datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Deepfake
DeepGuard Deepfake Dataset
A paired deepfake dataset for training and benchmarking deepfake detection models.
Generated as part of the DeepGuard AI Project (2026).
📥 2,301+ all-time downloads
Dataset Statistics
Fake images: 5426 (face-swapped using InsightFace inswapper_128)
Real images: 5426 (original paired faces)
Total: 10852 images
Format: JPEG, high quality (95%)
Generation method: InsightFace inswapper_128 (ONNX runtime)
Why This Dataset is… See the full description on the dataset page: https://huggingface.co/datasets/Sowaiba01/Deepfake.NTIRE-RobustAIGenDetection-train
Training set for NTIRE 2026 Robust AI-Generated Image Detection in the Wild
Robust AI-Generated Image Detection in the Wild Challenge is organized as a part of the New Trends in Image Restoration and Enhancement Workshop in conjunction with CVPR 2026.
Challenge overview
Text-to-image (T2I) models have made synthetic images nearly indistinguishable from real photos in many cases, which creates serious challenges for trust, authenticity, forensics, and content… See the full description on the dataset page: https://huggingface.co/datasets/deepfakesMSU/NTIRE-RobustAIGenDetection-train.20K_real_and_deepfake_images_PCAThis dataset contains the test images used to evaluate our deepfake detection framework. It originally contained 20,000 real and deepfake images, but as some 2600 files are protected by the UK Crown and we do not have a permission to reproduced them, so these files were removed.
Our framework contains 4 machine learning models, which feed in the original images, error-level analysis (ELA) images, noise analysis (NA) images and Principal Component Analysis (PCA) images.
The models were created… See the full description on the dataset page: https://huggingface.co/datasets/ts0pwo/20K_real_and_deepfake_images_PCA.deepfake-celeba
MetaFLOS Deepfake Dataset (CelebA Face Deepfake Generated Images)
Flux2-Klein generated deepfake images from CelebA face descriptions, for deepfake detection / comparison research.
Content
19,867 generated images (full CelebA validation split)
Resolution: 256×256
Each corresponds to a CelebA real face (see real_orig field in manifest)
Generation style: snapshot/crop candid-photo look (not studio portrait) — off-center framing, subject possibly touching or cut by… See the full description on the dataset page: https://huggingface.co/datasets/tjw/deepfake-celeba.NTIRE-RobustAIGenDetection-val
Validation set for NTIRE 2026 Robust AI-Generated Image Detection in the Wild (updated)
Note: This is an updated version of the dataset. For challenge submissions, please make sure you use this version.
Robust AI-Generated Image Detection in the Wild Challenge is organized as a part of the New Trends in Image Restoration and Enhancement Workshop in conjunction with CVPR 2026.
Challenge overview
Text-to-image (T2I) models have made synthetic images nearly… See the full description on the dataset page: https://huggingface.co/datasets/deepfakesMSU/NTIRE-RobustAIGenDetection-val.Deepfake-Identity-Isolated-Dataset-PreP
Identity-Isolated Deepfake Face Images Dataset.
A rigorously preprocessed, identity-aware deepfake detection dataset of 142,837 face images, constructed for training generalized deepfake detectors across Face Swap and Entire Face Synthesis manipulation categories.
Built as part of the DFDS project - an open-source deepfake detection API for FinTech KYC identity verification. The full system is available on GitHub at DFDS-XAI
Dataset Summary.
Property
Value… See the full description on the dataset page: https://huggingface.co/datasets/ThinothW/Deepfake-Identity-Isolated-Dataset-PreP.deepfake_face_classification
Deepfake Face Image Classification Dataset
This dataset is curated for the classification of deepfake face images, 16060 each real and fake images. It is derived from the DF40 dataset(only the test data), which includes 40 distinct deepfake techniques, facilitating the detection of state-of-the-art deepfakes and AI-generated content. (paperswithcode.com)
Dataset Overview
The dataset contains 32134 total images.
The dataset is divided into two categories:
Fake Images: 16… See the full description on the dataset page: https://huggingface.co/datasets/pujanpaudel/deepfake_face_classification.Sap_Kush_Med_Deepfake
Sap_Kush_Med_Deepfake Dataset
Paired medical-image forgery lineages across six modalities. Every lineage is
one source image, one mask, one seed: the arms differ only in what was done
inside the mask, so a comparison between arms isolates the manipulation rather
than an encoding artefact.
3956 lineages, 30385 files, 8.49 GiB.
v2 adds a removal arm grounded in human annotation for three more modalities
(endoscopy, ultrasound, MRI). v1 had removal for CT only.
What the… See the full description on the dataset page: https://huggingface.co/datasets/Kanhaiyya/Sap_Kush_Med_Deepfake.Deepfake-vs-Real-v2
Deepfake-vs-Real-v2
Deepfake-vs-Real-v2 is a dataset designed for image classification, distinguishing between deepfake and real images. This dataset includes a diverse collection of high-quality deepfake images to enhance classification accuracy and improve the model’s overall efficiency. By providing a well-balanced dataset, it aims to support the development of more robust deepfake detection models.
Label Mappings
Mapping of IDs to Labels: {0: 'Deepfake', 1:… See the full description on the dataset page: https://huggingface.co/datasets/dappai/Deepfake-vs-Real-v2.NTIRE-RobustAIGenDetection-test-public
Test set for NTIRE 2026 Robust AI-Generated Image Detection in the Wild
Robust AI-Generated Image Detection in the Wild Challenge is organized as a part of the New Trends in Image Restoration and Enhancement Workshop in conjunction with CVPR 2026.
Challenge overview
Text-to-image (T2I) models have made synthetic images nearly indistinguishable from real photos in many cases, which creates serious challenges for trust, authenticity, forensics, and content safety. At… See the full description on the dataset page: https://huggingface.co/datasets/deepfakesMSU/NTIRE-RobustAIGenDetection-test-public.20K_real_and_deepfake_imagesThis dataset contains the test images used to evaluate our deepfake detection framework. It originally contained 20,000 real and deepfake images, but as some 2600 files are protected by the UK Crown and we do not have a permission to reproduced them, so these files were removed.
Our framework contains 4 machine learning models, which feed in the original images, error-level analysis (ELA) images, noise analysis (NA) images and Principal Component Analysis (PCA) images.
The models were created… See the full description on the dataset page: https://huggingface.co/datasets/ts0pwo/20K_real_and_deepfake_images.MAHE_deepfake_recognition_datasetThis dataset is for different educational experiments of deepfake images classification.
It consists of the small datasets with various original and deepfake scenes (faces, animation, urban scenes and others):
Columbia Uncompressed Image Splicing Dataset
Kaggle Deepfake Dataset Challenge
CommunityForensics-Small
AI-vs-Deepfake-vs-Real-Resized-Aug
🧠 AI vs Deepfake vs Real — Processed Version
This dataset is the result of preprocessing and augmentation applied to the original datasetprithivMLmods/AI-vs-Deepfake-vs-Real.
📘 Overview
This dataset contains a collection of images categorized into three main classes:
🟩 AI-generated
🟥 Deepfake
🟦 Real (authentic human faces)
It is designed for image classification tasks that aim to distinguish between AI-generated, deepfake, and real faces.… See the full description on the dataset page: https://huggingface.co/datasets/chintalaswathi/AI-vs-Deepfake-vs-Real-Resized-Aug.Deepfake_dataset_cybersentinal
Dataset Card for OpenFake
OpenFake is a dataset and benchmark for detecting AI-generated images, with a focus on politically and socially salient content where misinformation risk is highest. It pairs real photographs with synthetic counterparts produced by a wide range of frontier proprietary generators, open-source diffusion models, and community fine-tunes. A separate in-the-wild test set is sourced from Reddit to evaluate detector performance on naturally circulated… See the full description on the dataset page: https://huggingface.co/datasets/vjjoshi23/Deepfake_dataset_cybersentinal.20K_real_and_deepfake_images_ELAThis dataset contains the test images used to evaluate our deepfake detection framework. It originally contained 20,000 real and deepfake images, but as some 2600 files are protected by the UK Crown and we do not have a permission to reproduced them, so these files were removed.
Our framework contains 4 machine learning models, which feed in the original images, error-level analysis (ELA) images, noise analysis (NA) images and Principal Component Analysis (PCA) images.
The models were created… See the full description on the dataset page: https://huggingface.co/datasets/ts0pwo/20K_real_and_deepfake_images_ELA.Deepfakes-QA-15K
Deepfake Quality Assessment
Deepfake QA is a Deepfake Quality Assessment model designed to analyze the quality of deepfake images & videos. It evaluates whether a deepfake is of good or bad quality, where:
0 represents a bad-quality deepfake
1 represents a good-quality deepfake
This classification serves as the foundation for training models on deepfake quality assessment, helping improve deepfake detection and enhancement techniques.
Citation
If you use our… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Deepfakes-QA-15K.deepfake-detection-dataset-v3
Deepfake Detection Dataset V3
This dataset contains images and detailed explanations for training and evaluating deepfake detection models. It includes original images, manipulated images, confidence scores, and comprehensive technical and non-technical explanations.
Dataset Structure
The dataset consists of:
Original images (image)
CAM visualization images (cam_image)
CAM overlay images (cam_overlay)
Comparison images (comparison_image)
Labels (label): Binary… See the full description on the dataset page: https://huggingface.co/datasets/saakshigupta/deepfake-detection-dataset-v3.dear-lsun-inpaint
DEAR Diagnostic Inpaint Set (dear-lsun-inpaint)
Diagnostic data for the dissection step of DEAR ("Dissect and Prune:
Enhancing Robustness in AI-Generated Image Detection", ICML 2026). Each real
LSUN image has a random rectangular region inpainted with Stable Diffusion
1.5, so real and generated pixels coexist in one image under a known mask.
DEAR uses these paired images and masks to measure per-channel Regional
Activation Discrepancy (RAD).
The image folders are shipped as tar… See the full description on the dataset page: https://huggingface.co/datasets/k-aisi-anti-deepfake/dear-lsun-inpaint.DeepFakeDetection
DeepFakeDetection Dataset
This repository contains a comprehensive dataset for DeepFake detection research and development. The dataset consists of 140,000 high-quality images split between real and fake categories.
Dataset Overview
Total Images: 140,000
Real Images: 70,000
Fake Images: 70,000
Data Splits
Training Set (80%): 112,000 images
56,000 real
56,000 fake
Validation Set (10%): 14,000 images
7,000 real
7,000 fake
Test Set (10%): 14,000 images
7… See the full description on the dataset page: https://huggingface.co/datasets/yashduhan/DeepFakeDetection.Deepfake-vs-Real-60K
Deepfake-vs-Real-60K
Deepfake-vs-Real-60K is a large-scale image classification dataset designed to distinguish between deepfake and real facial images. The dataset includes approximately 60,000 high-quality images, comprising 30,000 fake (deepfake) and 30,000 real images, to support the development of robust deepfake detection models.
By providing a well-balanced and diverse collection, Deepfake-vs-Real-60K aims to enhance classification accuracy and improve generalization for… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Deepfake-vs-Real-60K.Deepfake-vs-Real-v2
Deepfake-vs-Real-v2
Deepfake-vs-Real-v2 is a dataset designed for image classification, distinguishing between deepfake and real images. This dataset includes a diverse collection of high-quality deepfake images to enhance classification accuracy and improve the model’s overall efficiency. By providing a well-balanced dataset, it aims to support the development of more robust deepfake detection models.
Label Mappings
Mapping of IDs to Labels: {0: 'Deepfake', 1:… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Deepfake-vs-Real-v2.20K_real_and_deepfake_images_NAThis dataset contains the test images used to evaluate our deepfake detection framework. It originally contained 20,000 real and deepfake images, but as some 2600 files are protected by the UK Crown and we do not have a permission to reproduced them, so these files were removed.
Our framework contains 4 machine learning models, which feed in the original images, error-level analysis (ELA) images, noise analysis (NA) images and Principal Component Analysis (PCA) images.
The models were created… See the full description on the dataset page: https://huggingface.co/datasets/ts0pwo/20K_real_and_deepfake_images_NA.AI-vs-Deepfake-vs-Real
AI vs Deepfake vs Real
AI vs Deepfake vs Real is a dataset designed for image classification, distinguishing between artificial, deepfake, and real images. This dataset includes a diverse collection of high-quality images to enhance classification accuracy and improve the model’s overall efficiency. By providing a well-balanced dataset, it aims to support the development of more robust AI-generated and deepfake detection models.
Label Mappings
Mapping of IDs to… See the full description on the dataset page: https://huggingface.co/datasets/zeeshanaliicreativez/AI-vs-Deepfake-vs-Real.AI-vs-Deepfake-vs-Real
AI vs Deepfake vs Real
AI vs Deepfake vs Real is a dataset designed for image classification, distinguishing between artificial, deepfake, and real images. This dataset includes a diverse collection of high-quality images to enhance classification accuracy and improve the model’s overall efficiency. By providing a well-balanced dataset, it aims to support the development of more robust AI-generated and deepfake detection models.
Label Mappings
Mapping of IDs to… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/AI-vs-Deepfake-vs-Real.deepfake-v13-dataset
Deepfake Detection V13 Dataset
Overview
Total: 60,000 images
Real: 30,000
Fake: 30,000
Size: 224×224 RGB
Usage
from datasets import load_dataset
ds = load_dataset('ash12321/deepfake-v13-dataset')
Labels
0 = Real
1 = Fake
aigi-detection-ldm
DEAR-c Training Data (aigi-detection-ldm)
Training data for the base Corvi detector used in DEAR ("Dissect and
Prune: Enhancing Robustness in AI-Generated Image Detection", ICML 2026).
This repository hosts the LSUN real images and the LDM fake images. The
image folders are shipped as tar archives (Hugging Face allows at most 10000
files per folder). The DEAR detectors also use COCO real images, which are
not re-hosted here (download COCO 2017 from the official site, see below).… See the full description on the dataset page: https://huggingface.co/datasets/k-aisi-anti-deepfake/aigi-detection-ldm.Deepfake-vs-Real
Deepfake vs Real
Deepfake vs Real is a dataset designed for image classification, distinguishing between deepfake and real images. This dataset includes a diverse collection of high-quality deepfake images to enhance classification accuracy and improve the model’s overall efficiency. By providing a well-balanced dataset, it aims to support the development of more robust deepfake detection models.
Label Mappings
Mapping of IDs to Labels: {0: 'Deepfake', 1: 'Real'}… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Deepfake-vs-Real.selfiekyc_deepfake_detection
selfiekyc_deepfake_detection
Descripción General
Este repositorio contiene el código fuente, datasets procesados y experimentos desarrollados para la detección de imágenes faciales sintéticas o manipuladas mediante Inteligencia Artificial en escenarios de verificación de identidad corporativa tipo selfie-KYC.
El proyecto analiza la robustez de diferentes arquitecturas de Deep Learning frente a degradaciones operativas reales, incluyendo:
Compresión JPEG fuerte… See the full description on the dataset page: https://huggingface.co/datasets/juliopaciello/selfiekyc_deepfake_detection.deepfake-detection-dataset-v3
Deepfake Detection Dataset V3
This dataset contains images and detailed explanations for training and evaluating deepfake detection models. It includes original images, manipulated images, confidence scores, and comprehensive technical and non-technical explanations.
Dataset Structure
The dataset consists of:
Original images (image)
CAM visualization images (cam_image)
CAM overlay images (cam_overlay)
Comparison images (comparison_image)
Labels (label): Binary… See the full description on the dataset page: https://huggingface.co/datasets/KubasadNisha/deepfake-detection-dataset-v3.deepfake-detection-dataset-v3
Deepfake Detection Dataset V3
This dataset contains images and detailed explanations for training and evaluating deepfake detection models. It includes original images, manipulated images, confidence scores, and comprehensive technical and non-technical explanations.
Dataset Structure
The dataset consists of:
Original images (image)
CAM visualization images (cam_image)
CAM overlay images (cam_overlay)
Comparison images (comparison_image)
Labels (label): Binary… See the full description on the dataset page: https://huggingface.co/datasets/guglothmahipal007/deepfake-detection-dataset-v3.