MLFLOW-DEMO

Getting Started with MLflow

Step 1: Install MLflow

Install MLflow using pip or conda or uv:

pip install mlflow
or
conda install -c conda-forge mlflow
or 
uv add mlflow

Step 2: Initialize MLflow

Create a new Jupyter notebook and run the following code to initialize MLflow:

import mlflow

# Initialize MLflow
mlflow.set_tracking_uri("http://127.0.0.1:5000")

This sets the tracking URI to http://127.0.0.1:5000, which is the default URI for the MLflow UI.

Step 3: Create an Experiment

Create a new experiment with a name of your choice:

# Create an experiment
experiment_id = mlflow.create_experiment("my_experiment")

print(f"Experiment ID: {experiment_id}")

This creates a new experiment and prints the experiment ID.

Step 4: Log Metrics and Parameters

Log some metrics and parameters to the experiment:

# Log metrics and parameters
with mlflow.start_run(experiment_id=experiment_id):
    mlflow.log_metric("accuracy", 0.9)
    mlflow.log_param("learning_rate", 0.01)

This logs the accuracy metric and learning rate parameter to the experiment.

Setting an Experiment as Default

You can also set an experiment as the default experiment, so that all subsequent mlflow.start_run() calls will log to that experiment:

# Set an experiment
mlflow.set_experiment("Check localhost connection")

# Log metrics
with mlflow.start_run():
    mlflow.log_metric("test", 1)
    mlflow.log_metric("random", 2)

Step 5: View the Experiment

View the experiment in the MLflow UI:

mlflow ui

This starts the MLflow UI and displays the experiment.

Example Use Case

Suppose we want to train a simple linear regression model using scikit-learn and log the model to MLflow. We can do this in a Jupyter notebook:

import pandas as pd
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split

# Load the data
data = pd.read_csv("data/raw/data.csv")

# Split the data into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(data.drop("target", axis=1), data["target"], test_size=0.2, random_state=42)

# Train the model
model = LinearRegression()
model.fit(X_train, y_train)

# Log the model to MLflow
with mlflow.start_run(experiment_id=experiment_id):
    mlflow.sklearn.log_model(model, "model")

This trains a linear regression model and logs it to MLflow.

Tips and Variations

• You can also use mlflow.set_tracking_uri() to set the tracking URI to a different location, such as a remote server.
• You can use mlflow.create_experiment() to create an experiment with a specific name and description.
• You can use mlflow.log_artifact() to log artifacts, such as images or text files, to the experiment.
• You can use mlflow.search_runs() to search for runs based on specific criteria, such as metrics or parameters.