datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Wake-Vision
Dataset Card for Wake Vision
Dataset Description
"Wake Vision" is a large, high-quality dataset featuring over 6 million images, significantly exceeding the scale and diversity of
current tinyML datasets (100x). This dataset includes images with annotations of whether each image contains a person. Additionally,
it incorporates a comprehensive fine-grained benchmark to assess fairness and robustness, covering perceived gender, perceived age,
subject distance, lighting… See the full description on the dataset page: https://huggingface.co/datasets/Harvard-Edge/Wake-Vision.EpiBench-NeurIPS2026
EpiBench
Anonymous release for NeurIPS 2026 Evaluations & Datasets Track review (paper ID 1899). All methodology, ablations, and analyses are in the companion paper; this card lists only what reviewers and downstream users need to load the data.
A 25,737-patient ILAE-aligned multimodal epilepsy benchmark derived from PubMed Central case reports + 192 EpiRAG textbook vignettes.
6 tasks: epilepsy_type, seizure_type, ez_localization, aed_response, surgery_outcome, status_epilepticus… See the full description on the dataset page: https://huggingface.co/datasets/NeurIPS-1899-ED-2026/EpiBench-NeurIPS2026.edugraph-exercises
EduGraph Exercises Dataset
EduGraph Exercises is a synthetic ML dataset of math-related visual problems, precisely labeled for training AI models in the education sector.
Every image in this dataset is programmatically generated using the EduGraph Ontology to ensure that visual features are mathematically bound to their pedagogical labels.
Quick Links
Generation Engine: GitHub Repository (Contribute new generators or views!)
Ontology: EduGraph Ontology (Semantic… See the full description on the dataset page: https://huggingface.co/datasets/christian-bick/edugraph-exercises.coffee-lamp
Synthetic Image-Classification Dataset
Synthetic image-classification dataset generated with stable diffusion
(zerogpu_sdxl_turbo) using text-to-image from class names + short descriptions.
Classes
Label
Images
background
20
coffee-mug
20
lamp
20
Layout
train/<label>/<label>.<id>.jpg
test/<label>/<label>.<id>.jpg
metadata.csv
Loading
from datasets import load_dataset
ds = load_dataset("imagefolder"… See the full description on the dataset page: https://huggingface.co/datasets/edgeimpulse/coffee-lamp.edgeimpulse-test-image-classification
Edgeimpulse Test Image Classification
This dataset is an integration-test fixture for Edge Impulse's "Import from Hugging Face" flow.
Structure
Splits: train, validation, test
Main fields: image, label
Extra metadata columns (from metadata.csv):
source_split
source_file
source_stem
source_path
Important note
Label source mode: source-metadata.
ImageNet-ES
📸 ImageNet-ES
Unlike conventional robustness benchmarks that rely on digital perturbations, we directly capture 202k images by using a real camera in a controllable testbed. The dataset presents a wide range of covariate shifts caused by variations in light and camera sensor factors.
📄 Read the paper (CVPR 2024)
🗂️ ImageNet-ES Strucuture
ImageNet-ES
├── es-train
│ └── tin_no_resize_sample_removed
│ # 8K original validation samples of Tiny-ImageNet without… See the full description on the dataset page: https://huggingface.co/datasets/edw2n/ImageNet-ES.celebahq_512_id_clusters
celebahq_512 with SRK identity labels
Summary
This dataset is a derived version of jxie/celeba-hq. It keeps the original image set and adds automatically generated identity-group labels derived from face-embedding clustering.
As explained in our experimental setup, we use CelebA-HQ from Karras et al. (2018), specifically the Hugging Face snapshot at revision 7ecc6a45edfb5483ccf2f7df1035d298ffe7c76b. The referenced CelebA-HQ version provides gender labels but no identity… See the full description on the dataset page: https://huggingface.co/datasets/edgarcancinoe/celebahq_512_id_clusters.mb-atmospheric_dust_cls_edr
mb-atmospheric_dust_cls_edr
A Mars image classification dataset for planetary science research.
Dataset Metadata
License: CC-BY-4.0 (Creative Commons Attribution 4.0 International)
Version: 1.0
Date Published: 2025-05-15
Cite As: TBD
Classes
This dataset contains the following classes:
0: dusty
1: not_dusty
Statistics
train: 9817 images
test: 5214 images
val: 4969 images
few_shot_train_2_shot: 4 images
few_shot_train_1_shot: 2 images… See the full description on the dataset page: https://huggingface.co/datasets/Mirali33/mb-atmospheric_dust_cls_edr.benchmark
EditJudge-Bench
EditJudge-Bench is a synthetic benchmark for auditing vision-language models used as
automated judges for image-edit verification. Each row contains a source image,
an edited image, a factual edit instruction, counterfactual instructions, and
ground-truth scene parameters produced by a controlled Blender/Infinigen
generation pipeline.
This repository is an anonymous review release for a NeurIPS Evaluations and
Datasets submission.
Dataset Contents
1… See the full description on the dataset page: https://huggingface.co/datasets/EDAnonSubmission/benchmark.CNNBiasDataset
CNNBiasDataset
This dataset contains images of 5 different objects, which are used to quantify the bias that certain CNNs may have.It was created by Eder Martinez, a student volunteer at Kansas State University.
Contents
Formatted in JPG.Three main experiments were performed in controlled conditions, from which 5 crops were created.The "Original Image" folder contains the unedited 720x480 images, while the "Cropped" folder contains the cropped versions used in our paper… See the full description on the dataset page: https://huggingface.co/datasets/EderMartinez/CNNBiasDataset.EDID
EDID — Eye Disease Fundus Image Dataset
A curated dataset of retinal fundus images for automated eye disease classification.
The original Dataset shared in paper "A dataset of color fundus images for the detection and classification of eye diseases",
doi: https://doi.org/10.1016/j.dib.2024.110979.
Link to original dataset: https://data.mendeley.com/datasets/s9bfhswzjb/1
All credits goes to authors of mentioned paper.
Dataset Details
Please visit links above for… See the full description on the dataset page: https://huggingface.co/datasets/bekzatn/EDID.flymyai-ffhq-edit-bench
Face Identity Preservation Benchmark
A comprehensive evaluation dataset for face transformation APIs measuring identity preservation across complexity levels and transformation categories.
🔗 Complete Repository: https://github.com/FlyMyAI/bench_M1
Dataset Summary
This benchmark evaluates identity preservation in face image transformations using 8,832 transformation pairs across three major APIs. The dataset provides systematic evaluation of face editing quality using… See the full description on the dataset page: https://huggingface.co/datasets/flymy-ai/flymyai-ffhq-edit-bench.edge-inference-benchmarks
TinyEdge edge-inference benchmarks
Independently measured latency and accuracy for well-known vision models on
real edge devices (phones, tablets — fleet growing), produced by
TinyEdge, a device cloud for edge-AI benchmarking.
Nothing here is taken from papers or spec sheets: every row is a job executed
on the physical device through TinyEdge's production agent, with accuracy
measured on a fixed 500-image stratified sample of
ImageNet-V2 (matched-frequency)
using a standardized… See the full description on the dataset page: https://huggingface.co/datasets/TinyEdge/edge-inference-benchmarks.ImageNet-ES-Diverse
📸 ImageNet-ES Diverse
ImageNet-ES Diverse is a benchmark dataset of 192,000 real-world images captured with a physical camera in a controlled testbed (ES-Studio Diverse), under various sensor parameters and lighting conditions. It complements the original ImageNet-ES by introducing more diverse and realistic covariate shifts for robustness evaluation.
This dataset is introduced in the ICLR 2025 paper: Adaptive Camera Sensor for Vision Models.📄 Read the paper (ICLR 2025)… See the full description on the dataset page: https://huggingface.co/datasets/edw2n/ImageNet-ES-Diverse.kyc-passport-deepfake-pad-hidream-o1-qwen-image-edit
KYC Passport Deepfake / Presentation-Attack Benchmark
HiDream-O1 · Qwen-Image-Edit · 23 document-reproduction conditions
A presentation-attack detection (PAD) evaluation set for identity-document face
imagery: real reference faces and generator-attributed synthetic faces, each carried
through 23 acquisition conditions that emulate how a passport photo actually reaches a
KYC system: print and scan, photocopy, fax, phone recapture, JPEG chains, and
background… See the full description on the dataset page: https://huggingface.co/datasets/danb21/kyc-passport-deepfake-pad-hidream-o1-qwen-image-edit.Handwritten_KhattArabic Handwritten dataset.birdsnap_liteThis is a version of BirdSnap that will be easier on your free Google Colab quota.
chest-xrays-evaluation_cnn-cls
🩻 Chest X-Ray: Detección de Anomalías
Dataset de imágenes de radiografías de tórax procesadas para tareas de clasificación binaria (Normal vs Anomalía). Este dataset forma parte del curso de Deep Learning de inGeniia, utilizado para enseñar Redes Convolucionales (CNN) y Transfer Learning con modelos como YOLO11 (modo clasificación).
🖼️ Descripción del Dataset
Las imágenes han sido extraídas originalmente de Kaggle y procesadas con técnicas de Data Augmentation para… See the full description on the dataset page: https://huggingface.co/datasets/edgardoporto/chest-xrays-evaluation_cnn-cls.CNNBiasDataset_E4
CNNBiasDataset
This dataset contains images of 5 different objects, which are used to quantify the bias that certain CNNs may have.It was created by Eder Martinez, a student volunteer at Kansas State University.
Contents
Formatted in JPG.Three main experiments were performed in controlled conditions, from which 5 crops were created.The "Original Image" folder contains the unedited 720x480 images, while the "Cropped" folder contains the cropped versions used in our paper… See the full description on the dataset page: https://huggingface.co/datasets/EderMartinez/CNNBiasDataset_E4.