---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: label
    dtype: string
  - name: full_character
    sequence: string
  - name: base_character
    sequence: string
  - name: type
    sequence: string
  - name: diacritic_type
    sequence: string
  splits:
  - name: train
    num_bytes: 1019664521.2
    num_examples: 22300
  download_size: 457633803
  dataset_size: 1019664521.2
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---