---
dataset_info:
  features:
  - name: query
    dtype: string
  - name: document
    dtype: string
  - name: query_tokens
    sequence: int64
  - name: document_tokens
    sequence: int64
  - name: target_mask
    sequence: float32
  - name: is_trimmed
    dtype: bool
  splits:
  - name: train
    num_bytes: 527351957
    num_examples: 85924
  - name: val
    num_bytes: 27828986
    num_examples: 4523
  download_size: 309783650
  dataset_size: 555180943
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: val
    path: data/val-*
---