---
dataset_info:
  features:
  - name: input_ids
    sequence: int32
  - name: overflow_to_sample_mapping
    dtype: int64
  splits:
  - name: train
    num_bytes: 38831189208
    num_examples: 4733202
  download_size: 17991389032
  dataset_size: 38831189208
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---