PersianPunc: A Large-Scale Dataset and BERT-Based Approach for Persian Punctuation Restoration

Mohammad Javad Ranjbar Kalahroodi, Heshaam Faili, Azadeh Shakery


Abstract
Punctuation restoration is essential for improving the readability and downstream utility of automatic speech recognition (ASR) outputs, yet remains underexplored for Persian despite its importance. We introduce PersianPunc, a large-scale, high-quality dataset of 17 million samples for Persian punctuation restoration, constructed through systematic aggregation and filtering of existing textual resources. We formulate punctuation restoration as a token-level sequence labeling task and fine-tune ParsBERT to achieve strong performance. Through comparative evaluation, we demonstrate that while large language models can perform punctuation restoration, they suffer from critical limitations: over-correction tendencies that introduce undesired edits beyond punctuation insertion (particularly problematic for speech-to-text pipelines) and substantially higher computational requirements. Our lightweight BERT-based approach achieves a macro-averaged F1 score of 91.33% on our test set while maintaining efficiency suitable for real-time applications. We make our dataset and model publicly available to facilitate future research in Persian NLP and provide a scalable framework applicable to other morphologically rich, low-resource languages.
Anthology ID:
2026.silkroadnlp-1.11
Volume:
The Proceedings of the First Workshop on NLP and LLMs for the Iranian Language Family
Month:
March
Year:
2026
Address:
Rabat, Morocco
Editors:
Rayyan Merchant, Karine Megerdoomian
Venues:
SilkRoadNLP | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
105–113
Language:
URL:
https://aclanthology.org/2026.silkroadnlp-1.11/
DOI:
Bibkey:
Cite (ACL):
Mohammad Javad Ranjbar Kalahroodi, Heshaam Faili, and Azadeh Shakery. 2026. PersianPunc: A Large-Scale Dataset and BERT-Based Approach for Persian Punctuation Restoration. In The Proceedings of the First Workshop on NLP and LLMs for the Iranian Language Family, pages 105–113, Rabat, Morocco. Association for Computational Linguistics.
Cite (Informal):
PersianPunc: A Large-Scale Dataset and BERT-Based Approach for Persian Punctuation Restoration (Kalahroodi et al., SilkRoadNLP 2026)
Copy Citation:
PDF:
https://aclanthology.org/2026.silkroadnlp-1.11.pdf
Poster:
 2026.silkroadnlp-1.11.poster.pdf