README.md

---
dataset_info:
  features:
  - name: image
    dtype: binary
  - name: question
    dtype: string
  - name: ground_truth
    dtype: string
  splits:
  - name: train
    num_bytes: 1159320886
    num_examples: 4451
  - name: validation
    num_bytes: 105325482
    num_examples: 460
  - name: test
    num_bytes: 95351744
    num_examples: 423
  download_size: 128220786
  dataset_size: 1359998112
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---