Introduction

Playground

The quickest way to get started with Unweave is the Unweave Playground. The Playground lets you create and manage Jupyter Lab environments completely serverlessly, and with access to GPUs and persistent storage.

Launching a new Jupyter Lab

  1. Create a new project from the sidebar. You can name it whatever you want.

  2. Once your project is ready, you can select it from the sidebar and launch your first Jupyter Lab. Choose the type of processor you want to use. You can choose between CPU, or 5 different types of GPUs. If you choose the Default option, Unweave will assign you the first available GPU.

  3. Click Launch Jupyter Lab. This will create a new Jupyter Lab environment for you and give you a link to it within a few seconds. Click the Open button once it's ready to access your Jupyter Lab in a new tab.

Managing Data

Unweave comes with a persistent and versioned data store mounted at the base of the repository at ./uwstore. Any files you add to this path are persisted across Jupter Lab sessions and are available to all Jupyter Lab environments in the project.

Dependencies

All Jupyter Lab environments come with a set of pre-installed dependencies. If you'd like to add some of your own, you can install them through the terminal and persist them to a requirements.txt file.

❯ pip install pandas
❯ pip freeze > requirements.txt

Every time you start a new Jupyter Lab, Unweave will install all the dependencies listed in requirements.txt or environment.yml files.

Shutting Down

When you're done with your Jupyter Lab, you can shut it down by clicking the Stop button. If a Lab is idle for 20 minutes, it will automatically shut down. "Idle" means that it does not have any active kernels executing code.

Previous
Introduction