This is a Hugging Face Parquet input dataset for an audio pipeline.
Source dataset: archivartaunik/uladzimir-karatkevich-dzikae-paliavanne-karalia-stakha-aleg-garbuz
Format
config: default
split: train
format: parquet
id column: id
audio column: audio
rows: 19
shards: 7
language: be
The audio column is embedded into Parquet as Hugging Face Audio:
audio = {
"path": "file.mp3",
"bytes": b"..."
}
