metadata
dataset_info:
features:
- name: image
dtype: image
splits:
- name: train
num_bytes: 8975751122.008
num_examples: 20682
download_size: 17596132848
dataset_size: 8975751122.008
configs:
- config_name: default
data_files:
- split: train
path: data/train-*