1. Added custom fine tuned model to provide item explanations is specific format.

Feature branch to add feature for asynchronous inference with cv and llm models, for improving latency.

Added docker files for dev and prod.

Changed folder structure to align with hugging face streamlit spaces, where it has to be flat.

Renamed main to app, since huggingface spaces only allows app as the entry script name.

Copied entire codebase from working streamlit application from personal git. This application lets users upload images of item menus ( only supports indian item menus for now ); using opencv and llms, this application will explain each item in the menu in specific format.

initial commit