# su-openslr (Sundanese ASR subset)

This dataset bundles Sundanese audio with a metadata.csv.

## Files
- `audio/` — WAV files (`su_id_female/`, `su_id_male/`)
- `metadata.csv` — columns like `path`, `sentence`, etc.

## Usage
```python
from datasets import load_dataset, Audio
ds = load_dataset("novalalthoff/su-openslr", data_files="metadata.csv", split="train")
ds = ds.cast_column("path", Audio(sampling_rate=16000))
```