---
dataset_info:
  features:
  - name: image
    dtype: string
  - name: question
    dtype: string
  - name: question_id
    dtype: int64
  - name: image_id
    dtype: int64
  - name: ground_truth
    list: string
  - name: majority_vote
    dtype: string
  splits:
  - name: train
    num_bytes: 34330024
    num_examples: 214354
  download_size: 11738950
  dataset_size: 34330024
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---