---
dataset_info:
  features:
  - name: input_ids
    list: int32
  - name: labels
    list: int64
  - name: num_mask_tokens
    dtype: int64
  - name: original
    list: int64
  splits:
  - name: train
    num_bytes: 19474815000
    num_examples: 2870669
  - name: test
    num_bytes: 198018600
    num_examples: 28999
  download_size: 3268994346
  dataset_size: 19672833600
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---