---
dataset_info:
  features:
  - name: ground_truth
    dtype: image
  - name: caption
    dtype: string
  - name: qa_list
    list:
    - name: ground_truth_answer
      dtype: string
    - name: question
      dtype: string
  - name: category
    dtype: string
  splits:
  - name: train
    num_bytes: 69850628.625
    num_examples: 2043
  download_size: 49246498
  dataset_size: 69850628.625
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---