Indicates whether the inference component caches model artifacts as part of the auto scaling process.