Loading...
Loading...
Image classification, object detection, segmentation, face recognition, OCR, image generation, video understanding
16,092 datasets
YOLOData is a dataset published on Kaggle, likely containing images for training and evaluating YOLO (You Only Look Once) object detection models. The dataset's specific contents, size, and origin are not detailed in the provided metadata. Further verification after download is required to confirm the exact number of images, annotation types, and licensing terms.
Red Cowpeas Image Dataset is a collection of images intended for quality grade classification tasks. The dataset is hosted on Kaggle, but details about its size, author, and creation date are unknown. Its primary purpose is to support machine learning models for visual assessment of agricultural produce.
GAN Real Faces 50k FFHQ is a dataset of human face images, likely containing 50,000 samples. It is hosted on Kaggle and appears to be derived from the Flickr-Faces-HQ (FFHQ) dataset, a known benchmark for generative models. The specific author, collection method, and exact contents require verification after download.
A collection of fine-tuned models for text-to-speech synthesis. The dataset is hosted on Kaggle. The specific architecture combines Matcha_TTS and HiFi GAN components.
ViT-P Cityscapes dataset is a collection of urban scene images for semantic segmentation tasks. It is published on Kaggle and likely contains street-view imagery with pixel-level annotations. The dataset's specific size, annotation details, and creation methodology require verification after download.
A computer vision dataset likely containing images of cash for counting tasks. It is published on Kaggle. The specific number of images, creator, and update date are unknown.
Keypoint_No_! is a dataset published on Kaggle. The dataset's title suggests it likely contains data for keypoint detection, a common computer vision task. Metadata is minimal; actual content, scale, and specifics require verification after download.
Keypoint_no_exclamation is a dataset hosted on Kaggle. Its title suggests it likely contains images annotated with keypoints, which are specific points of interest on objects. The dataset's size, origin, and specific content are not detailed in the available metadata.
A dataset titled 'ckp_GuwenBert_nomnaocr_repair_stage2_aug_e1_30_v11' is hosted on Kaggle. Its name suggests it is likely a Chinese-language text corpus, possibly augmented for training a BERT model on OCR correction tasks. The dataset's author, organization, and specific content details are unknown.
Kaggle hosts this dataset for semantic segmentation tasks. The title indicates it is based on the COCO dataset and likely contains images annotated for segmentation. Metadata is minimal; actual content requires verification after download.
35dim Morgan fingerprints are a common cheminformatics representation for molecular structures. The dataset is hosted on Kaggle, a platform for sharing data science projects. Specific details about its size, source, and creation date are not provided in the available metadata.
A dataset titled 'exp21a_25morgan_11节点' is hosted on Kaggle. The title suggests it may contain data related to 11 nodes, potentially from an experiment or simulation. No further metadata is available to confirm its specific contents, size, or origin.
Terrestrial palaeo-environmental proxy data from the Upper Pliensbachian Margaritatus Zone, spanning approximately 350,000 years. The dataset includes high-resolution measurements of charcoal abundance, palynofacies, X-ray fluorescence, mass spectrometry, and clay mineralogy from the Mochras borehole in Wales. Data was collected and analyzed by researchers including Teuntje Hollaar and Claire Belcher.
2015 to 2017 data includes geochemical measurements quantifying carbon source and amount in organic matter from a 6-meter Lake Paringa sediment core and soil samples from Mt. Fox and Alex Knob elevation transects in Westland, New Zealand. The dataset contains measurements of organic carbon concentration, stable isotopic composition (δ13C), total nitrogen content and its isotopic composition (δ15N), radiocarbon activity (F14C), and biomarker abundances for n-alkanes.
This social science dataset by Kelebogile Zvobgo (Journal of Human Rights) tracks treaty reservation withdrawals across autocratic and democratic regimes. It contains statistical observations comparing state responses to technical treaty body reviews and political peer objections.
Functioning as identified as 'analcatdata_michiganacc' from the OpenML platform. No information is available regarding its contents, size, or structure.
Comprising between 10,000 and 100,000 multi-object scenes designed for training text-to-image models with explicit occlusion modeling. Developed by va1bhavagrawa1 for a CVPR 2026 publication, it introduces an Occlusion-aware 3D Scene Representation (OSCR) using translucent 3D boxes.
Aganaanooru is a dataset of Tamil poetry. The collection likely contains 120 poems, as suggested by the title 'Aganaanooru 1-120'. It is published on Kaggle, but the original author and compilation date are unknown.
A model checkpoint from the GuwenBERT project, likely from a training stage involving data augmentation. The title suggests it is associated with training for 30 to 60 epochs. It is hosted on the Kaggle platform.
PPOCRv5 Vietnamese Dataset is a collection for optical character recognition tasks. It is published on Kaggle, but its size, creation date, and specific contents are not detailed in the provided metadata. The dataset's nature suggests it contains images of Vietnamese text paired with transcriptions.