---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: alt_text
    list: string
  - name: sentids
    list: string
  - name: split
    dtype: string
  - name: img_id
    dtype: string
  - name: filename
    dtype: string
  - name: original_alt_text
    sequence: string
  splits:
  - name: train
    num_bytes: 4047847550.654543
    num_examples: 29000
  - name: validation
    num_bytes: 141534807.51488647
    num_examples: 1014
  - name: test
    num_bytes: 139580694.02257046
    num_examples: 1000
  download_size: 4307950179
  dataset_size: 4328963052.191999
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---