CyberHarem/iori_bluearchive

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

# Dataset of iori/銀鏡イオリ/伊织 (Blue Archive)

This is the dataset of iori/銀鏡イオリ/伊织 (Blue Archive), containing 500 images and their tags.

The core tags of this character are `grey_hair, long_hair, twintails, red_eyes, dark_skin, dark-skinned_female, pointy_ears, hair_over_one_eye, breasts, ribbon, halo, very_long_hair, hair_ribbon, tail, demon_tail, medium_breasts, sunglasses, eyewear_on_head`, 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 | 1021.17 MiB | [Download](https://huggingface.co/datasets/CyberHarem/iori_bluearchive/resolve/main/dataset-raw.zip) | Waifuc-Raw | Raw data with meta information (min edge aligned to 1400 if larger). |
| 1200 | 500 | 832.68 MiB | [Download](https://huggingface.co/datasets/CyberHarem/iori_bluearchive/resolve/main/dataset-1200.zip) | IMG+TXT | dataset with the shorter side not exceeding 1200 pixels. |
| stage3-p480-1200 | 1393 | 1.72 GiB | [Download](https://huggingface.co/datasets/CyberHarem/iori_bluearchive/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/iori_bluearchive',
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 | 17 |  |  |  |  |  | 1girl, black_bikini, black_choker, looking_at_viewer, multi-strapped_bikini_bottom, navel, official_alternate_costume, solo, stomach, blush, side-tie_bikini_bottom, blue_sky, day, outdoors, armpits, black_gloves, bare_shoulders, wet, small_breasts, closed_mouth, cloud, cowboy_shot, ocean, string_bikini, thigh_strap, arms_up, water |
| 1 | 5 |  |  |  |  |  | 1girl, armpits, arms_behind_head, arms_up, black_bikini, black_choker, blush, looking_at_viewer, multi-strapped_bikini_bottom, navel, official_alternate_costume, side-tie_bikini_bottom, simple_background, solo, stomach, string_bikini, white_background, closed_mouth, cowboy_shot, bare_shoulders, black_ribbon, thigh_strap, black_gloves, eyes_visible_through_hair, small_breasts, thighs |
| 2 | 14 |  |  |  |  |  | 1girl, black_bikini, black_choker, black_gloves, looking_at_viewer, multi-strapped_bikini_bottom, navel, official_alternate_costume, solo, bare_shoulders, side-tie_bikini_bottom, blush, hair_bow, stomach, thigh_strap, white_background, closed_mouth, collarbone, simple_background, thighs, small_breasts, adjusting_eyewear, black_ribbon, cowboy_shot, string_bikini, wet |
| 3 | 6 |  |  |  |  |  | 1girl, ass, bare_shoulders, black_bikini, black_gloves, looking_at_viewer, multi-strapped_bikini_bottom, official_alternate_costume, sideboob, solo, string_bikini, thigh_strap, thighs, wet, black_choker, black_ribbon, blue_sky, blush, day, from_behind, looking_back, outdoors, side-tie_bikini_bottom, small_breasts, water, cloud, halterneck, closed_mouth, hair_bow, open_mouth, v-shaped_eyebrows, wading |
| 4 | 5 |  |  |  |  |  | 1girl, ass, black_bikini, black_gloves, blush, from_behind, looking_at_viewer, looking_back, official_alternate_costume, solo, multi-strapped_bikini_bottom, on_stomach, sideboob, untied_bikini, clenched_teeth, choker, small_breasts, thigh_strap |
| 5 | 6 |  |  |  |  |  | 1boy, 1girl, blush, hetero, navel, nipples, penis, sex, solo_focus, spread_legs, sweat, vaginal, cum_in_pussy, cum_overflow, open_mouth, bar_censor, collarbone, completely_nude, looking_at_viewer, missionary, on_back, on_bed, small_breasts, arms_up, bed_sheet, female_pubic_hair, heart, pillow, pov, speech_bubble, stomach |
| 6 | 9 |  |  |  |  |  | blush, solo_focus, 1boy, 1girl, completely_nude, hetero, sweat, black_gloves, ass, white_background, anus, clenched_teeth, heart, on_stomach, censored, fingering, nipples, pussy_juice, small_breasts, trembling |
| 7 | 16 |  |  |  |  |  | 1girl, armband, black_necktie, black_skirt, holding_gun, long_sleeves, solo, white_shirt, bolt_action, looking_at_viewer, black_gloves, pleated_skirt, collared_shirt, black_ribbon, black_socks, simple_background, kneehighs, boots, full_body, white_background, black_footwear, high-waist_skirt, open_mouth |
| 8 | 26 |  |  |  |  |  | 1girl, black_necktie, black_skirt, long_sleeves, white_shirt, collared_shirt, looking_at_viewer, solo, black_gloves, blush, closed_mouth, pleated_skirt, black_ribbon, black_socks, simple_background, white_background, red_armband |
| 9 | 9 |  |  |  |  |  | blush, fake_animal_ears, looking_at_viewer, playboy_bunny, rabbit_ears, strapless_leotard, alternate_costume, bare_shoulders, detached_collar, 1girl, black_leotard, simple_background, small_breasts, white_background, covered_navel, solo, bowtie, wrist_cuffs, gloves, high_heels, highleg |

### Table Version

| # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | 1girl | black_bikini | black_choker | looking_at_viewer | multi-strapped_bikini_bottom | navel | official_alternate_costume | solo | stomach | blush | side-tie_bikini_bottom | blue_sky | day | outdoors | armpits | black_gloves | bare_shoulders | wet | small_breasts | closed_mouth | cloud | cowboy_shot | ocean | string_bikini | thigh_strap | arms_up | water | arms_behind_head | simple_background | white_background | black_ribbon | eyes_visible_through_hair | thighs | hair_bow | collarbone | adjusting_eyewear | ass | sideboob | from_behind | looking_back | halterneck | open_mouth | v-shaped_eyebrows | wading | on_stomach | untied_bikini | clenched_teeth | choker | 1boy | hetero | nipples | penis | sex | solo_focus | spread_legs | sweat | vaginal | cum_in_pussy | cum_overflow | bar_censor | completely_nude | missionary | on_back | on_bed | bed_sheet | female_pubic_hair | heart | pillow | pov | speech_bubble | anus | censored | fingering | pussy_juice | trembling | armband | black_necktie | black_skirt | holding_gun | long_sleeves | white_shirt | bolt_action | pleated_skirt | collared_shirt | black_socks | kneehighs | boots | full_body | black_footwear | high-waist_skirt | red_armband | fake_animal_ears | playboy_bunny | rabbit_ears | strapless_leotard | alternate_costume | detached_collar | black_leotard | covered_navel | bowtie | wrist_cuffs | gloves | high_heels | highleg |
|----:|----------:|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------|:---------------|:---------------|:--------------------|:-------------------------------|:--------|:-----------------------------|:-------|:----------|:--------|:-------------------------|:-----------|:------|:-----------|:----------|:---------------|:-----------------|:------|:----------------|:---------------|:--------|:--------------|:--------|:----------------|:--------------|:----------|:--------|:-------------------|:--------------------|:-------------------|:---------------|:----------------------------|:---------|:-----------|:-------------|:--------------------|:------|:-----------|:--------------|:---------------|:-------------|:-------------|:--------------------|:---------|:-------------|:----------------|:-----------------|:---------|:-------|:---------|:----------|:--------|:------|:-------------|:--------------|:--------|:----------|:---------------|:---------------|:-------------|:------------------|:-------------|:----------|:---------|:------------|:--------------------|:--------|:---------|:------|:----------------|:-------|:-----------|:------------|:--------------|:------------|:----------|:----------------|:--------------|:--------------|:---------------|:--------------|:--------------|:----------------|:-----------------|:--------------|:------------|:--------|:------------|:-----------------|:-------------------|:--------------|:-------------------|:----------------|:--------------|:--------------------|:--------------------|:------------------|:----------------|:----------------|:---------|:--------------|:---------|:-------------|:----------|
| 0 | 17 |  |  |  |  |  | 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 | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 1 | 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 | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 2 | 14 |  |  |  |  |  | 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 | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 3 | 6 |  |  |  |  |  | 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 | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 4 | 5 |  |  |  |  |  | X | X | | X | X | | X | X | | X | | | | | | X | | | X | | | | | | X | | | | | | | | | | | | X | X | X | X | | | | | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 5 | 6 |  |  |  |  |  | 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 | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 6 | 9 |  |  |  |  |  | X | | | | | | | | | X | | | | | | X | | | X | | | | | | | | | | | X | | | | | | | X | | | | | | | | X | | X | | X | X | X | | | X | | X | | | | | X | | | | | | X | | | | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 7 | 16 |  |  |  |  |  | X | | | X | | | | X | | | | | | | | X | | | | | | | | | | | | | X | X | X | | | | | | | | | | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | |
| 8 | 26 |  |  |  |  |  | X | | | X | | | | X | | X | | | | | | X | | | | X | | | | | | | | | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | X | X | | X | X | | X | X | X | | | | | | X | | | | | | | | | | | | | |
| 9 | 9 |  |  |  |  |  | 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：iori/銀鏡イオリ/伊织 (Blue Archive)

数据集基本信息

• 许可证: MIT
• 任务类别: 文本到图像
• 标签: 艺术, 不适合所有观众
• 大小类别: 小于1K

数据集内容

• 包含内容: 500张图像及其标签
• 核心标签: 灰发, 长发, 双马尾, 红眼, 深肤色, 深肤色女性, 尖耳朵, 头发遮住一只眼, 胸部, 丝带, 光环, 非常长发, 头发丝带, 尾巴, 恶魔尾巴, 中等胸部, 太阳镜, 头戴眼镜

数据集来源

• 图像来源: 从多个网站爬取，如danbooru, pixiv, zerochan等
• 爬虫系统: 由DeepGHS Team开发

数据集包列表

名称	图像数量	大小	下载链接	类型	描述
raw	500	1021.17 MiB	下载	Waifuc-Raw	原始数据，包含元信息（最小边对齐到1400像素，如果更大）
1200	500	832.68 MiB	下载	IMG+TXT	数据集，短边不超过1200像素
stage3-p480-1200	1393	1.72 GiB	下载	IMG+TXT	三阶段裁剪数据集，区域不小于480x480像素

数据集加载

• 加载工具: 使用waifuc加载原始数据集
• 加载代码示例: 提供Python代码示例，用于下载和提取数据集，并使用waifuc加载

数据集标签集群

• 集群列表: 提供标签集群的结果，可能包含可挖掘的服装信息
• 集群示例: 展示不同集群的样本图像及其标签，有助于理解数据集内容的多样性和特定主题的集中度

搜集汇总

数据集介绍

构建方式

在数字内容生成领域，高质量、结构化的图像数据集是驱动文本到图像模型性能提升的关键。CyberHarem/iori_bluearchive 数据集聚焦于游戏《碧蓝档案》中的角色“伊织”，系统性地收录了500张图像及其对应的标签信息。该数据集的构建依托于由 DeepGHS 团队开发的自动化爬取系统，从 Danbooru、Pixiv、Zerochan 等多个主流图像平台广泛采集原始素材。采集过程中，图像的最小边被对齐至1400像素以确保质量。随后，数据集被组织为多个版本：raw 版本保留原始元数据；1200 版本将图像短边限制在1200像素以内；stage3-p480-1200 版本则经过三级裁剪处理，确保每张图像的有效区域不小于480×480像素，从而适配不同训练需求。

使用方法

为便于研究者高效利用，该数据集提供了多种加载路径。用户可通过 Hugging Face Hub 直接下载预打包的 ZIP 文件，并根据任务选择 raw、1200 或 stage3-p480-1200 版本。对于希望深度集成工作流的用户，推荐使用 Waifuc 库加载 raw 数据集：首先通过 hf_hub_download 函数下载 dataset-raw.zip，解压至本地目录后，利用 waifuc.source.LocalSource 接口即可逐项访问图像及其对应的文件名与标签元数据。这种设计使得数据集能够无缝嵌入到基于 Python 的机器学习管线中，无论是用于训练扩散模型、进行图像风格分析，还是开展角色识别任务，均能快速启动。

背景与挑战

背景概述

在生成式人工智能与数字艺术交叉融合的前沿领域，高质量、精细标注的图像数据集成为驱动文本到图像（text-to-image）模型发展的核心基石。CyberHarem/iori_bluearchive数据集由DeepGHS团队于近期构建，聚焦于热门游戏《蔚蓝档案》中角色“伊织”（Iori）的形象数据。该数据集的核心研究问题在于如何系统性地收集、清洗并组织特定动漫角色的多风格图像，以支撑角色一致性生成与风格迁移等下游任务。通过整合来自Danbooru、Pixiv、Zerochan等多平台的500张图像及其语义标签，该数据集为社区提供了标准化的训练资源，对角色定制化图像生成领域产生了积极的推动作用，尤其为小样本学习与精细标签建模提供了宝贵的数据基础。

当前挑战

该数据集面临的核心挑战首先源于领域问题的复杂性：文本到图像任务要求模型能从稀疏的文本描述中精准还原角色的核心视觉特征（如灰发、红瞳、暗色皮肤等），但动漫角色在不同画师笔下存在风格变异，对模型泛化能力构成严峻考验。其次，在构建过程中，数据收集面临多源异构图像的版权与质量统一难题，自动爬取系统需要应对不同网站的访问限制与格式差异。此外，标签的自动标注与去重（如核心标签的剪枝）容易出现语义歧义，而图像裁剪（如stage3-p480-1200）需在保持内容完整性前提下满足尺寸约束，这些技术细节均对数据集的实用性与一致性提出了更高要求。

常用场景

经典使用场景

在文本到图像生成领域，CyberHarem/iori_bluearchive 数据集为基于扩散模型的二次元角色定制化生成提供了高质量的训练素材。该数据集包含500张经过精细标注的角色图像，涵盖多种姿态、服饰与场景，尤其适用于训练能够根据文本描述精准生成特定角色形象的模型，如Stable Diffusion的微调版本。研究者常利用其丰富的标签信息进行多模态对齐学习，以提升生成图像在角色特征保持上的忠实度。

解决学术问题

该数据集有效缓解了动漫角色生成研究中数据稀缺与标注不一致的难题。通过系统化爬取与清洗，它提供了统一的图像分辨率与标签体系，支持对角色核心属性（如发色、瞳色、服饰）的细粒度控制。这为探索条件生成模型中的属性解耦与组合泛化能力提供了基准，推动了面向虚拟角色的人物图像生成从单一模板向多样化、可编辑方向的发展。

实际应用

在实际应用中，该数据集被广泛用于游戏角色立绘的自动化生成、虚拟主播形象的快速定制以及同人创作辅助工具的开发。开发者可基于此数据集训练轻量化模型，实现用户通过自然语言描述即可生成符合特定风格的角色头像或全身图。此外，其提供的多尺度裁剪版本（如1200像素与阶段裁剪）适配了不同计算资源下的部署需求，降低了产业落地的门槛。

数据集最近研究

最新研究方向

在二次元角色图像生成领域，基于高质量、精细标注的角色数据集进行文本到图像（Text-to-Image）的微调训练已成为前沿热点。CyberHarem/iori_bluearchive数据集聚焦于《蔚蓝档案》中人气角色伊织，通过从Danbooru、Pixiv等多源平台自动化爬取并清洗500张图像，辅以核心标签（如灰发、长双马尾、深色皮肤、恶魔尾等）的精准剪枝与多分辨率版本（raw、1200、stage3裁剪）的标准化处理，为社区提供了可直接用于LoRA、DreamBooth等轻量化微调方法的素材。该数据集不仅支持waifuc框架的便捷加载，还通过标签聚类分析（如泳装、战斗服、兔女郎等风格簇）揭示了角色在不同装扮下的视觉语义分布，推动了可控角色生成、风格迁移及多模态对齐等研究。其开源与结构化设计，为二次元AI创作生态中角色一致性的保持、少样本学习能力的提升以及个性化生成模型的迭代提供了关键支撑，呼应了当前生成式AI在娱乐与创意产业中落地应用的热潮。