Domain Expert Model Fine-Tuning with Meta Llama 2 7B

This project was completed to fulfill the requirements of the Nanodegree course titled "Introducing Generative AI with AWS," offered by Udacity in collaboration with AWS

Project Overview

Welcome to the domain expert model fine-tuning project using the Meta Llama 2 7B foundation model. In this project, a specialized language model capable of generating contextually relevant text in healthcare/medical domain is created. This model will serve as a knowledgeable assistant, capable of enhancing user experience and streamlining information delivery.

Project Objective

The main objective of this project is to fine-tune the Meta Llama 2 7B model on a domain-specific dataset; that is, healthcare/medical dataset. By doing so, a language model proficient in understanding and completing/generating domain-specific text is developed. This model can be utilized for applications like chatbots, internal knowledge bases, and text content generation.

Project Tasks

Fine-Tuning the Language Model

1. Environment Setup: Configure AWS Sagemaker resources and necessary Python libraries.
2. Model Deployment: Deploy the Meta Llama 2 7B foundation model on the AWS platform.
3. Dataset Integration: Fine-tune the model using a selected domain-specific dataset.
4. Testing and Evaluation: Evaluate the model's performance on domain-specific text generation tasks.

Outcomes

• Gain advanced skills in machine learning and natural language processing.
• Hands-on experience with AWS Sagemaker and deploying models on cloud platforms.
• Insights into training domain-specific language models and their applications.
• Practical application of AI in solving real-world business challenges.