metadata
dataset_info:
features:
- name: image
dtype: image
- name: text
sequence: string
- name: kanji
dtype: string
- name: image_path
dtype: string
splits:
- name: train
num_bytes: 20176738.08
num_examples: 2860
- name: test
num_bytes: 4079955
num_examples: 574
download_size: 24402583
dataset_size: 24256693.08
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*