---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 21390128617.0
    num_examples: 5522808
  - name: special
    num_bytes: 7763651741.5
    num_examples: 1480380
  download_size: 29304931962
  dataset_size: 29153780358.5
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: special
    path: data/special-*
---