Modelldriftsättning

Introduktion till MLflow

Weston Bassler

Senior MLOps Engineer

ML-livscykeln

Steg i ML-livscykeln

1 datacamp.com
Introduktion till MLflow

Modellversioner och stadier

  • Modellversioner

    • Version +1

  • Modellstadier

    • Staging
    • Production
    • Archived

Modellversioner och stadier

Introduktion till MLflow

Sätt att driftsätta modeller

Läs in modell

# MLflow flavor
mlflow.FLAVOR.load_model()

Servera modell

# MLflow serve command-line
mlflow models serve
Introduktion till MLflow

Modellers URI

Konvention

models:/

Modellversion

models:/model_name/version

Modellstadium

models:/model_name/stage
Introduktion till MLflow

Läs in modeller

# Import flavor
import mlflow.FLAVOR


# Load version
mlflow.FLAVOR.load_model("models:/model_name/version")


# Load stage
mlflow.FLAVOR.load_model("models:/model_name/stage")
Introduktion till MLflow

Exempel på att läsa in modeller

# Import flavor
import mlflow.sklearn


# Load Unicorn model in Staging
model = mlflow.sklearn.load_model("models:/Unicorn/Staging")

# Print model
model

LogisticRegression()

# Inference
model.predict(data)
Introduktion till MLflow

Servera modeller

# Serve Unicorn model in Production stage
mlflow models serve -m "models:/Unicorn/Production"

2023/03/26 15:07:00 INFO mlflow.models.flavor_backend_registry: 
Selected backend for flavor 'python_function'
2023/03/26 15:07:00 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'
[2023-03-26 15:07:00 -0400] [86409] [INFO] Starting gunicorn 20.1.0
[2023-03-26 15:07:00 -0400] [86409] [INFO] Listening at: http://127.0.0.1:5000
[2023-03-26 15:07:00 -0400] [86409] [INFO] Using worker: sync
[2023-03-26 15:07:00 -0400] [86410] [INFO] Booting worker with pid: 86410
Introduktion till MLflow

Invocations-endpoint

api

http://localhost:5000/invocations

  • Format
    • CSV
    • JSON
1 Flaticon.com
Introduktion till MLflow

CSV-format

pandas_df.to_csv()

JSON-format

{
  "dataframe_split": {
      "columns": ["R&D Spend", "Administration", "Marketing Spend", "State"],
      "data": [["165349.20", 136897.80, 471784.10, 1]]
  }
}
Introduktion till MLflow

Modellprediktion

# Send payload to invocations endpoint
curl http://127.0.0.1:5000/invocations -H 'Content-Type: application/json' -d
{
  "dataframe_split": {
      "columns": ["R&D Spend", "Administration", "Marketing Spend", "State"],
      "data": [["165349.20", 136897.80, 471784.10, 1]]
  }
}

[[104055.1842384]]
Introduktion till MLflow

Nu kör vi en övning!

Introduktion till MLflow

Preparing Video For Download...