | | | X | | X | X | | | | X | | | | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 3 | 19 |  |  |  |  |  | X | X | | X | X | | X | | | | | | | | | | | | | | | | | | | | X | | | | | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4 | 9 |  |  |  |  |  | X | X | | X | X | | 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 | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 6 | 14 |  |  |  |  |  | X | X | X | X | X | | X | | | | | X | 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 | | | | | | | | | | | | | | 8 | 7 |  |  |  |  |  | X | | X | | | | | | X | | | | | | | | | | | | | | X | | X | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | X | X | X | X | X | X | X | X | X | X |This is the dataset of hayashimo/早霜/早霜 (Kantai Collection), containing 500 images and their tags. The core tags of this character are black_hair, long_hair, hair_over_one_eye, ribbon, very_long_hair, hair_ribbon, bow, brown_eyes, 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. The dataset is available in multiple formats including raw data, images of various sizes, and cropped datasets, along with a list of clusters based on tag clustering, which may be useful for further analysis or mining of specific outfits or themes.

原始信息汇总

数据集概述

数据集信息

• 名称: Dataset of hayashimo/早霜/早霜 (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: black_hair, long_hair, hair_over_one_eye, ribbon, very_long_hair, hair_ribbon, bow, brown_eyes

数据包列表

名称	图像数量	大小	类型	描述
raw	500	428.63 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	281.71 MiB	IMG+TXT	短边不超过800像素的数据集。
stage3-p480-800	1071	558.06 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素。
1200	500	390.21 MiB	IMG+TXT	短边不超过1200像素的数据集。
stage3-p480-1200	1071	723.97 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	11		1girl, solo, blush, looking_at_viewer, simple_background, sitting, white_background, barefoot, medium_breasts, navel, nude, purple_eyes, smile, bikini, yellow_eyes
1	6		1girl, black_bra, black_panties, solo, looking_at_viewer, navel, simple_background, small_breasts, underwear_only, blush, white_background, cleavage
2	6		1girl, blush, looking_at_viewer, open_shirt, solo, black_panties, large_breasts, open_mouth, purple_eyes, grey_pantyhose, navel, nipples, no_bra, smile, white_shirt, dakimakura_(medium), full_body, on_back, skirt
3	19		1girl, bowtie, school_uniform, solo, white_shirt, long_sleeves, looking_at_viewer, purple_dress, halterneck, simple_background, white_background, white_ribbon, cowboy_shot
4	9		1girl, bowtie, grey_pantyhose, looking_at_viewer, school_uniform, solo, white_shirt, lace-up_boots, simple_background, long_sleeves, white_background, full_body, skirt, smile
5	5		1girl, bowtie, holding_umbrella, long_sleeves, looking_at_viewer, school_uniform, solo, white_shirt, rain, smile, sleeveless_dress, hydrangea, upper_body
6	14		1girl, enmaided, solo, black_dress, maid_headdress, looking_at_viewer, simple_background, maid_apron, smile, white_apron, long_sleeves, white_background, frilled_apron, puffy_sleeves, twitter_username, blush, one-hour_drawing_challenge, breasts, dated, pantyhose, purple_eyes
7	21		playboy_bunny, 1girl, bowtie, detached_collar, fake_animal_ears, rabbit_ears, wrist_cuffs, solo, strapless_leotard, purple_leotard, looking_at_viewer, rabbit_tail, grey_pantyhose, simple_background, fishnet_pantyhose, adapted_costume, small_breasts, blush, cleavage, white_background
8	7		1boy, 1girl, blush, hetero, penis, solo_focus, fellatio, medium_breasts, open_mouth, closed_eyes, mosaic_censoring, nipples, shirt, bowtie, breasts_out, cum_on_tongue, facial, male_pubic_hair

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，该数据集聚焦于《舰队Collection》中的角色早霜，收录了500张图像及对应的标签信息。数据采集自Danbooru、Pixiv、Zerochan等多个知名图像平台，依托DeepGHS团队开发的自动化爬取系统完成。原始数据经过预处理，裁剪了角色的核心标签如黑发、长发、遮眼发、蝴蝶结等，并提供了多种分辨率版本，包括短边不超过800像素和1200像素的标准化数据集，以及基于三阶段裁剪策略生成的图像块数据集，确保图像区域不小于480×480像素。

使用方法

数据集支持通过Waifuc框架加载原始压缩包中的图像与标签信息，用户可利用HuggingFace Hub的API下载dataset-raw.zip文件，解压后通过LocalSource接口逐项访问图像及其元数据。此外，数据集提供了多个预处理的ZIP包，包括不同分辨率版本和三阶段裁剪版本，用户可根据需求直接下载使用。这些包内包含图像与对应的TXT标签文件，适合直接用于文本到图像模型的训练与评估，无需额外处理。

背景与挑战

背景概述

在文本到图像生成领域，高质量、标注精细的角色数据集是驱动模型学习特定视觉概念与风格的关键资源。CyberHarem/hayashimo_kantaicollection 数据集由 DeepGHS 团队于近期构建并发布，专注于收录来自热门游戏《舰队Collection》中的角色“早霜”。该数据集的核心研究问题在于如何系统性地从多个网络图库（如 Danbooru、Pixiv 等）自动采集并整理出具有一致性角色特征（如黑发、长发、蝴蝶结等）的图像集合，以服务于二次元角色定制化生成任务。通过提供原始图像、多尺度裁剪版本及聚类标签信息，该数据集为后续的模型微调与风格迁移研究奠定了坚实基础，在动漫风格生成社区中展现出显著的影响力。

当前挑战

该数据集所面临的挑战主要体现在两个层面。首先，在领域问题方面，其致力于解决动漫角色图像生成中因数据稀缺导致的模型泛化能力不足问题，需确保在有限样本（500张图像）条件下，模型能准确捕捉角色核心外观与多种服饰变体，避免过拟合或特征混淆。其次，在构建过程中，挑战包括从异构网站自动爬取图像时需应对版权与内容审核风险，以及清洗和标注海量噪声标签的复杂性。此外，多尺度裁剪策略（如 stage3-p480-800）需平衡图像分辨率与计算资源，聚类算法则需有效区分不同服饰主题（如校服、女仆装、泳装），以提升数据集的实用性与可解释性。

常用场景

经典使用场景

在二次元图像生成与风格迁移领域，CyberHarem/hayashimo_kantaicollection数据集为训练基于扩散模型的文本到图像生成系统提供了高质量的素材。该数据集收录了来自《舰队Collection》角色早霜的500张标注图像，涵盖校服、女仆装、泳装等多种服饰变体，尤其适用于LoRA（Low-Rank Adaptation）微调技术，用于在Stable Diffusion等预训练模型中注入特定角色概念，实现精准的角色再现与风格化创作。

解决学术问题

该数据集有效解决了动漫角色个性化生成中的样本稀缺与概念对齐难题。以往模型在面对非主流或细节丰富的角色时，常出现特征混淆或生成质量不稳定的问题。通过提供经过严格筛选与标注的高分辨率图像，并配套多尺度裁剪版本，研究者得以系统性地探索角色核心标签（如黑发、长刘海、蝴蝶结）与生成结果之间的映射关系，推动了可控图像生成与细粒度特征解耦研究的发展。

实际应用

在实际应用中，该数据集被广泛用于虚拟角色创作、游戏美术资产辅助设计及动漫同人作品自动化生产。创作者可利用微调后的模型快速生成风格统一、服饰多样的角色立绘，显著降低手绘成本。此外，该数据集支持的批量图像标注与聚类功能，有助于构建面向特定IP的定制化生成管线，在数字娱乐与内容营销领域展现出显著的应用价值。

数据集最近研究

最新研究方向

在二次元角色数据集构建与文本到图像生成领域，CyberHarem/hayashimo_kantaicollection数据集聚焦于《舰队Collection》中早霜角色的多模态表征学习。该数据集包含500张标注图像，通过自动爬取Danbooru、Pixiv等平台并经DeepGHS团队清洗，提供了raw、800、1200及三级裁剪等多种分辨率版本，适配不同训练需求。当前前沿方向集中于利用标签聚类（如制服、兔女郎、女仆装等服饰类别）进行角色概念分解，以提升生成模型对特定属性（如发色、表情、姿势）的细粒度控制能力。同时，该数据集在少样本个性化生成、风格迁移及角色一致性保持等热点研究中具有重要价值，为二次元文化中的虚拟角色数字化提供了标准化资源，推动了社区驱动的生成式AI在动漫领域的应用边界。