Palette-FastAPI-model-deployment

Repository for serving models of the Palette service using FastAPI framework on EC2 server

Description

Palette is a date course recommendation platform for couples.
This repository covers how we serve the recommender AI model used in the Palette.

Method

• Implement model serving by loading the FastAPI web framework on EC2
• Provides real-time recommendation inference results
• Access the database and parse the returned restaurant IDs by category

Contributions

• Version management in virtual environments using pyenv and poetry
• Implementing very fast response times suitable for real-time services
• Runs in server background for continuous operation

How to use

There are instructions to run the server

Reference

• FastAPI - A web framework for building APIs in Python
• Amazon EC2 - Server to load model API