Dataset · dolly-vn
1,000-hour Vietnamese speech corpus
dolly-vn/dolly-audio-1000h-vietnamese
Cleaned audio recordings and texts in Vietnamese from 152 speakers for speech synthesis and recognition tasks.
What's inside
Contains audio recordings with Vietnamese text transcripts — totaling between 100 thousand and 1 million examples. Total volume is nearly 1,000 hours of professionally cleaned audio from 152 different speakers. The dataset was collected by the Dolly AI Team.
What you can do with it
- Training and fine-tuning Vietnamese speech recognition (ASR) models
- Creating speech synthesis (TTS) systems with multi-voice support
- Voice modeling and cloning
- Languages
- Vietnamese
- Size
- 100K–1M rows
- Downloads this month
- 1.1K
- Likes
- 63
Models trained on it
- splendor1811/omnivoice-vietnameseText → speech 1.6K
- dolly-vn/viterboxText → speech 64
- Hacht/omnivoice-vietnameseText → speech 25
- Supbatomic/omnivoice-vietnameseText → speech 25
- vongocanhthi/omnivoice-vietnameseText → speech 15
- JayLL13/VoxCPM-0.5B-VNText → speech 14
- fnooub/VoxCPM-1.5-VNText → speech 13
- lukelucky29/viterboxText → speech 13
How to load it
from datasets import load_dataset
ds = load_dataset("dolly-vn/dolly-audio-1000h-vietnamese")