CyberHarem/noire_azurlane

该数据集是关于《碧蓝航线》中的角色诺瓦露（Noire/ノワール/诺瓦露）的图像数据集，包含500张图像及其标签。核心标签包括长头发、黑发、红眼、双马尾等，这些标签在数据集中进行了修剪。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，并展示了标签聚类结果。

This is an image dataset centered on the character Noire (alternatively denoted as ノワール and Nuòwǎlù) from the mobile game *Azur Lane*. The dataset contains 500 images paired with their corresponding labels. Core labels include long hair, black hair, red eyes, twin tails, and other attributes, which have been pruned in the dataset. All images were crawled from multiple online platforms including danbooru, pixiv, zerochan, among others, with the crawling system supported by the DeepGHS Team. The dataset provides multiple download options, including raw data, images with various resolutions, and cropped images. Additionally, code examples for loading the raw dataset using waifuc are provided, along with the label clustering results.

原始信息汇总

数据集概述：noire/ノワール/诺瓦露 (Azur Lane)

数据集基本信息

• 名称: noire/ノワール/诺瓦露 (Azur Lane)
• 许可证: MIT
• 任务类别: text-to-image
• 标签: art, not-for-all-audiences
• 大小类别: n<1K

数据集内容

• 描述: 包含500张图像及其标签。
• 核心标签: long_hair, black_hair, red_eyes, twintails, breasts, ribbon, hair_ribbon, medium_breasts, very_long_hair, hair_ornament, hair_between_eyes

数据集包

名称	图像数量	大小	下载链接	类型	描述
raw	500	776.65 MiB	下载	Waifuc-Raw	包含元信息的原始数据（如果较大，最小边对齐至1400像素）。
800	500	421.47 MiB	下载	IMG+TXT	短边不超过800像素的数据集。
stage3-p480-800	1310	943.48 MiB	下载	IMG+TXT	三阶段裁剪数据集，区域不小于480x480像素。
1200	500	684.36 MiB	下载	IMG+TXT	短边不超过1200像素的数据集。
stage3-p480-1200	1310	1.35 GiB	下载	IMG+TXT	三阶段裁剪数据集，区域不小于480x480像素。

数据集加载

• 加载工具: waifuc

• 加载代码示例: python import os import zipfile

  from huggingface_hub import hf_hub_download from waifuc.source import LocalSource

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

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

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

数据集集群

• 集群列表: 提供了多个集群的详细信息，包括样本数量、图像示例及其相关标签。
• 集群示例:
  • 集群0: 12个样本，标签包括1girl, bare_shoulders, looking_at_viewer, solo, cleavage, blush, open_mouth, simple_background, white_background, detached_sleeves。
  • 集群1: 8个样本，标签包括1girl, bare_shoulders, solo, looking_at_viewer, cleavage, smile, open_mouth。
  • 集群2: 17个样本，标签包括1girl, solo, looking_at_viewer, blush, bare_shoulders, midriff, navel, skirt, thighhighs, crop_top, detached_sleeves, smile。
  • 集群3: 13个样本，标签包括1girl, blush, detached_collar, looking_at_viewer, rabbit_ears, playboy_bunny, solo, wrist_cuffs, bare_shoulders, leotard, cleavage, fake_animal_ears, bowtie, pantyhose, rabbit_tail, smile。
  • 集群4: 5个样本，标签包括1girl, blush, looking_at_viewer, nipples, pussy, solo, completely_nude, navel, open_mouth, uncensored, smile, anus, large_breasts, simple_background, white_background。
  • 集群5: 16个样本，标签包括1girl, hetero, nipples, sex, solo_focus, vaginal, 1boy, blush, penis, open_mouth, navel, large_breasts, looking_at_viewer, cum_in_pussy, thighhighs, completely_nude, cowgirl_position, girl_on_top, uncensored。
  • 集群6: 11个样本，标签包括1boy, blush, hetero, penis, 1girl, solo_focus, paizuri, large_breasts, nipples, looking_at_viewer, mosaic_censoring, bare_shoulders, pov, bangs, breasts_squeezed_together, collarbone, cum_on_breasts, fellatio, smile, sweat。
  • 集群7: 7个样本，标签包括school_uniform, blazer, blush, pleated_skirt, alternate_costume, 1girl, bow, looking_at_viewer, open_mouth, plaid_skirt, school_bag, 2girls, loafers, socks, solo。