---
dataset_info:
  features:
  - name: image
    struct:
      dtype: image
  - name: conversations
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 7016691265
    num_examples: 56443
  download_size: 6533969401
  dataset_size: 7016691265
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---