VieNeu-TTS-140h / README.md
pnnbao-ump's picture
Update README.md
a5f8845 verified
metadata
language:
  - vi
task_categories:
  - text-to-speech
  - automatic-speech-recognition
size_categories:
  - 10K<n<100K
tags:
  - vietnamese
  - tts
  - speech
  - phonemized
  - multi-speaker
license: apache-2.0

pnnbao-ump/VieNeu-TTS-140h

Mô tả Dataset

A high-quality Vietnamese Text-to-Speech (TTS) dataset containing 74,858 audio samples with phonemized transcripts. This benchmark dataset is designed for fine-tuning modern TTS models with maximum synthesis quality. The text corpus is completely phonemized using standard international phonetic alphabet (IPA) representations suitable for neural acoustic modeling.

Quick Facts

  • Language: Vietnamese 🇻🇳
  • Tasks: Text-to-Speech (TTS), Automatic Speech Recognition (ASR)
  • Total Samples: 74,858
  • Audio Format: WAV, 24kHz
  • Average Duration: ~6.76 seconds per sample
  • Total Duration: ~140.7 hours
  • Multi-speaker: 193 unique speakers

Gender Distribution

  • 👨 Male: 50,200 samples (67.1%)
  • 👩 Female: 24,658 samples (32.9%)

Cấu trúc Dataset

Dataset bao gồm các trường sau:

{
    "_id": "jellyfish1010_0041_5",
    "audio": {
        "path": "...",
        "array": [...],
        "sampling_rate": 24000
    },
    "text": "Người người băn khoăn về AI, nhà nhà cùng nhau sử dụng AI.",
    "phonemized_text": "ŋˈyə2j ŋˈyə2j bˈan xwˈan vˈe2 ˌaːˈi, ɲˈaː2 ɲˈaː2 kˈu2ŋ ɲˈaw sˈy4 zˈu6ŋ ˌaːˈi.",
    "duration": 8.39,
    "speaker": "jellyfish1010_0041",
    "gender": "male",
    "language": "vi"
}

Các trường dữ liệu:

  • _id: ID duy nhất của mẫu (tên đã chuẩn hóa)
  • audio: File audio WAV (24kHz) - tên file gốc
  • text: Văn bản tiếng Việt gốc
  • phonemized_text: Phiên âm IPA của văn bản
  • duration: Độ dài audio (giây)
  • speaker: ID của speaker
  • gender: Giới tính speaker (male/female)
  • language: Mã ngôn ngữ (vi)

Cách sử dụng

Tải dataset

from datasets import load_dataset

# Load toàn bộ dataset
dataset = load_dataset("pnnbao-ump/VieNeu-TTS")

# Xem một mẫu
sample = dataset['train'][0]
print(f"ID: {sample['_id']}")
print(f"Text: {sample['text']}")
print(f"Phonemized: {sample['phonemized_text']}")
print(f"Duration: {sample['duration']}s")
print(f"Speaker: {sample['speaker']}")
print(f"Gender: {sample['gender']}")
print(f"Audio shape: {sample['audio']['array'].shape}")
print(f"Sample rate: {sample['audio']['sampling_rate']} Hz")

Thống kê Dataset

  • Tổng số mẫu: 74,858
  • Số speaker: 193
  • Độ dài trung bình: ~6.76 giây/mẫu
  • Tổng thời lượng: ~140.7 giờ
  • Chất lượng audio: WAV format, 24kHz
  • Phonemization: IPA (International Phonetic Alphabet) with stress markers

Giấy phép

Dataset này được phát hành dưới giấy phép Apache 2.0.

Bạn có thể:

✅ Sử dụng cho mục đích thương mại
✅ Sửa đổi và phân phối
✅ Cấp phép lại
✅ Sử dụng trong dự án private

Với điều kiện:

📝 Giữ nguyên thông báo bản quyền và giấy phép
📝 Ghi rõ những thay đổi đã thực hiện
📝 Bao gồm bản sao giấy phép Apache 2.0

⚠️ Lưu ý: Dataset được cung cấp "nguyên trạng" không có bất kỳ bảo hành nào.

Xem chi tiết tại: https://www.apache.org/licenses/LICENSE-2.0

Vui lòng trích dẫn nguồn khi sử dụng.

Citation

@dataset{vieneu_tts_140h,
  author = {Pham Nguyen Ngoc Bao},
  title = {VieNeu-TTS: Vietnamese Multi-Speaker Text-to-Speech Dataset},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/datasets/pnnbao-ump/VieNeu-TTS}},
  note = {140.7 hours, 193 speakers, Apache 2.0 License}
}

Liên hệ

Nếu có bất kỳ câu hỏi hoặc vấn đề nào, vui lòng tạo issue trên repository.


Tác giả: Pham Nguyen Ngoc Bao
Năm: 2025