---
dataset_info:
  features:
  - name: input_ids
    sequence: int32
  - name: token_type_ids
    sequence: int8
  - name: attention_mask
    sequence: int8
  - name: start_positions
    dtype: int64
  - name: end_positions
    dtype: int64
  splits:
  - name: train
    num_bytes: 16151564
    num_examples: 21593
  - name: validation
    num_bytes: 1921612
    num_examples: 2569
  download_size: 4291723
  dataset_size: 18073176
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---