---
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-*
---