---
dataset_info:
  features:
  - name: messages
    dtype: string
  - name: num_tokens
    dtype: int64
  - name: params
    struct:
    - name: extra_body
      struct:
      - name: ignore_eos
        dtype: bool
    - name: max_tokens
      dtype: int64
  splits:
  - name: train
    num_bytes: 1162906796
    num_examples: 2043
  download_size: 1030257190
  dataset_size: 1162906796
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---