New Features in June: Spot Instances, Private GitHub Repos, Runs on the Web, and more!

Spell
3 min readJun 17, 2019

--

Our new features in June make experiments much cheaper, saving models and run outputs 50% faster, and put more of your work at your fingertips on the web.

Read on for more!

Spot Instances

Screenshot of the create cluster modal with spot instance machine type selected

Save 80% or more on your cloud hardware costs with spot instances. Users on our Teams plan can now run regular runs, Jupyter workspaces, and distributed runs on spot instances. Simply select the “spot” option when creating a new machine type to save on cost.

GitHub App

Our new GitHub app allows Teams plan subscribers to create runs and Jupyter Workspaces using private GitHub repositories. This means you can create runs from the web and with the --github-url flag without having to publish code publicly to GitHub.

Interested in learning more? Email us at support@spell.run and we’ll get you started!

Create Runs and Re-Run Runs on the Web

Animated screen capture of re-running a run from the web console

Now you can create and re-run runs directly from the web via the runs page. Just go to web.spell.run/runs and click the big blue button on the top right to create a new run. To re-run a completed run, select “re-run” from the runs list menu (on the right side of the run row).

Saving Performance Improvements

We’ve made improvements during our save process that’s allowed a 2x speedup in performance. We won’t take up any extra space during the save step, so it will succeed even if your run has taken up the entire disk space.

New Tutorials: DeepDreams and Image Captioning

Learn how to create your own dreams using Google’s DeepDream in this Jupyter based tutorial.

Follow a new step-by-step tutorial to caption any image using a pre-trained neural net.

💡 Tip of the Month

Want to run a run within a public GitHub repository, and not planning to make any changes to the files? Instead of cloning it and initializing your run from within the directory, you can use the --github-url parameter to get the code within your run.

The captioning images tutorial has more info on how to use the --github-urlparameter.

Stay Tuned

There are many more exciting features in the works. Be the first to know by following us on Twitter or here on Medium!

Also, we’re hiring!

--

--

Spell

MLOps platform to streamline machine learning. Sign up to our new MLOps newsletter at www.Spell.ml