README.md

metadata

license: apache-2.0
dataset_info:
  features:
    - name: question_id
      dtype: int64
    - name: image
      dtype: image
    - name: question_ml
      dtype: string
    - name: answer_ml
      dtype: string
    - name: reason_ml
      dtype: string
  splits:
    - name: train
      num_bytes: 4657449488
      num_examples: 29459
    - name: validation
      num_bytes: 231826615
      num_examples: 1459
    - name: test
      num_bytes: 307451135
      num_examples: 1968
  download_size: 4496686992
  dataset_size: 5196727238
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*

This is the Malayalam translated version of the VQA-X dataset.