metadata
dataset_info:
features:
- name: im
dtype: image
- name: roman
dtype: string
- name: manchu
dtype: string
splits:
- name: train
num_bytes: 203797830.4
num_examples: 60000
- name: validation
num_bytes: 50949457.6
num_examples: 15000
download_size: 256334316
dataset_size: 254747288
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*