Dataset · MohammadJRanjbar
Multi-speaker Persian speech corpus for TTS and ASR
MohammadJRanjbar/ParsVoice
Largest open audio-text corpus in Persian, collected from audiobooks for speech synthesis and recognition.
What's inside
The dataset contains 1 to 10 million audio-text pairs in Persian, along with tabular metadata. Records are extracted from long Persian audiobooks using an automated pipeline: sentence segmentation, ASR transcription, and filtering through the ParsBERT model. Accepted at the EMNLP 2026 main conference.
What you can do with it
- Training multi-speaker Persian text-to-speech (TTS) models
- Training and fine-tuning speech recognition (ASR) systems for the Persian language
- Audio classification and speaker identification
- Tasks
- Text → speech, Automatic speech recognition, Audio classification
- Languages
- Persian
- Size
- 1M–10M rows
- Downloads this month
- 2.6K
- Likes
- 45
- Access
- on request on Hugging Face
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("MohammadJRanjbar/ParsVoice")