Skip to content

Deploy Models

Once we upload our model, create and configure our showcase project, we are all set to deploy it. Clouderizer Showcase offers to deploy projects in following ways

Figure 1 - Deploy model button

AWS

Deploying on AWS requires connecting your AWS account with Clouderizer. Please follow instruction here for this setup.

  1. Go to your Showcase project and press Deploy.

  2. Select AWS from the menu below. Figure 2 - Deploy on AWS

  3. Confirm and press "Yes lets do it now"

This will deploy your Showcase project on AWS infrastructure. Behind the scene, Clouderizer will spin EC2 CPU instances to run your models on. In case you wish to deploy your models on GPU instances, please contact info@clouderizer.com for instructions.

GCP

Deploying on GCP requires connecting your GCP account with Clouderizer. Please follow instruction here for this setup.

  1. Go to your Showcase project and press Deploy.

  2. Select GCP from the menu below. Figure 3 - Deploy on GCP

  3. Confirm and press "Yes lets do it now"

This will deploy your Showcase project on GCP infrastructure. Behind the scene, Clouderizer will spin GCP CPU instances to run your models on. In case you wish to deploy your models on GPU instances, please contact info@clouderizer.com for instructions.

Ubuntu

Showcase projects can be deployed to any Ubuntu machine. This can be a local Ubuntu machine or a remote / cloud machine as well. Pre-requisites

  • Access to terminal on Ubuntu machine
  • Docker should be installed and running
  • Min. 4GB RAM

  • Go to your Showcase project and press Deploy.

  • Select Ubuntu from the menu below. Figure 4 - Deploy on Ubuntu

  • This will show a command that needs to be run on Ubuntu terminal. Press the copy button to copy the command to your clipboard. Figure 5 - Ubuntu command

  • Open a terminal on the ubuntu machine and paste the above command and press enter. This will deploy the Showcase Clouderizer project in a docker container on that machine.

Mac

Showcase projects can be deployed to any local Mac machine. Pre-requisites

  • Access to terminal on Mac machine
  • Docker should be installed and running
  • Min. 4GB RAM

  • Go to your Showcase project and press Deploy.

  • Select Mac from the menu below. Figure 6 - Deploy on Mac

  • This will show a command that needs to be run on Mac terminal. Press the copy button to copy the command to your clipboard. Figure 7 - Mac command

  • Open a terminal on the Mac machine and paste the above command and press enter. This will deploy the Showcase Clouderizer project in a docker container on that machine.

Windows (Windows Subsystem for Linux)

Following are the pre-requisites to run Showcase projects on Windows

  • Install and enable Docker Desktop WSL 2 backend
  • Docker service should be running
  • Min. 4GB RAM

Once this setup is done, instruction above for Ubuntu can be followed on WSL terminal window.

Deployment Status

Once deployment is started using any of the methods above, status on the project page updates to indicate deployment is in progress. We can see details about the progress from bottom left portion of screen.

Figure 8 - Deployment in progress

When deployment is complete, status changes to Running state and we can see deployed model's URL in the bottom left portion of screen. This URL can be used for scoring the model. More details on scoring can be found here and here.

Figure 9 - Deployment complete

Scalable Deployment

For AWS and GCP, by default deployment occurs on a single cloud instance. Admins can choose to deploy on a Kubernetes cluster. Configuration to select Kubernetes deployment and specify number of nodes in cluster are in works and will be updated here soon. In case you want early access to this feature, plesae write to us at info@clouderizer.com