Model Training and Serving on Kubeflow deep-dive - Architecture and Artifacts
Model Training and Serving on Kubeflow deep-dive. This talk will discuss the architecture behind Kubeflow in the areas of model training and serving. It'll review Kubernetes CRDs (Custom Resource Definitions) in general, and specific Kubeflow CRDs involved in this process. Additionally, we'll cover how you can extend Kubeflow by defining your own training and serving components.