metadata
dataset_info:
features:
- name: image
dtype: image
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 4215984699
num_examples: 12000
- name: validation
num_bytes: 126870900
num_examples: 400
download_size: 4504169676
dataset_size: 4342855599
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*