metadata
dataset_info:
features:
- name: char
dtype: string
- name: unicode
dtype: string
- name: images
dtype: image
- name: stroke
dtype: int32
- name: strokes_sequence
sequence: int32
- name: components
sequence: int32
- name: font
dtype: string
splits:
- name: train
num_bytes: 538831637.375
num_examples: 85469
- name: test
num_bytes: 429704947.125
num_examples: 67095
download_size: 696955047
dataset_size: 968536584.5
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*