---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: log_weight
    dtype: float32
  - name: sampling_p_scaled
    dtype: float64
  splits:
  - name: train
    num_bytes: 3592885712.0
    num_examples: 1500000
  - name: validation
    num_bytes: 2436848
    num_examples: 1000
  download_size: 1475723812
  dataset_size: 3595322560.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---