Model serving

Introductie tot MLflow

Weston Bassler

Senior MLOps Engineer

MLflow Models

Standaardiseer modelpackaging
Log modellen
Modellevaluatie

Modelimplementatie

ml-levenscyclus

¹ datacamp.com

REST-API

/ping - voor health checks
/health - voor health checks
/version - om de MLflow-versie op te vragen
/invocations - voor modelscoring
Poort 5000

Invocations-endpoint

/invocations

No,Name,Subject
1,Bill Johnson,English
2,Gary Valentine,Mathematics

Content-Type:application/json of application/csv

{
    "1": {
        "No": "1",
        "Name": "Bill Johnson",
        "Subject": "English"
    },
    "2": {
        "No": "2",
        "Name": "Gary Valentine",
        "Subject": "Mathematics"
    }
}

CSV-indeling

Pandas DataFrame
pandas_df.to_csv()

JSON-indeling

dataframe_split - pandas DataFrame in split-orientatie
dataframe_records - pandas DataFrame in records-orientatie

DataFrame split

# DataFrame split-orientatie
{
  "dataframe_split": {
      "columns": ["sex", "age", "weight"],
      "data": [["male", 23, 160], ["female", 33, 120]]
  }
}

Modellen serven

# MLflow serve-commando
mlflow models serve --help
Usage: mlflow models serve [OPTIONS]

Serve-URI

# Lokaal bestandssysteem
mlflow models serve -m relative/path/to/local/model

# Run-ID
mlflow models serve -m runs:/<mlflow_run_id>/artifacts/model

# AWS S3
mlflow models serve -m s3://my_bucket/path/to/model

Serve-voorbeeld

# Serve model from run
mlflow models serve -m runs:/e84a122920de4bdeaedb54146deeb429/artifacts/model

2023/03/12 16:28:28 INFO mlflow.models.flavor_backend_registry: 
Selected backend for flavor 'python_function'
2023/03/12 16:28:28 INFO mlflow.pyfunc.backend: === Running command 
'exec gunicorn --timeout=60 -b 127.0.0.1:5000 -w 1 ${GUNICORN_CMD_ARGS} -- 
mlflow.pyfunc.scoring_server.wsgi:app'
[2023-03-12 16:28:29 -0400] [48431] [INFO] Starting gunicorn 20.1.0
[2023-03-12 16:28:29 -0400] [48431] [INFO] Listening at: http://127.0.0.1:5000 
(48431)
[2023-03-12 16:28:29 -0400] [48431] [INFO] Using worker: sync
[2023-03-12 16:28:29 -0400] [48432] [INFO] Booting worker with pid: 48432

Invocations-verzoek

# Stuur payload in dataframe_split-orientatie naar MLflow
curl http://127.0.0.1:5000/invocations -H 'Content-Type: application/json' -d '{
  "dataframe_split": {
      "columns": ["sex", "age", "weight"],
      "data": [["male", 23, 160], ["female", 33, 120]]
  }
}'

{"predictions": [1, 0]}

Laten we oefenen!

Introductie tot MLflow