metadata
dataset_info:
features:
- name: image
dtype: image
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 656033920
num_examples: 3000
- name: validation
num_bytes: 65892669
num_examples: 300
download_size: 640778389
dataset_size: 721926589
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*