24,167 datasets found

Organizations: No Organization Formats: JSON

Filter Results

  • FashionMNIST

    FashionMNIST is an image dataset of clothing items, used here to evaluate the performance of STN and P-STN models in recovering transformations and augmentations.

  • Stanford 3D Semantic Parsing Dataset

    The Stanford 3D semantic parsing dataset contains 3D scans from Matterport scanners in indoor areas with full annotations for various semantic classes including structural...

  • ShapeNet Part Segmentation Dataset

    ShapeNet part segmentation dataset contains 16,881 shapes from 16 categories, annotated with 50 parts in total, where each part category label is defined for each 3D point.

  • KITTI Object Detection

    The KITTI Object Detection dataset provides a comprehensive set of 2D object detection data for real-world driving scenarios, particularly focusing on car detection.

  • Bosch Small Traffic Lights Dataset

    The Bosch Small Traffic Lights Dataset presents a challenge for detecting small objects with partial occlusions, especially useful for localizing small objects under weak...

  • MIOvision Traffic Camera Dataset (MIO-TCD)

    MIOvision Traffic Camera Dataset (MIO-TCD) is the largest public benchmark for object detection in traffic surveillance images, with a vast array of annotated images for training...

  • WMT2017

    The WMT17 dataset is used for neural machine translation tasks, providing data for multiple language pairs.

  • IWSLT2014

    The IWSLT14 dataset is used for neural machine translation tasks, containing various parallel text translations.

  • CIFAR10

    The CIFAR10 dataset is used for training and evaluating deep neural networks, specifically in this study for assessing the performance of decision gates in ResNet-101 and...

  • SQuAD dataset

    The dataset used for training BERT consists of a concatenation of Wikipedia and BooksCorpus, specifically focused on the SQuAD task.

  • VIPeR

    VIPeR dataset consists of 632 persons with two images captured from different cameras, used for person re-identification tasks.

  • CUHK01

    CUHK01 is a medium-sized dataset containing 3,884 images of 971 identities, intended for testing person re-identification methods.

  • CUHK03

    CUHK03 is a large dataset containing 13,164 images for 1,360 identities captured by 6 cameras. It includes both detected and labeled images for training and testing.

  • PennTreebank

    The PennTreebank dataset is used for language modeling, containing a large annotated corpus of English text to evaluate the task of predicting the next character or word based...

  • Nottingham

    The Nottingham dataset contains British and American folk tunes and is used to evaluate models' capabilities in polyphonic music modeling.

  • Pano2Vid

    Pano2Vid is a real-world 360-degree video dataset containing videos from several categories. The frames are sampled and resized to 640x320 resolution for training and testing,...

  • Spherical MNIST

    Spherical MNIST is constructed from the MNIST dataset by back projecting the digits into equirectangular projection with a resolution of 160x80. The digit labels are used to...

  • 1 Billion Word Language Model Benchmark

    The 1 Billion Word Language Model Benchmark is a dataset used for measuring progress in statistical language modeling, consisting of a large collection of text data.

  • Caltech-UCSD Birds 200 dataset (CUB-200)

    The 2011 Caltech-UCSD Birds 200 dataset (CUB-200) contains 11,788 images of 200 different types of birds, widely used as a benchmark for text-to-image generation.

  • TaoMultimodal Dataset

    A large-scale dataset for multi-modal pretraining in Chinese, consisting of 3.1M image-text pairs from the mobile Taobao platform.