ZLSCompLing/LOD_Claude

---
license: cc0-1.0
task_categories:
- automatic-speech-recognition
language:
- lb
tags:
- audio
- speech
- luxembourgish
- synthetic
size_categories:
- 10K<n<100K
---

# LOD_Claude Dataset

## Dataset Description

LOD_Claude is a Luxembourgish speech dataset containing audio recordings paired with transcriptions. The audio features a synthetic voice named Claude reading example sentences from the LOD (Lëtzebuerger Online Dictionnaire) available at lod.lu.

## Dataset Statistics

- **Total samples**: 39,034
- **Training samples**: 37,084
- **Validation samples**: 1,950
- **Language**: Luxembourgish (Lëtzebuergesch)
- **Audio format**: WAV files
- **Sample rate**: 24,000 Hz

## Dataset Structure

Each sample contains:
- `audio`: Audio file in WAV format
- `text`: Transcription text
- `split`: Indicates whether sample is from "train" or "val" set
- `filename`: Original filename identifier

## Example

```python
from datasets import load_dataset

dataset = load_dataset("ZLSCompLing/LOD_Claude")

# Access a sample
sample = dataset[0]
print(f"Text: {sample['text']}")
# Audio can be accessed via sample['audio']
```

## License

This dataset is released under the CC0 license - fully public domain with free use and no attribution required.

## Contact

For questions or issues regarding this dataset, please contact the repository maintainers.