Skip to main content

What's New? ✨

Release Notes

Version 2.0#

Hub API (New Deployment Platform)#

Building and Training Models is one thing and Deploying it for the world to use is another. AI Deployment has always been a challenging and complicated task, so we at CellStrat decided to squash the complications and make it much simpler with the CellStrat Hub API Platform, a simple and highly flexible and customizable platform for deploying your AI Models as REST APIs which can be called from anywhere like a web application, server-side application or from the edge.

With CellStrat Hub API, you can,

  • Deploy your AI Models in under 15 minutes in only 3 steps.
  • As simple as it may look, you still have complete control over your deployment to customize it in any way you want.
  • The Models are deployed as Serverless APIs, which means you only pay for the number of invocations, unlike traditional deployment options. But that is if you pay at all! Did we mention its completely free to get started?
  • APIs are auto-scaled based on traffic, so you don't need to worry about the deployment infrastructure and can focus on building your model.
  • All Models are deployed using a docker container under the hood, that means you get complete control over your execution environment, allowing you to optimize your inference in low-level languages like C++ also.
  • You can deploy a maximum of 5 Model APIs for free. But hold on! Everything is completely customizable, so no one is stopping you from deploying multiple models in a single API! For all you know, you can deploy 10 models in a single API while staying in your free plan. This is the most cost effective option you can find to deploy your models.
  • Enterprise plans for Custom Deployment are also available with Premium Support.

Get Started with Deploying Your First Model Now!

Workspace (JupyterLab)#

We have rebranded the development environment based on Jupyter Lab to Workspace

New Features#

  • Get Access to a Dedicated Ubuntu Virtual Machine
  • One-Click Realtime Collaborative Coding (think Google Docs for Code)
  • Visual Debugger to easily debug your swamps of code
  • Code-Snippets Extension to save your frequently used snippets of code and reuse them in any of your projects
  • Seamless Git Integration via both, SSH and Personal Access Token
  • Improved Terminal Experience
  • Simplified List of Kernels i.e. Python 3.8, PyTorch 1.9, Tensorflow 2.4
  • Support for Matplotlib widgets
  • Personalize the appearance of your Workspace like theme, fonts, style, key bindings, etc. that persist across your sessions

Bug Fixes#

  • Relatively Faster Start Times (Precisely 1 min 40 seconds)
  • Smoother and Bug Free Development Experience
  • Squashed a lot of bugs and issues from previous releases