NaolBM/amharic-audio

🎵 Amharic Bible Audio Dataset

📋 Dataset Description

This dataset contains 59K audio chunks derived from Amharic Bible readings, split into 5-second segments for optimal training of speech models.

Audio Specifications

• Format: WAV (16-bit PCM)
• Sample Rate: 24 kHz
• Duration: 5 seconds per chunk
• Total Hours: ~82.5 hours

🚀 Usage

Load with HuggingFace Datasets

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("NaolBM/amharic-audio", split="train")

# Access audio files
for example in dataset:
    audio = example["audio"]  # Audio array
    id = example["id"]     # id of the audio file

🎯 Intended Uses

• Speech Recognition (ASR) training
• Text-to-Speech (TTS) synthesis
• Voice cloning experiments
• Speech representation learning
• Amharic language modeling