---
dataset_info:
  features:
  - name: caption
    dtype: string
  splits:
  - name: train
    num_bytes: 4858831
    num_examples: 21805
  download_size: 1021737
  dataset_size: 4858831
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---