---
dataset_info:
  features:
  - name: instruction_seed
    dtype: string
  - name: _source
    dtype: string
  - name: generated_text
    dtype: string
  - name: conversations
    list:
    - name: from
      dtype: string
    - name: value
      dtype: string
  splits:
  - name: train
    num_bytes: 159918219256
    num_examples: 3500000
  download_size: 62933872349
  dataset_size: 159918219256
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---