Similarly to the task for configuring NVIDIA CUDA machines, create a new task doing the same for AMD ROCm.

Integrate the task in testing pipeline.

Acceptance criteria:

• Pipeline can reliably run RHAIIS on ROCm VM using podman

Notes

• https://docs.redhat.com/en/documentation/red_hat_ai_inference_server/3.2/html/getting_started/inference-rhaiis-with-podman-amd-rocm_getting-started