We've added a new terminal interface to the Gradient IDE! 

Gradient Notebooks now offer terminal access announcement 

Previously you needed to open JupyterLab to access a unix terminal -- but now you can do so directly from the Gradient IDE! 

Just look for the terminal icon in the left sidebar of any Gradient Notebook. 

Terminals are great for managing files, installing or uninstalling libraries and packages, and for anything else that you'd like to do quickly without needing to run notebook cells.

Terminals are available today for G1 and higher subscribers. 

Other Improvements

• We shipped a basic version history viewer to Gradient Notebooks. Look for the History button in the left sidebar of any notebook
• We shipped a backend improvement that decreased notebook load times across all regions by ~1 second

Bugfixes

• We fixed a bug that sometimes produced excessive line breaks in cell outputs
• We now focus files correctly when creating/deleting in the IDE file manager
• We fixed a bug that sometimes caused stopped notebooks to show empty previews 

We've added file uploads as well as a number of useful improvements and bugfixes to Gradient Notebooks.

File uploader arrives in Gradient Notebooks announcement 

You can now upload files directly from the Gradient Notebook IDE file manager. To use the uploader select Upload File from the sidebar. Upload progress is visible at the top of the sidebar and uploaded files are available for immediate use. 

Other Improvements

• We've added a prominent Start Instance button to stopped notebooks to make it easier to pick up where you left off when returning to a notebook
• We've increased the visibility of the Stop Instance button
• We now provide CPU, GPU, and RAM system metrics in the notebook footer
• We added a link to the Billing  view from the main menu which makes it easier to find account-related information

Bugfixes

• We fixed a visual bug that caused notebook file delete confirmations to render as a button rather than as a text block
• We fixed an issue where files within a notebook would sometimes not have syntax highlighting
• We made it more difficult to run a container that requires a GPU on a CPU instances
• We fixed a bug that was causing some users to accidentally create multiple notebooks
• We fixed a bug that occasionally caused some notebooks to get stuck in a saving state 

For more info on this release be sure to check out the blogpost. 

We've added improved file management functionality to the IDE!

Better file management in Notebooks improvement 

Now you can add files and folders directly from the IDE.

You can also download files using the file manager.

As more file management enhancements come to the IDE we recommend you swap over to JupyterLab if you need something like a terminal with root access.

Other Improvements

• We now send GPU utilization metrics to the status bar (bottom left) in the notebook IDE
• When creating a new notebook with a specified workspace, we now load a preview of the files in the IDE while the notebook initializes
• When you use the Share feature in a notebook, we now include in the link the specific file you are sharing

Bugfixes

• We fixed a bug that sometimes caused notebooks nested in folders to have an incorrect working directory which caused issues when using relative paths
• We fixed a bug that sometimes caused long cell outputs to degrade notebook performance
• We fixed a bug that caused some notebooks to restart after being manually stopped by the user
• We fixed a bug that occasionally prevented large notebooks from loading offline

This week we released a complete rewrite of Gradient Notebooks!

This is an important release for Gradient. Right away you'll notice that we've brought notebooks up to the new Paperspace design standards while taking big steps to bring simple and scalable ML infrastructure into the notebook itself.

In addition to the new compute instance selection menu, we've also integrated system metrics as part of a status bar and a new file browser which will soon unlock new capabilities around data management.

Last but not least, the notebook experience is now substantially more performant across the board -- from spin-up to teardown to rendering and executing cells. 

There's plenty more to unpack so head over to the blogpost to see what else has changed!

We've released autoscaling for model serving and deployments.

To get more details on autoscaling we invite you to checkout the blogpost and read the docs!

You can now provision and manage machines on Paperspace programmatically using the new Terraform provider. To get started watch the demo below, read the blogpost, or head over to the GitHub repo.

Managed private clusters have launched on Gradient!

You can now set up a managed private cluster for your machine learning team in the web console or on the command line.

Read the full blogpost here.

You'll see a few new options under "Workspaces" that enable you to pull down data (eg from a Git repository) when creating a new notebook.  Check out the docs to learn more. 

The latest version of the Paperspace desktop app is now shipping with 4K support on all Dedicated GPU machines.

All you have to do is plug-in to your 4K-enabled display, check your settings, and go!

Read the blogpost for more information.

Two additional quality of life updates have also been released: the bandwidth adjustment algorithm has been enhanced to improve streaming quality, and key/mouse input capture has been improved to provide a more reliable experience.

You should notice an overall snappier and more responsive streaming experience.

Gradient now includes the ability to tag entities for improved navigation, search, and management of your ML workflow. Tags can be created and updated in both the user interface and CLI.

You can add tags to several elements including Projects, Experiments, Models, and Deployments.