Music generation using LSTMs
As we saw in the previous section, music is a continuous signal, which is a combination of sounds from various instruments and voices. Another characteristic is the presence of structural recurrent patterns which we pay attention to while listening. In other words, each musical piece has its own characteristic coherence, rhythm, and flow.
Such a setup is similar to the case of text generation we saw in Chapter 9, The Rise of Methods for Text Generation. In the case of text generation, we saw the power and effectiveness of LSTM-based networks. In this section, we will extend a stacked LSTM network for the task of music generation.
To keep things simple and easy to implement, we will focus on a single instrument/monophonic music generation task. Let's first look at the dataset and think about how we would prepare it for our task of music generation.
Dataset preparation
MIDI is an easy-to-use format which helps us extract a symbolic...
