datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
viet-cultural-vqaVietnamese Cultural VQA Dataset is a comprehensive multimodal dataset focusing on Vietnamese cultural heritage.
It contains 28,505 images across 12 cultural categories with 119,012 question-answer pairs in Vietnamese and English.
The dataset covers diverse aspects of Vietnamese culture including architecture, cuisine, traditional clothing,
landscapes, festivals, folk culture, traditional games, sports, handicrafts, musical instruments, daily life,
and transportation.danbooru2026
Danbooru2026: A Large-Scale Crowdsourced and Tagged Anime Illustration Dataset [WIP]
Dataset Description
Danbooru2026 is a large-scale anime illustration dataset containing over 10 million community-annotated images. It is intended for research and development in anime-style image generation, image classification, multimodal learning, and related tasks.
Danbooru is a long-running image board known for its extensive tagging system and community-maintained… See the full description on the dataset page: https://huggingface.co/datasets/nyanko-devs/danbooru2026.danbooru2023
Danbooru2023: A Large-Scale Crowdsourced and Tagged Anime Illustration Dataset
Danbooru2023 is a large-scale anime image dataset with over 5 million images contributed and annotated in detail by an enthusiast community. Image tags cover aspects like characters, scenes, copyrights, artists, etc with an average of 30 tags per image.
Danbooru is a veteran anime image board with high-quality images and extensive tag metadata. The dataset can be used to train image classification… See the full description on the dataset page: https://huggingface.co/datasets/nyanko7/danbooru2023.danbooru2024
Danbooru2024 Dataset
8058098 original images in total.
🎨 Danbooru2024 Dataset
📊 Dataset Overview
The Danbooru2024 dataset is a comprehensive collection focused on animation and illustration artwork, derived from the official Danbooru platform. It contains approximately 8.05 million high-quality, user-annotated images with corresponding tags and textual descriptions.
✨ Features
📋 Metadata Support
Includes a Parquet format… See the full description on the dataset page: https://huggingface.co/datasets/deepghs/danbooru2024.danbooru2023_indexTar index files for nyanko7/danbooru2023.
You can download images from both nyanko7/danbooru2023 and deepghs/danbooru_newest with cheesechaser.
from cheesechaser.datapool import DanbooruNewestDataPool
pool = DanbooruNewestDataPool()
# download danbooru original images from 7200000-7201000, to directory /data/danbooru_original
pool.batch_download_to_directory(
resource_ids=range(7200000, 7201000),
dst_dir='/data/danbooru_original',
max_workers=12,
)
danbooru2024-webp-4Mpixel
🎨 Danbooru2024 Webp 4MPixel Dataset
📊 Dataset Overview
The Danbooru2024-Webp dataset is a comprehensive collection focused on animation and illustration artwork, derived from the official Danbooru platform. It contains approximately 8.05 million high-quality, user-annotated images with corresponding tags and textual descriptions.
This dataset is 4MP-focused webp resized-dataset of Danbooru2024.
✨ Features
📋 Metadata Support
Includes a Parquet… See the full description on the dataset page: https://huggingface.co/datasets/deepghs/danbooru2024-webp-4Mpixel.danbooru2024-sfw
🎨 Danbooru2024 Dataset
📊 Dataset Overview
The Danbooru2024 dataset is a comprehensive collection focused on animation and illustration artwork, derived from the official Danbooru platform. It contains approximately 6.5 million high-quality, user-annotated images with corresponding tags and textual descriptions.
This dataset is filtered from an original set of 8.3 million entries, excluding NSFW-rated, opt-out entries to create a more accessible and… See the full description on the dataset page: https://huggingface.co/datasets/deepghs/danbooru2024-sfw.Danbooru-WD-EVA-EmbeddingsThis dataset includes WD EVA v2 large embeddings for danbooru images. Pixiv will be added later. Tensors under w are direct outputs and match indexes for WD EVA model. Tensors under e are from WD EVA as well however these strips the classifiaction head, they are smaller and suitable for deduplication computing for example.
The indexes of w, e and f (filename) match.
WD EVA Model: https://huggingface.co/SmilingWolf/wd-eva02-large-tagger-v3
danbooru2025-metadata
🎨 Danbooru 2025 Metadata
Latest Post ID: 9,158,800
(as of Apr 16, 2025)
📁 About the DatasetThis dataset provides structured metadata for user-submitted images on Danbooru, a large-scale imageboard focused on anime-style artwork.
Scraping began on January 2, 2025, and the data are stored in Parquet format for efficient programmatic access.Compared to earlier versions, this snapshot includes:
More consistent tag history tracking
Better coverage of older or previously… See the full description on the dataset page: https://huggingface.co/datasets/trojblue/danbooru2025-metadata.patfig
PatFig Dataset
Introduction
The PatFig Dataset is a curated collection of over 18,000 patent images from more than 7,000 European patent applications, spanning the year 2020. It aims to provide a comprehensive resource for research and applications in image captioning, abstract reasoning, patent analysis, and automated documentprocessing.
The overarching goal of this dataset is to advance the research in visually situated language understanding towards more… See the full description on the dataset page: https://huggingface.co/datasets/danaaubakirova/patfig.chexpert
CheXpert
CheXpert is a large dataset of chest X-rays and competition for automated chest x-ray interpretation, which features uncertainty labels and radiologist-labeled reference standard evaluation sets.
https://stanfordmlgroup.github.io/competitions/chexpert/
Warning on AP/PA label
I could not find in the paper a mapping from the 0/1 label to AP/PA, so I assumed 0=AP and 1=PA. Looking at a few images this seems to be correct, but I'm not a radiologist.… See the full description on the dataset page: https://huggingface.co/datasets/danjacobellis/chexpert.danbooru2023
[Mirror]Danbooru2023: A Large-Scale Crowdsourced and Tagged Anime Illustration Dataset
Danbooru2023 is an extension of Danbooru2021, featuring over 6.8 million anime-style images, totaling more than 8.3 TB.
Each image is accompanied by community-contributed tags that provide detailed descriptions of its content, including characters,
artists, copyright information, concepts, and attire.
This makes it a crucial resource for stylized computer vision tasks and transfer learning.… See the full description on the dataset page: https://huggingface.co/datasets/zenless-archive/danbooru2023.Danbooru-2024-Filtered-1Mdanbooru2024-latents-sdxl-1ktar
Danbooru 2024 SDXL VAE latents in 1k tar
Dedicated dataset to align deepghs/danbooru2024-webp-4Mpixel. "4MP-Focus" for average raw image resolution.
Latents are ARB with maximum size of 1024x1024 as the recommended setting in kohyas. Major reason is to make sure I can finetune with RTX 3090. VRAM usage will raise drastically after 1024.
Generated from prepare_buckets_latents_v2.py, modified from prepare_buckets_latents.py.
Used for kohya-ss/sd-scripts. In theory it may replace… See the full description on the dataset page: https://huggingface.co/datasets/6DammK9/danbooru2024-latents-sdxl-1ktar.damaged-media
Dataset Card for "ARTeFACT"
ARTeFACT: Benchmarking Segmentation Models on Diverse Analogue Media Damage
Here we provide example code for downloading the data, loading it as a PyTorch dataset, splitting by material and/or content, and visualising examples.
Housekeeping
!pip install datasets
!pip install -qqqU wandb transformers pytorch-lightning==1.9.2 albumentations torchmetrics torchinfo
!pip install -qqq requests gradio
import os
from glob import glob
import cv2… See the full description on the dataset page: https://huggingface.co/datasets/danielaivanova/damaged-media.danbooru2023-webp-4Mpixel_indexIndex files of KBlueLeaf/danbooru2023-webp-4Mpixel.
You can download images from KBlueLeaf/danbooru2023-webp-4Mpixel with cheesechaser.
from cheesechaser.datapool import DanbooruWebpDataPool
pool = DanbooruWebpDataPool()
# download danbooru images with webp format, to directory /data/danbooru_webp
pool.batch_download_to_directory(
resource_ids=range(6000000, 6001000),
dst_dir='/data/danbooru_webp',
max_workers=12,
)
DANI
DANI: Discrepancy Assessing for Natural and AI Images
Paper: D-Judge: How Far Are We? Evaluating the Discrepancies Between AI-synthesized Images and Natural Images through Multimodal Guidance
Code: https://github.com/RenyangLiu/DJudge
A Large-Scale Dataset for Visual Research on AI-Synthesized and Natural Images
Overview
DANI (Discrepancy Assessing for Natural and AI Images) is a large-scale, multimodal dataset for benchmarking and broad visual research on both… See the full description on the dataset page: https://huggingface.co/datasets/Renyang/DANI.recaptcha-57k-images-dataset
recaptcha-57k-images-dataset
Image dataset of reCAPTCHA tile images for image classification. This dataset
is used to train recaptcha-classification-57k
and powers vision-ai-recaptcha-solver.
Dataset summary
Task: image classification
Size: ~57k labeled images
Labels: 14 classes (13 target classes + other)
Labels
Target classes in this dataset:
bicycle
bridge
bus
car
chimney
crosswalk
fire hydrant
motorcycle
mountain
palm tree
stairs
tractor
traffic light… See the full description on the dataset page: https://huggingface.co/datasets/DannyLuna/recaptcha-57k-images-dataset.danbooru2023
Danbooru2023: A Large-Scale Crowdsourced and Tagged Anime Illustration Dataset
Danbooru2023 is a large-scale anime image dataset with over 5 million images contributed and annotated in detail by an enthusiast community. Image tags cover aspects like characters, scenes, copyrights, artists, etc with an average of 30 tags per image.
Danbooru is a veteran anime image board with high-quality images and extensive tag metadata. The dataset can be used to train image classification… See the full description on the dataset page: https://huggingface.co/datasets/jpft/danbooru2023.danbooru2023-webp-4Mpixel
Danbooru 2023 webp: A space-efficient version of Danbooru 2023
This dataset is a resized/re-encoded version of danbooru2023.
Which removed the non-image/truncated files and resize all of them into smaller size.
This dataset already be updated to latest_id = 7,832,883.
Thx to DeepGHS!
Notice: content of updates folder and deepghs/danbooru_newest-webp-4Mpixel have been merged to 2000~2999.tar, You can ignore all the content in updates folder safely!
Details
This… See the full description on the dataset page: https://huggingface.co/datasets/KBlueLeaf/danbooru2023-webp-4Mpixel.indian_dance_formsThis dataset is taken from https://www.kaggle.com/datasets/aditya48/indian-dance-form-classification but is originally from the Hackerearth deep learning contest of identifying Indian dance forms. All the credits of dataset goes to them.
Content
The dataset consists of 599 images belonging to 8 categories, namely manipuri, bharatanatyam, odissi, kathakali, kathak, sattriya, kuchipudi, and mohiniyattam. The original dataset was quite unstructured and all the images were put together.… See the full description on the dataset page: https://huggingface.co/datasets/tanmaykm/indian_dance_forms.mr-rateConversion of Forithmus/MR-RATE to Parquet:
Inserted all images ('series') into the row of their corresponding study, along with metadata, into the images column.
Allows iterating over the rows, and reading the images (as nifti_bytes) along with their report text.
Limited number of files to ~11k instead of the ~800k in the original repo. A smaller version of this dataset, for debugging purposes, is available under danielnobbe/mr-rate-toy-4000.
NOTE: The images in this repo are the ones from… See the full description on the dataset page: https://huggingface.co/datasets/danielnobbe/mr-rate.danbooru-tags-20260518Danbooru Dataset collected with my script.
Collected post ids: 1 ~ 11403815
Usage:
from datasets import load_dataset
dataset = load_dataset("u-haru/danbooru-tags-20260518", split="train")
danbooru-artists-10k
Danbooru Artists 10k
Dataset Summary
This dataset contains images from 10,000 artists collected from Danbooru. The images are primarily sourced from a mirror here. However, there are six images that are missing/broken in this mirror, which have been directly sourced from Danbooru. Each artist is represented by 30 images. This dataset is intended for use in metric learning or fine-grained classification tasks related to art styles. The artists are split into train… See the full description on the dataset page: https://huggingface.co/datasets/gustproof/danbooru-artists-10k.Full-Danbooru-Complement
Full Danbooru Complement
发布状态 / Release status（2026-08-06）：基线版本、2026-06 月度归档和 2026-07 月度归档均已完成整理、验证并发布。
The baseline, 2026-06 monthly archive, and 2026-07 monthly archive have all been assembled, verified, and published.
概述 / Overview
Full Danbooru Complement 是 deepghs/danbooru2024-webp-4Mpixel 的后续补充数据集，用于延伸其 Danbooru 图像与 post metadata 覆盖范围。数据按不可变发布目录组织；每个 Parquet 行对应一个 Danbooru post，并在行内保存图像字节与相关 metadata。
Full Danbooru Complement is a follow-up complement to… See the full description on the dataset page: https://huggingface.co/datasets/Xuness/Full-Danbooru-Complement.danyig-pedri-binary-script-classifier
Danyig vs Pedri Binary Script Classification Dataset
Stage-2 binary classifier for distinguishing Danyig (5 subscripts: DraDring, DraRing, Drathung, Gongshabma, Tsegdrig) from Pedri (2 subscripts: Peri, Petsuk). Real-only, all images human-reviewed.
Images per class
Class
train
val
test
All
Danyig
480
60
60
600
Pedri
480
60
60
600
Total
960
120
120
1,200
Splits
Manuscript-stratified split — each manuscript work appears in exactly… See the full description on the dataset page: https://huggingface.co/datasets/BDRC/danyig-pedri-binary-script-classifier.danbooru-wdtagger-v4-w640-ws-full
Danbooru WDTagger V4 Webdataset Full Dataset
This is the webdataset dataset for animetimm/danbooru-wdtagger-v4-w640.
Metadata here are cleaned by @SmilingWolf, can be used for multi-label classification training.
Images here are resized to min(width, height) <= 640.
How to Use It
from datasets import load_dataset
dataset = load_dataset('animetimm/danbooru-wdtagger-v4-w640-ws-full')
print(dataset["train"][0])
Images
5914596 images in total.
Split… See the full description on the dataset page: https://huggingface.co/datasets/animetimm/danbooru-wdtagger-v4-w640-ws-full.danbooru2023-metadata-database
Metadata Database for Danbooru2023
Danbooru 2023 datasets: https://huggingface.co/datasets/nyanko7/danbooru2023
The latest entry of this database is id 7,866,491. Which is newer than nyanko7's dataset.
This dataset contains a sqlite db file which have all the tags and posts metadata in it.
The Peewee ORM config file is provided too, plz check it for more information. (Especially on how I link posts and tags together)
The original data is from the official dump of the posts info.… See the full description on the dataset page: https://huggingface.co/datasets/KBlueLeaf/danbooru2023-metadata-database.danbooru-2023-sqlite-fixed-7110548
SQLITE-DB for Danbooru 2023 (until 7110548)
This is an cleaned-up version (almost totally recreated) of https://huggingface.co/datasets/KBlueLeaf/danbooru2023-sqlite
The previous sqlite database had major defects, especially with tag ids being mismatched, which was causing data to be actually different from server.
Note that minor information, such as uploader id, are not fixed.
Most of the discrepancy has been detected from id 5139963-6859952.
The additional scripts and example… See the full description on the dataset page: https://huggingface.co/datasets/AngelBottomless/danbooru-2023-sqlite-fixed-7110548.Israel-Photos
Israel Photos Dataset
A collection of 369 photographs captured across Israel between 2024 and 2025, with LLM-generated captions and location annotations. The images are sourced from the photographer's Pexels gallery.
About This Collection
This dataset was deliberately curated to provide a diverse visual representation of Israel, encompassing:
Varied locations: From the historic streets of Jerusalem's Old City to Tel Aviv's urban landscape, desert vistas in the Negev, and… See the full description on the dataset page: https://huggingface.co/datasets/danielrosehill/Israel-Photos.