metadata
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype:
class_label:
names:
'0': 0B
'1': 1B
'2': 2B
'3': background
splits:
- name: train
num_bytes: 4619588
num_examples: 522
download_size: 4621243
dataset_size: 4619588
configs:
- config_name: default
data_files:
- split: train
path: data/train-*