Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
16,039 datasets
An informational dataset from data.colorado.gov explains Socrata's user roles and their associated actions. It details the capabilities of roles like Viewer, Publisher, Administrator, Editor, and Designer. The dataset was last updated in February 2026.
Field data from Queen Elizabeth National Park, Uganda, examines the effect of prenatal food provisioning on female banded mongooses and their offspring. The study experimentally manipulated the condition of pregnant females between 2013 and 2016 to measure impacts on weight change, cooperative care allocation, and pup survival.
Sediment trap and core measurements from Rostherne Mere, Cheshire, UK, quantify organic carbon burial. Data includes loss-on-ignition (LOI), calcium carbonate (CaCO3), net flux, and net sedimentation rate calculations. The collection spans from circa 1360 AD to August 2016, derived from a NERC-funded PhD project.
A dataset contains nitrogen offtake, emissions, soil parameters, and herbage quality measurements from a three-cut silage plot trial at two UK grassland sites in 2016. Measurements were taken from 16 plots per site following fertiliser applications by researchers from Bangor University and Rothamsted Research. Data was collected for the Newton Fund project 'UK-China Virtual Joint Centre for Improved Nitrogen Agronomy'.
October 3-7, 2022 surface water measurements from 23 locations in Loch Etive, Scotland, including nine open loch stations, four fish-farm sites, and six river inflows. The dataset encompasses in-field parameters like water temperature and pH, plus laboratory analyses for major nutrients, trace elements, stable isotopes, and phytoplankton community composition via flow cytometry.
2000-2019 data from Queen Elizabeth National Park, Uganda, documents mortality costs and reproductive success from intergroup conflict in a wild banded mongoose population. The dataset includes observations of aggressive interactions, individual mortality from fighting, and long-term records of reproductive events and oestrus timing.
Marine multibeam backscatter data from around the UK, collected from 2005 onwards primarily for the Civil Hydrography Programme. The British Geological Survey archives this data, which is useful for seabed characterization in geological and habitat mapping. Data files are available for download in GeoTIFF format.
Insurance complaint records handled by the Texas Department of Insurance (TDI). The dataset covers complaints against licensed entities such as companies, agents, and adjusters, but excludes the identity of the party complained against.
FD-GAN-Final-FR is a dataset hosted on Kaggle. Its title suggests a focus on financial data generation using Generative Adversarial Networks (GANs). The dataset's specific content, scale, and authorship are not detailed in the available metadata.
A text corpus likely intended for training or fine-tuning BERT models on Chinese language data. The title suggests a focus on OCR correction or repair tasks, possibly involving classical or modern Chinese text. It was published on Kaggle, but the author, organization, and specific collection details are unknown.
GAN3D-Normalized is a dataset related to 3D data generation using Generative Adversarial Networks. It is hosted on Kaggle, but its specific contents, size, and creation details are not documented. The dataset's structure and intended use must be verified after download.
PaddleOCR Arabic Invoice Kaggle Dataset Cleaned is a dataset for optical character recognition tasks. It likely contains images of Arabic-language invoices and corresponding text annotations. The dataset was cleaned and is hosted on Kaggle.
A dataset titled 'FD_GAN_PleaseFinal' hosted on Kaggle. The title suggests a connection to Generative Adversarial Networks (GANs), a machine learning technique for generating synthetic data. No further metadata, such as author, size, or description, is provided.
GAN3D-Unnormalized is a dataset hosted on Kaggle. Its title suggests it contains 3D object data, likely intended for training or evaluating generative adversarial networks. The specific content, size, and creation details are not provided in the available metadata.
vvc_efficientnet_qp_32_2 is a dataset published on Kaggle. Its title suggests a focus on video quality analysis, likely involving metrics from the Versatile Video Coding (VVC) standard and the EfficientNet model architecture. The dataset's specific content, scale, and authorship are not detailed in the available metadata.
A dataset likely containing video data for object detection tasks. It is hosted on Kaggle, but specific details about its size, creation date, and authorship are not provided. The title suggests it is designed for use with the YOLOv8 model architecture.
bc26-local-gpu-mel-cnn-v2-cap40-public is a dataset published on Kaggle. The title suggests it contains performance metrics, likely for benchmarking convolutional neural networks (CNNs) on local GPU hardware. The specific data volume, collection method, and update date are not provided in the available metadata.
Grant awards from the New York State Council on the Arts, beginning in 2003, are detailed in this dataset. It includes funding for non-profit organizations, Indian tribes, and government units within the state. The data is provided by data.ny.gov and was last updated in January 2026.
Meganetpth is a dataset published on Kaggle. The dataset's title suggests a focus on computer vision, likely containing images for model training. Specific details regarding size, content, and creation are unavailable from the provided metadata.
Thirty-four time-series stations were occupied in the Salish Sea during a July 2014 cruise aboard the R/V Clifford A. Barnes. The dataset contains discrete profile measurements of dissolved inorganic carbon, total alkalinity, temperature, salinity, oxygen, and nutrients, collected to study ocean acidification. This effort was conducted by the Washington Ocean Acidification Center and NOAA, adhering to climate-quality monitoring guidelines.