Tutorial Lengkap MLflow: Dari Setup hingga Production

## Pendahuluan MLflow adalah platform open-source untuk mengelola end-to-end machine learning lifecycle. Dikembangkan oleh Databricks, MLflow membantu data scientist dan ML engineer untuk tracking ex...

By Ruby Abdullah · · tutorial
MLflowMLOpsMachine LearningModel VersioningPython

Pendahuluan

MLflow adalah platform open-source untuk mengelola end-to-end machine learning lifecycle. Dikembangkan oleh Databricks, MLflow membantu data scientist dan ML engineer untuk tracking experiments, packaging code, managing models, dan deploying ke production.

Mengapa MLflow?
  • Reproducibility: Track semua experiment dengan detail
  • Collaboration: Share results dengan team
  • Model versioning: Kelola berbagai versi model
  • Deployment ready: Deploy model dengan mudah ke berbagai platform
  • Framework agnostic: Bekerja dengan TensorFlow, PyTorch, Scikit-learn, dll

Komponen Utama MLflow

MLflow terdiri dari 4 komponen utama:

  • MLflow Tracking: Record dan query experiments
  • MLflow Projects: Package ML code untuk reproducibility
  • MLflow Models: Deploy models ke berbagai platform
  • MLflow Registry: Centralized model store untuk versioning

    • Instalasi dan Setup

    Instalasi Dasar

    # Install MLflow

    pip install mlflow
    


    Install dengan extras untuk berbagai backend
    

    pip install mlflow[extras]
    


    Verify instalasi
    

    mlflow --version

    Setup Database Backend (PostgreSQL)

    Untuk production, gunakan database backend:

    # Install dependencies

    pip install psycopg2-binary
    


    Setup PostgreSQL (contoh menggunakan Docker)
    

    docker run -d \
    

      --name mlflow-db \
    

      -e POSTGRESUSER=mlflow \
    

      -e POSTGRESPASSWORD=mlflow \
    

      -e POSTGRESDB=mlflow \
    

      -p 5432:5432 \
    

      postgres:13

    Setup Artifact Store (MinIO/S3)

    # Install boto3 untuk S3 compatibility

    pip install boto3
    


    Setup MinIO (S3-compatible storage)
    

    docker run -d \
    

      --name mlflow-minio \
    

      -p 9000:9000 \
    

      -p 9001:9001 \
    

      -e MINIOROOTUSER=minioadmin \
    

      -e MINIOROOTPASSWORD=minioadmin \
    

      minio/minio server /data --console-address ":9001"

    Jalankan MLflow Server

    # Development mode (local file store)

    mlflow server --host 0.0.0.0 --port 5000
    


    Production mode (dengan database dan S3)
    

    mlflow server \
    

      --backend-store-uri postgresql://mlflow:mlflow@localhost:5432/mlflow \
    

      --default-artifact-root s3://mlflow-artifacts \
    

      --host 0.0.0.0 \
    

      --port 5000

    Setup Environment Variables

    Buat file .env:

    # MLflow Tracking

    MLFLOWTRACKINGURI=http://localhost:5000
    


    S3/MinIO Configuration
    

    AWSACCESSKEYID=minioadmin
    

    AWSSECRETACCESSKEY=minioadmin
    

    MLFLOWS3ENDPOINTURL=http://localhost:9000

    MLflow Tracking: Experiment Tracking

    Basic Tracking

    import mlflow

    import mlflow.sklearn
    

    from sklearn.ensemble import RandomForestClassifier
    

    from sklearn.datasets import loadiris
    

    from sklearn.modelselection import traintestsplit
    

    from sklearn.metrics import accuracyscore, f1score
    


    Set tracking URI
    

    mlflow.settrackinguri("http://localhost:5000")
    


    Set experiment
    

    mlflow.setexperiment("iris-classification")
    


    Load data
    

    iris = loadiris()
    

    Xtrain, Xtest, ytrain, ytest = traintestsplit(
    

        iris.data, iris.target, testsize=0.2, randomstate=42
    

    )
    


    Start MLflow run
    

    with mlflow.startrun(runname="random-forest-v1") as run:
    

        # Log parameters
    

        params = {
    

            "nestimators": 100,
    

            "maxdepth": 5,
    

            "randomstate": 42
    

        }
    

        mlflow.logparams(params)
    


        # Train model
    

        model = RandomForestClassifier(params)
    

        model.fit(Xtrain, ytrain)
    


        # Make predictions
    

        ypred = model.predict(Xtest)
    


        # Log metrics
    

        metrics = {
    

            "accuracy": accuracyscore(ytest, ypred),
    

            "f1score": f1score(ytest, ypred, average="weighted")
    

        }
    

        mlflow.logmetrics(metrics)
    


        # Log model
    

        mlflow.sklearn.logmodel(
    

            model, 
    

            "model",

    Artikel Terkait