Tarjoman Podcast 2023 ASR Dataset

This datasets consists of a collection of 507 articles from the Tarjoman website until the end of 2023, each accompanied by corresponding audio recordings.

Dataset Description

The dataset includes complete articles along with their audio counterparts. Each article is presented in its entirety, without sentence segmentation. Every entry contains the following metadata fields:

Title: The main title of the article.
Subtitle: A secondary title that may provide additional context.
Introduction: A brief introductory section summarizing the article’s content.
Content: The body of the Article.

Usage

Huggingface datasets library:

from datasets import load_dataset
dataset = load_dataset('PerSets/tarjoman2023asr')