Kubeflow Pipelines examples
Kubeflow is an OSS project to support a machine learning stack on Kubernetes, to make deployments of ML workflows on Kubernetes simple, portable and scalable.
Kubeflow Pipelines is a new component of Kubeflow that makes it easy to compose, deploy and manage end-to-end machine learning workflows. The Kubeflow Pipelines documentation is here.
This directory tree contains code for several different groups of Kubeflow Pipelines examples. The examples highlight how Kubeflow and Kubeflow Pipelines can help support portability, composability and reproducibility, scalability, and visualization and collaboration in your ML lifecycle; and make it easier to support hybrid ML solutions.
- A pipeline that implements an AutoML Tables end-to-end workflow.
- Distributed Keras Tuner + KFP example
- A pipeline that shows how you can make calls to the AutoML Vision API to build a pipeline that creates an AutoML dataset and then trains a model on that dataset: samples/automl/README.md.
- Example pipeline for Scale by the Bay workshop (2019)
Deprecated examples
These examples are not currently maintained and most likely don't work properly.
- README_taxidata_examples.md
- README_github_summ.md: going forward, the current version of this example lives here: https://github.com/kubeflow/examples/tree/master/github_issue_summarization/pipelines.