datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
winoground
Dataset Card for Winoground
Dataset Description
Winoground is a novel task and dataset for evaluating the ability of vision and language models to conduct visio-linguistic compositional reasoning. Given two images and two captions, the goal is to match them correctly—but crucially, both captions contain a completely identical set of words/morphemes, only in a different order. The dataset was carefully hand-curated by expert annotators and is labeled with a rich set of… See the full description on the dataset page: https://huggingface.co/datasets/facebook/winoground.140k-Real-and-Fake-Faces
140k Real and Fake Faces Dataset
📝 Dataset Description
Dataset Summary
The 140k Real and Fake Faces Dataset is a balanced face image collection designed to benchmark models for detecting StyleGAN-generated faces. The dataset contains 70,000 real human face photographs sourced from the Flickr-Faces-HQ (FFHQ) dataset compiled by NVIDIA, and 70,000 fake faces sampled from the 1 Million Fake Faces dataset generated by StyleGAN and originally provided by Bojan.… See the full description on the dataset page: https://huggingface.co/datasets/TheKernel01/140k-Real-and-Fake-Faces.human-nonhuman-face-classification
Human vs Non-Human Face Dataset
A robust dataset for binary classification between real human faces and non-human face-like objects (statues, art, gaming, anime).
📊 Dataset Statistics
Split
Human
Non-Human
Total
Train
3,024
2,949
5,973
Validation
864
842
1,706
Test
433
422
855
Total
8,534
📁 Format
Labels: 0: human, 1: non_human.
🚀 Quick Start
from datasets import load_dataset
ds =… See the full description on the dataset page: https://huggingface.co/datasets/LakoreAI/human-nonhuman-face-classification.Realistic-Face-Portrait-1024px
Realistic-Face-Portrait-1024px
Dataset Summary
Realistic-Face-Portrait-1024px is a high-resolution image dataset containing 6,712 realistic portrait images of male and female individuals. Each image is standardized to 1024×1024 pixels, making it suitable for tasks involving high-fidelity facial analysis, face generation, and image-to-image transformation tasks such as super-resolution or inpainting.
Dataset Structure
Split: train
Number of rows: 6,712… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Realistic-Face-Portrait-1024px.emonet-face-hq
EmoNet-Face: A Fine-Grained, Expert-Annotated Benchmark for Facial Emotion Recognition
Dataset Summary
EmoNet-Face is a comprehensive benchmark suite designed to address critical gaps in facial emotion recognition (FER). Current benchmarks often have a narrow emotional spectrum, lack demographic diversity, and use uncontrolled imagery. EmoNet-Face provides a robust foundation for developing and evaluating AI systems with a deeper, more nuanced understanding of human… See the full description on the dataset page: https://huggingface.co/datasets/laion/emonet-face-hq.synthetic-watch-faces-dataset
Synthetic Watch Faces Dataset
A synthetic dataset of analog watch faces displaying various times for training vision models in time recognition tasks.
Dataset Description
This dataset consists of randomly generated analog watch faces showing different times. Each image contains a watch with hour and minute hands positioned to display a specific time. The dataset is designed to help train and evaluate computer vision models and Vision-Language Models (VLMs) for time… See the full description on the dataset page: https://huggingface.co/datasets/elischwartz/synthetic-watch-faces-dataset.face-or-not
Face or Not
theoriclabs/face-or-not is a balanced binary image-classification dataset of
4,000 128x128 RGB crops. The label answers one narrow question: does this crop
contain an Open Images Human face annotation?
This is classification, not face localization, identification, recognition,
or biometric matching. It has no names or identity labels.
Splits
split
no_face
face
total
train
1,600
1,600
3,200
validation
200
200
400
test
200
200
400… See the full description on the dataset page: https://huggingface.co/datasets/theoriclabs/face-or-not.face-celeb-vietnamese
Dataset Card for "face-celeb-vietnamese"
Dataset Summary
This dataset contains information on over 8,000 samples of well-known Vietnamese individuals, categorized into three professions: singers, actors, and beauty queens. The dataset includes data on more than 100 celebrities in each of the three job categories.
Languages
Vietnamese: The label is used to indicate the name of celebrities in Vietnamese.
Dataset Structure
The image and Vietnamese… See the full description on the dataset page: https://huggingface.co/datasets/fptudsc/face-celeb-vietnamese.Face_Generation_Benchmark
Rapidata Human Face Generation Alignment
This T2I dataset contains over ~22'000 human responses, collected in less than 1h using the Rapidata Python API, accessible to anyone and ideal for large scale evaluation.
Evaluating 12 different image generation models on which one can generate faces more accurately.
The question that the annotators get asked is: "Which Image follows the description of the human better?"
To evaluate your own models and create leaderboard check out our… See the full description on the dataset page: https://huggingface.co/datasets/Rapidata/Face_Generation_Benchmark.Face-Age-10K
Face-Age-10K Dataset
The Face-Age-10K dataset consists of over 9,000 facial images annotated with age group labels. It is designed for training machine learning models to perform age classification from facial features.
Dataset Details
Total Images: 9,165
Image Size: 200x200 pixels
Format: Parquet
Modality: Image
Split:
train: 9,165 images
Labels
The dataset includes 8 age group classes:
labels_list = [
'age 01-10',
'age 11-20',
'age 21-30'… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Face-Age-10K.vietnam_celeb_faceface-detection
Face Detection Dataset
Small grayscale face crops + a large pool of natural-image negatives, packaged
for classical face detectors (Viola-Jones, Haar cascades, sliding-window
classifiers).
At a glance
Split
Rows
Faces / Non-faces
train
106,977
102,429 / 4,548
test
24,045
472 / 23,573 (CBCL benchmark)
negatives
29,879
— / 29,879 (Caltech-256)
Same schema across all splits: image, label (0/1), source, category.
Quick start
from datasets… See the full description on the dataset page: https://huggingface.co/datasets/salvacarrion/face-detection.facepass_eval
FacePass Evaluation Dataset (Real LFW Faces)
This dataset contains real face images from the LFW (Labeled Faces in the Wild) dataset, curated for face recognition evaluation.
⚠️ IMPORTANT: This is the corrected version with actual face photographs (not colored squares).
Key Features
✅ Real faces: Actual photographs of people, not synthetic images✅ Balanced dataset: All individuals have 20+ images✅ Proper splits: 80/20 train/test split per person✅ Standardized: Resized to… See the full description on the dataset page: https://huggingface.co/datasets/besartshyti/facepass_eval.face-celeb-vietnamese
Dataset Card for "face-celeb-vietnamese"
Dataset Summary
This dataset contains information on over 8,000 samples of well-known Vietnamese individuals, categorized into three professions: singers, actors, and beauty queens. The dataset includes data on more than 100 celebrities in each of the three job categories.
Languages
Vietnamese: The label is used to indicate the name of celebrities in Vietnamese.
Dataset Structure
The image and Vietnamese… See the full description on the dataset page: https://huggingface.co/datasets/thuanML/face-celeb-vietnamese.synthetic-face-sdxl-instantid-bench
Synthetic Face Detection Benchmark — SDXL+InstantID
Version: v1.0.0 · Build date: 2026-05-16 · Rows: 26492
Evaluation benchmark for synthetic-face detection under platform-realistic
conditions. Sampled to satisfy the ISO/IEC 19795 floor of 300 samples per
demographic subgroup across a 6×2 (skin tone × gender) cell grid. Not
training data; not licensed for commercial use.
See release.json for build provenance, manifest.csv for per-row
license attestation, LICENSES.csv for the… See the full description on the dataset page: https://huggingface.co/datasets/danb21/synthetic-face-sdxl-instantid-bench.