Skip to content

Dataset · doof-ferb

Unofficial mirror of FPT Open Speech Dataset

doof-ferb/fpt_fosd

Vietnamese audio dataset for automatic speech recognition and synthesis.

What's inside

An unofficial mirror of the FOSD dataset, containing about 100 hours and 25.9 thousand audio samples with transcripts in Vietnamese. Semantic noise and several files with missing transcriptions were removed from the set.

What you can do with it

  • Training automatic speech recognition (ASR) systems
  • Developing speech synthesis (TTS) models in Vietnamese
Tasks
Automatic speech recognition, Text → speech
Languages
Vietnamese
Size
10K–100K rows
Downloads this month
812
Likes
6

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("doof-ferb/fpt_fosd")