---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: image_path
    dtype: string
  - name: query
    dtype: string
  splits:
  - name: train
    num_bytes: 20165951597
    num_examples: 200000
  download_size: 20050680877
  dataset_size: 20165951597
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: mit
task_categories:
- image-to-text
language:
- en
size_categories:
- 100K<n<1M
---