README.md

---
dataset_info:
  config_name: default
  features:
  - name: image
    dtype:
      image:
        decode: false
  - name: xml_content
    dtype: large_string
  - name: filename
    dtype: large_string
  - name: project_name
    dtype: large_string
  - name: inference_xml_20260825_113642
    dtype: large_string
  splits:
  - name: train
    num_examples: 6
    num_bytes: 52313352
  download_size: 52313352
  dataset_size: 52313352
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train/**/*.parquet
tags:
- xml
- pagexml
- inference
- htr
license: mit
---

# Dataset Card for anothertest

This dataset is derived from [jwidmer/anothertest](https://huggingface.co/datasets/jwidmer/anothertest) and enriched with raw XML inference derived from [jwidmer/anothertest-lines](https://huggingface.co/datasets/jwidmer/anothertest-lines).

## Dataset Summary

This dataset contains 6 samples across 1 split(s).

### Projects Included

1505-02-10_Hanserezess,_Lübeck_Dienstag_nach_Scholastice_1505_(SAHST_Rep__2,_I_040-4)

## Dataset Structure

### Data Splits

- **train**: 6 samples

### Dataset Size

- Approximate total size: 49.89 MB
- Total samples: 6

### Features

- **image**: `Image(mode=None, decode=False)`
- **xml_content**: `Value('large_string')`
- **filename**: `Value('large_string')`
- **project_name**: `Value('large_string')`
- **inference_xml_20260825_113642**: `Value('large_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/anothertest")

# Load specific split
dataset_split = load_dataset("jwidmer/anothertest", split="train")
```