VirtuTA is an AI teaching assistant that delivers quick, accurate responses to student queries directly on Piazza. Powered by agentic workflows, Google Gemini, and Langchain, it automates both conceptual and logistical course queries.

Giving RecurrentGemma sight.

Multi-Modal Representational Learning for Social Media Popularity Prediction

A framework streamlining Training, Finetuning, Evaluation and Deployment of Multi Modal Language models

Localized multimodal large language model integrated with Streamlit and Ollama for interactive text and image processing tasks.

Multimodal RAG and comparisons between language models. (Project for Deep Learning Module at the FHSWF)

A curated list of awesome Image captioning strudies, aimed at annotating and reporting CT / MRI scans

Implementation of "Arcana: Improving Multi-modal Large Language Model through Boosting Vision Capabilitie"

A Streamlit-based AI assistant generates custom Streamlit app code from user-provided images or text using the Google Gemini model.

Voice assistant using Multimodal LLMs - LLaVA-NeXT (Mistral 7B) finetuned & PhoWhisper

Composition of Multimodal Language Models From Scratch

Pressure Testing Large Video-Language Models (LVLM): Doing multimodal retrieval from LVLM at any video lengths to measure accuracy

Official implementation of "Holmes-VAD: Towards Unbiased and Explainable Video Anomaly Detection via Multi-modal LLM"

Official implementation for "MJ-BENCH: Is Your Multimodal Reward Model Really a Good Judge?"

up-to-date and curated list of awesome state-of-the-art LVLMs hallucinations research work, papers & resources

VideoHallucer, The first comprehensive benchmark for hallucination detection in large video-language models (LVLMs)

Advances in recent large vision language models (LVLMs)