---
dataset_info:
  features:
  - name: multiple_choice_answer
    dtype: string
  - name: question
    dtype: string
  - name: image
    dtype: image
  splits:
  - name: train
    num_bytes: 16427957.954119343
    num_examples: 100
  download_size: 16332826
  dataset_size: 16427957.954119343
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

A small VQAv2 Dataset For Testing Script or Notebook on Google Collab. Use split='train'.