---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: bboxes
    sequence:
      sequence: int64
  - name: text
    sequence: string
  - name: language
    dtype: string
  - name: source
    dtype: string
  splits:
  - name: train
    num_bytes: 127423305.0
    num_examples: 143
  download_size: 127285086
  dataset_size: 127423305.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

Benchmark for [surya OCR](https://github.com/VikParuchuri/surya).  Data originally from [tapuscorpus](https://github.com/HTR-United/tapuscorpus).