---
dataset_info:
  features:
  - name: subset
    dtype: string
  - name: anchor
    dtype: string
  - name: positive
    dtype: string
  - name: anchor_token_count
    dtype: int64
  - name: positive_token_count
    dtype: int64
  splits:
  - name: train
    num_bytes: 4200948328
    num_examples: 1000603
  download_size: 2433805825
  dataset_size: 4200948328
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---