kcna video for you have a Kubernetes cluster with nodes that have specialized hardware accelerators (e.g., GPUs) and nodes without accelerators. You want to
You have a Kubernetes cluster with nodes that have specialized hardware accelerators (e.g., GPUs) and nodes without accelerators. You want to ensure that a specific set of pods, labeled as app=ml-app , always run on nodes with GPUs to maximize performance. Which Kubernetes feature should you use for this purpose?