---
dataset_info:
  features:
  - name: im
    dtype: image
  - name: roman
    dtype: string
  - name: manchu
    dtype: string
  splits:
  - name: train
    num_bytes: 203797830.4
    num_examples: 60000
  - name: validation
    num_bytes: 50949457.6
    num_examples: 15000
  download_size: 256334316
  dataset_size: 254747288.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---