---
dataset_info:
  features:
  - name: messages
    dtype: string
  - name: images
    sequence: image
  splits:
  - name: train
    num_bytes: 197196363.0
    num_examples: 4000
  download_size: 191694906
  dataset_size: 197196363.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---