README.md

AI-Music

Repositorio para el TFG: AI-Music

Alumno: López Estrada, José Ángel

Tutores: César Ignacio García Osorio, Alicia Olivares Gil

---

Description

AI-Music is an application that predicts music genres from audio files. The application includes a user-friendly interface that enables the users to upload files and visualizate the predicted genres with the main features of the song.

External files

Dataset and metadata used to train the model is availiable in this repository: FMA: A Dataset For Music Analysis

The latest version of the application has been trained using:

1. fma_small.zip: audio tracks used to train the model.
2. fma_metadata.zip: audio tracks information to obtain the genres asociated with the audio files.

Information about training process

To begin with the training process the first step is process the data and extract features using /app/src/extract_features.py script.

Later, to train the model it is neccesary to download and extract a dataset from FMA: A Dataset For Music Analysis in /data/raw/

It can be used fma_small.zip, fma_medium.zip, fma_large.zip or fma_full.zip. The latest version of the application has been trained using fma_small.zip.

Usage

AI-Music project can be used online or locally. For local use, follow the next steps:

1. Clone the repository

git clone https://github.com/jle1001/AI-Music.git

2. Create a virtual environment

python3 -m venv venv

3. Activate the virtual environment

Linux

source venv/bin/activate

Windows

venv\Scripts\activate

4. Install dependencies

pip install -r requirements.txt

5. Run the application

flask run

License