README.md

---
dataset_info:
  config_name: default
  features:
    - name: image
      dtype:
        image:
          decode: false
    - name: xml_content
      dtype: string
    - name: filename
      dtype: string
    - name: project_name
      dtype: string
  splits:
  - name: train
    num_examples: 27
    num_bytes: 412007872
  - name: test
    num_examples: 2
    num_bytes: 412007872
  download_size: 824015744
  dataset_size: 824015744
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train/**/*.parquet
  - split: test
    path: data/test/**/*.parquet
tags:
  - image-to-text
  - htr
  - trocr
  - transcription
  - pagexml
license: mit
---

# Dataset Card for rawxl-test-overwrite

This dataset was created using pagexml-hf converter from Transkribus PageXML data.

## Dataset Summary

This dataset contains 29 samples across 2 split(s).

### Projects Included

- 1611-02-25_Rezess_(HAStK-RBA_Best__82_A_51)
- B_IX_490_duplicated

## Dataset Structure

### Data Splits

- **train**: 27 samples
- **test**: 2 samples

### Dataset Size

- Approximate total size: 785.84 MB
- Total samples: 29

### Features

- **image**: `Image(mode=None, decode=False)`
- **xml_content**: `Value('string')`
- **filename**: `Value('string')`
- **project_name**: `Value('string')`

## Data Organization

Data is organized as parquet shards by split and project:
```
data/
├── <split>/
│   └── <project_name>/
│       └── <timestamp>-<shard>.parquet
```

The HuggingFace Hub automatically merges all parquet files when loading the dataset.

## Usage

```python
from datasets import load_dataset

# Load entire dataset
dataset = load_dataset("jwidmer/rawxl-test-overwrite") 

# Load specific split
train_dataset = load_dataset("jwidmer/rawxl-test-overwrite", split="train")
```