See create-cluster.sh to create GKE standard clustr, with Nvidia-tesla-t4 nodepool or Nvidia-l4 nodepool Validations
gcloud container clusters get-credentials skypilotcluster --zone us-central1-c --project $PROJECT_ID
Build cloud workstation container image: update cloudbuild.yaml, with proper artifact registry path
cd workstation
gcloud builds submit .
Build Workbench instance container image: update workbench/cloudbuild.yaml, with proper artifact registry path
cd workbench
gcloud builds submit .
Create a Workbench instance with built container image:
gcloud workbench instances create skypilot-wb-instance --container-repository=us-east1-docker.pkg.dev/$PROJECT_ID/gke-llm/skypilot-bench --container-tag=latest --machine-type=n1-standard-4 --location=us-central1-b
or with GPU
gcloud workbench instances create skypilot-wb-instance --container-repository=us-east1-docker.pkg.dev/$PROJECT_ID/gke-llm/skypilot-bench --container-tag=latest --machine-type=n1-standard-4 --location=us-central1-b --accelerator-type=NVIDIA_TESLA_T4 --accelerator-core-count=1
Open a terminal in Cloud Workstation or Workbench instance,
gcloud init
gcloud auth application-default login
gcloud config set project PROJECT_ID
gcloud auth application-default set-quota-project PROJECT_ID
sky check
run SKy check and make sure Enabled clouds GCP(Checked)
gcloud container clusters get-credentials testclusterromil --zone us-central1-c --project rick-vertex-ai
sky check
🎉 Enabled clouds 🎉 ✔ GCP ✔ Kubernetes
Single instance
HF_TOKEN=xxx sky launch gemma.yaml -c gemma --env HF_TOKEN --cloud kubernetes
SkyServe:
HF_TOKEN=xxx sky serve up llama3.yaml -n llama3 --env HF_TOKEN --cloud kubernetes
HF_TOKEN=xxx sky launch ray-train.yaml -c ray --env HF_TOKEN --cloud kubernetes
open ray-dist-train notebook , and run through all the steps
sky down cluster_name
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent