k8s_ros_sample
k8s_ros_sample copied to clipboard
Support Nvidia Device-Plugin
It is likely to have Nvidia GPU as hardware acceleration on edge devices, in that case, it should be managed as custom resource so that Kubernetes can control those resource with deployment. Usually specific application requires GPU acceleration for AI processing. Cluster administrator should be able to bind Nvidia GPU to those containers.
Reference: https://github.com/NVIDIA/k8s-device-plugin