---
dataset_info:
  features:
  - name: source
    dtype: string
  - name: tokens
    sequence: int64
  - name: index
    dtype: int64
  - name: text
    dtype: string
  - name: transcript
    dtype: string
  - name: conversations
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 33956402461.0
    num_examples: 2420047
  download_size: 7263983300
  dataset_size: 33956402461.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---