parsbert

🤗 ParsBERT: Transformer-based Model for Persian Language Understanding

ALBERT: A Lite BERT for Self-supervised Learning of Language Representations for the Persian Language

A well-structured summarization dataset for the Persian language!

A Bert2Bert model which able to summarize articles!

🤗 ParsBERT Persian NER Tasks

Persian spelling correction based on Parsbert

ALBERT Persian Playground

A sentiment analysis project for Persian-language SnappFood reviews using Transformer models (ParsBERT), traditional machine learning methods, and deep learning architectures (RNN, LSTM, GRU). Includes an interactive UI built with Streamlit for review prediction.

This is an NLP project, where I am attempting to detect sarcasm in social networks in Persian language.

Hooshvare Research Lab

Identify if each of the words in a Persian sentence need a kasr-e-ezafeh tag or not.

This repo applies sentiment analysis on Persian text using the ArmanEmo dataset. Various transformer models, including ParsBert and XLM-RoBERTa, were tested to classify emotions. The persian_xlm_roberta_large model achieved the best accuracy. Preprocessing included text normalization and cleaning, with the use of modern NLP techniques for SA in fa.

Corpus building and NLP analysis for Persian Telegram channel messages. Includes a notebook to parse and clean channel_messages.json, stopword normalization, word cloud with a silhouette mask, and CSV outputs (filtered_messages.csv, final_results.csv). Reproducible pipeline for EDA and basic modeling.

Fine-tuned ParsBERT for Persian text author classification with 5-fold cross-validation achieving F1 ~91.8.

Persian sentence-level stylistic scoring prototype using linguistic features, ParsBERT embeddings, and Ridge Regression.

In this project, we attempted to analyze users' sentiments on Twitter. For this purpose, we utilized the parsBERT model

A chatbot to work as an assistant using telegram bots, Scoped in university of Kerman.

an IR engine that uses sentence embedding instead of tf-idf vectorization

Sentiment analysis and NER on Taaghche Persian comments. Includes a Taaghche book-page crawler + metadata extractor, plus document-level (TF‑IDF/LogReg, ParsBERT) and token-level (BiLSTM, BERT) classification notebooks.