metadata
dataset_info:
features:
- name: id
dtype: string
- name: bbox_list
sequence:
sequence: int32
- name: label_list
sequence: string
- name: size
sequence: int32
- name: text_list
sequence: string
- name: caption_list
sequence: string
- name: images
sequence: image
splits:
- name: train
num_bytes: 2703067774.416
num_examples: 16176
- name: test
num_bytes: 338980832.737
num_examples: 2003
download_size: 2838248781
dataset_size: 3042048607.153
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*