---
dataset_info:
  features:
  - name: question
    dtype: string
  - name: image
    dtype: image
  - name: answer
    dtype: string
  - name: valid
    dtype: bool
  - name: answer_type
    dtype: string
  - name: source
    dtype: string
  splits:
  - name: train
    num_bytes: 246478780
    num_examples: 5600
  download_size: 245045331
  dataset_size: 246478780
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---