模型 API

MLflow 入門

Weston Bassler

Senior MLOps Engineer

MLflow REST API

api

1 istock.com
MLflow 入門

Model API

  • 儲存

  • 紀錄

  • 載入

Scikit-Learn

1 wikipedia.org
MLflow 入門

Model API 函式

# 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)
MLflow 入門

載入模型

  • 本機檔案系統 - relative/path/to/local/model 或 /Users/me/path/to/local/model

  • MLflow Tracking - runs:/<mlflow_run_id>/run-relative/path/to/model

  • 支援 S3 - s3://my_bucket/path/to/model

MLflow 入門

儲存模型

# 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
MLflow 入門

載入本機模型

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


# Show model
model

LogisticRegression()
MLflow 入門

紀錄模型

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


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

追蹤介面

追蹤介面

MLflow 入門

最後活動的 run

# 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'>>
MLflow 入門

最後活動的 run id

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

# Show run_id of last run
run.info.run_id

'8c2061731caf447e805a2ac65630e70c'
MLflow 入門

設定 run id

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

# Set run_id variable
run_id = run.info.run_id

run_id

'8c2061731caf447e805a2ac65630e70c'
MLflow 入門

從 MLflow Tracking 載入模型

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

# Show model
model

LogisticRegression()
MLflow 入門

一起來練習吧！

MLflow 入門

Preparing Video For Download...