metadata
dataset_info:
features:
- name: image
dtype: image
- name: bboxes
sequence:
sequence: int32
- name: labels
sequence: int32
- name: layout
dtype: image
splits:
- name: train
num_bytes: 202471471
num_examples: 126
download_size: 199176894
dataset_size: 202471471
configs:
- config_name: default
data_files:
- split: train
path: data/train-*