metadata

language:
  - en
license: mit
task_categories:
  - text-to-speech
  - automatic-speech-recognition
  - audio-to-audio
pretty_name: Hailuo AI Voices Dataset
size_categories:
  - 10K<n<100K
tags:
  - audio
  - speech
  - voice
  - text-to-speech
  - UnlimitedBytes
dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: transcription
      dtype: string
    - name: voice
      dtype: string
    - name: phonemes
      dtype: int64
    - name: language
      dtype: string
    - name: gender
      dtype: string

Hailuo AI Voices Dataset 🎤

A curated collection of high-quality voice recordings with corresponding transcriptions and phoneme analysis. This dataset is designed for speech recognition, text-to-speech, and voice analysis tasks.

📊 Dataset Overview

The dataset provides a comprehensive collection of voice samples with the following features:

Feature	Description
Audio Files	High-quality WAV format recordings
Transcription	Accurate transcriptions of each recording
Voice IDs	Unique identifiers for different speakers
Phoneme Counts	Number of phonemes in each utterance
Language	Language of the audio recording
Gender	Speaker's gender

📊 Dataset Statistics

Audio Duration Distribution

Text Length Distribution

Samples per Voice

Key Metrics

Total Samples: 68034
Total Duration: 320.64 hours
Average Duration: 16.97 seconds
Average Text Length: 256.3 characters

🚀 Quick Start

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("unlimitedbytes/hailuo-ai-voices", split="full")

# Explore a sample
example = dataset['train'][0]
print("📝 Text:", example['text'])
print("🎙️ Voice:", example['voice'])
print("🔤 Phonemes:", example['phonemes'])

📖 Dataset Structure

Each entry in the dataset contains:

{
    'audio': {
        'path': str,      # Path to the audio file
        'array': np.array # Audio data as numpy array
    },
    'text': str,         # Transcription of the audio
    'voice': str,        # Voice identifier (e.g., 'voice1')
    'phonemes': int      # Number of phonemes in the utterance
}

🔍 Data Composition

Multiple voice recordings from different speakers
Clean transcriptions for each audio sample
Phonetic analysis with phoneme counts
Consistent audio format (WAV)

🛠️ Use Cases

Text-to-Speech (TTS) training
Speech Recognition
Voice Analysis
Phoneme Studies
Speech Synthesis

📄 License

This project is licensed under the MIT License - see the LICENSE for details.

📚 Citation

If you use this dataset in your research, please cite:

@dataset{hailuo_ai_voices,
    author = {UnlimitedBytes},
    title = {Hailuo AI Voices Dataset},
    year = {2025},
    publisher = {Hugging Face},
    url = {https://huggingface.co./datasets/unlimitedbytes/hailuo-ai-voices}
}