How to run ML Inference with Apache Beam

Опубликовано 3 февраля 2023, 0:00

RunInference → goo.gle/3kWnkC5
Machine Learning → goo.gle/3XR73wD
Dataflow ML → goo.gle/3kWwMVL

You’ve built a machine learning model and run it on your laptop for development and testing. However, you are ready for more significant tasks like running predictions against a large batch of data or continuously running predictions at scale. In this video, Head of Decision Intelligence Cassie Kozyrkov discusses using a data processing framework like Apache Beam to put a machine learning model into production. Watch along to get started with the new Run-Inference utility in Apache Beam.

Chapters:
0:00 - Intro
0:48 - Using Apache Beam for data processing framework
2:26 - The process before the Apache Beam RunInference utility
4:09 - A faster method with Apache Beam RunInference utility
5:12 - Multiple models in the same pipeline
6:00 - RunInference + Beam expressiveness Branched A/B models
6:36 - Wrap up

Subscribe to Google Cloud Tech → goo.gle/GoogleCloudTech

#GoogleCloudTech