Loading...
Loading...
Student performance, MOOC logs, knowledge tracing, standardized tests, learning analytics
13,904 datasets
Omniact Grounding Filtered is a subset of the Omniact dataset, containing only grounding task samples. It was created by InfiX-ai as an example dataset for training their InfiGUI-G1 model. The dataset was last updated on August 18, 2025.
State of Connecticut provides a listing of licensed automobile dealers and repairers, updated nightly. The data includes business names and potential Doing-Business-As (DBA) names for facilities.
Soundata is a Python library for downloading and loading environmental, urban, and bioacoustic datasets, maintained by the Soundata organization and updated in September 2025. It provides a unified API to access raw audio and metadata from multiple research repositories, standardizing the interface for diverse sound data sources.
314,000 synthetic variational problems generated by the Qwen2.5-32B-Instruct model during 600 steps of RLVR training. Each entry pairs a synthesized mathematical problem with a reference answer generated via the SvS strategy on the DAPO-17k dataset.
Svg Metadata is a dataset created by gitcat-404 and last updated on September 7, 2025. The dataset's specific content and size are not detailed in the provided metadata. It is licensed under CC BY-NC 4.0, permitting research and educational use while prohibiting commercial use without explicit permission.
Japan's Ministry of Finance published this document in June 2021. The content likely contains educational information about the Japanese tax system. The author is listed as the Tax Bureau's General Affairs Division.
Contains approximately 13,100 examples formatted in Parquet for efficient processing. Provided by MiroMind AI, this dataset is curated for reinforcement learning fine-tuning within the MiroRL framework and was last updated in August 2025.
MOSE is a video object segmentation benchmark introduced at ICCV 2023 by Henghui Ding for tracking and segmenting objects in challenging environments. The dataset provides pixel-level masks for objects within video sequences characterized by heavy occlusions and crowded scenes.
4,649 examples of repair and teardown procedures for smartwatches and other wearable devices. The dataset includes metadata and images, and was created by AS-Robotics. It was last updated on August 15, 2025.
Humanity's Last Exam (HLE) Bio/Chem Gold is a curated subset of a challenging question-answering benchmark for advanced academic fields. The dataset contains biology and chemistry questions audited by FutureHouse, which found approximately 30% of original answers were contradicted by peer-reviewed sources. It was created by FutureHouse and last updated in July 2025.
MLX is a machine learning asset catalog and execution engine developed by the claimed-framework organization, with updates recorded through September 2025. It provides a structured repository for models, pipelines, and Jupyter notebooks specifically designed for Kubernetes and OpenShift environments.
School campus centroids throughout Pierce County, Washington. The dataset includes public and private primary and secondary schools, universities, colleges, and technical schools, sourced from open.piercecountywa.gov. It was last updated on July 11, 2025.
2023 school district boundaries in effect for the 2022-2023 school year, as of January 1, 2023. The dataset contains polygon shapefiles for unified school districts in Arkansas, extracted from the U.S. Census Bureau's national Master Address File/TIGER Database. It provides geographic and administrative information for single-purpose administrative units providing public educational services.
100 candidate documents per query from the MS MARCO train split annotated with re-ranking scores from bi-encoder and cross-encoder teacher models. The dataset utilizes the opensearch-neural-sparse-encoding-doc-v1 model for initial retrieval to identify relevant hard negatives for training retrieval systems. These annotations facilitate the development of neural search models through knowledge distillation and contrastive learning.
This dataset contains information on projects supported by Emergency Management Performance Grants (EMPG) recipients as reported in the Grant Reporting Tool (GRT). It is maintained by the Department of Homeland Security and FEMA, with data from fiscal year 2016 to the present, and is updated twice a year.
The medical-exams-PES-PL-2007-2024 dataset is described in a 2024 arXiv preprint by researchers including Εukasz Grzybowski. It is a Polish medical exam dataset intended for assessing cross-lingual medical knowledge transfer. The dataset page is hosted on Hugging Face by the author amu-cai and was last updated in August 2025.
A collection of YouTube video clips with associated speaker identification annotations. The dataset, authored by 'dorni', includes a list of video IDs and unique clip identifiers for locating annotations. It was last updated on August 4, 2025.
Live-WhisperX-526K contains 526,000 video-text records developed by chenjoya in 2025 to train the LiveCC-7B-Instruct model. The collection features synthetic user prompts generated by GPT-4o to support multimodal instruction tuning for video-to-text tasks.
Gopher-Lab provides an example dataset demonstrating a TikTok scraper tool. The tool generates AI-ready transcripts and metadata from TikTok videos via URL input. The dataset page was last updated on August 6, 2025.
EdSight, a data portal from the State of Connecticut, integrates information from over 30 sources including district reports and external data. It provides information on key performance measures for the Next Generation Accountability System, school finance, special education, staffing levels, and school enrollment.