homelab/k8s/apps/llm-serving/reasoning.yaml

apiVersion: serving.kserve.io/v1beta1
kind: InferenceService
metadata:
  annotations:
    serving.kserve.io/deploymentMode: RawDeployment
  labels:
    app.kubernetes.io/name: llm-reasoning
    app.kubernetes.io/part-of: llm-serving
  name: reasoning
  namespace: llm-serving
spec:
  predictor:
    containers:
    - args:
      - --model=unsloth/DeepSeek-R1-Distill-Qwen-32B-bnb-4bit
      - --served-model-name=reasoning
      - --quantization=bitsandbytes
      - --dtype=float16
      - --kv-cache-dtype=auto
      - --tensor-parallel-size=1
      - --max-model-len=16384
      - --gpu-memory-utilization=0.90
      - --max-num-seqs=4
      - --enable-chunked-prefill
      - --enable-prefix-caching
      - --reasoning-parser=deepseek_r1
      # Without these, any client sending tool_choice="auto" (pi does, for
      # Read/Bash/etc.) gets a 400: "auto" tool choice requires
      # --enable-auto-tool-choice and --tool-call-parser to be set. Verified
      # against this exact deployed vLLM v0.11.0's registered parsers
      # (ToolParserManager.tool_parsers) -- deepseek_v3 matches, same family
      # as the deepseek_r1 reasoning-parser already set above (this model
      # emits DeepSeek's own special-token format, not Qwen's Hermes-style,
      # despite being a Qwen-base distillation).
      - --enable-auto-tool-choice
      - --tool-call-parser=deepseek_v3
      - --host=0.0.0.0
      - --port=8080
      env:
      - name: VLLM_USE_FLASHINFER_SAMPLER
        value: '0'
      - name: VLLM_ATTENTION_BACKEND
        value: TRITON_ATTN
      - name: HF_HOME
        value: /mnt/models
      image: vllm/vllm-openai:v0.11.0@sha256:014a95f21c9edf6abe0aea6b07353f96baa4ec291c427bb1176dc7c93a85845c
      name: kserve-container
      ports:
      - containerPort: 8080
        protocol: TCP
      readinessProbe:
        httpGet:
          path: /health
          port: 8080
        periodSeconds: 10
      resources:
        limits:
          cpu: '16'
          memory: 16Gi
          nvidia.com/gpu: '1'
        requests:
          cpu: '8'
          memory: 8Gi
          nvidia.com/gpu: '1'
      startupProbe:
        failureThreshold: 80
        httpGet:
          path: /health
          port: 8080
        periodSeconds: 15
      volumeMounts:
      - mountPath: /mnt/models
        name: models
      - mountPath: /dev/shm
        name: shm
    deploymentStrategy:
      type: Recreate
    maxReplicas: 2
    minReplicas: 2
    nodeSelector:
      kubernetes.io/hostname: worker-1
    runtimeClassName: nvidia
    volumes:
    - name: models
      persistentVolumeClaim:
        claimName: llm-models
    - emptyDir:
        medium: Memory
        sizeLimit: 2Gi
      name: shm