Building a music genre classification model for the McGill AI Bootcamp. Using the GTZAN Genre Collection dataset consisting of 1000 audio tracks each 30 seconds long. It contains 10 genres, each represented by 100 tracks. The tracks are all 22050Hz Mono 16-bit audio files in .wav format. The dataset consists of 10 genres: Blues, Classical, Country, Disco, Hip-hop, Jazz, Metal, Pop, Reggae, Rock. Each genre contains 100 audio tracks. This is the Web Application that will be used to demonstrate my project.
- Flask (1.1.1)
- Werkzeug (0.16.0)
- Numpy (1.12.1)
- Librosa (0.7.1)
- Pydub (0.18.0)
- ffmpeg (4.2.1)
git clone https://github.com/xinruili07/MusicGenreClassifierWebsite.git
pip install -r requirements.txt
brew install ffmpeg
python3 app.py
- App will run on
localhost:8080
The web application is written in Python using Flask. It uses a saved model (more details can be found MusicGenreClassifier for finding the genre of input song.
With 10 genre classes, we are getting a test accuracy of 77%