---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: image_id
    dtype: int64
  - name: new_image_id
    dtype: int64
  - name: answers
    sequence: string
  splits:
  - name: train
    num_bytes: 8209032325.0
    num_examples: 316219
  download_size: 8106455712
  dataset_size: 8209032325.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---