metadata
dataset_info:
features:
- name: image
dtype: image
- name: image_id
dtype: int64
- name: area
sequence: float32
- name: bbox
sequence:
sequence: float32
- name: label
sequence: float64
splits:
- name: train
num_bytes: 1845303844.116
num_examples: 16102
download_size: 1840009692
dataset_size: 1845303844.116
configs:
- config_name: default
data_files:
- split: train
path: data/train-*