Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
17,661 datasets
195 flag images for all sovereign nations, formatted as PNG files at 640x427 pixels. The dataset is intended for image classification tasks, using English country names as labels. It was created by Aniket96 and last updated on 2025-04-21.
A synthetic video dataset created by Max-Ploter for object detection and tracking tasks. It features moving MNIST digits with 1-10 digits per sequence, 20 frames per video, and per-frame annotations including digit labels and center coordinates. The dataset was last updated on Hugging Face on 2025-04 09.
Composed of image files organized into two distinct categories: AI-generated images and real-world photographs. The data is structured into multiple batches to facilitate efficient processing and hosting within the Hugging Face ecosystem.
A specialized collection of tool images combines visual data with detailed safety and usage metadata. The dataset, created by akameswa, includes bounding box annotations and is split into train, test, and validation sets. It was last updated on April 16, 2025.
Video-R1 provides instruction-tuning data for reinforcing video reasoning in Multimodal Large Language Models. The repository includes a 165k example JSON file for supervised fine-tuning and a 260k example file for reinforcement learning training, sourced from datasets like CLEVRER and LLaVA-Video-178K. It was uploaded by the Video-R1 organization on April 11, 2025.
NuScenes Depth Estimation is a dataset hosted on HuggingFace by the author 'oges'. The dataset was last updated on 2025-05-31. Its specific content and scale must be verified after download, as metadata is minimal.
Annual records from fiscal years 2014 to 2023 detail tons of food scraps and other organic waste collected via drop-off programs in New York City. The dataset tracks contributions from both the Department of Sanitation (DSNY) and nonprofit partners, with separate columns for each. It is published by the City of New York's open data portal.
Coderonion curated this repository of YOLO object detection projects and datasets, updated through May 2025. It indexes resources for multiple YOLO versions, including YOLOv5 and YOLOv8, alongside specialized implementations for few-shot and open-world detection.
A collection of hyperspectral image datasets, curated by Sellifake and hosted on GitHub. The datasets are licensed under the MIT license and were last updated on May 30, 2025.
A multiview sports video dataset released by ember-lab-berkeley on 2025.04.14. The dataset likely contains clipped match videos alongside synchronized tracking data for balls, paddles, and human keypoints in 2D and 3D. The data structure suggests it is intended for computer vision tasks in sports analytics.
ContactPose contains 2.9 million RGB-D grasp images paired with hand and object pose data, developed by Facebook Research. The collection includes synchronized contact maps and MANO hand model parameters for various hand-object interactions. It serves as a benchmark for understanding how humans manipulate objects using visual and tactile data.
Over 2.16 million labeled images of Arabic text extracted from diverse sources, created by mssqpi and last updated on March 14, 2025. The dataset is designed to enhance Optical Character Recognition capabilities for the Arabic language and is stored in Parquet format with a total size of 1.87 GB.
Socratic is a text dataset for instruction tuning and question answering, created by FreedomIntelligence and hosted on Hugging Face. The dataset falls within the 10k to 100k sample size category and was last updated in June 2025. It is released under the Apache 2.0 license.
UniDataPro's License Plate Detection dataset features over 1,200,000 images of license plates from more than 32 countries. It includes OCR data, bounding box labels, and corresponding masks for recognition tasks, focusing on plate detection and character recognition systems. The dataset was last updated on Hugging Face on April 4, 2025.
A telephone directory from the Department of Economics of the Kyiv Regional State (Military) Administration lists enterprises, institutions, and organizations under its management. The dataset includes identification codes, official websites, email addresses, telephones, and physical addresses. It was last updated on 2025-04-09 and is hosted on the States site of Ukraine.
Over 5,000 images of water meters, including segmentation masks and OCR labels for meter readings. The dataset is designed for research in water consumption analysis and smart meter technology, providing insights into residential and commercial water usage. It was created by UniDataPro and last updated on 2025-04-04.
EPFL-ECEO released Coralscapes in 2025, providing 2,075 high-resolution images for dense semantic segmentation of coral reef environments. The dataset adopts the Cityscapes structure to facilitate benchmarking in the domain of underwater scene understanding.
Fatdove's Iris Database contains 17,695 high-quality synthetic colored iris images generated using diffusion models. The dataset is designed to be biometrically unique from its training data while maintaining realistic iris pigmentation distributions. The repository was last updated on March 24, 2025.
Ukraine's official dataset contains information on normative legal acts and acts of individual action adopted by the information manager, as well as draft decisions for discussion and documents defining responsible persons. The data is provided by the States site of Ukraine and was last updated on April 1, 2025. The dataset is available in CSV format.
ASCII Art is a text-to-image dataset where the images are composed of ASCII characters. The dataset aggregates content from multiple sources, including independent artists, common Twitch emotes, styled text samples, and conversions from the DataCompDR-12M dataset. It was created by apehex and last updated on March 30, III.