Deploying an End-to-End Machine Learning Solution on Kubeflow Pipelines

Abstract

A Kubeflow pipeline component is an implementation of a pipeline task. A component is a step in the workflow. Each task takes one or more artifacts as input and may produce one or more artifacts as output. Each component usually includes two parts: • Client code: The code that talks to endpoints to submit jobs, for example, code to connect with the Google Cloud Machine Learning Engine. • Runtime code: The code that does the actual job and usually runs in the cluster, for example, the code that prepares the model for training on Cloud MLE. A component consists of an interface (inputs/outputs), the implementation (a Docker container image and command-line arguments), and metadata (name, description).