Datasets:
metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train.parquet
It is a dataset of ukrainian audiobooks
Each sample contain an approximately 8 secod seconds od ukrainian speech