Modell-API

Introduktion till MLflow

Weston Bassler

Senior MLOps Engineer

MLflow REST API

api

¹ istock.com

Modell-API

Spara
Logga
Ladda

Scikit-Learn

¹ wikipedia.org

Modell-API:ets funktioner

# Save a model to the local filesystem
mlflow.sklearn.save_model(model, path)

# Log a model as an artifact to MLflow Tracking.
mlflow.sklearn.log_model(model, artifact_path)

# Load a model from local filesystem or from MLflow Tracking.
mlflow.sklearn.load_model(model_uri)

Ladda modell

Lokalt filsystem – relative/path/to/local/model eller /Users/me/path/to/local/model
MLflow Tracking – runs:/<mlflow_run_id>/run-relative/path/to/model
S3-stöd – s3://my_bucket/path/to/model

Spara modell

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


# Save model locall
mlflow.sklearn.save_model(lr, "local_path")

ls local_path/

MLmodel            model.pkl        requirements.txt        python_env.yaml

Ladda lokal modell

# Load model from local path
model = mlflow.sklearn.load_model("local_path")


# Show model
model

LogisticRegression()

Logga modell

# Model
lr = LogisticRegression(n_jobs=n_jobs)
lr.fit(X, y)


# Log model
mlflow.sklearn.log_model(lr, "tracking_path")

Tracking UI

Senast aktiva körning

# Format for runs
runs:/<mlflow_run_id>/run-relative/path/to/model

# Get last active run
run = mlflow.last_active_run()

run

<Run: data=<RunData: metrics={}, params={}, 
tags={'mlflow.runName': 'run_name'}>, 
 info=<RunInfo: artifact_uri='uri', end_time='end_time', 
 experiment_id='0', lifecycle_stage='active', run_id='run_id', 
 run_name='name', run_uuid='run_uuid', start_time=start_time, 
 status='FINISHED', user_id='user_id'>>

ID för senast aktiva körning

# Get last active run
run = mlflow.last_active_run()

# Show run_id of last run
run.info.run_id

'8c2061731caf447e805a2ac65630e70c'

Ange körnings-ID

# Get last active run
run = mlflow.last_active_run()

# Set run_id variable
run_id = run.info.run_id

run_id

'8c2061731caf447e805a2ac65630e70c'

Ladda modell från MLflow Tracking

# Pass run_id as f-string literal
model = mlflow.sklearn.load_model(f"runs:/{run_id}/tracking_path")

# Show model
model

LogisticRegression()

Nu kör vi en övning!

Introduktion till MLflow