Metis Machine's Skafos

Machine Learning Delivered. A Machine Learning automation platform for those that want to focus on their work.

Welcome to the Metis Machine documentation hub. You'll find comprehensive guides and documentation to help you start working with Metis Machine's Skafos platform as quickly as possible, as well as support if you get stuck. Fire it up!

Get Started    

Jupyter IDE

Skafos integrates with JupyterLab as a way to interactively build and test your ML pipelines.

Skafos + Jupyter = Better Together

Jupyter hosted on Skafos gives you the full power of the Skafos platform powering JupyterLab.

Introduction

Project Jupyter is a non-profit, open-source project to support interactive data science and scientific computing. JupyterLab is an interactive development environment that has full support for Jupyter Notebooks, is highly customizable, and provides integration between notebooks and code.

Skafos provisions a JupyterLab instance for every Skafos Project you create. Using JupyterLab is not required, but it can be an incredibly powerful tool for writing machine learning models.

When you create a new Project, a fresh lab instance is automatically provisioned for you to begin your development. To access this lab instance, navigate to the Dashboard and click on the Launch JupyterLab button next to the project of interest. Once inside your lab workspace, you can:

  • write and test new code for jobs within your project
  • easily install your favorite packages or dependencies in the terminal
  • interact with the Skafos SDK and Data Engine
  • track changes using git

This allows you to leverage the power of Skafos while working in a familiar development environment synced up to your personal code repository.

Jupyter Lab Features

Jupyter Lab has many features that many users are familiar with, and are well documented in the JupyterLab Documentation. These include:

Skafos Features within Jupyter Lab

Because each of the Jupyter Lab instances provisioned by Skafos is built on our infrastructure, you have full access to the Skafos SDK, including Data & ML Model Handling through the Skafos Data Engine. This is an excellent way to prototype data access and handling prior to delivering into a production pipeline.

Deployment

When you're ready to deploy your project from a notebook session, you have two options:
1) Use the terminal to push your code changes to the master branch of your associated github repostitory. (You will need to add the Skafos App to your github repository first).
2) Use the Skafos Deploy option directly from the notebook, as described below.

To deploy directly from the notebook you are using, click the Skafos link. Two options will be present:

  • Open Project Config
  • Deploy

Begin by opening the project config.

Make sure that the entry point and resources reflect what is needed for this job. (Note that if you need an additional job or dependency, this can only be achieved if you use the Command Line Interface to add jobs.

Once this file has been edited and saved, if needed, go back to the Skafos menu and select Deploy. You will receive a pop-up

Click Deploy, and you will see another pop up, and a link for where you can monitor your job.

Jupyter IDE


Skafos integrates with JupyterLab as a way to interactively build and test your ML pipelines.

Suggested Edits are limited on API Reference Pages

You can only suggest edits to Markdown body content, but not to the API spec.