Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
17,661 datasets
3.4 million video clips of 61 frames each, categorized into Cut (C), Transition (T), and Empty (E) classes. The data is sourced from AutoShot, ClipShots, and Pexels, featuring both real-world footage and synthetic transitions centered on the 31st frame.
Approximately 1.8 million synthetically generated CAPTCHA images created by szili2011. Each image contains a random, case-sensitive sequence of letters (a-z, A-Z) and numbers (0-9). The dataset was last updated on June 26, 2025.
This collection of ground-truth forest datasets, curated by blutjens as of August 2025, targets machine learning applications in forestry and climate change. It organizes external resources by domain, including carbon sequestration, biodiversity, and ecosystem health.
20,000 examples form a training subset from a larger merged collection of mathematical datasets. The dataset, created by weijiezz, combines multiple sources of mathematical problems and solutions for AI training and evaluation. It was last updated on June 25, 2025.
Pavlograd City Council's directory of enterprises, institutions, and organizations under its financial management. The dataset likely contains registration and contact details, including USREOU identification codes, names, head information, postal addresses, email addresses, and phone numbers. It was published on the States site of Ukraine and last updated on June 18, 2025.
A 240 GB collection of full image files for GUI grounding training with bounding box supervision, released by Microsoft researchers in 2025. The dataset supports projects focused on user interface understanding and visual grounding tasks.
Niconico Video is a Japanese video sharing service that began in 2006. The dataset is hosted by DSULT-Core and was last updated on July 2, 2025. It likely contains video content from the platform, including the first video ever posted, identified as 'sm9'.
T2I-Diversity Evaluation Prompt Set is a collection of 4,340 English prompts for evaluating text-to-image models. The dataset was created by AIML-TUDA and last updated on June 23, 2025. It contains 1,085 base prompts sourced from DrawBench and Parti-Prompts, each expanded into four variants of increasing descriptive density using GPT-4o.
Llama-4-Maverick-17B-128E-Instruct-FP8 model generated summaries for CNN and DailyMail news articles. The dataset was created by PursuitOfDataScience and last updated on June 24, 2025. Each summary aims to provide a concise and accurate overview of the main story.
Pascal VOC is a benchmark dataset for visual object recognition tasks. It was uploaded to Hugging Face by DerrickUnleashed and last updated on August 1, 2025. The dataset likely contains annotated images for tasks such as classification, detection, and segmentation.
151,000 samples focused on video spatial reasoning designed to reinforce Multimodal Large Language Models (MLLMs). The dataset supports the SpaceR framework, providing data to improve model performance in understanding spatial relationships and dynamics within video sequences.
A YOLO v8 model achieved a mean Average Precision (mAP) of 0.68 on the Eyesea optical dataset and up to 0.65 mAP on unseen sonar data. The submission includes trained model weights, code for five experiments, and documentation for replicating fish detection research. The project aims to improve monitoring around marine energy facilities using optical and sonar imagery.
Chemical-Protein interaction data identifies chemical and protein entities and classifies their likely relations, such as agonists or antagonists. The dataset was created by the bigbio team for the BioCreative VI challenge track. It was last updated on the platform in June 2025.
ViViD-5k is a large-scale vineyard image dataset for grape cluster analysis. It contains 5,000 images across 13 grape varieties with over 648,000 annotated berry centroid keypoints and more than 18,000 grape cluster instance masks and bounding boxes. The dataset was created by XZhi and last updated on Hugging Face in June 2025.
The National Library of Medicine maintains this international standard for arranging medical and scientific library materials, updated biannually in January and August. The classification is a product of the U.S. Department of Health & Human Services, with publication of printed editions ceasing in 1999. It is now distributed in PDF and HTML formats.
Six collections of Scanning Electron Microscope images are provided, augmented with Poisson noise and contrast variations. The dataset includes three trained U-Net AI models for segmentation, alongside CSV files containing image quality metrics and model accuracy evaluations. This work was funded by the CHIPS Metrology Program and published by the National Institute of Standards and Technology in July 2025.
A dataset of 100,000 images of Chinese license plates, published on HuggingFace by zenitsu09. The dataset is formatted for YOLO (You Only Look Once) object detection models. It was last updated on July 28, 2025.
A high-quality dataset for ultra-high-resolution image generation, featuring carefully selected images and captions generated by GPT-4o. The dataset was introduced by author zhang0jhon and was last updated on the Hugging Face platform on June 4, 2025. Low-quality images with motion blur, focus issues, or mismatched text prompts were filtered out through manual inspection.
A dataset hosted by gulucaptain on Hugging Face, focusing on human-centric tasks for generative AI. The data is intended for training models in areas like pose-driven animation and audio-driven action generation. It was last updated on June 13, 2025.
The dataset contains information about fairs operating within the city of Cherkasy. It includes details such as the term of holding, place, number and cost of places, organizers, and contracts concluded with organizers. The data was published by the States site of Ukraine and last updated on June 20, 2025.