LLM-Financial-Advisor: Fine-Tuning a LLM with Finance Data in AWS

Overview

This project explores the application of generative AI in enhancing customer experience through the fine-tuning of a large language model (Meta Llama 2 7B) using Amazon SageMaker and AWS tools. The aim is to develop a domain expert model capable of generating informative, accurate, and contextually relevant text responses for specific domains such as finance (already fine-tuned), medical, or IT.

Objective

The goal is to train (fine-tune) the Meta Llama 2 7B foundation model to become proficient in a chosen domain, functioning as a knowledgeable consultant for generating text content that aids internal and customer-facing activities.

Technologies Used

• Amazon SageMaker: For model training, testing, and deployment.
• AWS IAM: For managing permissions and access to AWS services.
• Jupyter Notebook: For executing Python code and model training scripts.
• Python: Programming language used for model training and evaluation.

Setup Instructions

1. Configure AWS Environment:

   • Set up an AWS SageMaker IAM Role with necessary permissions.
   • Create a SageMaker Notebook Instance using the created IAM role.
   • Ensure the region is set to US-west-2 (Oregon).

2. Download Project Files:

   • Clone the repository and navigate to the Notebooks directory.
   • Start the SageMaker Notebook instance and open JupyterLab to upload the notebook files.

3. Run the Notebooks:

   • Use the Model_Evaluation.ipynb to deploy and evaluate the pre-trained model.
   • Use the Model_FineTuning.ipynb to fine-tune the model on your selected domain-specific dataset.

Project Steps

1. Select the Dataset: Choose a dataset from the Datasets directory relevant to your chosen domain.
2. Fine-tune the Model: Utilize the notebook to fine-tune the model using the selected dataset, already fine-tuned on financial dataset.
3. Deploy and Evaluate the Model: Deploy the fine-tuned model and evaluate its performance on domain-specific tasks.
4. Documentation and Submission: Document the process, challenges, and outcomes in the Project Documentation Report.

Budget Management

• Ensure to manage AWS resource usage within your limit by stopping instances and deleting unnecessary resources.