README.md

metadata

dataset_info:
  features:
    - name: image
      dtype: image
    - name: question
      dtype: string
    - name: answer
      dtype: string
    - name: dataset
      dtype: string
  splits:
    - name: train
      num_bytes: 11216749228.01
      num_examples: 48723
    - name: test
      num_bytes: 2729350601.82
      num_examples: 12186
  download_size: 2286399320
  dataset_size: 13946099829.83
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
task_categories:
  - visual-question-answering
  - image-feature-extraction
language:
  - fr
license: apache-2.0

Description

cmarkea/doc-vqa dataset that we processed.

Citation

@online{SoSoDocvqa,
  AUTHOR = {Loïc SOKOUDJOU SONAGU, Yoann SOLA},
  URL = {https://huggingface.co/datasets/cmarkea/doc-vqa},
  YEAR = {2024},
  KEYWORDS = {NLP ; Multimodal}
}