metadata
dataset_info:
features:
- name: im
dtype: image
- name: roman
dtype: string
- name: manchu
dtype: string
splits:
- name: train
num_bytes: 305015678
num_examples: 80306
- name: validation
num_bytes: 67762210
num_examples: 18359
- name: test
num_bytes: 789711
num_examples: 753
download_size: 370723474
dataset_size: 373567599
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*