metadata
dataset_info:
features:
- name: char
dtype: string
- name: unicode
dtype: string
- name: images
dtype: image
- name: target_images
dtype: image
- name: stroke
dtype: int32
- name: strokes_sequence
sequence: int32
- name: components
sequence: int32
splits:
- name: train
num_bytes: 128716533
num_examples: 9196
- name: test
num_bytes: 488705813.5
num_examples: 67086
download_size: 482905461
dataset_size: 617422346.5
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*