Logo
HomeAboutProjectsCreationsContact
Hire Me
Back to projects
Ayush BhandariCompletedAcademicFeatured

Nepali Spelling Correction Using Transformer Model

A research project implementing transformer-based sequence-to-sequence models for Nepali language spelling correction, comparing mT5-small, mBART, and varta-T5 models.

2023 - 2024
4 Contributors
9 Technologies

Project Overview

As Nepal is embracing digitization across various sectors, the need for robust language tools like spelling correction tool for Nepali language is very important. Spelling errors, the fundamental errors in any language, hinder effective communication, knowledge creation, and further linguistic studies. The complex nature of its Devanagari script and nuanced grammar poses challenges in terms of spelling accuracy. While a previous research has employed sequence-to-sequence technique using GRU-based model with Levenshtein distance for Nepali spelling correction. This project takes a significant step forward by leveraging the power of transfer learning of transformer based sequence-to-sequence pre-trained models and explores their effectiveness by employing and comparing the performance of three distinct transformer-based models of different sizes. A parallel dataset is prepared from A Large Scale Nepali Text Corpus, using pseudo-random synthetic data generation technique, to fine tune these models. The three models varta-t5, mt5-small and mbart obtained accuracy of 75.47%, 76% and 75% on their test set, and 48.22%, 39.11% and 51.67% on the test set of different domain, showing the higher performance of the mt5-small on specific domain and better generalization capacity of mbart on dataset of other domain.

Key Features

1Transformer-based sequence-to-sequence spelling correction
2Comparison of mT5-small, mBART, and varta-T5 models
3Synthetic parallel corpus generation for training
4Web-based user interface using Streamlit
5Model deployment on Hugging Face Spaces
6Accuracy metrics including precision, recall, and F1-score
7Cross-domain evaluation on TV program scripts
8Error analysis and correction highlighting
9Support for multiple model selection in UI
10Real-time spelling correction with visual feedback

Tech Stack

PythonTransformer models (mT5-small, mBART, varta-T5)Hugging Face Transformers libraryPyTorch frameworkStreamlit for web UIHugging Face Spaces for deploymentScikit-learn for evaluation metricsDataset library for data handlingPandas for data manipulation

Tags

ResearchNLPTransformerMachine LearningNepaliSpelling CorrectionDeep Learning

Team

  • Ayush Bhandari
  • Amar Dura
  • Harish Joshi
  • Sugam Pokharel

More Projects

Unified Devanagari Rendering Engine for Nepali Language

A research report on developing a unified rendering engine for the Devanagari script specifically tailored for the Nepali language, addressing character shaping and text rendering challenges.

Completed

Hostel Finder

A C++/Qt application that connects hostel seekers with hostel owners, allowing users to search for hostels by location and category while enabling owners to manage their hostel listings through a secure dashboard.

Completed

Library Management System - LibMAX

A console-based library management system implemented in C programming language featuring secure admin authentication, book and student management, issue/return tracking with fine calculation, and rule configurability.

Completed

Simplex Method Implementation

A C++ implementation of the simplex method for solving linear programming problems with graphical visualization of the solution process.

Completed

Library Management System

A QBASIC-based library management system for tracking books, managing issuance/returns, and maintaining library records with fine calculation.

Completed
View All Projects ->

Subscribe to Newsletter

Get the latest updates and insights delivered to your inbox.

Protected by reCAPTCHA

Ayush Bhandari

Building digital experiences

Follow Me

Quick Links

  • Home->
  • About->
  • Projects->
  • Creations->

Creations

  • Blog Posts->
  • Poems->
  • Stories->
  • Articles->

Connect

  • Contact->
  • Email->

Legal & SEO

  • Privacy Policy->
  • HTML Sitemap->
  • XML Sitemap->

© 2026 Ayush Bhandari. All rights reserved.

Made withby Ayush Bhandari