Datasets:

lbourdois
/

VQA-vistext

dataset_info:
  features:
    - name: image
      sequence: image
    - name: question
      dtype: string
    - name: answer
      dtype: string
    - name: dataset
      dtype: string
  splits:
    - name: train
      num_bytes: 550798005.875
      num_examples: 9969
  download_size: 386323418
  dataset_size: 550798005.875
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - visual-question-answering
language:
  - fra
license: gpl-3.0

Description

This dataset is a processed version of the VisText dataset by Tang, Boggust and Satyanarayan.
We converted the images to PIL and translated question and answers from Englist to French.

Citation

VisText

@inproceedings{2023-vistext,
  title = {{VisText: A Benchmark for Semantically Rich Chart Captioning}},
  author = {Benny J. Tang AND Angie Boggust AND Arvind Satyanarayan},
  booktitle = {The Annual Meeting of the Association for Computational Linguistics (ACL)},
  year = {2023},
  doi = {10.18653/v1/2023.acl-long.401},
  url = {https://vis.csail.mit.edu/pubs/vistext}
}