MLflow 3 Tutorials

This is a collection of tutorials for MLflow 3.

• MLflow 3 Tutorials
  • Installation (development)
  • Running projects
    • First MLflow Model
    • Hyperparameter Tuning & Deployment
    • Deep Learning Quickstart
    • Model Registry Quickstart

Installation (development)

You can install the project with uv, using the following command:

# Sync all dependency groups and download `uv`-managed Python
uv sync --managed-python --all-groups

Note

Running scripts

• Check pyproject.toml for available scripts
• Run them using uv run <script_name>

Running projects

All the information about running the individual projects can be found here.

Before running any projects, be sure to start the tracking server:

# Start the MLflow tracking server
uv run start_server

Tip

You can run rm -rf .venv/ uv.lock after each project, and before running the next required uv sync command to cut down on the size of your .venv folder.

First MLflow Model

Important

Ensure you have run the initial uv sync command.

• Code: first_model.py
  • Run: uv run first_model
    • If you encounter an error that the experiment already exists:
      • Run: uv run remove_experiments

Hyperparameter Tuning & Deployment

Important

Run:

• uv sync --managed-python --all-groups --extra tensorflow --extra mlflow_extras

• Code: tuning_deployment.py
  • Run: uv run tuning_deployment
  • Serve: uv run serve_wine_model

Deep Learning Quickstart

Important

Run:

• uv sync --managed-python --all-groups --extra pytorch

• Code: deep_learning_quickstart.py
  • Run: uv run deep_learning

Model Registry Quickstart

• Work in progress