GQA-REX

Name: GQA-REX
Published: 2022-03-11
License: MIT license

Dataset Information

Modalities

Images, Texts

Languages

English

Introduced

2022

License

MIT license

Homepage

Official Website

Contents

Overview

A GQA-based dataset with 1,040,830 multi-modal explanations of visual reasoning processes.

Variants: GQA-REX

This dataset is used in 1 benchmark:

Explanatory Visual Question Answering - Metrics: BLEU-4, CIDEr, GQA-test, GQA-val, Grounding, METEOR, ROUGE-L, SPICE

Task	Model	Paper	Date
Explanatory Visual Question Answering	REX-LXMERT	REX: Reasoning-aware and Grounded Explanation	2022-03-11
Explanatory Visual Question Answering	REX-VisualBert	REX: Reasoning-aware and Grounded Explanation	2022-03-11
Explanatory Visual Question Answering	EXP	Faithful Multimodal Explanation for Visual …	2018-09-08
Explanatory Visual Question Answering	VQAE	VQA-E: Explaining, Elaborating, and Enhancing …	2018-03-20

Recent papers with results on this dataset:

External Links: