This repository contains code and resources for a medical question-answering system using LangChain and Hugging Face's Transformers library. The system is designed to answer medical questions based on the MASHQA dataset and demonstrates the integration of advanced NLP models for medical domain-specific tasks.
-
data_processing.ipynb: Jupyter Notebook providing code for data processing and setup.
-
dataset.py: Python script for handling the MASHQA dataset.
-
finetune_flan_t5.py: Script for fine-tuning a medical question-answering model based on 'google/flan-t5-small'.
-
gradio_app.py: Python script to deploy the question-answering model using the Gradio library for interactive web-based access.
-
langchain_rag_medical_qa.ipynb: A Jupyter Notebook that explains the LangChain pipeline for retrieval-augmented generation in medical question-answering.
- Clone this repository to your local machine. git clone https://github.com/kolubex/Megathon/
Link to the presentation: presentation
Megathon Submission repository, Team: BannedOnStackOverflow