---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: 'Unnamed: 0'
    dtype: int64
  - name: caption
    dtype: string
  - name: url
    dtype: string
  splits:
  - name: train
    num_bytes: 4928465872.772
    num_examples: 24086
  download_size: 6855021237
  dataset_size: 4928465872.772
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---


PART 2