metadata
dataset_info:
features:
- name: image
dtype: image
- name: bbox
sequence:
sequence: float32
- name: categories
dtype: string
- name: bbox_str
dtype: string
splits:
- name: train
num_bytes: 18181685481.55
num_examples: 23530
download_size: 17704342721
dataset_size: 18181685481.55
configs:
- config_name: default
data_files:
- split: train
path: data/train-*