How deeply does Distributed Machine Learning Patterns cover Kubeflow training operator?

Defines custom resources, installs the operator, submits a TFJob, and checks Pods; it defers workflow orchestration.

Kubeflow training operator

As explained in Distributed Machine Learning Patterns

The Kubeflow Training Operator is a Kubernetes operator that manages distributed and nondistributed machine learning training jobs through custom resources.

Kubeflow Training Operator is a Kubernetes operator for running distributed and nondistributed machine-learning jobs. It exposes Kubernetes custom resources for frameworks and job types including TensorFlow, PyTorch, Apache MXNet, XGBoost, and MPI, allowing workloads to be created with kubectl like built-in Kubernetes resources. section 8.3 section 8.3.1

A typical local-cluster workflow is:

Create and activate a dedicated kubeflow namespace. section 8.3.1
Apply manifests that install the Training Operator and related tools. section 8.3.1

Distributed TensorFlow job

A TFJob declares two Worker replicas running a TensorFlow MNIST container. The operator creates the corresponding worker Pods, which can be observed through the TFJob and Pod listings; failed workers use an OnFailure restart policy.

How Distributed Machine Learning Patterns covers it

Cover of Distributed Machine Learning Patterns

liveBook manning.com

Read Distributed Machine Learning Patterns on liveBook Get the book at manning.com

Also explained in

liveBook manning.com

Designing Deep Learning Systems: A guide for software engineersA Kubeflow controller that launches, monitors, and reconciles Kubernetes resources for machine learning training jobs in a specific framework.

Compare how 2 books explain Kubeflow training operator →

Related concepts

Kubernetes operatoruses · A Kubernetes operator is a controller pattern that extends Kubernetes to automate application or service management by reconciling desired and actual state through the Kubernetes API.
Kubeflowpart of · Kubeflow is an open-source platform for developing, deploying, and orchestrating machine-learning workflows on Kubernetes.
custom objectuses · A Kubernetes custom object is an API object whose resource type and schema are defined by a custom resource definition or another extension mechanism.
custom resource definitionuses · A CustomResourceDefinition is a Kubernetes API object that defines a custom resource type and its schema.

This explainer was generated by Manning's knowledge graph from the text of Distributed Machine Learning Patterns. The section links point to the parts of the book it draws on.