---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: loss_mask
    sequence: int64
  - name: input_len
    dtype: int64
  splits:
  - name: train
    num_bytes: 182952952
    num_examples: 1000000
  download_size: 9940230
  dataset_size: 182952952
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---