datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
civitai-top-nsfw-images-with-metadata
CivitAI Top NSFW Images Dataset
This dataset contains 6k+ top NSFW 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-NSFW-Images-Dataset-6k-images
Dataset collection date
June 2025
Dataset structure:
├── 📂 images/
│ ├── 1.jpg
│ ├── 2.jpg
│ ├── 3.jpg
│ ├── ....
├──… See the full description on the dataset page: https://huggingface.co/datasets/wallstoneai/civitai-top-nsfw-images-with-metadata.nsfw-dataset
Dataset Card: NSFW Image Classification Dataset
Dataset Summary
This dataset is designed for binary image classification to detect NSFW (Not Safe For Work) vs. Safe content. It is intended for training and evaluating computer vision models, such as Vision Transformers (ViT) and CNN-based architectures, to aid in automated content moderation.
Task: Image Classification
Labels: safe, nsfw
Modality: Images (RGB)
Use cases: Content moderation, filtering, safety systems… See the full description on the dataset page: https://huggingface.co/datasets/theusamaaslam/nsfw-dataset.nsfw_detect
NSFW Detection Dataset
Summary
This dataset is specifically designed for training NSFW (Not Safe For Work) detection models in the context of artistic content and image classification. The collection follows the established categorization format from popular NSFW detection implementations, providing a comprehensive benchmark for content moderation systems. The dataset contains images organized into five distinct classes that represent different levels of appropriateness… See the full description on the dataset page: https://huggingface.co/datasets/deepghs/nsfw_detect.civitai-top-nsfw-images-with-metadata
CivitAI Top NSFW Images Dataset
This dataset contains 6k+ top NSFW 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-NSFW-Images-Dataset-6k-images
Dataset collection date
June 2025
Dataset structure:
├── 📂 images/
│ ├── 1.jpg
│ ├── 2.jpg
│ ├── 3.jpg
│ ├── ....
├──… See the full description on the dataset page: https://huggingface.co/datasets/pasindu29/civitai-top-nsfw-images-with-metadata.cashbox-nsfw-image-level
Cashbox NSFW Image Level (v6)
四分类图像敏感内容分级数据集，用于训练与评估 L1–L4 内容安全分类模型。
⚠️ 18+ / Not For All Audiences：含成人向与裸露内容，仅限研究与合规场景使用。
项目
说明
版本
v6
样本数
40,928
类别
L1 / L2 / L3 / L4（4 类）
格式
Parquet 分片（图像 bytes 内嵌）
图像规格
长边 ≤ 768 px，保持宽高比
配套模型
jiangchengchengNLP/image_l_class
标签分布
等级
数量
占比
L1
11,233
27.4%
L2
5,468
13.4%
L3
10,838
26.5%
L4
13,389
32.7%
合计
40,928
100%
label 为整数索引：0→L1, 1→L2, 2→L3, 3→L4（按类名字母序）。
分级标准… See the full description on the dataset page: https://huggingface.co/datasets/jiangchengchengNLP/cashbox-nsfw-image-level.NSFW-MultiDomain-Classification
NSFW_MultiDomain
The NSFW_MultiDomain dataset is a curated image classification dataset focused on multi-domain adult content recognition. It consists of 5 distinct categories aimed at facilitating the development of robust NSFW (Not Safe For Work) image classification models. This dataset enables training and benchmarking of models that can distinguish between subtle variations in explicit and non-explicit content across artistic, animated, and real-world imagery.… See the full description on the dataset page: https://huggingface.co/datasets/strangerguardhf/NSFW-MultiDomain-Classification.civitai-top-nsfw-images-with-metadata
CivitAI Top NSFW Images Dataset
This dataset contains 6k+ top NSFW 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-NSFW-Images-Dataset-6k-images
Dataset collection date
June 2025
Dataset structure:
├── 📂 images/
│ ├── 1.jpg
│ ├── 2.jpg
│ ├── 3.jpg
│ ├── ....
├──… See the full description on the dataset page: https://huggingface.co/datasets/Avanish11/civitai-top-nsfw-images-with-metadata.NSFW-MultiDomain-Classification-v2.0
NSFW-MultiDomain-Classification-v2.0
This dataset is curated for multi-domain NSFW (Not Safe For Work) classification tasks. It provides a diverse range of images categorized into multiple NSFW and SFW (Safe For Work) domains, allowing for the training and evaluation of models focused on NSFW detection across different content types.
Label Categories
The dataset includes images labeled into the following categories:
Anime Picture
Illustrations or animations in anime… See the full description on the dataset page: https://huggingface.co/datasets/strangerguardhf/NSFW-MultiDomain-Classification-v2.0.nsfw
NSFW/SFW Image Classification Dataset
This dataset contains 16,800 images:
NSFW: 11,200
SFW: 5,600
NSFW-T2I
Introduction (Version 1)
About 38k image-text pairs(10k from LAION and 28k from nsfw_detect), and captions are generated by LLaVA-NeXT with prompt "Describe the photo in detail (attributes of person)".
The "txt" column shown in the dataset viewer is originated from LAION, not the captions yielded by LLaVA-NeXT.
Caption Codes
pretrained = "lmms-lab/llama3-llava-next-8b"
model_name = "llava_llama3"
device = "cuda:2"
device_map = "auto"
tokenizer, model, image_processor… See the full description on the dataset page: https://huggingface.co/datasets/zxbsmk/NSFW-T2I.nsfw_detection_xlNSFW-T2I
Introduction (Version 1)
About 38k image-text pairs(10k from LAION and 28k from nsfw_detect), and captions are generated by LLaVA-NeXT with prompt "Describe the photo in detail (attributes of person)".
The "txt" column shown in the dataset viewer is originated from LAION, not the captions yielded by LLaVA-NeXT.
Caption Codes
pretrained = "lmms-lab/llama3-llava-next-8b"
model_name = "llava_llama3"
device = "cuda:2"
device_map = "auto"
tokenizer, model, image_processor… See the full description on the dataset page: https://huggingface.co/datasets/K00B404/NSFW-T2I.NSFW-MultiDomain-2CNSFW-filter-DecentScanNSFW-T2I
Introduction (Version 1)
About 38k image-text pairs(10k from LAION and 28k from nsfw_detect), and captions are generated by LLaVA-NeXT with prompt "Describe the photo in detail (attributes of person)".
The "txt" column shown in the dataset viewer is originated from LAION, not the captions yielded by LLaVA-NeXT.
Caption Codes
pretrained = "lmms-lab/llama3-llava-next-8b"
model_name = "llava_llama3"
device = "cuda:2"
device_map = "auto"
tokenizer, model… See the full description on the dataset page: https://huggingface.co/datasets/Arcanius/NSFW-T2I.NSFW-T2I
Introduction (Version 1)
About 38k image-text pairs(10k from LAION and 28k from nsfw_detect), and captions are generated by LLaVA-NeXT with prompt "Describe the photo in detail (attributes of person)".
The "txt" column shown in the dataset viewer is originated from LAION, not the captions yielded by LLaVA-NeXT.
Caption Codes
pretrained = "lmms-lab/llama3-llava-next-8b"
model_name = "llava_llama3"
device = "cuda:2"
device_map = "auto"
tokenizer, model… See the full description on the dataset page: https://huggingface.co/datasets/premfrd123/NSFW-T2I.