CyberHarem/mishiro_idolmastercinderellagirls

这是名为mishiro/美城常務 (THE iDOLM@STER: Cinderella Girls)的数据集，包含58张图片及其标签。图片从多个网站（如danbooru, pixiv, zerochan等）爬取，爬取系统由DeepGHS团队提供支持。数据集的核心标签包括black_hair, long_hair, ponytail, earrings, breasts, large_breasts, green_eyes，这些标签在数据集中被修剪。数据集提供了不同版本的下载链接，包括原始数据、不同分辨率的图片集以及三阶段裁剪的数据集。此外，还提供了如何使用waifuc加载原始数据集的代码示例，并展示了标签聚类结果的示例。

This is the dataset named mishiro/美城常務 (THE iDOLM@STER: Cinderella Girls), which contains 58 images and their associated labels. The images were crawled from various platforms such as danbooru, pixiv, zerochan and more, with the crawling infrastructure supported by the DeepGHS team. The core tags of the dataset include black_hair, long_hair, ponytail, earrings, breasts, large_breasts and green_eyes, and these tags have been pruned in the dataset. Multiple download links are offered for different variants of the dataset, including the raw data, image collections with different resolutions, and the three-stage cropped dataset. In addition, code examples for loading the raw dataset via waifuc are provided, alongside demonstrations of tag clustering results.

原始信息汇总

数据集概述

基本信息

• 名称: Dataset of mishiro/美城常務 (THE iDOLM@STER: Cinderella Girls)
• 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: 58张
• 核心标签: black_hair, long_hair, ponytail, earrings, breasts, large_breasts, green_eyes

数据包列表

名称	图像数量	大小	类型	描述
raw	58	42.92 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	58	29.15 MiB	IMG+TXT	短边不超过800像素的数据集。
stage3-p480-800	108	52.62 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素。
1200	58	39.47 MiB	IMG+TXT	短边不超过1200像素的数据集。
stage3-p480-1200	108	67.10 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素。

集群列表

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	15						1girl, formal, hair_pulled_back, solo, suit, necklace, makeup, cleavage

表格版本

#	样本数量	图像1	图像2	图像3	图像4	图像5	1girl	formal	hair_pulled_back	solo	suit	necklace	makeup	cleavage
0	15						X	X	X	X	X	X	X	X