Is this your dataset? Claim it with the Hugging Face account that owns it.

acidente/
fala_pb

Name: fala_pb
Creator: acidente

Views

No views yet

fala_pb — Paraíba Speech Corpus

Unified corpus of Paraíba (Brazil) speech for ASR/TTS. 34,866 audio files (~111.8 h) in wavs/part_1/ … wavs/part_4/ (max 10,000 files per folder, split by numeric range: part_1 = 1–10000, part_2 = 10001–20000, part_3 = 20001–30000, part_4 = 30001–34866), with metadata in metadata.csv.

	Sources

source subset audios description

colingpb segments 33,649 Corpus Linguístico da Paraíba — segmented interviews, with transcriptions… See the full description on the dataset page: https://huggingface.co/datasets/acidente/fala_pb.