To generate music, we will need a good size set of training data of music files. These will be used to extract sequences while building our training dataset. To simplify this process, in this chapter, we are using the soundtrack of a single instrument. We collected some melodies and stored them in MIDI files. The following sample of a MIDI file shows you what this looks like:
The image represents the pitch and note distribution for a sample MIDI file
We can see the intervals between notes, the offset for each note, and the pitch.
To extract the contents of our dataset, we will be using music21. This also takes the output ...
