CoolFace
Datasetpublic

porupski/ParlaSpeech-HR-benchmark_v1

ParlaSpeechHR Benchmark v1 A benchmark dataset of 17,083 Croatian parliamentary speech clips from ParlaSpeech-HR v1. Each clip includes audio (WAV) and rich metadata for speaker profiling tasks. Contents 17,083 audio segments Metadata: speaker info, party affiliation, birth year, gender Word-level transcriptions and per-word start times (both raw and normalized variants) Note: ~2,414 audio files referenced in the manifest are currently unavailable in this version… See the full description on the dataset page: https://huggingface.co/datasets/porupski/ParlaSpeech-HR-benchmark_v1.

sourceHugging Faceupdated 3mo agoView on Hugging Face
0likes122downloads
Dataset Card

ParlaSpeechHR Benchmark v1

A benchmark dataset of 17,083 Croatian parliamentary speech clips from ParlaSpeech-HR v1. Each clip includes audio (WAV) and rich metadata for speaker profiling tasks.

Contents

  • —17,083 audio segments
  • —Metadata: speaker info, party affiliation, birth year, gender
  • —Word-level transcriptions and per-word start times (both raw and normalized variants)
  • —Note: ~2,414 audio files referenced in the manifest are currently unavailable in this version

Data format

Two datasets configs, each with a single train split. Task splits are per-task columns, not HF splits — the same utterance can appear in gender/train and age/dev at the same time.

`default` — what the viewer opens; everything needed to browse, filter, and train on utterance-level tasks.

ColumnTypeNotes
instance_idstringderived from path (seg.<hash>_<s>-<e>.flac → <hash>_<s>-<e>)
audioAudio(sampling_rate=16000)16 kHz mono, inlined bytes
textstringjoined from the words list
path, orig_file, start, end, utterance_id_start, utterance_id_end, splitscalarv1 provenance. split is the corpus-wide split, kept for cross-reference
speaker_role, speaker_type, speaker_party, speaker_party_name, party_status, speaker_name, speaker_gender, speaker_birthscalarspeaker & party (v1's speaker_info is thinner than v3's — no ParlaMint metadata)
benchmark_{gender,speaker_id,power_status,age}_labelscalartask labels; nullable when the row isn't in that task
benchmark_{...}_splitstringtrain / dev / test per task

`alignments` — opt-in via load_dataset(repo, "alignments"). Join back on instance_id.

ColumnTypeNotes
instance_idstringjoin key
wordslist[{word, time_s}]word text + start time
norm_wordslist[{word, time_s}]normalized-form variants

v1 does not carry chars_align, primary_stress, linguistic_annotation, or TextGrids — those layers only exist in v3.

Quick example:

python
from datasets import load_dataset

ds = load_dataset("porupski/ParlaSpeech-HR-benchmark_v1", "default", split="train")
gender_train = ds.filter(lambda r: r["benchmark_gender_split"] == "train")
print(len(gender_train), gender_train[0]["speaker_gender"])

# Attach word-level times when needed
align = load_dataset("porupski/ParlaSpeech-HR-benchmark_v1", "alignments", split="train")

Benchmark tasks

All four v1 tasks are classification (v1's age target is a young/old group, not a continuous year — that's v3's job):

  • —Gender classification (M/F)
  • —Speaker identification (50-class)
  • —Power status classification (Coalition/Opposition)
  • —Age group classification (Young/Old)

Use cases

  • —Speaker profiling and verification
  • —Gender classification from speech
  • —Political affiliation prediction
  • —Fine-tuning Wav2Vec2 models for Slavic speech

References

  • —Full ParlaSpeech-HR v1 corpus: https://www.clarin.si/repository/xmlui/handle/11356/1494
  • —Paper: https://doi.org/10.5281/zenodo.13953438

License

CC BY-SA 4.0 (inherited from ParlaSpeech-HR).