metadata
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: 1063466238.75
num_examples: 2186
download_size: 1061913732
dataset_size: 1063466238.75
configs:
- config_name: default
data_files:
- split: train
path: data/train-*