Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
16,317 datasets
Swiss data supports a New Empirical Industrial Organisation method to estimate market power in the petrol retailing industry. The analysis, conducted by author Nils Herger, suggests a low level of market power held by Swiss petrol retailers. The dataset underpins research on competition and alleged collusive conduct in this sector.
This dataset documents $7 billion in state and local government spending on 134 minor league baseball stadiums for MLB affiliates. It was created by John Charles Bradbury to analyze the prevalence and policy implications of public subsidies for these facilities. The data covers all Minor League Baseball stadiums operating in 2025.
Datadocred is a dataset published on Kaggle. Its title suggests a focus on document redaction, likely involving the removal or masking of sensitive information from text. The dataset's specific size, columns, and creation details are not provided in the available metadata.
YOLO_4 is a dataset of model weights for a YOLO (You Only Look Once) object detection system. The raw description references 'yolo11l.pt' and 'yolo_5k', suggesting it likely contains a pre-trained model file. It is published on Kaggle.
PHYSIODYN_1DUNET_WAVEGAN_V1 is a dataset hosted on Kaggle. Its title suggests a focus on physiological dynamics and the use of 1D U-Net and WaveGAN architectures, likely for generating or processing synthetic waveform data. The dataset's specific content, scale, and authorship are not detailed in the provided metadata.
Deepfake Video Dataset provides a collection of videos organized into real and fake categories. The dataset is described as multilingual, suggesting it includes content from multiple languages. Its origin and specific scale are unknown.
This dataset supports an analysis of U.S. military production during World War II, examining the relationship between production scale and unit cost declines. It is described as an unprecedentedly large collection of data, compiled by author François Lafond. The data is used to quantify the roles of exogenous technological progress and endogenous production experience effects.
8,575 digitized herbarium specimens are documented in this collection from the Eloy Valenzuela Botanical Garden Herbarium. The collection was founded on September 22, 1987 and is managed by the Subdirección de Ordenamiento y Planificación Integral del Territorio. Specimens were gathered from field excursions for forest inventories, watershed studies, and ecological research in various Colombian ecosystems.
115,650 records of European noble lifespans spanning the years 800 to 1800, compiled by Neil Cummins. The data tracks adult age at death and identifies trends in longevity, including the impact of violent deaths from battle across different European regions.
FLUX-Reason-6M features 6 million synthetic images generated by the FLUX.1-dev model paired with 20 million bilingual descriptions. Created by LucasFang and updated in early 2026, the collection targets the improvement of complex reasoning capabilities in generative text-to-image systems.
1770–1810 newspaper page records from the Library of Congress Chronicling America collection. The dataset includes images, OCR text, and publication metadata, providing a foundation for research and machine learning projects. It was created by RevolutionCrossroads and last updated on Hugging Face in February 2026.
This dataset supports research on developing molecular tests for Chlamydia trachomatis detection in wastewater, employing PCR platforms including qPCR, ddPCR, and NextGenPCR. The work aims to enable specific, sensitive, and rapid pathogen detection for public health surveillance.
A collection of retinal Optical Coherence Tomography (OCT) images for classification tasks. The dataset, titled C8, is hosted on Kaggle. Specific details on volume, collection dates, and authorship are not provided in the available metadata.
FACE MASK YOLOV8 is a dataset hosted on Kaggle. The title suggests it contains images for training or evaluating a YOLOv8 object detection model, likely for identifying whether individuals are wearing face masks. No further metadata on size, origin, or specific contents is available.
ResNet_weights is a dataset of pre-trained model parameters for the ResNet architecture, published on Kaggle. The dataset likely contains the learned weights from training on a large-scale image corpus, enabling transfer learning. Specific details on the training data, model variant, and performance metrics are not provided in the available metadata.
Oferta de Trámites, Otros Procedimientos Administrativos (OPA) y Consultas de Acceso a Información Pública lists administrative procedures and public information requests offered by entities within Colombia's Executive Branch. The data is registered in the Sistema Único de Información de Trámites (SUIT) and was last updated in December 2025. It is published by the Colombian government's open data portal, www.datos.gov.co.
A directory lists enterprises, institutions, and organizations managed by an information manager. The dataset is provided by the States site of Ukraine and was last updated on February 20, 2026. It is available as an Excel XLSX file.
11 worksheets catalog aerial photographs and satellite imagery held at the BGS Murchison House office. The inventory was fully catalogued by late 2001 and updated through 2003 to include digital holdings. Metadata quality varies across the nine aerial photograph sub-collections.
Reports dating back to the 1950s document ground investigations across Great Britain, with concentrations around urban areas and transportation routes. The collection contains information on boreholes, trial pits, laboratory tests, and chemical analyses, held by the British Geological Survey. Scanning of paper, microfilm, and digital records is ongoing, with images stored in TIFF format.
A collection of medical images for detecting and classifying canine tumors. The dataset includes data formatted for YOLO object detection and CNN classification tasks, along with an independent test set. It was sourced from Kaggle, but the author, organization, and specific creation date are unknown.