fine-tune-llama3-price-predictor
vasiliskou/fine-tune-llama3-price-predictor
Jupyter Notebook
This project fine-tunes the open-source LLaMA 3.1–8B model using QLoRA to predict product prices from descriptions. It includes data preprocessing, model training, evaluation against baseline and frontier models, deployment via Modal, and a Gradio-based UI for real-time inference.