Dataset · vhdm
Cleaned Persian speech corpus based on Common Voice 17
vhdm/persian-voice-v1
Version of the Mozilla Common Voice 17 Persian section with corrected spelling errors for speech recognition tasks.
What's inside
The dataset combines audio recordings and Persian text transcripts (10K to 100K recordings). The authors took the Persian portion of Mozilla Common Voice 17 and manually cleaned spelling inaccuracies and typos in the transcripts to ensure more accurate audio-text alignment during model training.
What you can do with it
- Training and fine-tuning automatic speech recognition (ASR) models in Persian
- Improving transcription accuracy through clean annotations
- Testing Persian voice systems
- Tasks
- Automatic speech recognition
- Languages
- Persian
- Size
- 10K–100K rows
- Downloads this month
- 150
- Likes
- 32
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("vhdm/persian-voice-v1")