---
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: image
    dtype: image
  - name: conversations
    list:
    - name: role
      dtype: string
    - name: content
      dtype: string
  - name: dataset_name
    dtype: string
  - name: task_name
    dtype: string
  splits:
  - name: train
    num_bytes: 338292153802.75
    num_examples: 3035954
  - name: validation
    num_bytes: 1151589139.875
    num_examples: 9449
  download_size: 330069408828
  dataset_size: 339443742942.625
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---