---
dataset_info:
  features:
  - name: input_ids
    sequence: int32
  - name: attention_mask
    sequence: int8
  - name: special_tokens_mask
    sequence: int8
  splits:
  - name: train
    num_bytes: 16995635700
    num_examples: 1381759
  download_size: 4366533537
  dataset_size: 16995635700
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---