CyberHarem/elice_fireemblem

这是一个名为Dataset of elice/エリス (Fire Emblem)的数据集，包含500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团队开发。数据集的核心标签包括`blue_hair, blue_eyes, long_hair, hair_between_eyes, breasts, bangs`。此外，README还提供了数据集的下载链接、加载方法以及标签聚类结果的示例。

This is a dataset titled "Dataset of elice/エリス (Fire Emblem)", which contains 500 images and their corresponding labels. The images were crawled from multiple websites such as danbooru, pixiv, zerochan, and other platforms, with the crawling system developed by the DeepGHS team. The core tags of this dataset include `blue_hair, blue_eyes, long_hair, hair_between_eyes, breasts, bangs`. Additionally, the README file provides the dataset's download link, loading method, and examples of tag clustering results.

原始信息汇总

数据集概述

基本信息

• 数据集名称: Dataset of elice/エリス (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: 500
• 核心标签: blue_hair, blue_eyes, long_hair, hair_between_eyes, breasts, bangs

数据包列表

名称	图像数量	大小	类型	描述
raw	500	652.44 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	500	370.28 MiB	IMG+TXT	短边不超过800像素的数据集。
stage3-p480-800	1132	754.38 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素。
1200	500	574.49 MiB	IMG+TXT	短边不超过1200像素的数据集。
stage3-p480-1200	1132	1.05 GiB	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	6		1girl, cape, falchion_(fire_emblem), fingerless_gloves, solo, sword, tiara, smile, armor, blush, looking_at_viewer
1	12		1girl, cape, falchion_(fire_emblem), solo, tiara, armor, fingerless_gloves, holding_sword, looking_at_viewer, simple_background, belt
2	5		1girl, blue_gloves, falchion_(fire_emblem), fingerless_gloves, holding_sword, long_sleeves, looking_at_viewer, solo, tiara, white_background, blue_cape, closed_mouth, red_cape, simple_background, blue_footwear, brown_belt, shoulder_armor, sweater, thigh_boots
3	7		1girl, bare_shoulders, hair_flower, looking_at_viewer, official_alternate_costume, solo, white_dress, sleeveless_dress, smile, blush, cleavage, closed_mouth, collarbone, symbol-shaped_pupils, armlet, small_breasts, upper_body, white_flower
4	5		1girl, crop_top, looking_at_viewer, midriff, navel, short_shorts, solo, tiara, bare_shoulders, blush, official_alternate_costume, small_breasts, open_mouth, simple_background, sleeveless, thighs, white_background, :d, arm_up, armpits, belt, bikini, blue_shorts, innertube
5	7		1girl, day, navel, smile, solo, tiara, blue_bikini, crop_top, looking_at_viewer, midriff, outdoors, armpits, bare_shoulders, cloud, ocean, short_shorts, small_breasts, water, alternate_costume, arm_up, beach, belt, blue_sky, blush, closed_mouth, cowboy_shot, innertube, sleeveless, thighs, tree, wet
6	7		1girl, large_breasts, smile, solo, alternate_breast_size, looking_at_viewer, navel, tiara, blue_bikini, patreon_username
7	19		1girl, looking_at_viewer, solo, alternate_costume, orange_shorts, short_shorts, tiara, waitress, beer_mug, smile, cleavage, employee_uniform, blush, medium_breasts, open_mouth, white_tank_top, holding_plate, chicken_(food), holding_cup, navel, tray
8	12		1girl, official_alternate_costume, playboy_bunny, rabbit_ears, smile, solo, fake_animal_ears, looking_at_viewer, white_pantyhose, leotard, rabbit_tail, simple_background, cleavage, open_mouth, blush, small_breasts, easter_egg, frilled_choker, puffy_short_sleeves, white_background, white_gloves
9	8		1girl, tiara, uncensored, blush, hetero, nipples, penis, pussy, 1boy, looking_at_viewer, navel, sex, solo_focus, spread_legs, large_breasts, open_mouth, vaginal, completely_nude, cum, lying, alternate_breast_size, clitoris, pov, sweat, tongue_out

搜集汇总

背景与挑战

背景概述

该数据集由CyberHarem发布，专注于文本到图像生成任务，规模在1千到1万条数据之间。数据集内容涉及艺术领域，但包含敏感信息，标注为'Not-For-All-Audiences'，使用时需注意潜在风险，并遵循MIT许可证。