遇见数据集

CyberHarem/shez_fireemblem

收藏
Hugging Face2024-01-17 更新2024-03-04 收录
官方服务:

资源简介:

--- license: mit task_categories: - text-to-image tags: - art - not-for-all-audiences size_categories: - n<1K --- # Dataset of shez (Fire Emblem) This is the dataset of shez (Fire Emblem), containing 393 images and their tags. The core tags of this character are `purple_hair, hair_over_one_eye, purple_eyes, long_hair, breasts, bangs, hair_bun, large_breasts, single_hair_bun`, 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 | 393 | 589.63 MiB | [Download](https://huggingface.co/datasets/CyberHarem/shez_fireemblem/resolve/main/dataset-raw.zip) | Waifuc-Raw | Raw data with meta information (min edge aligned to 1400 if larger). | | 800 | 393 | 317.47 MiB | [Download](https://huggingface.co/datasets/CyberHarem/shez_fireemblem/resolve/main/dataset-800.zip) | IMG+TXT | dataset with the shorter side not exceeding 800 pixels. | | stage3-p480-800 | 926 | 660.82 MiB | [Download](https://huggingface.co/datasets/CyberHarem/shez_fireemblem/resolve/main/dataset-stage3-p480-800.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. | | 1200 | 393 | 513.39 MiB | [Download](https://huggingface.co/datasets/CyberHarem/shez_fireemblem/resolve/main/dataset-1200.zip) | IMG+TXT | dataset with the shorter side not exceeding 1200 pixels. | | stage3-p480-1200 | 926 | 979.72 MiB | [Download](https://huggingface.co/datasets/CyberHarem/shez_fireemblem/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/shez_fireemblem', 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 | 24 | ![](samples/0/clu0-sample0.png) | ![](samples/0/clu0-sample1.png) | ![](samples/0/clu0-sample2.png) | ![](samples/0/clu0-sample3.png) | ![](samples/0/clu0-sample4.png) | 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, gloves, looking_at_viewer, simple_background, smile, solo | | 1 | 7 | ![](samples/1/clu1-sample0.png) | ![](samples/1/clu1-sample1.png) | ![](samples/1/clu1-sample2.png) | ![](samples/1/clu1-sample3.png) | ![](samples/1/clu1-sample4.png) | 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, gloves, looking_at_viewer, simple_background, solo | | 2 | 5 | ![](samples/2/clu2-sample0.png) | ![](samples/2/clu2-sample1.png) | ![](samples/2/clu2-sample2.png) | ![](samples/2/clu2-sample3.png) | ![](samples/2/clu2-sample4.png) | 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, looking_at_viewer, simple_background, smile, solo | | 3 | 10 | ![](samples/3/clu3-sample0.png) | ![](samples/3/clu3-sample1.png) | ![](samples/3/clu3-sample2.png) | ![](samples/3/clu3-sample3.png) | ![](samples/3/clu3-sample4.png) | 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, looking_at_viewer, simple_background, solo | | 4 | 14 | ![](samples/4/clu4-sample0.png) | ![](samples/4/clu4-sample1.png) | ![](samples/4/clu4-sample2.png) | ![](samples/4/clu4-sample3.png) | ![](samples/4/clu4-sample4.png) | 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, looking_at_viewer, simple_background, solo, sword, gloves, holding | | 5 | 6 | ![](samples/5/clu5-sample0.png) | ![](samples/5/clu5-sample1.png) | ![](samples/5/clu5-sample2.png) | ![](samples/5/clu5-sample3.png) | ![](samples/5/clu5-sample4.png) | 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, dual_wielding, looking_at_viewer, simple_background, solo, gloves, holding_sword | | 6 | 5 | ![](samples/6/clu6-sample0.png) | ![](samples/6/clu6-sample1.png) | ![](samples/6/clu6-sample2.png) | ![](samples/6/clu6-sample3.png) | ![](samples/6/clu6-sample4.png) | 2girls, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, simple_background, armor, gloves, medium_hair, smile, looking_at_viewer, closed_eyes | | 7 | 5 | ![](samples/7/clu7-sample0.png) | ![](samples/7/clu7-sample1.png) | ![](samples/7/clu7-sample2.png) | ![](samples/7/clu7-sample3.png) | ![](samples/7/clu7-sample4.png) | 1girl, black_bikini, choker, cleavage, hair_flower, navel, official_alternate_costume, smile, solo, closed_mouth, looking_at_viewer, simple_background, bare_shoulders, white_background | | 8 | 6 | ![](samples/8/clu8-sample0.png) | ![](samples/8/clu8-sample1.png) | ![](samples/8/clu8-sample2.png) | ![](samples/8/clu8-sample3.png) | ![](samples/8/clu8-sample4.png) | 1girl, black_bikini, choker, cleavage, fingerless_gloves, hair_flower, looking_at_viewer, navel, official_alternate_costume, solo, smile | | 9 | 11 | ![](samples/9/clu9-sample0.png) | ![](samples/9/clu9-sample1.png) | ![](samples/9/clu9-sample2.png) | ![](samples/9/clu9-sample3.png) | ![](samples/9/clu9-sample4.png) | 1girl, hetero, nipples, penis, choker, sex, 1boy, armor, asymmetrical_clothes, blush, cum_in_pussy, spread_legs, vaginal, looking_at_viewer, mosaic_censoring, gloves, rape, breasts_out, cleavage, cum_on_breasts, cum_on_hair, facial, navel, open_mouth, thighhighs | ### Table Version | # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | 1girl | armor | asymmetrical_clothes | cape | choker | cleavage | closed_mouth | gloves | looking_at_viewer | simple_background | smile | solo | sword | holding | dual_wielding | holding_sword | 2girls | medium_hair | closed_eyes | black_bikini | hair_flower | navel | official_alternate_costume | bare_shoulders | white_background | fingerless_gloves | hetero | nipples | penis | sex | 1boy | blush | cum_in_pussy | spread_legs | vaginal | mosaic_censoring | rape | breasts_out | cum_on_breasts | cum_on_hair | facial | open_mouth | thighhighs | |----:|----------:|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------|:--------|:-----------------------|:-------|:---------|:-----------|:---------------|:---------|:--------------------|:--------------------|:--------|:-------|:--------|:----------|:----------------|:----------------|:---------|:--------------|:--------------|:---------------|:--------------|:--------|:-----------------------------|:-----------------|:-------------------|:--------------------|:---------|:----------|:--------|:------|:-------|:--------|:---------------|:--------------|:----------|:-------------------|:-------|:--------------|:-----------------|:--------------|:---------|:-------------|:-------------| | 0 | 24 | ![](samples/0/clu0-sample0.png) | ![](samples/0/clu0-sample1.png) | ![](samples/0/clu0-sample2.png) | ![](samples/0/clu0-sample3.png) | ![](samples/0/clu0-sample4.png) | X | X | X | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 1 | 7 | ![](samples/1/clu1-sample0.png) | ![](samples/1/clu1-sample1.png) | ![](samples/1/clu1-sample2.png) | ![](samples/1/clu1-sample3.png) | ![](samples/1/clu1-sample4.png) | X | X | X | X | X | X | X | X | X | X | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 2 | 5 | ![](samples/2/clu2-sample0.png) | ![](samples/2/clu2-sample1.png) | ![](samples/2/clu2-sample2.png) | ![](samples/2/clu2-sample3.png) | ![](samples/2/clu2-sample4.png) | X | X | X | X | X | X | X | | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 3 | 10 | ![](samples/3/clu3-sample0.png) | ![](samples/3/clu3-sample1.png) | ![](samples/3/clu3-sample2.png) | ![](samples/3/clu3-sample3.png) | ![](samples/3/clu3-sample4.png) | X | X | X | X | X | X | X | | X | X | | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4 | 14 | ![](samples/4/clu4-sample0.png) | ![](samples/4/clu4-sample1.png) | ![](samples/4/clu4-sample2.png) | ![](samples/4/clu4-sample3.png) | ![](samples/4/clu4-sample4.png) | X | X | X | X | X | X | X | X | X | X | | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 5 | 6 | ![](samples/5/clu5-sample0.png) | ![](samples/5/clu5-sample1.png) | ![](samples/5/clu5-sample2.png) | ![](samples/5/clu5-sample3.png) | ![](samples/5/clu5-sample4.png) | X | X | X | X | X | X | X | X | X | X | | X | | | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 6 | 5 | ![](samples/6/clu6-sample0.png) | ![](samples/6/clu6-sample1.png) | ![](samples/6/clu6-sample2.png) | ![](samples/6/clu6-sample3.png) | ![](samples/6/clu6-sample4.png) | | X | X | X | X | X | X | X | X | X | X | | | | | | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | 7 | 5 | ![](samples/7/clu7-sample0.png) | ![](samples/7/clu7-sample1.png) | ![](samples/7/clu7-sample2.png) | ![](samples/7/clu7-sample3.png) | ![](samples/7/clu7-sample4.png) | X | | | | X | X | X | | X | X | X | X | | | | | | | | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | 8 | 6 | ![](samples/8/clu8-sample0.png) | ![](samples/8/clu8-sample1.png) | ![](samples/8/clu8-sample2.png) | ![](samples/8/clu8-sample3.png) | ![](samples/8/clu8-sample4.png) | X | | | | X | X | | | X | | X | X | | | | | | | | X | X | X | X | | | X | | | | | | | | | | | | | | | | | | | 9 | 11 | ![](samples/9/clu9-sample0.png) | ![](samples/9/clu9-sample1.png) | ![](samples/9/clu9-sample2.png) | ![](samples/9/clu9-sample3.png) | ![](samples/9/clu9-sample4.png) | 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 |

提供机构:
CyberHarem
原始信息汇总

数据集概述

基本信息

  • 数据集名称: Dataset of shez (Fire Emblem)
  • 许可证: MIT
  • 任务类别: text-to-image
  • 标签: art, not-for-all-audiences
  • 数据量: n<1K

数据内容

  • 图像数量: 393张
  • 核心标签: purple_hair, hair_over_one_eye, purple_eyes, long_hair, breasts, bangs, hair_bun, large_breasts, single_hair_bun

数据包列表

名称 图像数量 大小 类型 描述
raw 393 589.63 MiB Waifuc-Raw 包含元信息的原始数据(最小边对齐到1400像素,如果更大)
800 393 317.47 MiB IMG+TXT 短边不超过800像素的数据集
stage3-p480-800 926 660.82 MiB IMG+TXT 3阶段裁剪数据集,区域不小于480x480像素
1200 393 513.39 MiB IMG+TXT 短边不超过1200像素的数据集
stage3-p480-1200 926 979.72 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1 图像2 图像3 图像4 图像5 标签
0 24 样本0 样本1 样本2 样本3 样本4 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, gloves, looking_at_viewer, simple_background, smile, solo
1 7 样本0 样本1 样本2 样本3 样本4 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, gloves, looking_at_viewer, simple_background, solo
2 5 样本0 样本1 样本2 样本3 样本4 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, looking_at_viewer, simple_background, smile, solo
3 10 样本0 样本1 样本2 样本3 样本4 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, looking_at_viewer, simple_background, solo
4 14 样本0 样本1 样本2 样本3 样本4 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, looking_at_viewer, simple_background, solo, sword, gloves, holding
5 6 样本0 样本1 样本2 样本3 样本4 1girl, armor, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, dual_wielding, looking_at_viewer, simple_background, solo, gloves, holding_sword
6 5 样本0 样本1 样本2 样本3 样本4 2girls, asymmetrical_clothes, cape, choker, cleavage, closed_mouth, simple_background, armor, gloves, medium_hair, smile, looking_at_viewer, closed_eyes
7 5 样本0 样本1 样本2 样本3 样本4 1girl, black_bikini, choker, cleavage, hair_flower, navel, official_alternate_costume, smile, solo, closed_mouth, looking_at_viewer, simple_background, bare_shoulders, white_background
8 6 样本0 样本1 样本2 样本3 样本4 1girl, black_bikini, choker, cleavage, fingerless_gloves, hair_flower, looking_at_viewer, navel, official_alternate_costume, solo, smile
9 11 样本0 样本1 样本2 样本3 样本4 1girl, hetero, nipples, penis, choker, sex, 1boy, armor, asymmetrical_clothes, blush, cum_in_pussy, spread_legs, vaginal, looking_at_viewer, mosaic_censoring, gloves, rape, breasts_out, cleavage, cum_on_breasts, cum_on_hair, facial, navel, open_mouth, thighhighs

表格版本

# 样本数量 图像1 图像2 图像3 图像4 图像5 1girl armor asymmetrical_clothes cape choker cleavage closed_mouth gloves looking_at_viewer simple_background smile solo sword holding dual_wielding holding_sword 2girls medium_hair closed_eyes black_bikini hair_flower navel official_alternate_costume bare_shoulders white_background fingerless_gloves hetero nipples penis sex 1boy blush cum_in_pussy spread_legs vaginal mosaic_censoring rape breasts_out cum_on_breasts cum_on_hair facial open_mouth thighhighs
0 24 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X X
1 7 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X
2 5 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X
3 10 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X
4 14 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X X X
5 6 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X X
6 5 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X X X
7 5 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X X X X
8 6 样本0 样本1 样本2 样本3 样本4 X X X X X X X X X X X
9 11 样本0 样本1 [样本2](samples
搜集汇总
数据集介绍
CyberHarem/shez_fireemblem 数据集图片
构建方式
在数字艺术与游戏同人创作领域,高质量的角色图像数据集对于文本到图像生成模型的训练至关重要。本数据集聚焦于《火焰之纹章》系列中的角色谢兹(Shez),通过自动化爬取系统从Danbooru、Pixiv、Zerochan等多个知名图像平台收集原始素材,共收录393张图像及其对应标签。数据采集由DeepGHS团队开发的工具驱动,确保了来源的多样性与广度。数据集提供了多种处理版本,包括原始未裁剪版本(raw)、短边不超过800或1200像素的缩放版本,以及经过三阶段裁剪、确保区域不小于480x480像素的增强版本,以适应不同训练需求。核心标签如紫色头发、遮眼发、紫色眼睛等已被精简,便于聚焦角色特征。
使用方法
使用本数据集时,研究者可根据训练目标选择合适的版本。对于需要原始图像与完整元信息的场景,推荐下载raw版本,并通过Waifuc库的LocalSource加载,示例代码展示了如何解压并逐项访问图像及标签。若关注计算效率,可选择800或1200像素的缩放版本,它们平衡了分辨率与存储需求。对于需要聚焦角色主体、减少背景干扰的任务,三阶段裁剪版本(stage3-p480-800或stage3-p480-1200)更为合适,其通过智能裁剪保留了关键区域。所有版本均以IMAG+TXT格式组织,可直接用于多数图像生成框架。用户可通过Hugging Face Hub直接下载对应ZIP文件,或利用提供的Python代码实现自动化加载与处理。
背景与挑战
背景概述
在文本到图像生成领域,高质量、细粒度的角色数据集对于提升模型对特定角色特征的捕捉与再现能力至关重要。CyberHarem/shez_fireemblem数据集由DeepGHS团队于近期构建,聚焦于《火焰之纹章》系列中的角色“谢兹”。该数据集包含393张图像及其对应的标签,核心标签包括紫色头发、遮眼发、紫色眼睛等,旨在为动漫角色生成任务提供标准化训练素材。通过从Danbooru、Pixiv、Zerochan等多个平台自动爬取并精心筛选,该数据集不仅丰富了角色视觉特征的多样性,还通过标签聚类分析揭示了不同着装与姿态模式,为后续角色定制化生成研究奠定了数据基础。其影响力体现在为开源社区提供了一个可直接用于微调扩散模型的角色专用数据集,推动了二次元角色生成领域的发展。
当前挑战
该数据集面临的核心挑战集中在领域问题解决与构建过程两方面。在领域层面,文本到图像生成模型在处理特定动漫角色时,常因数据稀疏性导致特征混淆或细节丢失,例如谢兹的“遮眼发”与“紫色眼睛”等细微特征难以被精准还原,且不同着装(如铠甲与泳装)下的角色一致性保持困难。在构建过程中,自动爬取系统面临多源数据质量参差不齐的难题,包括图像分辨率差异、标签噪声(如部分图像包含无关背景或多人场景)以及版权合规性审查。此外,数据集的规模(<1K)限制了模型对角色复杂姿态与交互场景的学习能力,而标签聚类中出现的敏感内容(如第9簇)进一步增加了数据筛选与伦理审查的复杂度。
常用场景
经典使用场景
该数据集聚焦于《火焰之纹章》系列中“谢兹”这一角色的视觉素材,收录了393张经过精细标注的图片,涵盖其标志性的紫色长发、单侧遮眼、紫色眼眸等核心特征。在文生图(text-to-image)领域,它常被用于训练和评估基于扩散模型或生成对抗网络的角色定制化生成任务,研究者可借助其丰富的标签体系(如铠甲、比基尼、双持武器等变体)探索条件控制生成、风格迁移以及多模态对齐等经典问题。数据集的多种分辨率版本(如800px、1200px)及裁剪策略(stage3-p480-800)进一步适配了不同模型架构的输入需求,使其成为二次元角色生成研究的基准资源之一。
解决学术问题
该数据集的核心学术价值在于解决了动漫角色图像生成中精细属性解耦与多模态一致性的难题。通过提供结构化标签(如服饰类型、动作姿态、场景背景),它使研究者能够量化分析生成模型对特定视觉概念的捕捉能力,例如区分“盔甲”与“比基尼”等风格化衣着的表征差异。同时,其标注的“双持武器”“闭眼微笑”等复合标签为评估模型在复杂语义组合下的鲁棒性提供了测试床。此外,数据集中包含的聚类结果(如基于标签共现的分组)可辅助研究无监督特征学习与角色身份保持机制,推动少样本生成、可控图像编辑等前沿方向的进展。
实际应用
在实际应用中,该数据集为游戏开发、虚拟偶像制作及创意设计领域提供了高效的角色资产复用方案。开发者可基于此数据集微调预训练的文生图模型,快速生成谢兹在不同服饰(如官方换装)、武器配置或互动场景下的概念图,显著降低手绘成本。其多分辨率版本(如480p裁剪版)直接适配移动端实时渲染管线,支持游戏内动态立绘生成。在二次元社区中,该资源亦被用于个性化角色创作工具,用户可通过输入文本描述(如“泳装、持剑、微笑”)直接获得符合角色原设的视觉输出,打通了从文本创意到视觉落地的自动化流程。
数据集最近研究
最新研究方向
在文本到图像生成领域,尤其是针对动漫游戏角色如《火焰纹章》中Shez的精细化建模,该数据集通过多尺度裁剪与标签聚类技术,为角色一致性生成提供了高质量训练资源。当前前沿研究方向聚焦于利用此类标注精细的二次元角色数据集,结合扩散模型实现零样本或少样本的角色定制化生成,例如在保持核心特征(如紫色长发、单眼刘海等)的前提下进行姿态、服饰或场景的灵活变换。与此相关的热点事件包括AI绘画社区对特定角色版权与风格迁移的争议,以及HuggingFace等平台推动的开放数据集生态建设,这些努力正加速角色生成技术从泛化模型向垂直领域落地,对于游戏资产自动化生产与虚拟偶像产业具有深远影响。
以上内容由遇见数据集搜集并总结生成
二维码
社区交流群
二维码
科研交流群
商业服务