README.md

metadata

dataset_info:
  features:
    - name: anchor
      dtype: string
    - name: positive
      dtype: string
    - name: negative
      dtype: string
  splits:
    - name: train
      num_bytes: 20550568977
      num_examples: 25863649
    - name: validation
      num_bytes: 419477179
      num_examples: 527832
  download_size: 13638899915
  dataset_size: 20970046156
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*