This is a Keras implementation to classify genre of audio files from GTZAN Genre Collection using Neural Networks.
- Instead of applying dense neural networks, convert audio files to spectrogram images and apply cnn model
- Model is overfitting, therefore, apply regularisation
- Other dataset that can be used is Audio set dataset made available by Google Inc.
- Dropout: A Simple Way to Prevent Neural Networks from Overfitting
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- This notebook uses librosa library to extract features features from audio clips.
- Medium blog by Parul Pandey