Mobile Deep Learning with TensorFlow Lite, ML Kit and Flutter

Deploying a model using Flask

For this part of the project, you may use your local system or deploy the script in app.py elsewhere. We will be editing this file to create a Flask server that generates music and allows the MIDI files that are generated to be downloaded.

Some of the code in this file will be similar to the Jupyter notebook, since the audio samples will always require a similar treatment every time they're loaded and used with the model we've generated:

1. We use the following code to make the required module imports to this script:

import midofrom mido import MidiFile, MidiTrack, Messagefrom tensorflow.keras.models import load_modelfrom sklearn.preprocessing import MinMaxScalerimport numpy as npimport randomimport timefrom flask ...