@@ -190,6 +192,11 @@ func buildArgs(ms *mlv1.ModelService, args []string) []string { "--served-model-name": ms.Spec.ServedModelName, } vGPUNumber := getVGPUNumber(ms) if vGPUNumber > 0 { specArgs["--tensor-parallel-size"] = strconv.Itoa(vGPUNumber) } for k, v := rang...