---
dataset_info:
  features:
  - name: images
    dtype: image
  - name: contest_number
    dtype: string
  - name: question
    list: string
  - name: problem
    dtype: string
  - name: answer
    dtype: string
  - name: task
    dtype: string
  - name: visual_info
    dtype: string
  splits:
  - name: train
    num_bytes: 3765312116.692
    num_examples: 5163
  - name: test
    num_bytes: 513734565.0
    num_examples: 685
  - name: test_hard
    num_bytes: 256628678.0
    num_examples: 350
  - name: test_very_hard
    num_bytes: 256625514.0
    num_examples: 350
  download_size: 4091188525
  dataset_size: 4792300873.691999
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
  - split: test_hard
    path: data/test_hard-*
  - split: test_very_hard
    path: data/test_very_hard-*
---