Steps are automatically cached — rerunning only changes

Overview

ZenML is an MLOps framework for portable, reproducible ML pipelines. It provides a standardized pipeline abstraction with built-in tracking, caching, artifact management, and integration with major ML and cloud tools.

Installation

uv pip install zenml

Basic Pipeline

from zenml import pipeline, step

@step
def load_data() -> dict:
    return {"data": [1, 2, 3], "labels": [0, 1, 0]}

@step
def train_model(data: dict) -> str:
    return f"Trained on {len(data['data'])} samples"

@pipeline
def training_pipeline():
    data = load_data()
    model = train_model(data)

training_pipeline()

Caching

# Steps are automatically cached — rerunning only changes
@step(enable_cache=True)
def preprocess(raw: dict) -> dict:
    return {"features": [x * 2 for x in raw["data"]]}

# Changing parameters invalidates cache