72 datasets found

The dataset used in the paper is a image captioning task.

High Quality Image-Text Pairs (HQITP) dataset contains 134M high-quality image-caption pairs.

Image captioning is a fundamental task in vision-language understanding, where the model predicts a textual informative caption to a given input image.

The dataset used in the paper to evaluate the effectiveness of context-driven entity-aware captioning.

The Winoground dataset consists of 400 items, each containing two image-caption pairs (I0, C0), (I1, C1).

The VQA dataset is a large-scale visual question answering dataset that consists of pairs of images that require natural language answers.

CC12M dataset is used for training and testing the proposed method. It contains 12 million images with 12 million captions.

Flickr8K-Expert dataset is used for evaluating the proposed method. It contains 8,000 images with 8,000 captions.

The Concept Conjunction 500 (CC-500) dataset is a benchmark for text-to-image synthesis, consisting of 500 images with 500 corresponding text descriptions.

The Attribute Binding Contrast (ABC-6K) dataset is a benchmark for text-to-image synthesis, consisting of 6,000 images with 6,000 corresponding text descriptions.

The Composite dataset, containing 11,985 human judgments over Flickr 8K, Flickr 30K, and COCO captions.

The Flickr30K Entities dataset consists of 31,783 images each matched with 5 captions. The dataset links distinct sentence entities to image bounding boxes, resulting in 70K...

The COCO dataset is a large-scale dataset for object detection and image classification.

The GQA dataset is a visual question answering dataset that characterizes in compositional question answering and visual reasoning about real-world images.

The LAION-Improved-Aesthetics (v1.2) dataset used for training the Stable Diffusion model, which includes images with captions.

The Conceptual Captions 12M (CC-12M) dataset consists of 12 million diverse and high-quality images paired with descriptive captions and titles.

The COCO-captions dataset contains ∼120k RGB images with text captions.

The dataset used in the paper for text-only image captioning with synthetic pairs.

The MS COCO dataset is a large benchmark for image captioning, containing 328K images with 5 caption descriptions each.

Referring image segmentation aims at localizing all pixels of the visual objects described by a natural language sentence.