Topic details: Tensor Processing Units (TPUs) are hardware accelerators that greatly speed up the training of deep learning models. In independent tests conducted by Stanford University, the ResNet-50 model trained on a TPU was the fastest (30 minutes) to reach the desired accuracy on the ImageNet dataset. In this talk, I’ll walk you through the process of training a state-of-the-art image and text classification model on your own data using Google’s Cloud TPUs. I’ll finish with how you could adapt your own model for TPU training.
Speaker Bio: Lak is a Tech Lead for Big Data and Machine Learning Professional Services on Google Cloud Platform. His mission is to democratize machine learning so that it can be done by anyone anywhere using Google's amazing infrastructure (i.e., without deep knowledge of statistics or programming or ownership of lots of hardware). He is the author of the O'Reilly book on Data Science on Google Cloud Platform (http://shop.oreilly.com/product/0636920057628.do), an end-to-end look at building data pipelines (from ingest to machine learning), which is available on Kindle and paperback.
5:45 - 6:30 PM Networking/Social/Food
6:30 - 6:35 PM Introduction
6:35 - 7:30 PM Talk and Q&A
7:30 - 8:00 PM Networking/Social/Closing