For the April TensorFlow Meetup, the Seldon team, alongside speakers from Thoughtworks and PyImageSearch, were delighted to host attendees from all over the world as the event went virtual for only the second time.
The 28th May meetup welcomed almost 100 attendees from all over the world. Our first speaker Danilo Sato from Thoughtworks spoke about the challenges of testing to create high-quality ML systems. The second speaker, Sayak Paul from PyImageSearch talked about using Tensorflow Hub to create effective ML models.
You can watch their talks below, and we’ve also included some of the relevant links mentioned in their presentations.
Presentation #1: CD4ML and the challenges of testing and quality in ML systems
Speaker: Danilo Sato, principal consultant at ThoughtWorks
Danilo Sato (@dtsato) is a principal consultant at ThoughtWorks with experience in many areas of architecture and engineering: software, data, infrastructure, and machine learning. He is the author of “DevOps in Practice: Reliable and Automated Software Delivery”, a member of ThoughtWorks Technology Advisory Board, and ThoughtWorks Office of the CTO.
Continuous Delivery for Machine Learning (CD4ML) deals with the challenges of applying Continuous Delivery principles to ML systems to make the end-to-end process of developing and deploying them more repeatable and reliable. These systems are generally more complex than traditional software applications, and ML models are non-deterministic and hard to explain. In this talk we will discuss the challenges of testing and quality in ML systems, and share some practices for applying different types of tests to help overcome those issues.
Presentation #2:TensorFlow Hub: Models, Models, and Models
Speaker: Sayak Paul, PyImageSearch
Sayak works at PyImageSearch where he applies deep learning to solve problems in computer vision, and brings solutions to edge devices. He also provides Q&A support to PyImageSearch readers. Previously, Sayak developed projects and practice pools for DataCamp. Outside of work, Sayak enjoys writing technical articles and giving talks at developer meetups and conferences.
Using SoTA machine learning models to build great products can be challenging. This becomes more evident when the developers don’t have access to the SoTA models in the first place. TensorFlow Hub promises to bridge that gap by providing a comprehensive repository of SoTA models across a large variety of domains. This session is going to focus on the onboarding aspect of TensorFlow Hub.