README.md

metadata

dataset_info:
  config_name: parquet
  features:
    - name: text
      dtype: string
    - name: id
      dtype: string
    - name: url
      dtype: string
    - name: caption
      dtype: string
    - name: width
      dtype: int64
    - name: height
      dtype: int64
    - name: mime_type
      dtype: string
    - name: hash
      dtype: string
    - name: license
      dtype: string
    - name: source
      dtype: string
  splits:
    - name: train
      num_bytes: 8655889565
      num_examples: 12249454
  download_size: 3647461171
  dataset_size: 8655889565
configs:
  - config_name: parquet
    data_files:
      - split: train
        path: parquet/train-*