---
dataset_info:
  features:
  - name: image_id
    dtype: string
  - name: question
    dtype: string
  - name: answer
    dtype: string
  - name: task
    dtype: string
  - name: image
    dtype: image
  - name: conversations
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 6702592612.0
    num_examples: 50000
  download_size: 6638298618
  dataset_size: 6702592612.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---