JavaScript Insurance App LLM Finetuning

This project is a Node.js application that uploads a training dataset to OpenAI for fine-tuning a model. The fine-tuning job is initiated via the OpenAI API.

This app is designed to fine-tune a language model (GPT-3.5-turbo) to act as an insurance chatbot. It uploads an insurance-specific training dataset (stored in training_data.jsonl) to OpenAI and then initiates a fine-tuning job using that file. The resulting model is focused on answering insurance-related questions while filtering out unrelated topics. The main functionality is implemented in index.js

Project Structure

• .env - (To be created one for yourself after you clone) Contains environment variables like the OpenAI API key.
• .gitignore - Excludes node_modules and the .env file.
• index.js - Main application file (see code snippet below).
• package.json - Contains project metadata and dependencies.
• training_data.jsonl - Training dataset in JSON Lines format.

Prerequisites

• Node.js installed on your machine.
• An OpenAI API key. Obtain one from the OpenAI website.

Setup

1. Clone the repository.
2. Install dependencies by running:

   npm install

3. Create a .env file in the root directory and add your OpenAI API key: OPENAI_API_KEY="your_openai_api_key_here"

Usage

To start the fine-tuning process, run the following command in your terminal: node index.js

The application will:

Upload your training file (training_data.jsonl) to OpenAI. Initiate a fine-tuning job using the uploaded file. Log responses to the terminal, allowing you to track the job's status.