---
dataset_info:
  features:
  - name: jpeg_bytes
    sequence: uint16
  - name: caption
    sequence: uint16
  splits:
  - name: train
    num_bytes: 53991248422
    num_examples: 13759351
  download_size: 54378225122
  dataset_size: 53991248422
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---