Dataset · pnnbao-ump
140 hours of Vietnamese speech with phonemic notation
pnnbao-ump/VieNeu-TTS-140h
A benchmark dataset of 74,858 Vietnamese audio recordings with IPA transcriptions for speech synthesis and recognition.
What's inside
The dataset contains 74,858 audio samples (about 140 hours) in Vietnamese and their text transcripts. The entire text corpus is fully converted into a phonemic representation using the International Phonetic Alphabet (IPA). The set is positioned as a high-quality benchmark for acoustic modeling.
What you can do with it
- Fine-tuning modern speech synthesis (TTS) models for the Vietnamese language
- Training and testing phoneme-based speech recognition (ASR) systems
- Tasks
- Text → speech, Automatic speech recognition
- Languages
- Vietnamese
- Size
- 10K–100K rows
- Downloads this month
- 856
- Likes
- 40
- Access
- on request on Hugging Face
Models trained on it
- pnnbao-ump/VieNeu-TTS-0.3B-q4-ggufText → speech 177.5K
- pnnbao-ump/VieNeu-TTS-0.3BText → speech 1.5K
- pnnbao-ump/VieNeu-TTSText → speech 1.1K
- mradermacher/VieNeu-TTS-i1-GGUF 361
- mradermacher/VieNeu-TTS-0.3B-GGUF 170
- pnnbao-ump/VieNeu-TTS-q8-ggufText → speech 130
- pnnbao-ump/VieNeu-TTS-q4-ggufText → speech 127
- pnnbao-ump/VieNeu-TTS-0.3B-q8-ggufText → speech 109
How to load it
from datasets import load_dataset
ds = load_dataset("pnnbao-ump/VieNeu-TTS-140h")