CyberHarem/kousaka_umi_theidolmstermillionlive
收藏Hugging Face2024-01-16 更新2024-03-04 收录
下载链接:
https://hf-mirror.com/datasets/CyberHarem/kousaka_umi_theidolmstermillionlive
下载链接
链接失效反馈官方服务:
资源简介:
---
license: mit
task_categories:
- text-to-image
tags:
- art
- not-for-all-audiences
size_categories:
- n<1K
---
# Dataset of kousaka_umi/高坂海美 (THE iDOLM@STER: Million Live!)
This is the dataset of kousaka_umi/高坂海美 (THE iDOLM@STER: Million Live!), containing 268 images and their tags.
The core tags of this character are `long_hair, brown_hair, braid, blue_eyes, twin_braids, bangs, breasts, medium_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 | 268 | 298.27 MiB | [Download](https://huggingface.co/datasets/CyberHarem/kousaka_umi_theidolmstermillionlive/resolve/main/dataset-raw.zip) | Waifuc-Raw | Raw data with meta information (min edge aligned to 1400 if larger). |
| 800 | 268 | 195.22 MiB | [Download](https://huggingface.co/datasets/CyberHarem/kousaka_umi_theidolmstermillionlive/resolve/main/dataset-800.zip) | IMG+TXT | dataset with the shorter side not exceeding 800 pixels. |
| stage3-p480-800 | 620 | 398.30 MiB | [Download](https://huggingface.co/datasets/CyberHarem/kousaka_umi_theidolmstermillionlive/resolve/main/dataset-stage3-p480-800.zip) | IMG+TXT | 3-stage cropped dataset with the area not less than 480x480 pixels. |
| 1200 | 268 | 275.03 MiB | [Download](https://huggingface.co/datasets/CyberHarem/kousaka_umi_theidolmstermillionlive/resolve/main/dataset-1200.zip) | IMG+TXT | dataset with the shorter side not exceeding 1200 pixels. |
| stage3-p480-1200 | 620 | 520.28 MiB | [Download](https://huggingface.co/datasets/CyberHarem/kousaka_umi_theidolmstermillionlive/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/kousaka_umi_theidolmstermillionlive',
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, looking_at_viewer, solo, aqua_eyes, open_mouth, :d, bracelet, necklace, navel |
| 1 | 9 |  |  |  |  |  | 1girl, smile, striped_bikini, blush, cleavage, looking_at_viewer, navel, open_mouth, solo, collarbone, white_background, front-tie_top, simple_background |
| 2 | 10 |  |  |  |  |  | day, smile, 1girl, navel, solo, cleavage, outdoors, looking_at_viewer, open_mouth, beach, blue_sky, bracelet, cloud, cowboy_shot, palm_tree, side-tie_bikini_bottom, standing |
| 3 | 14 |  |  |  |  |  | 1girl, solo, open_mouth, white_shirt, looking_at_viewer, short_sleeves, blush, short_shorts, midriff, navel, tied_shirt, :d, belt, blue_shorts, collarbone, hair_between_eyes, side_braid, denim_shorts, standing, white_background, simple_background, striped_shirt, crop_top, holding, very_long_hair |
| 4 | 7 |  |  |  |  |  | looking_at_viewer, open_mouth, 1girl, beret, red_neckerchief, solo, white_headwear, white_shirt, blue_skirt, red_bow, sleeveless_shirt, standing, white_sailor_collar, cowboy_shot, hair_bow, holding, miniskirt, pleated_skirt, short_shorts, shorts_under_skirt, white_gloves, :d, ;d, blue_sky, cloud, day, hat_bow, one_eye_closed, white_shorts |
### Table Version
| # | Samples | Img-1 | Img-2 | Img-3 | Img-4 | Img-5 | 1girl | looking_at_viewer | solo | aqua_eyes | open_mouth | :d | bracelet | necklace | navel | smile | striped_bikini | blush | cleavage | collarbone | white_background | front-tie_top | simple_background | day | outdoors | beach | blue_sky | cloud | cowboy_shot | palm_tree | side-tie_bikini_bottom | standing | white_shirt | short_sleeves | short_shorts | midriff | tied_shirt | belt | blue_shorts | hair_between_eyes | side_braid | denim_shorts | striped_shirt | crop_top | holding | very_long_hair | beret | red_neckerchief | white_headwear | blue_skirt | red_bow | sleeveless_shirt | white_sailor_collar | hair_bow | miniskirt | pleated_skirt | shorts_under_skirt | white_gloves | ;d | hat_bow | one_eye_closed | white_shorts |
|----:|----------:|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------------------------------|:--------|:--------------------|:-------|:------------|:-------------|:-----|:-----------|:-----------|:--------|:--------|:-----------------|:--------|:-----------|:-------------|:-------------------|:----------------|:--------------------|:------|:-----------|:--------|:-----------|:--------|:--------------|:------------|:-------------------------|:-----------|:--------------|:----------------|:---------------|:----------|:-------------|:-------|:--------------|:--------------------|:-------------|:---------------|:----------------|:-----------|:----------|:-----------------|:--------|:------------------|:-----------------|:-------------|:----------|:-------------------|:----------------------|:-----------|:------------|:----------------|:---------------------|:---------------|:-----|:----------|:-----------------|:---------------|
| 0 | 14 |  |  |  |  |  | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 1 | 9 |  |  |  |  |  | X | X | X | | X | | | | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 2 | 10 |  |  |  |  |  | X | X | X | | X | | X | | X | X | | | X | | | | | X | X | X | X | X | X | X | X | X | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
| 3 | 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 | | | | | | | | | | | | | | | | |
| 4 | 7 |  |  |  |  |  | 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 |
提供机构:
CyberHarem
原始信息汇总
数据集概述
数据集信息
- 名称: kousaka_umi/高坂海美 (THE iDOLM@STER: Million Live!)
- 描述: 包含268张图片及其标签。
- 核心标签: long_hair, brown_hair, braid, blue_eyes, twin_braids, bangs, breasts, medium_breasts
- 标签分类: art, not-for-all-audiences
- 大小分类: n<1K
- 许可证: MIT
数据集包列表
| 名称 | 图片数量 | 大小 | 类型 | 描述 |
|---|---|---|---|---|
| raw | 268 | 298.27 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 | 268 | 195.22 MiB | IMG+TXT | 短边不超过800像素的数据集。 |
| stage3-p480-800 | 620 | 398.30 MiB | IMG+TXT | 3阶段裁剪数据集,区域不小于480x480像素。 |
| 1200 | 268 | 275.03 MiB | IMG+TXT | 短边不超过1200像素的数据集。 |
| stage3-p480-1200 | 620 | 520.28 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 | 图片示例 | 标签 |
|---|---|---|---|
| 0 | 14 | ![]() |
1girl, looking_at_viewer, solo, aqua_eyes, open_mouth, :d, bracelet, necklace, navel |
| 1 | 9 | ![]() |
1girl, smile, striped_bikini, blush, cleavage, looking_at_viewer, navel, open_mouth, solo, collarbone, white_background, front-tie_top, simple_background |
| 2 | 10 | ![]() |
day, smile, 1girl, navel, solo, cleavage, outdoors, looking_at_viewer, open_mouth, beach, blue_sky, bracelet, cloud, cowboy_shot, palm_tree, side-tie_bikini_bottom, standing |
| 3 | 14 | ![]() |
1girl, solo, open_mouth, white_shirt, looking_at_viewer, short_sleeves, blush, short_shorts, midriff, navel, tied_shirt, :d, belt, blue_shorts, collarbone, hair_between_eyes, side_braid, denim_shorts, standing, white_background, simple_background, striped_shirt, crop_top, holding, very_long_hair |
| 4 | 7 | ![]() |
looking_at_viewer, open_mouth, 1girl, beret, red_neckerchief, solo, white_headwear, white_shirt, blue_skirt, red_bow, sleeveless_shirt, standing, white_sailor_collar, cowboy_shot, hair_bow, holding, miniskirt, pleated_skirt, short_shorts, shorts_under_skirt, white_gloves, :d, ;d, blue_sky, cloud, day, hat_bow, one_eye_closed, white_shorts |
表格版本
| # | 样本数量 | 图片示例 | 1girl | looking_at_viewer | solo | aqua_eyes | open_mouth | :d | bracelet | necklace | navel | smile | striped_bikini | blush | cleavage | collarbone | white_background | front-tie_top | simple_background | day | outdoors | beach | blue_sky | cloud | cowboy_shot | palm_tree | side-tie_bikini_bottom | standing | white_shirt | short_sleeves | short_shorts | midriff | tied_shirt | belt | blue_shorts | hair_between_eyes | side_braid | denim_shorts | striped_shirt | crop_top | holding | very_long_hair | beret | red_neckerchief | white_headwear | blue_skirt | red_bow | sleeveless_shirt | white_sailor_collar | hair_bow | miniskirt | pleated_skirt | shorts_under_skirt | white_gloves | ;d | hat_bow | one_eye_closed | white_shorts |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 14 | ![]() |
X | X | X | X | X | X | X | X | X | |||||||||||||||||||||||||||||||||||||||||||||||
| 1 | 9 | ![]() |
X | X | X | X | X | X | X | X | X | X | X | X | X | |||||||||||||||||||||||||||||||||||||||||||
| 2 | 10 | ![]() |
X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | X | |||||||||||||||||||||||||||||||||||||||
| 3 | 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 | |||||||||||||||||||||||||||||||
| 4 | 7 | ![]() |
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 |
搜集汇总
数据集介绍

构建方式
在动漫角色数据集构建领域,针对《偶像大师:百万现场!》中的高坂海美角色,本数据集通过自动化爬取系统从Danbooru、Pixiv、Zerochan等多个图像平台采集了268张原始图像及其标签。数据采集由DeepGHS团队开发的Waifuc框架驱动,确保了来源的广泛性与效率。原始图像经过短边对齐至1400像素的预处理后,进一步衍生出多种规格的打包版本,包括短边不超过800像素和1200像素的常规缩放数据集,以及基于三阶段裁剪策略生成且面积不低于480×480像素的增强数据集,共计620张图像。核心标签如长褐发、编辫、蓝瞳等已被精简,以聚焦角色特征。
特点
该数据集的核心特色在于其多层次的规格设计与结构化组织。除了提供原始元数据版本外,还推出了IMG+TXT格式的标准化数据集,便于直接用于文本到图像生成模型的训练。尤为突出的是聚类分析结果的集成,通过标签聚类将图像划分为五个主题簇,每个簇包含5个样本预览图及关联标签(如泳装、校园风等),这为挖掘角色在不同场景下的服饰与姿态模式提供了直观的参考。此外,数据集的规模虽小(不足1K图像),但通过裁剪增强策略提升了样本多样性,兼顾了存储效率与训练质量。
使用方法
本数据集支持灵活的使用路径。对于需要原始元数据的用户,可通过Waifuc框架加载Raw版本:首先利用HuggingFace Hub下载dataset-raw.zip压缩包,解压至本地目录后,借助LocalSource接口即可遍历图像及其元数据(包括文件名和标签)。对于追求便捷性的用户,可直接下载IMG+TXT格式的打包文件(如800或1200版本),这些文件已包含对齐后的图像与对应标签文本,可直接导入常见的图像生成训练管线。同时,聚类结果以表格和文本形式提供,便于用户按主题筛选或进行风格迁移研究。
背景与挑战
背景概述
在文本到图像生成领域日益蓬勃发展的当下,高质量、标注精细的角色专属数据集成为驱动模型精准刻画特定虚拟形象的关键要素。由DeepGHS团队于近期构建并发布的CyberHarem/kousaka_umi_theidolmstermillionlive数据集,聚焦于《偶像大师:百万现场!》中的角色高坂海美,收录了268张图像及其对应的标签信息。该数据集旨在为二次元角色定制化生成任务提供标准化训练素材,其核心标签如长棕发、编辫、蓝瞳等精准捕捉了角色的视觉特征,为相关学术研究与工业应用奠定了数据基础。数据集通过自动化爬取系统从Danbooru、Pixiv等多源平台采集,并经由Waifuc工具提供多分辨率版本,显著降低了研究者获取优质动漫角色数据的门槛,对推动角色一致性生成、风格迁移等方向的发展具有重要参考价值。
当前挑战
该数据集所面临的挑战首先体现在领域问题层面:文本到图像生成任务要求模型在理解复杂语义的同时,保持角色身份特征的高度一致性与细节准确性,而当前数据集仅含268张样本,规模有限,难以覆盖角色在不同场景、服饰与姿态下的多样性,容易导致生成结果出现特征混淆或模式坍塌。其次,构建过程中遭遇的挑战不容忽视:多源图像爬取虽提高了数据丰富度,却也引入了版权归属模糊、图像质量参差不齐及标签噪声等问题;自动标注系统难以完美处理遮挡、光照变化及跨画风差异,部分标签可能缺失或与图像内容不匹配。此外,数据集的道德合规性标注为“not-for-all-audiences”,暗示其中包含敏感内容,这进一步限制了其在公开研究中的广泛使用与合规分发。
常用场景
经典使用场景
在二次元角色图像生成领域,该数据集为高坂海美这一特定角色的文本到图像(text-to-image)任务提供了高质量的素材。研究者可基于其包含的268张图像及精心筛选的标签(如长发、棕发、双马尾等核心特征),训练或微调扩散模型与生成对抗网络,实现对该角色形象的高保真复现与风格迁移。数据集提供的多分辨率版本(如800px、1200px)与裁剪版本,可适应不同模型输入尺寸需求,成为角色定制化图像生成研究的理想基准。
解决学术问题
该数据集有效解决了动漫角色细粒度图像生成中数据稀缺与标注不统一的学术难题。通过系统化爬取多源图像并统一标签体系,为角色一致性生成、属性解耦学习等研究提供了标准化训练语料。其标签聚类结果(如泳装、校园制服等服饰组合)支持了多模态特征对齐与概念分解的探索,推动了从通用图像生成向特定角色身份保持的技术演进,在少样本学习与个性化生成领域具有方法论参考价值。
衍生相关工作
该数据集衍生了一系列围绕动漫角色身份保持与风格控制的研究工作。典型工作包括基于LoRA的低秩适配微调方法,通过该数据集实现角色核心特征的高效植入;以及标签聚类驱动的可控生成框架,利用聚类结果引导模型生成特定服饰组合。此外,数据集的多源采集与清洗流程启发了自动化数据管线(如Waifuc)的优化,推动了大规模动漫图像数据集的构建标准与高效复用范式的发展。
以上内容由遇见数据集搜集并总结生成








