CyberHarem/kako_kantaicollection

这是kako/加古/加古 (Kantai Collection)的数据集，包含352张图像及其标签。核心标签包括黑发、长发、呆毛、马尾辫、蓝眼睛等。图像从多个网站（如danbooru、pixiv、zerochan等）爬取，爬取系统由DeepGHS团队提供。数据集提供了多种格式的下载链接，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，还提供了如何使用waifuc加载原始数据集的代码示例。最后，README文件还列出了标签聚类结果，展示了不同聚类中的图像样本及其标签。

原始信息汇总

数据集概述

数据集信息

• 名称: Dataset of kako/加古/加古 (Kantai Collection)
• 许可证: MIT
• 任务类别: text-to-image
• 标签: art, not-for-all-audiences
• 大小类别: n<1K
• 图像数量: 352
• 核心标签: black_hair, long_hair, ahoge, ponytail, blue_eyes, hair_over_one_eye, hair_ornament, bangs, ribbon, messy_hair, breasts

数据集包列表

名称	图像数量	大小	类型	描述
raw	352	243.89 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	352	188.14 MiB	IMG+TXT	短边不超过800像素的数据集
stage3-p480-800	698	341.62 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素
1200	352	230.32 MiB	IMG+TXT	短边不超过1200像素的数据集
stage3-p480-1200	698	407.96 MiB	IMG+TXT	3阶段裁剪数据集，区域不小于480x480像素

数据集加载

• 加载工具: waifuc

• 加载代码: python import os import zipfile from huggingface_hub import hf_hub_download from waifuc.source import LocalSource

  下载原始归档文件

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

  解压文件到指定目录

  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])

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	16						2girls, serafuku, pleated_skirt, open_mouth, hairclip, navel, blue_sailor_collar, red_neckerchief, blue_skirt, midriff, parted_bangs, short_hair, short_sleeves, smile, brown_hair, dated, solo_focus, twitter_username
  1	8						1girl, machinery, navel, serafuku, smile, solo, midriff, open_mouth, pleated_skirt, turret, chain, electricity, looking_at_viewer, cannon
  2	7						1girl, looking_at_viewer, serafuku, solo, cowboy_shot, navel, parted_bangs, pleated_skirt, red_neckerchief, smile, blue_skirt, hairclip, midriff, dated, open_mouth, short_sleeves, simple_background, blue_sailor_collar, hair_ribbon, white_background
  3	5						1girl, blue_skirt, parted_bangs, pleated_skirt, red_neckerchief, serafuku, simple_background, solo, twitter_username, blue_sailor_collar, dated, open_mouth, blush, cowboy_shot, one-hour_drawing_challenge, white_background, closed_eyes, drooling
  4	8						1girl, looking_at_viewer, navel, simple_background, solo, blush, large_breasts, open_mouth, white_background, naked_ribbon, hair_ribbon, smile, sweat, swimsuit, twitter_username
  5	11						1girl, solo, official_alternate_costume, white_background, white_skirt, looking_at_viewer, one-hour_drawing_challenge, simple_background, blue_shirt, smile, long_skirt, open_mouth, blouse, blush, hair_down, pleated_skirt
  6	5						1girl, holding_umbrella, hydrangea, official_alternate_costume, solo, rain, white_skirt, blouse, blue_shirt, blush, purple_shirt, shoes, smile, twitter_username, black_footwear, black_shirt, long_skirt, looking_at_viewer, outdoors