NVIDIA NIM on GKE quiz

1. Which Cloud Shell command is used to create a GKE cluster?

   gcloud container clusters create

   helm install

   docker run

   kubectl apply

2. What series of NVIDIA GPUs is not available in GKE?

   A100

   B200

   L4

   RTX5000

3. Why do you need to configure an NVIDIA NGC API Key to deploy NVIDIA NIM?

   To create the GKE cluster required.

   To install the necessary GPU drivers.

   To authenticate and fetch the Helm chart and container images from NVIDIA.

   It provides the login credentials for the Google Cloud project.

4. Before installing the NVIDIA NIM with Helm, what Kubernetes resources should you create to deploy the microservice on? (choose all that apply).

   Choose as many answers as you see fit.

   A new GKE cluster

   GPU node pool

   Namespace

   A deployment configuration

5. To pull the NIM container image from the nvcr.io registry, what type of resource should you create in your namespace using an NGC API key?

   docker-registry secret.

   GPU node pool.

   Custom YAML file.

   A port-forwarding rule.

6. In the nim_custom_value.yaml file, what is the purpose of the image.repository key?

   To set the NIM deployment version.

   To define which model container we want to deploy.

   To name the Kubernetes secret for pulling images.

   To define the Helm chart location.

7. Which command is used to install the NIM using a Helm chart, a custom configuration file, and a specific namespace?

   gcloud container clusters create

   kubectl apply

   helm install

   docker run

8. How do you make the NIM service accessible to your local Cloud Shell terminal for testing?

   Use kubectl port-forward.

   Assign a public IP to the pod.

   Edit the cluster's firewall rules.

   SSH into the GKE node.

Submit answers