Training a model using Kubeflow
Kubeflow (https://www.kubeflow.org) covers every step of model development, including data exploration, preprocessing, feature extraction, model training, model serving, inferencing, and versioning. Kubeflow allows you to easily scale from a local development environment to production clusters by leveraging containers and Kubernetes, a management system for containerized applications.
Kubeflow might be your first choice for distributed training if your organization is already using the Kubernetes ecosystem.
Introducing Kubernetes
Kubernetes is an open source orchestration platform that’s used to manage containerized workloads and services (https://kubernetes.io):
- Kubernetes helps with continuous delivery, integration, and deployment.
- It separates development environments from deployment environments. You can construct a container image and develop the application in parallel.
- The container...
