Introduction to MLflow Models

Introduction to MLflow

Weston Bassler

Senior MLOps Engineer

MLflow Models

Simplify ML library integration
Simplify Deployment
Convention called "Flavors"

flavors

¹ unsplash.com

Built-In Flavors

IMAGES OF FLAVORS

Write custom tools from ML libraries
Flavors simplify the new for custom code

# Import flavor from mlflow module
import mlflow.FLAVOR

¹ mlflow.org

Autolog

# Automatically log model and metrics
mlflow.FLAVOR.autolog()

# Scikit-learn built-in flavor
mlflow.sklearn.autolog()

Scikit-learn Flavor

# Import scikit-learn
import mlflow
from sklearn.linear_model import \
    LinearRegression


# Using auto-logging 
mlflow.sklearn.autolog()

# Train the model
lr = LinearRegression()
lr.fit(X, y)

Model will be logged automatically on model.fit()

Common Metrics

Regression
- mean squared error
- root mean squared error
- mean absolute error
- r2 score

Classification
- precision score
- recall score
- f1 score
- accuracy score

Common Parameters

MODEL.get_params()

Common parameters

# Train the model
lr = LinearRegression()
lr.fit(X, y)

# Get params
params = lr.get_params(deep=True)

params

{'copy_X': True, 'fit_intercept': True, 'n_jobs': None, 
    'normalize': 'deprecated', 'positive': False}

Autolog parameters

parameters

# Model
lr = LinearRegression()
lr.fit(X, y)

Autolog metrics

metrics

# Model
lr = LinearRegression()
lr.fit(X, y)

Storage format

Directory structure for a model:

model/
    MLmodel
    model.pkl
    python_env.yaml
    requirements.txt

# Autolog
mlflow.sklearn.autolog()

Artifacts

Contents of MLmodel

artifact_path: model
flavors:
  python_function:
    env:
      virtualenv: python_env.yaml
    loader_module: mlflow.sklearn
    model_path: model.pkl
    predict_fn: predict
    python_version: 3.10.8
  sklearn:
    code: null
    pickled_model: model.pkl
    serialization_format: cloudpickle
    sklearn_version: 1.1.3

MLmodel

Let's practice!

Introduction to MLflow