nutanix / kubeflow-manifests Goto Github PK

With Kubeflow 1.7 release, the progress is tracked for the following features on Nutanix Karbon

Kubernetes version - 1.24, Kustomize version - 5.0.1

• multi-user deployment with profiles
• distributed training using Training operator(TFJob)
• Hyperparameter optimization using Katib
• Testing Inference using KServe
• Workflow creation using Pipelines
• e2e-mnist manifest testing

With Kubeflow 1.8 release, the progress is tracked for the following features on Nutanix Kubernetes Engine.

Kubernetes version - 1.28.5, Kustomize version - 5.2.1, Nutanix Kubernetes Engine - 2.10.0

• multi-user deployment with profiles
• distributed training using Training operator(TFJob)
• Hyperparameter optimization using Katib
• Testing Inference using KServe
• Workflow creation using Pipelines
• Tensorboard using object store

(gptk8s) root@administrator:~# curl -v -H "Host: ${SERVICE_HOSTNAME}" -H "Content-Type: application/json" -u [email protected]:12341234 http://${INGRESS_HOST}:${INGRESS_PORT}/v2/models/llama2_7b_chat/infer -d @./NKE/data/qa/sample_text2.json

• Trying 192.168.30.113:31658...
• TCP_NODELAY set
• Connected to 192.168.30.113 (192.168.30.113) port 31658 (#0)
• Server auth using Basic with user '[email protected]'

POST /v2/models/llama2_7b_chat/infer HTTP/1.1
Host: llm27c-deploy.default.example.com
Authorization: Basic dXNlckBleGFtcGxlLmNvbToxMjM0MTIzNA==
User-Agent: curl/7.68.0
Accept: /
Content-Type: application/json
Content-Length: 218

• upload completely sent off: 218 out of 218 bytes
• Mark bundle as not supporting multiuse
  < HTTP/1.1 302 Found
  < location: /dex/auth?client_id=kubeflow-oidc-authservice&redirect_uri=%2Fauthservice%2Foidc%2Fcallback&response_type=code&scope=openid+profile+email+groups&state=MTcxNDM3MzI2OXxOd3dBTkUwelJsQTNRMWxHTlZKYVNqWkpORFUzTkZwQ01sQkdWbGxFUkVsRVZWZEZVMWcxVDBSTlMwTXpWa1JaU0VKUFdqZEtUVUU9fO27vIFKshkKp9TszfeY-WAP3qNfDBqvlXUoi4bVMxRL
  < set-cookie: oidc_state_csrf=MTcxNDM3MzI2OXxOd3dBTkUwelJsQTNRMWxHTlZKYVNqWkpORFUzTkZwQ01sQkdWbGxFUkVsRVZWZEZVMWcxVDBSTlMwTXpWa1JaU0VKUFdqZEtUVUU9fO27vIFKshkKp9TszfeY-WAP3qNfDBqvlXUoi4bVMxRL; Path=/; Expires=Mon, 19 Oct 2054 07:41:18 GMT; Max-Age=1200000000000
  < date: Mon, 29 Apr 2024 06:47:49 GMT
  < x-envoy-upstream-service-time: 4
  < server: istio-envoy
  < content-length: 0
  <
• Connection #0 to host 192.168.30.113 left intact
  (gptk8s) root@administrator:~#

With Kubeflow 1.8 release, the progress is tracked for the following features on Nutanix Kubernetes Engine.

Kubernetes version - 1.25, Kustomize version - 5.0.3, Nutanix Kubernetes Engine - 2.8.0

• multi-user deployment with profiles
• distributed training using Training operator(TFJob)
• Hyperparameter optimization using Katib
• Testing Inference using KServe
• Workflow creation using Pipelines
• e2e-mnist manifest testing
• Tensorboard using object store

With the change to DKP/NKP platform we should add necessary information when using that platform