---
dataset_info:
  features:
  - name: image_id
    dtype: string
  - name: caption
    dtype: string
  - name: true_triples
    list:
      list: string
  - name: objects
    list: string
  splits:
  - name: train
    num_bytes: 127309052
    num_examples: 566610
  download_size: 33805593
  dataset_size: 127309052
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---