diff --git a/deploy/kubernetes/llamacpp.yaml b/deploy/kubernetes/llamacpp.yaml index 695e3770..f8beb552 100644 --- a/deploy/kubernetes/llamacpp.yaml +++ b/deploy/kubernetes/llamacpp.yaml @@ -82,11 +82,6 @@ spec: configMapKeyRef: name: context-engine-config key: LLAMACPP_MODEL_NAME - ports: - - name: http - containerPort: 8080 - protocol: TCP - env: - name: LLAMA_ARG_MODEL value: "/models/model.gguf" - name: LLAMA_ARG_CTX_SIZE @@ -95,6 +90,10 @@ spec: value: "0.0.0.0" - name: LLAMA_ARG_PORT value: "8080" + ports: + - name: http + containerPort: 8080 + protocol: TCP command: ["llama-server"] args: - "--model"