Oriel — AI & technology worth noticing, every day.

nvidia/Granary-HiFi

Dataset · https://huggingface.co/datasets/nvidia/Granary-HiFi

Dataset Card for Granary-HiFi Dataset Description Granary-HiFi is a corpus of multilingual speech data to which complementary filtering approaches have been applied to identify high-fidelity samples. This release presents Granary-v2 enriched metadata for the VoxPopuli subset; audio bytes are not distributed in this repository. The processing pipeline includes: Sound event detection to identify speech coverage and filter samples contaminated by non-speech sounds.… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Granary-HiFi.

First seen by Oriel on .

AI summary

Granary-HiFi is a speech dataset containing audio segments paired with Bulgarian transcripts, phonetic transcriptions, and speaker identifiers.

  • The examples are short WAV segments from a 2009 parliamentary plenary session, with filenames and session-level identifiers.
  • Each segment includes Bulgarian text and an IPA-like phonetic transcription; speaker labels appear on many records.
  • The speech covers parliamentary remarks on energy policy, discrimination, migration, wine production, and international affairs.

#speech datasets #audio #bulgarian #phonetics

Generated by gpt-5.6-sol. Check the original for details.

Metric trends

Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained

Likes

Total likes reported by the source.

  • Hugging Face Datasets: 8, latest observation ; 0 over observed period since

Downloads (30d)

Rolling 30-day downloads. A change in this value is not new downloads during the observed period.

  • Hugging Face Datasets: 1,432, latest observation ; +155 over observed period since

Aliases

SourceExternal IDURL
Hugging Face Datasetsnvidia/Granary-HiFihttps://huggingface.co/datasets/nvidia/Granary-HiFi Read discussion