---
dataset_info:
  features:
  - name: input_ids
    sequence: int32
  - name: attention_mask
    sequence: int8
  - name: labels
    sequence: int64
  - name: position_ids
    sequence: int64
  - name: length
    dtype: int64
  splits:
  - name: train
    num_bytes: 947748827022
    num_examples: 70783733
  download_size: 205952677727
  dataset_size: 947748827022
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---