Generating Piano music using WaveNet

This repo uses the WaveNet architecture to generate piano music

WaveNet is a Deep Learning-based generative model for raw audio developed by Google DeepMind WaveNet takes the chunk of raw audio wave as input and tries to predict the scuccessive amplitude value as output.

The building block of WaveNet is Casual Dialated 1D Convolution layers.

1D Covolutions The objective of 1D convolution is similar to the Long Short Term Memory model. It is used to solve similar tasks to those of LSTM. In 1D convolution, a kernel or a filter moves along only one direction.

Casual 1D Convolutions This is defined as convolutions where output at time t is convolved only with elements from time t and earlier in the previous layer.

Dialated Casual 1D Convolutions A Causal 1D convolution layer with the holes or spaces in between the values of a kernel is known as Dilated 1D convolution. The Dialated 1D convolution network increases the receptive field by exponentailly increasing the dialation rate of every hidden layer.

Residual Block of WaveNet Residual and Skip Connections are used thoughout the network to speed up convergence and enable training of much deeper models.

Workflow of WaveNet

1. Input is fed into 1D convolutions.
2. The output is then fed to 2 different dilated 1D convolution layers with sigmoid and tanh activations
3. The element-wise multiplication of 2 different activation values results in a skip connection
4. And the element-wise addition of a skip connection and output of causal 1D results in the residual

Read this and this for more information on WaveNet

Dataset used can be downloaded from here