metadata
dataset_info:
features:
- name: image
dtype: image
- name: label
dtype: string
- name: full_character
sequence: string
- name: base_character
sequence: string
- name: type
sequence: string
- name: diacritic_type
sequence: string
splits:
- name: train
num_bytes: 1019664521.2
num_examples: 22300
download_size: 457633803
dataset_size: 1019664521.2
configs:
- config_name: default
data_files:
- split: train
path: data/train-*