---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: loss_mask
    sequence: int64
  - name: input_len
    dtype: int64
  splits:
  - name: train
    num_bytes: 3191732424
    num_examples: 900000
  - name: test
    num_bytes: 354656878
    num_examples: 100000
  download_size: 226528509
  dataset_size: 3546389302
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---