Overview

This project uses a Convolutional Neural Network architecture alongside spectrogram generation to identify the musical key of a song. Uses GiantSteps dataset with .wav files

Attribution

Input handling/application architecture modified from

https://github.com/francesclluis/source-separation-wavenet

Model inspired by

https://github.com/drscotthawley/audio-classifier-keras-cnn/blob/master/eval_network.py