Released: November 2020License: Apache License 2.0
Github open issues
Github Stars
07 April
Github last commit
Stackoverflow questions

What is ZenML?

Originating from Munich’s technological landscape, ZenML bursts onto the MLOps scene with a singular proposition. Where many platforms box you into specific infrastructures for model deployment and orchestration, ZenML unfetters you. It presents an unprecedented elasticity, allowing for MLOps workflows to execute on a multitude of infrastructures, be they cloud or on-premise. This flexibility fundamentally alters the MLOps paradigm, providing a liberating toolset for organizations and solo practitioners aiming to unleash machine learning models without the typical restrictions.

Key Features of ZenML

  • Infrastructure Agnosticism: Liberate your choices with ZenML; the platform is uncoupled from any specific tech stack, enabling work on any cloud or local setup.
  • Streamlined Orchestration: Navigate the labyrinthine pathways of ML workflows effortlessly; ZenML’s unified interface minimizes complexity.
  • Versatile Model Deployment: From lightweight mobile apps to expansive enterprise solutions, ZenML adjusts smoothly, meeting deployment needs across the spectrum.
  • Open-source Advantage: Unshackle yourself from subscription woes and hidden agendas; ZenML offers a transparent, tweakable core.
  • Integration Ecosystem: ZenML’s treasure trove of pre-configured hooks for mainstream services smooths out the integration bumps, making it a snug fit in your existing ecosystem.
  • Automated Workflow Monitoring: Real-time metrics served hot on a dashboard that empowers you to act swiftly when performance quirks appear.
  • Robust Documentation: ZenML’s knowledge base is a trove of tutorials, guides, and active community discussions designed to accelerate your onboarding process.

Getting Started with ZenML

Commencing your ZenML journey merely requires a sprinkle of initiative and a dash of curiosity. First, install ZenML directly via pip. Open your terminal, punch in `pip install zenml,` and witness the magic as the package materializes in your local environment.

Got Python projects sprawled across multiple environments? Fear not. ZenML plays nice with virtual environments as well, letting you keep your workspace pristine and uncluttered. Simply activate your chosen environment and proceed with the pip install command.

Once the package installation wraps up, initiate a new ZenML repository by navigating to your project directory and running `zenml init.` This pivotal step lays down the blueprint for your machine-learning pipelines and associated configurations.

But wait, there’s more. Before diving headlong into pipeline creation, introduce yourself to ZenML’s concept of ‘datasources.’ These are reusable chunks of your data, ready to be funneled into any pipeline you craft. To create one, use the `zenml datasource create` command, specifying the data location and type.

After setting up the datasource, it’s time to plunge into pipeline creation. ZenML’s CLI makes this process less of a chore. Utilize `zenml pipeline create` followed by your specified configurations and settings. Remember, these pipelines are fully customizable, catering to diverse tasks such as data preprocessing, model training, and post-training analyses.

With your pipeline now ready to shine, execute it using `zenml run.` ZenML will manage all the associated cloud or local resources, ensuring the pipeline completes successfully. You’ll be given real-time updates, allowing you to keep tabs on every step of the pipeline’s journey.

Finally, after you’ve had your fill of running pipelines locally, consider taking things to the cloud for more power and flexibility. ZenML’s cloud adapters for popular services like AWS, GCP, and Azure make this transition seamless. Activate these adapters by merely inputting `zenml cloud set` followed by the cloud service of your choice.

By diligently following these steps, you set the stage for an organized, efficient, and, more importantly, scalable MLOps strategy. ZenML aims to lift the shroud of confusion that often enshrouds the orchestration and deployment of machine learning models, offering you a robust and adaptable toolset for your AI endeavors.

Subscribe to Our Newsletter

Do you want to stay informed? Keep up-to-date with industry news, the latest trends in MLOps, and observability of ML systems.

Webinar Event
The Best LLM Safety-Net to Date:
Deepchecks, Garak, and NeMo Guardrails 🚀
June 18th, 2024    8:00 AM PST

Register NowRegister Now