---
dataset_info:
  features:
  - name: document_id
    dtype: int64
  - name: document
    dtype: string
  - name: description_candidates
    list: string
  - name: document_overflows_ctx
    dtype: bool
  splits:
  - name: train
    num_bytes: 2973441
    num_examples: 1638
  download_size: 1563035
  dataset_size: 2973441
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---