Datasets:

Codyfederer
/

jp-podcast-1

Tasks:

Automatic Speech Recognition

Modalities:

Formats:

Languages:

Size:

Libraries:

License:

Dataset card Data Studio Files Files and versions

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

Japanese dataset

Audio dataset for speech synthesis and recognition.

Dataset Details

Segments: 66,575
Total Duration: 182.7 hours
Language: ja
Quality Score Range: 50.0 - 100.0
Duration Range: 1.0s - 30.0s

Usage

from datasets import load_dataset

dataset = load_dataset("Codyfederer/jp-podcast-1")

Fields

audio: Audio file
text: Transcript text
duration_seconds: Audio duration
speaker: Speaker identifier
quality_score: Quality score (0-100)
episode_name: Source episode
show_name: Source show

Downloads last month: 23

Total file size: