Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
16,977 datasets
176,999 programming conversations translated into Somali, sourced from the glaiveai/glaive-code-assistant-v2 dataset. The dataset was created by michsethowusu and last updated on Hugging Face in October 2025. It aims to make coding education accessible to Somali speakers.
2022-2023 data containing all IRS 990 organizations by Employer Identification Number (EIN) and their matched OpenCorporates unique identifiers. The dataset was created by GivingTuesday and allows linking nonprofit tax data with state-level corporate records. All 990-CN, EZ, and PF organization filings for the period were tested for matches.
A dataset of Spanish financial annual report paragraphs from 2014 to 2018, annotated for causality detection. The data was created for the FinCausal 2023 shared task, with linguists labeling cause and effect elements within each paragraph. It is published by e-cienciaDatos Harvested Dataverse and was last updated in October 2025.
CausalVerse Image Dataset contains two families of splits: Physics splits (Fall, Refraction, Slope, Spring) and Static image generation splits (scene1-4). The dataset is associated with the paper 'CausalVerse: Benchmarking Causal Representation Learning with Configurable High-Fidelity Simulations' and was last updated on 2025-10-23.
IJCNN is a classic benchmark dataset from the 2001 International Joint Conference on Neural Networks competition. It contains data for a binary classification task, likely involving signal or pattern recognition. The dataset's origin and specific size are not detailed in the provided metadata.
176,999 programming conversations, originally from the glaive-code-assistant-v2 dataset, have been translated into the Kanuri language. The dataset was created by michsethowusu and last updated on October 30, 2025. It consists of multi-turn dialogues covering various programming concepts and algorithms.
Npix Open Image Dataset is hosted on Hugging Face by ArkAiLab-Adl. The dataset was last updated on December 13, 2025. Its specific content, scale, and annotations require verification after download.
Narrated (In)justice - Casus 3: Het slavernijverleden is a qualitative dataset containing 9 interviews. The interviews were conducted by J. Chung and archived by DANS Data Station Social Sciences and Humanities Collection, with a last update in October 2025. The data explores how narratives about historical injustice, specifically the Dutch slavery past, are transmitted across generations and groups.
Ukrainian-language prompts and paired responses designed to counter Russian propaganda narratives. The dataset, created by lapa-llm, trains counter-narrative reasoning with fact-based replies grounded in international law and reputable sources. Its last update was recorded in November 2025.
250,000 images sampled from the ImageNet-21K dataset form this multimodal reasoning collection. For each image, the dataset provides a prompt and two different step-by-step reasoning tokens and outputs. Created by krishnateja95 and last updated on October 8, 2025, it is designed for research on multimodal summarization.
Approximately 1.04 million per-frame 2D observations of 557 globally unique static objects across 8 categories, captured on a university campus. The dataset was created by Dongmeyong Lee and last updated on October 15, 2025. It is intended for improving long-term robot perception and navigation under varying environmental conditions.
High-resolution images of English handwritten notes cover algorithm explanations, code snippets, flowcharts, and theoretical content. The dataset is authored by HumynLabs and was last updated on October 23, 2025. It is designed to support AI research in document understanding for computer science education.
UniqueData provides a collection of document images annotated for text detection. The dataset includes scanned papers, forms, invoices, and handwritten notes with varied layouts, fonts, and styles. It was last updated on October 10, 2025.
Qari Arabic OCR 10K is a dataset for optical character recognition of Arabic text, hosted on HuggingFace. The dataset was uploaded by author melsiddieg and was last updated on December 7, 2025. The specific content, scale, and collection methodology are not detailed in the available metadata.
The EPFL-Smart-Kitchen-30 dataset provides a benchmark for action recognition. It likely contains video clips of kitchen activities annotated with action labels. The dataset was created by amathislab and was last updated on October 23, 2025.
Stpls3D provides synthetic and real-world 2D and 3D point cloud data for semantic and instance segmentation, published by meidachen for BMVC 2022. The dataset utilizes AirSim for synthetic generation alongside real-world photogrammetry captures to support 3D reconstruction benchmarks.
Long-term Occupational Projections from data.ny.gov provide a 10-year forecast of job openings across New York State and its 10 labor market regions. The dataset includes annual estimates for total openings, growth openings, and replacement openings for detailed occupations. It was last updated in September 2025.
SPSLingAnc provides a mixed-methods analysis of complete fictional narratives in English, focusing on linguistic tokens that shape readers' conceptualization of storyworld possible selves. The dataset comprises 91,328 words from two novels, with 5,980 tagged tokens analyzed using MAXQDA 2022 software. It was authored by María-Ángeles Martínez and last updated on October 14, 2025.
17,091,657 face images of 360,232 distinct individuals comprise this collection for training face recognition models. The dataset was cleaned and released by yayoimizuha, with the latest update recorded in October 2025. It is designed to support state-of-the-art performance on benchmarks like IJB-C and MegaFace.
The Kinetics-700 dataset is a large-scale collection of YouTube video URLs for human action recognition, containing 700 distinct action classes. It is an extension of the original Kinetics dataset, with each video clip being approximately 10 seconds long and depicting a single human action. The dataset was uploaded by author 'bitmind' to the Hugging Face platform, with a last recorded update in October 2025.