datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
OpenSDI_test
OpenSDI: Spotting Diffusion-Generated Images in the Open World
This dataset is designed to address the OpenSDI challenge: spotting diffusion-generated images in realistic, open-world scenarios. It is described in the paper:
Project Page: https://iamwangyabin.github.io/OpenSDI/
OpenSDID Dataset Highlights:
User Diversity: Simulates a wide range of user intentions and creative styles using diverse text prompts generated by VLMs.
Model Innovation: Includes images from multiple… See the full description on the dataset page: https://huggingface.co/datasets/nebula/OpenSDI_test.rlbenchfail_test_dataset
Guardian: RLBench-Fail Dataset
This dataset is part of the Guardian project: Detecting Robotic Planning and Execution Errors with Vision-Language Models. It contains annotated robotic manipulation failure data generated in the RLBench simulator for training and evaluating Vision-Language Models (VLMs) on failure detection tasks.
Failures are produced by an automated pipeline that procedurally perturbs successful scripted trajectories in simulation, generating diverse planning… See the full description on the dataset page: https://huggingface.co/datasets/paulpacaud/rlbenchfail_test_dataset.Mirage-Test
🌊 Mirage-Test Dataset
Mirage-Test is a modern test-only dataset for benchmarking AI-generated image detection models.
It contains real (0_real) and fake (1_fake) images across five distinct content domains, designed to evaluate generalization across diverse visual semantics.
The fake images are generated using state-of-the-art generative models specifically optimized for perceptual realism and visual fidelity.
📌 This dataset is for evaluation only. No training split is… See the full description on the dataset page: https://huggingface.co/datasets/Yunncheng/Mirage-Test.ur5fail_test_dataset
Guardian Failure Detection Dataset
This dataset is part of the Guardian project: Detecting Robotic Planning and Execution Errors with Vision-Language Models. It contains annotated robotic manipulation failure data for training and evaluating Vision-Language Models (VLMs) on failure detection tasks.
Guardian introduces an automated failure generation approach that procedurally perturbs successful robot trajectories to produce diverse planning failures and execution failures, each… See the full description on the dataset page: https://huggingface.co/datasets/paulpacaud/ur5fail_test_dataset.testing-goldstandard-cuthill
Dataset Card for Curated Gold Standard Hoyal Cuthill Dataset
Dataset Description
Dorsal full body images of subspecies of Heliconius erato and Heliconius melpomene (18 subspecies total).
There are 960 images with 320 specimens (3 images of each specimen: Original/ Bird transformed/ Butterfly transformed)
The original images are low-resolution RGB photographs (photographs were "cropped and resized to a height of 64 pixels (maintaining the original image aspect ratio and… See the full description on the dataset page: https://huggingface.co/datasets/jrw2989/testing-goldstandard-cuthill.winml-test-set
WinML Test Set
Dataset Summary
WinML Test Set is an evaluation‑only collection for validating model accuracy and stability on Windows ML / DirectML / ONNX Runtime pipelines. It aggregates several permissively‑licensed sources and harmonizes schema for reproducible, regression‑grade testing across backends and versions. Not intended for training.
Intended Use
Accuracy and regression benchmarking of Windows ML / DirectML / ONNX Runtime pipelines.… See the full description on the dataset page: https://huggingface.co/datasets/Futuremark/winml-test-set.bdv2fail_test_dataset
Guardian: BridgeDataV2-Fail Dataset
This dataset is part of the Guardian project: Detecting Robotic Planning and Execution Errors with Vision-Language Models. It contains annotated robotic manipulation failure data derived from the BridgeDataV2 real-robot dataset for training and evaluating Vision-Language Models (VLMs) on failure detection tasks.
Failures are produced by an automated pipeline that perturbs successful real-robot trajectories offline (without re-executing actions)… See the full description on the dataset page: https://huggingface.co/datasets/paulpacaud/bdv2fail_test_dataset.photo-test
Anime vs. Live Action Film Image Classification
OmarK211/photo-test
A binary image classification dataset designed to distinguish between Anime (Label 0) and Live Action (Label 1) film frames/imagery. Images are prepared as standardized square RGB files with multi-pass synthetic training variants.
Source and task
The images were manually imported from the web into my local computer. Then manually uploaded to Google Colab
Preparation source: 24-679 Image Data… See the full description on the dataset page: https://huggingface.co/datasets/OmarK211/photo-test.testset
Dataset Card for TreeOfLife-10M Captions
This dataset consists of generated captions, Wikipedia-derived descriptions and format examples for the TreeOfLife-10M. These captions were generated using InternVL3-38B based on biological contexts that help the model generate more accurate captions. It was used to train BioCAP, a CLIP-based model.
Dataset Details
This dataset is comprised of captions for the images in TreeOfLife-10M that were generated using InternVL3 38B.… See the full description on the dataset page: https://huggingface.co/datasets/ZihengZ/testset.test
Tiny overlapping ImageFolder demo
A tiny synthetic repository for testing Hugging Face Dataset Viewer.
12 lossless WebP images, each exactly 128×128.
full: 8 train, 2 validation, 2 test.
core: 4 train, 1 validation, 1 test.
core is a subset of full.
Images exist once in images/.
manifest.csv is the canonical manifest.
splits/ contains lightweight source indices.
Root full_*.csv and core_*.csv are materialized metadata files used by ImageFolder.
Expected Viewer… See the full description on the dataset page: https://huggingface.co/datasets/Krows7/test.GenAI-RealEstate-TestSet
🏙️ GenAI Real Estate Test Set (Track B)
Dataset for the MenaML Winter School 2026 Challenge.
📊 Dataset Structure
This dataset contains 1,000 images split evenly between:
Authentic: Real estate photography from the Places365 dataset.
Manipulated: Synthetically generated deepfake artifacts (Inpainting, Diffusion Noise, GAN Grids).
🕵️ How to Use
This dataset is designed for testing forensic detection models.
diagram-eval-access-test
Diagram Evaluation Access Test
This public one-image dataset tests the external-access workflow planned for the
Diagram Evaluation project. It is not a research dataset release.
Dataset structure
The repository uses Hugging Face's ImageFolder layout:
data/
train/
metadata.csv
sample-diagram.png
The metadata records the image's provenance, license, and intended use. A production
release can use the same contract with sharded Parquet or WebDataset files.… See the full description on the dataset page: https://huggingface.co/datasets/abhisheklalwani96/diagram-eval-access-test.HowToEat-test
HowToEat: Hand-Object Interaction and Eating Action in Eating Scenarios
HowToEat is an image dataset for analysing eating behaviour. It provides:
Hand-object interaction + eating face detection (hand_object_detection): 95,190 images with 190,333 hand instances (box, left/right side, contact state, and the box and category of the held object) and 151,620 face instances (box, eating / not eating).
Eating action recognition (eating_recognition): 6,280 manually labelled faces… See the full description on the dataset page: https://huggingface.co/datasets/thxplz/HowToEat-test.astrobridge-yse-test-dataset-v2
AstroBridge YSE external test dataset v2
This dataset contains 266 object-disjoint, spectroscopically labeled YSE DR1 transients. The broad-class counts are SN II: 71, SN Ia: 180, SN Ibc: 15.
V2 shortens the YSE forced-photometry time coverage to resemble the alert-photometry coverage of the AstroBridge BTS training dataset. For each object, it retains the smallest inclusive time interval that contains every positive measurement with flux/uncertainty at least 5 and at least five… See the full description on the dataset page: https://huggingface.co/datasets/BuildNg/astrobridge-yse-test-dataset-v2.ai-detector-benchmark-test-data
🎯 AI Detector Benchmark Test Dataset
A comprehensive benchmark dataset for testing AI image detection models.
📊 Dataset Summary
Total Images: 700
AI-Generated: 250 images (from 5 different generators)
Real Images: 450 images (from 9 diverse datasets)
Perfect for:
✅ Testing AI detection models
✅ Creating leaderboards
✅ Comparing model performance
✅ Benchmarking new approaches
🤖 AI Generators Included
Generator
Images
Accuracy Baseline
FLUX… See the full description on the dataset page: https://huggingface.co/datasets/Robo531/ai-detector-benchmark-test-data.test
Zoo Animal Re-Identification Dataset
A dataset for animal re-identification with 2,705 body images, 1,192 face crops, and 6 configurations.
Configurations
1. face_and_body
Individual frames with both face and body crops.
Features:
date: Date of capture (YYYY-MM-DD)
time: Time of capture (HH:MM:SS)
class: Animal name
video: Source video filename
frame_number: Frame number in video
camera: Camera ID
face_image: Cropped face image
body_image: Cropped body/full… See the full description on the dataset page: https://huggingface.co/datasets/Maxscha/test.VGGSound-EG2-100-test
Dataset Card for t1-vggsound-eg2-xmodal
This is a FiftyOne dataset with 100 samples.
Installation
If you haven't already, install FiftyOne:
pip install -U fiftyone
Usage
import fiftyone as fo
from fiftyone.utils.huggingface import load_from_hub
# Load the dataset
# Note: other available arguments include 'max_samples', etc
dataset = load_from_hub("harpreetsahota/VGGSound-EG2-100-test")
# Launch the App
session = fo.launch_app(dataset)… See the full description on the dataset page: https://huggingface.co/datasets/harpreetsahota/VGGSound-EG2-100-test.astrobridge-yse-test-dataset
AstroBridge YSE external test dataset
This dataset contains 266 spectroscopically labeled YSE DR1 transients that do not overlap the final AstroBridge BTS training dataset by normalized TNS identity or a two-arcsecond transient-coordinate match. The broad-class counts are SN II: 71, SN Ia: 180, SN Ibc: 15.
Each row has at least five valid ZTF g and five valid ZTF r observations, ordered by time. The light-curve flux arrays are in nJy. The ATCAT arrays retain SNANA FLUXCAL at… See the full description on the dataset page: https://huggingface.co/datasets/BuildNg/astrobridge-yse-test-dataset.ai-detector-benchmark-test-data
🎯 AI Detector Benchmark Test Dataset
A comprehensive benchmark dataset for testing AI image detection models.
📊 Dataset Summary
Total Images: 700
AI-Generated: 250 images (from 5 different generators)
Real Images: 450 images (from 9 diverse datasets)
Perfect for:
✅ Testing AI detection models
✅ Creating leaderboards
✅ Comparing model performance
✅ Benchmarking new approaches
🤖 AI Generators Included
Generator
Images
Accuracy Baseline
FLUX… See the full description on the dataset page: https://huggingface.co/datasets/ash12321/ai-detector-benchmark-test-data.test
Overview
lalalalalalalalanewwwwwwhhhhjyfjyf
This project aims to introduce and release a comprehensive food image dataset designed specifically for computer vision tasks, particularly food recognition, classification, and nutritional analysis. We hope this dataset will provide a reliable resource for researchers and developers to advance the field of food AI. By publishing on Hugging Face, we expect to foster community collaboration and accelerate innovation in applications… See the full description on the dataset page: https://huggingface.co/datasets/mkzheng/test.video-dataset-pre1_test
Video Dataset - pre1_test
Dataset Description
This dataset contains video frames extracted from annotated video segments, along with annotations, transcriptions, and corresponding video clips.
Dataset Structure
frames/ — extracted frames grouped by role (start, middle, end)
segments/ — video clips for each annotation interval
annotations/ — original JSON annotation
transcriptions/ — transcription files (full_transcription.txt + per segment)
dataset.csv —… See the full description on the dataset page: https://huggingface.co/datasets/Quazitron420/video-dataset-pre1_test.video-dataset-test2040
Video Dataset - test2040
Dataset Description
This dataset contains video frames extracted from annotated video segments, along with annotations, transcriptions, and corresponding video clips.
Dataset Structure
frames/ — extracted frames grouped by role (start, middle, end)
segments/ — video clips for each annotation interval
annotations/ — original JSON annotation
transcriptions/ — transcription files (full_transcription.txt + per segment)
dataset.csv —… See the full description on the dataset page: https://huggingface.co/datasets/Quazitron420/video-dataset-test2040.lineex-test
LineEX Test Split
This dataset is a structured Hugging Face mirror of the test split released with the
LineEX paper, "LineEX: Data Extraction from Scientific Line Charts" (WACV 2023).
Repo: 13point5/lineex-test
Split: test
Rows: 20,000
Source format: PNG images plus COCO-style chart-element and line annotations
Columns
image_id
file_name
image
width
height
data_type
chart_elements
lines
chart_elements fields
annotation_id
category_id
category_name… See the full description on the dataset page: https://huggingface.co/datasets/13point5/lineex-test.radiology-test-v2
Radiology Test v2
Evaluation dataset for the Radiology agent of an AI Medical Department,
paired with UngLong/radiology-ready-v2
(training set). CT images are in UngLong/openm3chest-npy-v2.
⚠️ Before using: CT scans for this test set must be uploaded to openm3chest-npy-v2 first.
See scripts/test_npy_needed.txt (154 scan keys) for the list to upload via build_npy_hub.py.
Dataset Summary
Total rows
154
Screening rows
74 (all 8 screening tasks)… See the full description on the dataset page: https://huggingface.co/datasets/UngLong/radiology-test-v2.botanical-vision-test
Botanical Vision
Fine-grained flowering-plant classification dataset: 495 research-grade
iNaturalist photos across 5 species (all flowering plants with at least
2,000 observations). Built for Advanced Computer Vision (UChicago ADSP 32023).
Splits
split
images
train
345
val
75
test
75
Split is stratified within each species (70/15/15). Exact and cross-species
duplicate images were removed before splitting.
Fields
image — the… See the full description on the dataset page: https://huggingface.co/datasets/dbabnigg/botanical-vision-test.video-dataset-anonim_test
Video Dataset - anonim_test
Dataset Description
This dataset contains video frames extracted from annotated video segments, along with annotations, transcriptions, and corresponding video clips.
Dataset Structure
frames/ — extracted frames (first frame from each segment)
segments/ — video clips for each annotation interval
annotations/ — original JSON annotation
transcriptions/ — transcription files (full_transcription.txt + per segment)
dataset.csv — mapping… See the full description on the dataset page: https://huggingface.co/datasets/Quazitron420/video-dataset-anonim_test.image-in-Words400_DOCCI_Testedgeimpulse-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.
openm3chest-cardiology-test-200
OpenM3Chest Cardiology Test 200
This dataset is a sampled test subset prepared from UngLong/openm3chest-labels.
Source
Source repo: UngLong/openm3chest-labels
Source split: test
Subsets/tasks: ['CVD_diagnosis', 'CVD_mortality']
Sampling
Number of samples: 200
Sampling mode: primary_ratio
Primary subset: CVD_diagnosis
Positive label: 1
Positive ratio: 0.5
Seed: 42
Columns
keys: CT scan key / series identifier from source dataset.… See the full description on the dataset page: https://huggingface.co/datasets/ChonJohn171105/openm3chest-cardiology-test-200.fish-vista-test
Dataset Card for Fish-Visual Trait Analysis (Fish-Vista)
Dataset Deetails
Dataset Description
The Fish-Visual Trait Analysis (Fish-Vista) dataset is a large, annotated collection of 60K fish images spanning 1900 different species; it supports several challenging and biologically relevant tasks including species classification, trait identification, and trait segmentation. These images have been curated through a sophisticated data processing… See the full description on the dataset page: https://huggingface.co/datasets/egrace479/fish-vista-test.