Katonic MLOPS Platform runs in your Kubernetes cluster, and the infrastructure can be managed with Kubernetes native tools like kubectl. To simplify the deployment and configuration of Katonic services, we provide an installation automation tool using Kustomize to deploy Katonic into your compatible cluster. If you already have a Kubenetes cluster go ahead and deploy Katonic Directly ( Link to Deploy )
In this Section, you got your first real taste of Katonic. We will give instructions to get the MLOps Platform and will covere a simple end-to-end example with the standard MNIST, allowing you to see the different core components of Katonicin action. We will introduce the pipeline, which ties all of Katonic together, and you used it to train your model.
To setup your account and start using the platform, follow these steps:
- Continue to try.katonic.ai and click on Sign up now to set up a new account and register with your credentials
On the sign-up page, enter your details and proceed to sign up. Verify your email address by clicking on the link sent to the registered email.
- Continue to try.katonic.ai and login by entering your credentials on the Login screen.
- Congratulations! You have successfully setup your account and are ready to use the platform.
Setting up cluster with Minikube.
Setting up cluster in Amazon Web Services
Setting up cluster in Microsoft Azure
Setting up cluster in Google cloud project
Once done with cluster follow deployment steps
Figure- Katonic web UI
Click pipeline you should see the Pipelines web UI, as in figure
Figure- Pipelines web UI
From here we can upload our pipeline.
There two ways to do that:
Download sample pipleine from here and upload
Import pipeline from the URL here
we can use the same web UI to create a run of the pipeline. After you click the upload pipeline you’ll be able to create a run, as shown in figure
Figure- Pipeline detail page
Finally, let’s query our model and monitor the results. A “housing” is a simple test to ensure our model is a simple test to ensure our model is making predictions that are theoretically reasonable. For examples-- we’re attempting to guess what is the price of house.
Figure- Katonic piplines UI: prepackaged pipelines
Clicking a specific pipeline will show its execution graph or source, as seen in Figure
Figure- Katonic pipelines UI: pipeline graph view
Clicking the YAML tab will show the pipeline’s compiled code, which is a YAML file
In this area you are welcome to experiment with running pipelines to get a better feel for their execution and the capabilities of the Pipelines UI.
Figure- Katonic pipelines UI: pipeline view
To run the pipeline, click the “Create run” button and follow the instructions on the screen.
Figure- Katonic visualizaton of preprocessing container in pipeline