Loading...
Loading...
Text classification, translation, QA, summarization, dialogue, sentiment analysis, language modeling, text corpora
46,399 datasets
October to December 2022 saw 2,296 alleged protection incidents perpetrated on individuals in five prefectures of the Central African Republic. UNHCR, as the lead of the protection cluster, collected this data continuously through its partner INTERSOS, leveraging a network of community relays and protection committees. The dataset covers the prefectures of Haute-Kotto, Kemo, Nana-Gribizi, Ouaka, and Ouham.
CV-18 NER is the first publicly available dataset for Named Entity Recognition from Arabic speech. It was created by augmenting the Arabic Common Voice 18 corpus with manual NER annotations following the fine-grained Wojood schema, which covers 21 entity types. The dataset was authored by Elyadata and last updated on 2026-05-19.
Liminal Dreamcore 1K is a collection of 1,000 AI-generated images curated to represent the Dreamcore aesthetic. The dataset was created by LukaDev13 and was last updated on Hugging Face in May 2026. All images are AI-generated, capturing the strange, liminal, and half-remembered visual language of dreams.
UNHCR's Socio-Economic Vulnerability Assessment Tool (SEVAT) provides household-level demographic and socio-economic data for refugee households in Iraq. The data was collected in 2022 to facilitate the identification of vulnerabilities for assistance targeting, comparing camp versus out-of-camp settings. The dataset is hosted on the HDX platform and was last updated in March 2026.
An oil and gas industry jack-up rig site survey acquired in April 2023 under licence P2321. The survey traversed block number 27/5 in the UK and was published by the British Geological Survey (BGS). Its record was last updated on 2026-05-28.
The Canadian Trade Commissioner Service provides support for businesses. The dataset is maintained by Global Affairs Canada and was last updated on May 5, 2026. It likely contains information about services for Canadian businesses seeking to start up, scale up, or export.
PNG image files containing the output of a Swin Transformer-based model designed for global damage restoration of historical maps. The dataset was created by author Giser F and is hosted on figshare under a CC-BY-4.0 license. It was last updated on 2026-05-07.
I.H.A. Knottnerus uploaded a video file demonstrating the rearrangement of laser beams into a 5th generation Sierpinski triangle pattern using a Spatial Light Modulator (SLM). The dataset is a single MP4 file sized at 584.4 KB and was last updated on May 8, 2026. It is shared under a CC-BY-4.0 license.
Operating Dry Cleaner Registrations in Texas with current and valid status are listed. The dataset includes self-reported information from facilities and drop stations required to register annually with the Texas Commission on Environmental Quality. Records include registration details, location data, and compliance indicators.
A dataset from figshare by Rebecca S. Kann, last updated in April 2026. It contains prevalence data for enteric pathogen infections, stratified by rainy and dry seasons. The dataset includes adjusted prevalence ratio (PR) and beta (b) estimates from generalized estimating equation models for a sample of 630 observations.
DiversityNet provides a unified benchmark for evaluating generative AI models in chemistry, addressing a gap in the flourishing literature. The benchmark continues the tradition of data science benchmarks like MoleculeNet and ImageNet, aiming to establish best practices for AI molecule generators. It was created by Mostapha Benhenda and is available under an Open Access license.
Multilingual tutoring dialogues designed to reflect realistic student language rather than engineered prompts. The dataset was produced by CanisAI for the Uncharted Data Challenge and last updated on May 17, 2026. It aims to train tutors to meet real student language.
The Low-Income Energy Affordability Data (LEAD) Tool is an interactive, web-based platform developed by the U.S. Department of Energy. Its goal is to help stakeholders understand energy burdens and household energy costs for low- and moderate-income populations. The platform was last updated on 2026-05-04.
A ClinVar variant-effect-prediction benchmark for evaluating DNA language models, containing two balanced binary classification subsets. The dataset includes 39,473 coding variants and 15,258 non-coding variants, labeled as pathogenic/likely pathogenic or benign/likely benign. It was created by HuggingFaceBio and was last updated on 2026-05-06.
A dataset for AlexandriaX Subtask 3, a dialectal Arabic machine translation evaluation task. It contains train and development splits with machine-translated outputs, source text, and reference-side information. The dataset was created by UBC-NLP and its record was last updated on 2026-05-20.
A dataset from the Government Digital Service detailing the Scheme of Councillors' Allowances and Expenses. The data likely contains recommendations and revisions from Independent Remuneration Panels convened every four-year Council term. It is available in multiple formats including XML, HTML, CSV, and JSON under an Open Government Licence.
Human V1 70F is a dataset of 85,856 episodes of human activity, containing 343,424 uploaded files including video and skeleton data. The dataset was created by zywu2115 and last updated on 2026-05-18. It is filtered based on criteria such as a minimum of 70 frames, a minimum visible ratio of 0.3, and a minimum hand motion of 0.005.
SciDraw-6K contains 6,291 scientific illustrations synthesized by Google Gemini image-generation models, paired with prompts in eleven languages. The dataset spans eight scientific categories, with biomedical illustrations making up 44.9% of the collection. It was created by Chen, Davie and published via Harvard Dataverse in April 2026.
A dataset developed by Salvà Salvà, Macià for cyber threat intelligence research. It combines metadata, reputation scores, security reports, community assessments, and textual threat intelligence from VirusTotal and AlienVault OTX. The dataset enables the study of malicious domain detection using statistical, metadata-based, and text-based features.
Public data for constructing, calibrating, and running a SWAT+ hydrological model for a watershed. Input data includes freely available climate data from AEMET and geographic coverages for elevation, land use from SIOSE, and soil types from DSOLMap. Calibration data uses river flow gauges from the CEDEX Yearbook and sampled parameter values, while output data consists of hydrological variables estimated by the SWAT+ model. The dataset was authored by José Manuel Rodríguez-Castellanos and harvested by e-cienciaDatos.