---
dataset_info:
  features:
  - name: qid
    dtype: int64
  - name: perm
    dtype: string
  - name: dataset
    dtype: string
  - name: image
    dtype: image
  - name: question
    dtype: string
  - name: answer
    dtype: string
  splits:
  - name: train
    num_bytes: 556250395
    num_examples: 39858
  download_size: 389472063
  dataset_size: 556250395
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---