16 datasets found

Fine-grained image classification, which aims to distinguish images with subtle distinctions, is a challenging task due to two main issues: lack of sufficient training data for...

The dataset used in the paper is SUN Attribute, which consists of 717 classes of images with annotations.

Automatic Pronunciation Assessment (APA) is vital for computer-assisted language learning. Prior methods rely on annotated speech-text data to train Automatic Speech Recognition...

The dataset used in the paper for zero-shot medical image classification, pneumonia, tuberculosis, retinopathy, and brain tumor.

Zero-Shot Temporal Action Detection via Vision-Language Prompting (STALE) model for the under-studied yet practically useful zero-shot temporal action detection (ZS-TAD)

The CRL model is based on class-level classifiers, built class-by-class, that would be a representative of instances of a specific class by utilizing activation features of...

The paper discusses the use of multi-objective Bayesian optimization for hyperparameter transfer in topic models.

GeneCIS is a benchmark for Compositional Image Retrieval, which aims to retrieve the relevant target image in a database.

CIRCO is a benchmark for Compositional Image Retrieval, which aims to retrieve the relevant target image in a database.

Compositional Image Retrieval through Vision-by-Language (CIReVL) is a training-free approach for Zero-Shot Compositional Image Retrieval (CIR). Utilizing off-the-shelf...

ODIN is an innovative approach that addresses the problem of dataset constraints by integrating generative AI models.

CIRR is a general image dataset that comprises 36,554 triplets derived from 21,552 images from the popular natural language inference dataset NLVR2.

The FashionIQ dataset contains images of fashion products over 3 categories: Dress, Toptee, and Shirt, with 46,609 images in the training and 31,075 images in the validation set.

The Imagenette dataset is a zero-shot image classification dataset, containing 13,394 images from ten easily separable classes in ImageNet.

The proposed SC VALL-E model uses a Korean grapheme-to-phoneme converter to extract phonemes from Korean text and use them for training.

The C-GQA dataset is a large-scale dataset for compositional zero-shot learning, containing 413 attributes and 674 objects.