TFX: Production ML pipelines with TensorFlow

Robert Crowe (Google), Pedram Pejman (Google)

13:30–17:00 Tuesday, 15 October 2019

Location: Blenheim Room - Palace Suite

Implementing AI

Secondary topics: Deep Learning tools, Machine Learning tools

Average rating:

(2.90, 10 ratings)

Download slides (1-PDF)

Download slides (2-PDF)

Who is this presentation for?

Data scientists, ML engineers, ML Ops practitioners, and ML management

Level

Intermediate

Description

Putting machine learning models into production is now mission critical for every business—no matter what size.

TensorFlow is the industry-leading platform for developing, modeling, and serving deep learning solutions. But putting together a complete pipeline for deploying and maintaining a production application of AI and deep learning is much more than training a model. Google has taken its years of experience in developing production ML pipelines and offered the open source community TensorFlow Extended (TFX), an open source version of the tools and libraries that it uses internally.

Robert Crowe and Pedram Pejman walk you through working code in an example pipeline so you can learn what’s involved in creating a production pipeline. You’ll be able take what you learn and get started on creating your own pipelines for your applications.

Prerequisite knowledge

Experience with machine learning, Python, and Linux

Materials or downloads needed in advance

A laptop - Linux or MacOS (recommended) or Windows (optional)
Git (or Git for Windows if using Windows, which includes Git Bash)
Docker Desktop
For Windows, please make sure that you can share your drive with Docker and open ports 445, 8080, 8888, and 6006.
Min 3GB available disk space
Min 8GB DRAM
Before the workshop, please run: docker pull gcr.io/tfx-oss-public/tfx-workshop:latest
Before the workshop, also please run: docker pull tensorflow/serving:nightly

What you'll learn

Understand production ML pipelines, the components of TensorFlow Extended pipelines, and issues in ML production deployments

Robert Crowe

Google

Robert Crowe is a data scientist and TensorFlow Developer Advocate at Google with a passion for helping developers quickly learn what they need to be productive. He’s used TensorFlow since the very early days and is excited about how it’s evolving quickly to become even better than it already is. Previously, Robert deployed production ML applications and led software engineering teams for large and small companies, always focusing on clean, elegant solutions to well-defined needs. In his spare time, Robert sails, surfs occasionally, and raises a family.

Website

Pedram Pejman

Google

Pedram Pejman is a technical program manager on the TensorFlow Extended team at Google Brain, on a mission to create the best inference experience on TensorFlow. Previously, he managed some of Google Cloud’s internal efforts in the machine intelligence space while getting to work on distributed systems and Kubernetes. Aside from building infrastructure, he enjoys writing music, playing soccer, and scrolling through memes.