metadata
dataset_info:
features:
- name: image
dtype: image
- name: bboxes
sequence:
sequence: float32
- name: rows
sequence:
sequence: int64
- name: cols
sequence:
sequence: int64
splits:
- name: train
num_bytes: 460197789.446
num_examples: 9679
download_size: 450209455
dataset_size: 460197789.446
configs:
- config_name: default
data_files:
- split: train
path: data/train-*