---
dataset_info:
  features:
  - name: image_id
    dtype: string
  - name: image
    dtype: image
  - name: objects
    dtype: string
  - name: relationships
    dtype: string
  splits:
  - name: train
    num_bytes: 22844626962.625
    num_examples: 46563
  download_size: 22814353825
  dataset_size: 22844626962.625
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---