Awesome

Music Genre Classification with LSTMs

• Classify music files based on genre from the GTZAN music corpus
• GTZAN corpus is included for easy of use
• Use multiple layers of LSTM Recurrent Neural Nets
• Implementations in PyTorch, PyTorch-Lightning, Keras

Test trained LSTM model

In the ./weights/ you can find trained model weights and model architecture.

To test the model on your custom audio file, run

 python3 predict_example.py path/to/custom/file.mp3

or to test the model on our custom files, run

 python3 predict_example.py audio/classical_music.mp3

Audio features extracted

• MFCC
• Spectral Centroid
• Chroma
• Spectral Contrast

Dependencies

• Python3
• numpy
• librosa → for audio feature extraction
• Keras
  • pip install keras
• PyTorch
  • pip install torch torchvision
  • brew install libomp

Ideas for improving accuracy:

• GTZAN dataset has problems, how do we use it with consideration?
• Normalize MFCCs & other input features (Recurrent BatchNorm?)
• Decay learning rate
• How are we initing the weights?
• Better optimization hyperparameters (too little dropout)
• Do you have avoidable bias? How's your variance?

Accuracy

At Epoch 400, training on a TITAN X GPU (October 2017):

At Epoch 400, training on a 2018 Macbook Pro CPU (May 2019):