Loading...
Loading...
Text classification, translation, QA, summarization, dialogue, sentiment analysis, language modeling, text corpora
46,429 datasets
Cuarzo-100K v2 contains 99,683 bidirectional, deterministically verified pairs of Python code and human language descriptions. Released by Cuarzo AI, this second public release from their Aether engine adds Mandarin Chinese as a full fourth language and expands verification to per-language roundtrip checks across English, Spanish, French, and Mandarin.
Biosiliceous sediments from a submarine valley system on the East Antarctic Shelf contain intervals of ripple cross-lamination. Based on radiocarbon dates from one core, the most intensely cross-laminated sediments were deposited between 6,000 and 3,500 years before present. The dataset is provided by Geoscience Australia Data and was last updated on 2026-04-30.
Northern Ireland's Ulsterbus and Goldliner timetable data, provided in the ATCO.cif standard format used widely within the public transport industry. The dataset is published by OpenDataNI under the OGL-UK-3.0 license and was last updated on 22 April 2026. The primary file can be opened as a text file, with downloadable formats including PDF, ZIP, and DOCX.
Over 15 years, Japanese, Australian, and Russian expeditions collected data used to compile surficial sediment, bathymetric, and geomorphological maps of the Prydz Bay region in East Antarctica. The underlying data includes approximately 30,000 kilometers of bathymetric data, 6,000 kilometers of sidescan sonar data, and more than 250 sediment grab and core samples. The dataset is provided by Geoscience Australia Data and was last updated in April 2026.
Global daily snow cover maps derived from the Visible Infrared Imager Radiometer Suite (VIIRS) on the JPSS-2 satellite. The data is processed using a cloud-gap-filled algorithm to estimate snow under cloud cover and is mapped to a 375-meter sinusoidal grid. This dataset is hosted on AWS Open Data and is provided under a CC-BY-4.0 license.
WindSat Polarimetric Radiometer data provides ocean surface wind vector measurements from space. The instrument was launched on January 6, 2003 aboard the Coriolis satellite and developed by the Naval Research Laboratory for the U.S. Navy and NPOESS. This Level 1C dataset contains Top of the Atmosphere brightness temperatures processed by RSS, resampled on a 0.125-degree Earth grid with calibration using the GMI sensor and other targets.
Fisheries and Oceans Canada provides two geotiff layers modeling the habitat of Dungeness crab (*Metacarcinus magister*). The first layer shows the mean coastwide probability of crab detection predicted from a geostatistical model, while the second layer represents the uncertainty in those predictions. The data was last updated on March 24, 2026, and is based on an analysis integrating dive, trawl, and baited-trap surveys.
Sections 2.1 and 2.2 of the official Land Use and Development Plan for the Montreal Agglomeration council, guiding sustainable urban planning decisions. The plan prioritizes consolidating commercial centers and axes from Map 4 and leverages unique regional economic clusters from Map 8 for growth. It outlines parameters for compact neighborhoods, increased public transit, and economic dynamism from a sustainable development perspective.
A training dataset associated with the paper 'Crosslingual On-Policy Self-Distillation for Multilingual Reasoning'. It is designed to improve mathematical reasoning in low-resource languages using the COPSD method. The dataset was created by yihongLiu and was last updated on Hugging Face in May 2026.
MECAT is a benchmark constructed on large-scale data to evaluate machine understanding of audio content. It focuses on two core tasks: generating textual descriptions for audio and answering questions about audio. The dataset was created by mispeech and last updated on 2026-05-11.
27 hauls were completed over 11 days of fishing trials in November and December 2017. The work, part of the Fisheries Science Partnership programme, compared catch rates of five different cod ends to develop an optimum construction for reducing small Nephrops catches. The trials were conducted in the Farne Deeps fishing grounds off the North-East coast of England.
The Fisheries Science Partnership study explores different approaches to fully documenting fisheries catch, as required by the proposed Common Fisheries Policy reform. The project aimed to establish data requirements, evaluate collection mechanisms like Remote Electronic Monitoring and observers, and assess the suitability of different approaches for various fishery types. The dataset originates from the Government Digital Service via the eu_open_data platform.
A Fisheries Science Partnership project chartered the vessel *Harmoni* to test the commercial and environmental viability of electro-fishing for razor clams. The project, delayed to winter 2013/14, conducted fieldwork off Southport and another site in May and later that year. Despite gear functioning as designed, no razor clams were caught, leading to the project's abandonment.
A bathymetry survey acquired for the Australian Hydrographic Office on 11 December 2022. The dataset provides separate grids in 0.5m and 2m resolution for two surveyed sites near Rottnest Island, WA, in MSL, LAT, and Ellipsoid vertical datums. The data was processed using Caris HIPS & SIPS and exported as 32-bit floating point GeoTIFF grids.
An investigation of beach-sand heavy-mineral deposits conducted by the Bureau of Mineral Resources between 1948 and 1950. The work involved detailed boring and sampling of beaches, coastal dunes, and plains along the coast from the Clarence River, New South Wales, to North Stradbroke Island, Queensland. Results related to changing sea levels are summarized in the publication.
An 800-square-mile area in the Western Highlands of New Guinea was mapped during a reconnaissance geological survey in 1959. The data describes a complex geological structure where Jurassic and Cretaceous rocks are faulted against an Eocene to Pliocene sequence, with folding and major faults. The youngest features are Pleistocene volcanics and alluvial deposits, and a large basic pluton is identified as a source of gold, platinum, and minor copper mineralization.
A critique of magnetostratigraphic coverage for the entire Phanerozoic eon is presented. The document, from Geoscience Australia Data, assesses constraints from the Cambrian to the Recent period, noting well-established data from the Triassic onward. It was last updated on 2026-04-20.
Geoscience Australia Data compiled disparate information on Australia's near-pristine estuaries into a national resource. The investigation synthesized data from scientific articles, reports, government sources, and expert advice, summarized on both state and national levels. The dataset was last updated on April 20, 2026.
A geological report from Geoscience Australia details results from 22,355 feet of scout boring over approximately 50 square miles on the western flank of the Muswellbrook anticline in New South Wales. The report provides a more complete section of the Upper Coal Measures than previously available, documenting 46 encountered coal seams and the influence of igneous intrusions. It was last updated on 2026-04-20.
Over 30 species and 16 genera of Permian Polyzoa are documented from the large, triangular Bowen Basin in eastern Queensland. Four species, including Fenestella bowenensis and Saffordotaxis multinodata, are described as new, suggesting a short-lived marine connection with Western Australia. The dataset details the basin's geological structure, composed of Permian and Triassic sediments over 20,000 square miles.