This project aims to predict the genre of a given song file using Data Science and Machine Learning techniques.

This repository contains the REST API built using Fast API that serves the Tensorflow model and librosa to extract features from uploaded MP3 files.

• Web application: https://classify.k8s.pouretadev.com/

• Frontend - Web App: https://github.com/andbluedev/Music-Genre-Classification-Model-Web-APP
• Project Notebooks: https://github.com/andbluedev/Music-Genre-Classification-Notebooks

It is recommended to use a virtual python environment in order to install every dependency.

Create a virtual environment the venv directory at the root of this project.

python3 -m venv venv

To activate the created virtual environment

source venv/bin/activate

Installing dependencies:

pip install -r requirements.txt

uvicorn app.main:app --reload --port 5000

note: any other availabel port can be used

The Fast API for production uses Docker, python and uvicorn and is deployed on a kubernetes cluster.

docker build -t music-genre-prediction-model-rest-api .

docker run -p 5000:5000  music-genre-prediction-model-rest-api

note: Add -d to this command if you wan't to run the app in detached mode (running in the background leaving the current shell available for other commands).

The API documentation can be found here while the app is running locally (assuming it is running on localhost:5000).