directtt/pl_debates_test

Polish Politicians Speech

Dataset in the public domain (2025)

Description

Studio-recorded short utterances by 16 Polish political figures (e.g. “Bartoszewicz Artur”, “Zandberg Adrian”), each reading ~15 sentences. Audio in .wav (16 kHz).

Files

• audio/ – per-speaker WAV files
• transcript/pl/test.tsv – tabular metadata & transcripts

Usage

from datasets import load_dataset
ds = load_dataset("USERNAME/pl-politicians-speech", "plain_audio")
print(ds["train"][0])
# {'path': ..., 'audio': {...}, 'sentence': ..., 'age': ..., 'gender': ...}