---
dataset_info:
  features:
  - name: query
    dtype: int64
  - name: positive
    dtype: int64
  - name: negative
    dtype: int64
  - name: score
    dtype: float64
  splits:
  - name: train
    num_bytes: 2554621344.0
    num_examples: 79831917
  - name: eval
    num_bytes: 320000.0
    num_examples: 10000
  download_size: 2240483855
  dataset_size: 2554941344.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: eval
    path: data/eval-*
---