We recently launched the Terraform-based Gradient MLOps Installer to let users turn any cloud into a super-powerful machine learning platform. Out of the box the Gradient Installer lets you run Gradient on the public cloud (AWS, GCP, Azure) or on your own private cloud or on-premise cluster.

Private Clusters 🌱

We are excited to announce that today we are introducing 1-click, fully-managed private clusters that you can provision and manage without ever leaving the web console!

Any team on a T1 or higher Gradient subscription can now provision their own private Gradient cluster. For this first release, the private cluster runs on the Paperspace CORE cloud – meaning you can leverage our high-performance, low-cost GPU infrastructure.

You can select single node or multi-node clusters and select machines located in US-West (CA1), US-East (NY2), or Europe-West (AMS1) locations.

New Installer CLI 👩🏽‍💻

Do you prefer to work in the command line? With Gradient Installer (see it on GitHub) we're also introducing a CLI to manage your clusters.

To get started, install Gradient-Installer from source:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/paperspace/gradient-installer/master/bin/install)"

Next, run the updater:

gradient-installer update

When prompted for an API key, visit the Paperspace console to generate an application-specific API token. Once you login to Paperspace, you can find your API keys here: https://console.paperspace.com/account/api

Enter your API key and that's it! You're ready to start building.

First go ahead and run gradient-installer clusters list to see your available private clusters.

Want to spin-up a new cluster? Run gradient-installer clusters up and follow the setup prompts – that's all there is to it!

When it's time to tear down your cluster, just use the command gradient-installer clusters down.

What This All Means ✨

We want your machine learning team to be able to take Gradient with you no matter where your data lives and no matter what cloud you want to use for training and inference. That's why we're so excited to make private clusters a reality.

From now on – if you want to setup a private cluster on AWS or Azure or on the GPUs under your desk – we can help.

You'll get all the same advantages of Gradient no matter where you run your compute:

  • Jupyter-backed notebooks
  • Collaborative team workspaces
  • CI/CD
  • Job runner
  • Lifecycle management
  • and more!

Moving forward, we'll be adding new configurations and capabilities to the managed private cluster offering.

If you want to dive right into private managed clusters, login to the console and get going!

If you want a demo for your team or want to consult with one of our Solutions Architects, drop us a line at sales@paperspace.com.