/model_serving_app

Primary LanguagePython

Model Serving App

This web app provides a platform for specific users to add, share, and edit ML models.

Usage

When on the FastAPI UI, to access some of the enpoints, click the lock on the top right and enter in username and password of an existing user.

Code Setup

  1. Run the server-side FastAPI app in one terminal window:

    $ cd backend
$ python3.9 -m venv venv
$ source venv/bin/activate
(venv)$ pip install -r requirements.txt
(venv)$ python main.py

  2. Starting up PostgreSQL:

    Navigate to [https://www.postgresql.org/download/] and select download depending on the OS

Config

The environment variables are in the 'backend/app/.env_backend' and some can be tweaked

Run with docker-compose

  • Checkout the 'docker-compose.yml' for detailed configuration.
# if it needs to be built,
$ docker-compose up -d --build

# if it only needs to run
$ docker-compose up -d
  • This will run the container for the backend side and the database
  • Access the site at 'localhost:48000/docs'.

Connect to PostgreSQL Database Server

$ docker exec -it model_serving_app_database_1 /bin/bash
:/# psql -U postgres
  • '\l' to list databases
  • '\c' to connect to a database
  • '\d' to list tables in the connected database

Project Support and Development

This project has been developed as part of my internship at the NCSOFT Vision AI Lab in summer 2021.