metadata
dataset_info:
features:
- name: image_id
dtype: int32
- name: image
dtype: image
- name: relationships_text
dtype: string
- name: objects
sequence: string
splits:
- name: train
num_bytes: 13891342065.484573
num_examples: 97269
- name: test
num_bytes: 1570308173.2344272
num_examples: 10808
download_size: 15325984725
dataset_size: 15461650238.719
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*