https://fortinetcloudcse.github.io/ai-101/01intro/index.htmlChoose how you will run the labs — Docker Compose on your own machine, or Kubernetes with Helm — then follow that path's prerequisites and setup steps.Hugoen-UShttps://fortinetcloudcse.github.io/ai-101/01intro/1_prereqs_docker.htmlMon, 01 Jan 0001 00:00:00 +0000https://fortinetcloudcse.github.io/ai-101/01intro/1_prereqs_docker.htmlDocker Compose path This whole page is the Docker Compose path. If you chose Kubernetes / Helm, use Kubernetes / Helm Setup instead. Prerequisites Requirement Version Check Docker Engine 24+ docker version Docker Compose v2.20+ docker compose version Git any git --version jq 1.6+ jq --version Free RAM 8 GB — Free disk 5 GB for model cache GPU optional A discrete GPU (NVIDIA or Apple Silicon) cuts model load time from ~60 s to ~5 s and speeds inference significantly. The labs work without one — first responses will just be slower.https://fortinetcloudcse.github.io/ai-101/01intro/2_prereqs_k8s.htmlMon, 01 Jan 0001 00:00:00 +0000https://fortinetcloudcse.github.io/ai-101/01intro/2_prereqs_k8s.htmlKubernetes / Helm path This whole page is the Kubernetes / Helm path. If you chose Docker Compose, use Docker Compose Setup instead. Prerequisites Requirement Version Check kubectl 1.28+ kubectl version --client Helm 3.14+ helm version jq 1.6+ jq --version A running cluster — kubectl cluster-info Default StorageClass — kubectl get storageclass The Helm chart creates a PersistentVolumeClaim for Ollama’s model cache. A default StorageClass is required unless you set ollama.storage.storageClassName explicitly.