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: 1871914440.98
      num_examples: 7654
  download_size: 787800310
  dataset_size: 1871914440.98
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
task_categories:
  - visual-question-answering
language:
  - fr

Description

To be honest, I don't remember where I found the data.
They were lying around on my computer in a file called “IA-Factory”, hence the name of the repo.
Of the 7,654 data, 6,420 were already in French and 1,234 were in English. I then translated these 1,234 data (these are the last 1,234 rows if you only want to work on data in native French rather than native+translation).
The dataset column is for internal use by the person who processed the dataset, and can therefore be ignored by the user.