---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: question
    dtype: string
  - name: answer
    dtype: string
  splits:
  - name: train
    num_bytes: 573795591.0
    num_examples: 100
  download_size: 573779925
  dataset_size: 573795591.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

crop 했어야 했는데, crop을 하지 못 함.