Created a custom music genre classification dataset to mirror the GTZAN dataset. Used chop.sh to convert the audio files from either .flac or .mp3 to .wav then divide the audio files into 3 second segments.
mgc.ipynb is a notebook for the GTZAN dataset and mgc_custom.ipynb is for our custom dataset.
Inside the notebooks, extracted MFCC features from the audio files and trained CNN models with it to classify music genres. Tested the models and performed analysis and visualizations with the results.