README.md

metadata

dataset_info:
  features:
    - name: image
      sequence: image
    - name: question
      dtype: string
    - name: answer
      dtype: string
    - name: dataset
      dtype: string
  splits:
    - name: train
      num_bytes: 6179907012
      num_examples: 21953
  download_size: 6175685752
  dataset_size: 6179907012
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - visual-question-answering
language:
  - fra
license: mit

Description

This dataset is a processed version of the TextCaps dataset by Sidorov, Hu, Rohrbach and Singh, particularly for a visual question answering task where answer is a caption.
We converted the images to PIL format (image column of the dataset); created 55 different questions in French (question column); and finally translated the original captions from English to French and used them as answers (answer column).

Citation

TextCaps

@misc{sidorov2020textcapsdatasetimagecaptioning,
      title={TextCaps: a Dataset for Image Captioning with Reading Comprehension}, 
      author={Oleksii Sidorov and Ronghang Hu and Marcus Rohrbach and Amanpreet Singh},
      year={2020},
      eprint={2003.12462},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2003.12462}, 
}