metadata
dataset_info:
features:
- name: image
dtype: image
- name: input_ids
sequence: int64
- name: base_character_indices
sequence: int64
- name: diacritic_indices
sequence: int64
- name: word
dtype: string
splits:
- name: train
num_bytes: 682976802.6636946
num_examples: 178774
download_size: 575275176
dataset_size: 682976802.6636946
configs:
- config_name: default
data_files:
- split: train
path: data/train-*