Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
17,661 datasets
Over 130,000 books published in the United States before 1929, which entered the public domain in 2024. The collection was identified using the HathiTrust bibliographic catalog and contains digitized OCR plain text files downloaded from the Internet Archive. It was compiled by common-pile and last updated on Hugging Face in June 2025.
Information on the structure of the Department of Case Management and Legal Support of Cherkasy City Council. The dataset is published on the EU Open Data portal by the States site of Ukraine and was last updated on June 9, 2025. The data is provided in CSV format.
LHPR-VLN is a benchmark for long-horizon planning and reasoning in Vision-Language Navigation, introduced in the paper 'Towards Long-Horizon Vision-Language Navigation: Platform, Benchmark and Method'. The dataset is hosted by author Starry123 on Hugging Face and was last updated on May 30, 2025. It is also available on the ModelScope platform.
A version of the peS2o dataset restricted to openly licensed articles, derived from the S2ORC corpus of academic papers. The dataset was created by the common-pile organization and was last updated on June 6, 2025. It contains papers converted to a structured format using Grobid, with filtering applied for quality and language.
10 million enriched events from 10,000 high-elo League of Legends matches, provided by the GPTilt open-source initiative. The dataset is aimed at democratizing access to high-quality LoL data for research and AI development. It was last updated on May 28, 2025.
H3DS provides 3D human head scans and Python utilities for 3D vision tasks, developed by CrisalixSA and updated in July 2025. The repository focuses on high-fidelity facial geometry for deep learning research and mesh processing.
A large-scale multi-style image translation dataset created by showlab and last updated on May 29, 2025. It contains aligned image pairs for 22 distinct artistic styles, each consisting of a source image, a stylized target image, and a descriptive text prompt. The dataset is suitable for style transfer and conditional generation tasks.
BioTrove is a large curated image dataset enabling AI for biodiversity, comprising well-processed metadata with full taxa information and URLs pointing to image files. The dataset is hosted by BGLab on HuggingFace and was last updated on 2025-05-13. The metadata can be used to filter specific categories, visualize data distribution, and manage imbalance effectively.
1 utility for transforming object detection annotations from JSON format into the YOLO-compatible text format. The tool processes bounding box data and class indices to ensure compatibility with YOLOv5, YOLOv8, and other Darknet-based models.
GSEval is a benchmark dataset containing 3,800 images, curated by hustvl and last updated on 2025-05-20. It is designed to evaluate the performance of AI systems in pixel-level and bounding box-level grounding based on natural language descriptions.
Cityscape-Adverse extends the original Cityscapes dataset by applying eight realistic environmental modifications—rainy, foggy, spring, autumn, snowy, sunny, night, and dawn—using diffusion-based image editing. All transformations preserve the original 2048×1024 semantic labels, enabling direct evaluation of model robustness in out-of-distribution scenarios. The dataset was created by author 'naufalso' and was last updated on Hugging Face on May 14, 2025.
The StoryReasoning dataset contains 4,178 cohesive visual stories derived from 52,016 images. It organizes temporally connected image sequences extracted from the same movie scenes to ensure narrative coherence. The dataset was created by author daniel3303 and was last updated on the Hugging Face platform in May 2025.
2,474,584 rows of molecular data pairing canonical SMILES strings with natural language descriptions. The dataset is structured into a single training split containing unique identifiers, chemical structures, and textual explanations for each compound.
Information on the quarantine state and spread of regulated pests in Ukraine. The dataset was published by the States site of Ukraine and was last updated on May 30, 2025. Its specific row count and column details are not provided in the metadata.
20.4 billion tokens of question-answer pairs from Stack Exchange, transformed into Markdown format. The dataset, created by marin-community, preserves technical discussion content organized into threads. It was last updated on 2025-05-18 and is sourced from the Stack Exchange archive.
A large-scale dataset of 2.5 million training and 100,000 test images generated using 25 diffusion models, created by lesc-unifi and released in May 2025. The dataset is designed to include images from both recent and older, well-established generative architectures.
299 real-world robot demonstrations and synthetic vision-language data across two tasks: Cocktail and Open-World Visual Grounding. The collection includes reasoning annotations for each demonstration and is formatted for the LeRobot ecosystem to support the development of adaptive Vision-Language-Action models.
Parallel sentences in English and Luganda designed for training machine translation models. The dataset was created by author 'kambale' and was last updated on May 30, 2025. Sentence pairs were extracted from a source document.
A directory of enterprises, institutions, and territorial bodies managed by the Department of Social Policy of the Kryvyi Rih City Council. The data is published by the city council's executive committee on the eu_open_data platform. It was last updated on 2025-05-30.
A synthetic dataset for training Optical Character Recognition models on the Myanmar language. The images were generated using a fork of TextRecognitionDataGenerator with fixes for Myanmar character splitting. The dataset was created by chuuhtetnaing and last updated on May 17, 2025.