This repository contains the system description and the codes that we implemented for participating in EACL-2024 Shared Task-5.

DravidaKavacham is an open-source tool for detecting abusive content in Dravidian (Tamil and Malayalam), focused on harmful language targeting women. Developed as part of the DravidianLangTech@NAACL 2025 shared task, it uses NLP and machine learning for accurate text classification and content analysis.

Implementation of Dense Fusion Network with Multimodal Residual (DFMR) for Multi-modal Sentiment Analysis(MSA) in native Indian Languages like Malayalam. The model processes the textual, visual, and auditory modalities of the video to classify the sentiment into five categories.

Advanced Telugu sentiment analysis for YouTube comments using Meta Llama 3.2 with 4-bit quantization. Features interactive dashboard, code-mixed text support, real-time processing, and comprehensive evaluation metrics. Perfect for analyzing Telugu social media content and movie reviews.

EACL 2021 paper (SJ_AJ@DravidianLangTech-EACL2021: Task-Adaptive Pre-Training of Multilingual BERT models for Offensive Language Identification)