---
dataset_info:
  features:
  - name: images
    sequence: image
  - name: texts
    list:
    - name: user
      dtype: string
    - name: assistant
      dtype: string
    - name: source
      dtype: string
  - name: image
    dtype: image
  - name: conversations
    list:
    - name: from
      dtype: string
    - name: value
      dtype: string
  splits:
  - name: train
    num_bytes: 36331050941.248
    num_examples: 56398
  download_size: 36000592447
  dataset_size: 36331050941.248
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---