---
dataset_info:
  features:
  - name: key
    dtype: int32
  - name: image
    dtype: image
  - name: caption
    dtype: string
  - name: qa
    list:
    - name: question
      dtype: string
    - name: answer
      dtype: string
  splits:
  - name: train
    num_bytes: 176256192686.875
    num_examples: 903817
  download_size: 50444416448
  dataset_size: 176256192686.875
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---