datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
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.db-sfw-512px-general-filter
Danbooru SFW 512px General Filter
This dataset is meant to be used for training a simple binary classifier that can filter the
Danbooru SFW 2021 dataset. There are two classes: "accepted" and "rejected", with "accepted"
representing samples that should pass through the filter and "rejected" representing samples
that should not.
Generallly, to be accepted a sample should meet the following criteria:
No horizontal/vertical bars, or bars that can be removed easily by filling in… See the full description on the dataset page: https://huggingface.co/datasets/hayden-donnelly/db-sfw-512px-general-filter.civitai-top-sfw-images-with-metadata
CivitAI Top SFW Images Dataset
This dataset contains 12k+ top SFW images from CivitAI filtered using top reactions. The dataset contains prompt & nsfw level metadata in prompts.json file. The nsfw levels are: Soft, Mature & X.
Original forum post:
https://diffused.to/Thread-CivitAI-Top-SFW-Images-Dataset-12k-images
Dataset collection date
July 2025
Dataset structure:
├── 📂 images/
│ ├── 1.jpeg
│ ├── 2.jpeg
│ ├── 3.jpeg
│ ├── ....
├──… See the full description on the dataset page: https://huggingface.co/datasets/wallstoneai/civitai-top-sfw-images-with-metadata.db-sfw-128px-filtered-and-cropped
Danbooru SFW 512 Filtered and Cropped
A version of Danbooru SFW which has been automatically filtered and cropped to 128x128 so that the resulting images focus almost entirely on characters.
First filtering was applied to remove non-character focused images, then cropping was applied to remove horizontal/vertical bars and further focus on the
central character(s) of each image. Both steps were performed automatically by two different vision models trained on manually labelled… See the full description on the dataset page: https://huggingface.co/datasets/hayden-donnelly/db-sfw-128px-filtered-and-cropped.db-sfw-512px-character-filter
Danbooru SFW 512px Character Filter
This dataset is meant to be used for training a simple binary classifier that can filter the
Danbooru SFW 2021 dataset. It is similar to db-sfw-512-general-filter-dataset
but it has different class criteria. Just like the general dataset there are two
classes: "accepted" and "rejected", with "accepted" representing samples that should pass
through the filter and "rejected" representing samples that should not.
To be accepted, a sample should… See the full description on the dataset page: https://huggingface.co/datasets/hayden-donnelly/db-sfw-512px-character-filter.