CyberHarem/amagi_azurlane

该数据集是关于《碧蓝航线》中的角色天城（Amagi）的图像数据集，包含500张图像及其标签。图像从多个网站（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`animal_ears, brown_hair, long_hair, fox_ears, breasts, purple_eyes, bangs, large_breasts, tail, animal_ear_fluff, thick_eyebrows, fox_girl, fox_tail, multiple_tails, blunt_bangs, hair_ornament`等。数据集提供了多种格式的下载选项，包括原始数据、不同分辨率的图像以及经过裁剪的图像。此外，还提供了使用waifuc加载原始数据集的代码示例。

原始信息汇总

数据集概述：amagi/天城/天城 (Azur Lane)

数据集基本信息

• 名称: amagi/天城/天城 (Azur Lane)
• 包含内容: 500张图像及其标签
• 核心标签: animal_ears, brown_hair, long_hair, fox_ears, breasts, purple_eyes, bangs, large_breasts, tail, animal_ear_fluff, thick_eyebrows, fox_girl, fox_tail, multiple_tails, blunt_bangs, hair_ornament
• 许可证: MIT
• 任务类别: text-to-image
• 标签: art, not-for-all-audiences
• 大小类别: n<1K

数据集包信息

包列表

名称	图像数量	大小	类型	描述
raw	500	1.01 GiB	Waifuc-Raw	原始数据，包含元信息，最小边对齐至1400像素（如果更大）。
800	500	485.57 MiB	IMG+TXT	数据集，短边不超过800像素。
stage3-p480-800	1281	1.05 GiB	IMG+TXT	三阶段裁剪数据集，区域不小于480x480像素。
1200	500	860.31 MiB	IMG+TXT	数据集，短边不超过1200像素。
stage3-p480-1200	1281	1.64 GiB	IMG+TXT	三阶段裁剪数据集，区域不小于480x480像素。

数据集加载

使用Waifuc加载原始数据集

提供原始数据集（包括标记的图像）供waifuc加载。加载代码示例：

python import os import zipfile

from huggingface_hub import hf_hub_download from waifuc.source import LocalSource

下载原始存档文件

zip_file = hf_hub_download( repo_id=CyberHarem/amagi_azurlane, repo_type=dataset, filename=dataset-raw.zip, )

提取文件到你的目录

dataset_dir = dataset_dir os.makedirs(dataset_dir, exist_ok=True) with zipfile.ZipFile(zip_file, r) as zf: zf.extractall(dataset_dir)

使用waifuc加载数据集

source = LocalSource(dataset_dir) for item in source: print(item.image, item.meta[filename], item.meta[tags])

数据集标签聚类结果

聚类列表（部分示例）

#	样本数	标签（部分）
0	36	1girl, solo, white_bikini, bare_shoulders, cleavage, looking_at_viewer, see-through, very_long_hair, choker, official_alternate_costume, thigh_strap, sash, thighs, sitting, kitsune
1	14	1girl, holding_umbrella, oil-paper_umbrella, wide_sleeves, looking_at_viewer, red_coat, solo, purple_kimono, kitsune, petals, cherry_blossoms, sakuramon, smile, brown_gloves
2	11	1girl, looking_at_viewer, off_shoulder, shimenawa, solo, twintails, wide_sleeves, black_pantyhose, kyuubi, sidelocks, long_sleeves, red_kimono, detached_sleeves, bell, cherry_blossoms, ribbon, wooden_floor, holding, seiza, simple_background
3	8	1girl, bare_shoulders, cleavage, solo, looking_at_viewer, collarbone, off_shoulder, red_kimono, wide_sleeves, brown_tail, sitting, very_long_hair
4	5	1girl, cleavage, kimono, red_eyes, red_skirt, smile, solo, wide_sleeves, black_gloves, black_thighhighs, outdoors, pleated_skirt, full_body, standing, black_hair, brown_tail, cherry_blossoms, cloud, holding, kyuubi, looking_at_viewer, makeup, parted_lips, sakuramon, sidelocks, sky
5	6	steam, 1girl, cleavage, collarbone, solo, bare_shoulders, naked_towel, looking_at_viewer
6	6	1girl, cleavage, folding_fan, holding_fan, looking_at_viewer, thighs, black_dress, indoors, solo, bare_shoulders, feather_boa, kitsune, official_alternate_costume, window, barefoot, nail_polish, on_couch, on_side, very_long_hair
7	8	1girl, hetero, solo_focus, blush, nipples, open_mouth, ejaculation, heart, navel, 1boy, completely_nude, cum_in_pussy, multiple_boys, multiple_penises, overflow, sweat, uncensored, ahegao, choker, collarbone, dark-skinned_male, gangbang, handjob, huge_breasts, lactation, mosaic_censoring, sky, spread_legs, vaginal