Add LoRA fine-tuning demo notebook

[andrefoo]

Description

This PR adds a basic Jupyter notebook demonstrating how to perform LoRA (Low-Rank Adaptation) fine-tuning using Fireworks AI. The notebook provides a step-by-step guide for customizing language models efficiently using customer support data as an example.

Features

• Complete setup instructions including prerequisites
• Dataset preparation and formatting
• LoRA fine-tuning with both default and custom parameters
• Model deployment and testing
• Resource cleanup guidance

Key Components

• Installation and credential setup
• Dataset loading and formatting for Fireworks' API
• Fine-tuning job creation and monitoring
• Model deployment and inference testing
• Resource management (deployment/undeployment)

Technical Details

• Uses llama-v3p1-8b-instruct as the base model
• Demonstrates LoRA parameter configuration (rank, learning rate, epochs)
• Includes error handling and troubleshooting guidance
• Compatible with Google Colab environment

Testing Done

• ✅ Notebook runs end-to-end in Google Colab
• ✅ All code cells execute successfully
• ✅ Resource cleanup steps verified

Documentation

• Added detailed comments and explanations throughout the notebook
• Included links to relevant documentation
• Added parameter explanations and best practices