Implementation:Kserve Kserve GPU Cluster Credentials

Knowledge Sources	KServe HuggingFace Hub
Domains	Infrastructure, GPU_Computing, Security
Last Updated	2026-02-13 00:00 GMT

Overview

Concrete YAML pattern for creating HuggingFace credentials and verifying GPU availability for LLM inference workloads.

Description

This pattern creates the credential resources required by LLMInferenceService for downloading gated or private models from HuggingFace Hub. The Secret stores the HF token, and the ServiceAccount is referenced by the LLMInferenceService spec.

Usage

Apply these credentials before deploying an LLMInferenceService that references an hf:// model URI.

Code Reference

Source Location

• Repository: kserve
• File: docs/samples/storage/hf/hf_secret.yaml, Lines 1-14

Signature

apiVersion: v1
kind: Secret
metadata:
  name: storage-config
type: Opaque
stringData:
  HF_TOKEN: <your-huggingface-token>
---
apiVersion: v1
kind: ServiceAccount
metadata:
  name: hfserviceacc
secrets:
  - name: storage-config

Import

kubectl apply -f hf_secret.yaml

I/O Contract

Inputs

Name	Type	Required	Description
HF_TOKEN	string	Yes	HuggingFace Hub access token
GPU nodes	kubernetes nodes	Yes	Nodes with nvidia.com/gpu resource

Outputs

Name	Type	Description
Secret	v1.Secret	storage-config with HF_TOKEN
ServiceAccount	v1.ServiceAccount	hfserviceacc bound to storage-config

Usage Examples

Setup and Verify

# 1. Create HF credentials
kubectl create namespace kserve-demo
kubectl apply -f - <<EOF
apiVersion: v1
kind: Secret
metadata:
  name: storage-config
  namespace: kserve-demo
type: Opaque
stringData:
  HF_TOKEN: hf_aBcDeFgHiJkLmNoPqRsTuVwXyZ
---
apiVersion: v1
kind: ServiceAccount
metadata:
  name: hfserviceacc
  namespace: kserve-demo
secrets:
  - name: storage-config
EOF

# 2. Verify GPU availability
kubectl get nodes -o custom-columns='NAME:.metadata.name,GPU:.status.capacity.nvidia\.com/gpu'

Related Pages

Implements Principle

• Principle:Kserve_Kserve_GPU_Cluster_Preparation

Requires Environment

• Environment:Kserve_Kserve_GPU_Accelerator
• Environment:Kserve_Kserve_Kubernetes_Cluster