---
dataset_info:
  features:
  - name: input_ids
    sequence: int32
  - name: token_type_ids
    sequence: int8
  - name: attention_mask
    sequence: int8
  - name: overflow_to_sample_mapping
    dtype: int64
  splits:
  - name: train
    num_bytes: 11872129248
    num_examples: 15074547
  - name: validation
    num_bytes: 1363453970
    num_examples: 1731193
  download_size: 4988001484
  dataset_size: 13235583218
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---