Workflows

Introduction to MLflow

Weston Bassler

Senior MLOps Engieer

MLflow Projects

Workflow

1 unsplash.com
Introduction to MLflow

MLproject

name: project_name
python_env: python_env.yaml
entry_points:

  step_1:
    command: "python train_model.py"

  step_2:
    command: "python evaluate_model.py {run_id}"
    parameters:
        run_id:
            type: str
            default: None
Introduction to MLflow

Workflows

import mlflow

# Step 1
step_1 = mlflow.projects.run(
    uri='./',
    entry_point='step_1'
)

# Step 2
step_2 = mlflow.projects.run(
    uri='./',
    entry_point='step_2'
)
Introduction to MLflow

Projects run

import mlflow

# Step 1
step_1 = mlflow.projects.run(
    uri='./',
    entry_point='step_1'
)


print(step_1)

<mlflow.projects.submitted_run.LocalSubmittedRun object at 0x125eac8b0>
Introduction to MLflow

Projects run

step_1.cancel() - Terminate a current run

step_1.get_status() - Get the status of a run

step_1.run_id - run_id of the run

step_1.wait() - Wait for the run to finish

Introduction to MLflow

Projects run

import mlflow

# Step 1
step_1 = mlflow.projects.run(
    uri='./',
    entry_point='step_1'
)


# Set variable for step_1 run_id
step_1_run_id = step_1.run_id

# Step 2
step_2 = mlflow.projects.run(
    uri='./',
    entry_point='step_2',

    parameters={
        'run_id': step_1_run_id
    }

)
Introduction to MLflow

ML Lifecycle

Model Engineering and Model Evaluation

Introduction to MLflow

Let's practice!

Introduction to MLflow

Preparing Video For Download...