This repository is workshop for exploring TensorFlow Extended (TFX) libraries and implementing ML training pipelines using TFX Components and Kubeflow Pipelines (KFP). The pipeline uses Google Cloud Platform (GCP) managed services. The following labs are included:
- TFX libraries deep-dive: This lab covers TFDV, TFT, and TFMA, as well as TF Estimator API
- TFX Pipeline walk-through: This lab covers TFX Components and ML Metadata
- TFX with KFP on GCP: This lab covers deploying and running TFX with KFP and using GCP managed services
This workshop uses Python 3.6 requires the following packages:
- TensorFlow 1.15
- Apache Beam 2.16
- TFX 0.15
- KFP SDK 0.1.36