This workshop focuses on building machine learning microservices for different uses cases in the music industry. To follow along, please go over the Set Up instructions below.

By the end of the workshop, you will be able to

• create different machine learning microservices
• build user interfaces for your applications
• connect different models together
• understand what music is and how to tweak it

1. At least a year of experience coding
2. Knows how to use virtual environments
3. Does not mind using jupyter notebooks 😎

• You will need to have ffmpeg and docker installed
• At least 15GB of disk space available
• At least 16 GB of memory ram available

For linux

sudo apt-get install ffmpeg

For Mac

brew install ffmpeg

Open up your terminal and navigate to a directory of your choosing in your computer. Once there, run the following command to get the code for the session.

 git clone https://github.com/ramonpzg/ml_microservices_workshop.git

Conversely, you can click on the green download button at the top and download all files to your desired folder/directory. Once you download it, unzip it and move on to the second step.

To get all dependencies, packages and everything else that would be useful in this tutorial, you can recreate the environment by first going into the directory for today.

cd ml_microservices_workshop

Then you will need to create an environment with all of the dependencies needed for the session by running the following command.

mamba create -n ml_micro_svc python=3.11
mamba activate ml_micro_svc
pip install -r requirements.txt

## OR

conda create -n ml_micro_svc python=3.11
conda activate ml_micro_svc
pip install -r requirements.txt

## OR

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Open up Jupyter Lab and you should be ready to go.

code .

## OR

jupyter lab

You're all set. Now navigate to 1_microservices directory and open the tutorial.ipynb notebook.

Micro-Services

• MLServer
• Gradio
• nicegui

Music/Audio

• Music information retrieval with Python — Mateusz Modrzejewski
• Neural Audio Effects || Christian Steinmetz
• Lessons Learned from a Decade of Audio Programming
• Working with Audio in Python (feat. Pedalboard) - presented by Peter Sobot

Tools to Try

• FoxDot
• librosa
• pedalboard
• demucs