---
license: mit
dataset_info:
  features:
  - name: input_ids
    sequence: int32
  - name: attention_mask
    sequence: int8
  splits:
  - name: train
    num_bytes: 52276032496
    num_examples: 30348917
  download_size: 24427035053
  dataset_size: 52276032496
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---