resproj007/librispeech_male

Librispeech Male Dataset

Overview

This dataset contains healthy speech samples from a male speaker (4014) in the LibriSpeech corpus, prepared for pathological speech synthesis research.

Speaker Information:

• Speaker ID: 4014
• Corpus: LibriSpeech
• Gender: Male
• Speech Status: Healthy

Dataset Statistics

• Total Samples: 160
• Total Duration: 0.41 hours
• Sampling Rate: 24,000 Hz
• Format: Audio arrays with transcriptions

Training Split

• Samples: 130
• Duration: 0.33 hours
• Avg Duration: 9.1s
• Duration Range: 2.1s - 16.7s
• Avg Text Length: 158 characters

Test Split

• Samples: 30
• Duration: 0.08 hours
• Avg Duration: 9.8s
• Duration Range: 2.3s - 17.0s
• Avg Text Length: 169 characters

Loading the Dataset

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("your-username/librispeech_male")

# Access train and test splits
train_data = dataset['train']
test_data = dataset['test']

# Each sample contains:
# - 'audio': {'array': numpy_array, 'sampling_rate': 24000}
# - 'text': str (normalized transcription)

# Example usage
sample = train_data[0]
audio_array = sample['audio']['array']
transcription = sample['text']
sampling_rate = sample['audio']['sampling_rate']

Direct Training with Transformers

from transformers import Trainer
from datasets import load_dataset

# Load and use directly with Trainer (no preprocessing needed)
dataset = load_dataset("your-username/librispeech_male")
trainer = Trainer(
    train_dataset=dataset['train'],
    eval_dataset=dataset['test'],
    # ... other trainer arguments
)