You can deploy this model and start earning money today!

duongtr/
ReasonVQA

agpl-3.0

1M<n<10M

image

Views

No views yet

ReasonVQA

This is the official repository for ReasonVQA, A Multi-hop Reasoning Benchmark with Structural Knowledge for Visual Question Answering. Our paper has been accepted at ICCV 2025. Visit homepage: https://duong-tr.github.io/ReasonVQA/

Dataset Description

Data about questions for the train.jsonl and val.jsonl in the following format:

question_id: (int) ID of the question question: (str) Content of the question image_id: (str) ID of the image image_name: (str) Name… See the full description on the dataset page: https://huggingface.co/datasets/duongtr/ReasonVQA.