DevFest Budapest - Improving ML models with data preprocessing, running inference with Coral EdgeTPU
Improving ML models with data pre-processing and running inference with Coral EdgeTPU by Markku Lepisto, Head of IoT/ML Solutions, APAC, Google Cloud, Google
Join this session to see a step by step example how training data pre-processing raised the Machine Learning model quality from 14% to over 95%. With code examples and live demos, using Google Coral EdgeTPU ML accelerator and AutoML custom model training.
Markku is a Cloud Solutions Architect at Google. He has worked in cloud computing for the past 10 years. Before joining Google, Markku led the cloud architecture of Nokia Siemens Networks’ global business units, and covered APAC as a technology evangelist at Amazon Web Services. He works out of Tokyo, and helps developers and customers across APAC understand how to best use and architect solutions for Google Cloud. Markku is the IoT/ML solutions lead for APAC for Google Cloud.
5th November 2020
https://www.devfest.hu