Release: April 2024

Hello everyone, the purpose of this repository is create a a simple program that will answer medical questions by using the latest technologies of IBM.

The aim of this program is help people who require help. This program does not replace a real doctor but help to identify the possible health solution.

The technologies which will use is WatsonX of IBM and Watson Assistant.

Watsonx.ai is part of the IBM watsonx platform that brings together new generative AI capabilities, powered by foundation models, and traditional machine learning into a powerful studio spanning the AI lifecycle. With watsonx.ai, you can train, validate, tune, and deploy generative AI, foundation models, and machine learning capabilities with ease and build AI applications in a fraction of the time with a fraction of the data.

We are going to use Foundation Models and test different models such as:

• flan-ul2-20b - It is an encoder decoder model based on the T5 architecture and instruction-tuned using the Fine-tuned LAnguage Net.Model by Google
• mt0-xxl-13b - An instruction-tuned iteration on mT5.Model by BigScience
• gpt-neox-20b - A 20 billion parameter autoregressive language model trained on the Pile.Model by EleutherAI
• flan-t5-xxl-11b - It is an 11 billion parameter model based on the Flan-T5 family.Model by Google
• mpt-7b-instruct - It is a decoder-style transformer pretrained from scratch on 1T tokens of English text and code.

The structure of the program contains 5 parts.

1. Environment creation

   Here we are going to create the environment to create the models locally that later can be used

2. Creation of the Medical Dataset.

   In this part we are going to build the Datasets that will be used create the Medical Model

3. Creation of the model by using RAG In this part we will perform feature engineering and create the model

4. Finetuning Models for the Medical Chatbot We create a custom model based on medical information

5. Multimodal Medical Chatbot We develop a medical chatbot multimodal, that from images can give you a description of the issue. We analazize different Medical Images Datasets.

Implementation of a chatbot with WatsonX in production.

Here we will create a chatbot with the capability to answer questions by using the Model created before. For Production in WatsonX you can checkout this repo

Watsonx-Assistant-with-Milvus-as-Vector-Database

We have also developed another version which uses a custom LLM

Medical-Chatbot-with-Langchain-with-a-Custom-LLM

Medical-Chatbot by RAG method.

https://huggingface.co/spaces/ruslanmv/Medical-Llama3-Chatbot

Medical Chatbot by using Medical-Llama3-8B

https://huggingface.co/spaces/ruslanmv/Medical-Llama3-Chatbot

Currently there are two base models that were pretrained with ai-medical-chatbot

This repository provides a fine-tuned version of the powerful Llama3 8B model, specifically designed to answer medical questions in an informative way. It leverages the rich knowledge contained in the AI Medical Chatbot dataset.

Medical-Llama3-8B

Fine-tuned Mixtral model for answering medical assistance questions. This model is a novel version of mistralai/Mistral-7B-Instruct-v0.2, adapted to a subset of 2.0k records from the AI Medical Chatbot dataset, which contains 250k records . The purpose of this model is to provide a ready chatbot to answer questions related to medical assistance.

Medical-Mixtral-7B-v2k

For more details how was pretrained you can visit this post here

Let us use the best technologies in the world to help us.

Please free to contribute following the standard guidelines for submitting patches and additions or solutions. Feel free to submit issues and enhancement requests.

To more information visit www.ruslanmv.com

Copyright 2024 Ruslan Magana Vsevolodovna This program is distributed under the terms of the GNU Lesser General Public License.