Your organization is deploying a high-performance distributed deep learning training workload that requires GPU-enabled bare metal Compute instances connected through an ultra-low-latency Oracle Cloud Infrastructure (OCI) RDMA cluster network. To maximize processing efficiency and comply with strict corporate OS requirements, your team must use a custom, pre-hardened Oracle Linux operating system image that is not supported by default node pools. The containerized training jobs must be orchestrated and scheduled using your existing Container Engine for Kubernetes (OKE) control plane.
Which worker node deployment model must you implement?
Submit