czyzi0/pwr-azon-speech-dataset

---
license: cc-by-sa-4.0
task_categories:
- automatic-speech-recognition
language:
- pl
pretty_name: PWr AZON
size_categories:
- 10K<n<100K
dataset_info:
features:
- name: audio
dtype:
audio:
sampling_rate: 44100
- name: transcript
dtype: string
- name: gender
dtype: string
- name: id
dtype: string
- name: id_og
dtype: string
splits:
- name: train
num_bytes: 8585221408.406
num_examples: 14491
- name: unsup
num_bytes: 1128648882
num_examples: 841
download_size: 9746452069
dataset_size: 9713870290.406
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: unsup
path: data/unsup-*
---
This speech dataset consists of 15332 short audio clips of multiple speakers speaking in Polish. Transcription is provided for 14491 audio clips (`train` split), and it is missing for 841 audio clips (`unsup` split). Gender of speaker is provided for the whole dataset. Clips have total length of almost 31 hours.

This dataset was created from _Korpus nagrań próbek mowy do celów budowy modeli akustycznych dla automatycznego rozpoznawania mowy w języku polskim_. The dataset was repackaged into easier to use format. If you are interested in the original data, please visit https://zasobynauki.pl/zasoby/korpus-nagran-probek-mowy-do-celow-budowy-modeli-akustycznych-dla-automatycznego-rozpoznawania-mowy,53293/

Also, if you find this resource helpful, kindly consider leaving a like.

原始信息汇总

数据集概述

基本信息

• 许可证: cc-by-sa-4.0
• 任务类别: automatic-speech-recognition
• 语言: pl
• 数据集名称: PWr AZON
• 数据量级: 10K<n<100K

数据集详情

特征

• 音频:
  • 采样率: 44100
• 转录文本: string
• 性别: string
• ID: string
• 原始ID: string

数据分割

• 训练集:
  • 字节数: 8585221408.406
  • 样本数: 14491
• 无监督集:
  • 字节数: 1128648882
  • 样本数: 841

下载与数据集大小

• 下载大小: 9746452069
• 数据集大小: 9713870290.406

配置

• 默认配置:
  • 训练集路径: data/train-*
  • 无监督集路径: data/unsup-*

数据集描述

该语音数据集包含15332个短音频片段，由多名波兰语说话者录制。其中14491个音频片段（训练集）提供了转录文本，841个音频片段（无监督集）未提供转录文本。整个数据集提供了说话者的性别信息，总时长近31小时。