hooshvare/parsbert
A Persian BERT model pre-trained on 1.3B words from 3.9M documents for downstream NLP tasks.
Not currently ranked — collecting fresh signals.
star history
ParsBERT is a transformer-based language model built on Google’s BERT architecture, specifically designed for Persian language understanding. It was pre-trained on a large corpus of Persian text including Wikipedia dumps, books, news articles, and various web sources. The model supports three downstream tasks: sentiment analysis, text classification, and named entity recognition (NER), achieving state-of-the-art results on Persian NLP benchmarks.
Frequently asked
- What is hooshvare/parsbert?
- A Persian BERT model pre-trained on 1.3B words from 3.9M documents for downstream NLP tasks.
- Is parsbert open source?
- Yes — hooshvare/parsbert is open source, released under the Apache-2.0 license.
- What language is parsbert written in?
- hooshvare/parsbert is primarily written in Jupyter Notebook.
- How popular is parsbert?
- hooshvare/parsbert has 417 stars on GitHub.
- Where can I find parsbert?
- hooshvare/parsbert is on GitHub at https://github.com/hooshvare/parsbert.