AI from the Data Center to the Edge

Summary

Whether you are an experienced data scientist or new to the field, this course takes a hands-on approach to help you understand the data science workflow, use Intel’s AI portfolio of processors and optimized software, and apply them to the challenge you are facing. The course uses an enterprise image classification problem and provides lectures for each stage in the process, accompanied by Jupyter* notebooks that walk you through the implementation.

Topics include:

• An overview of Intel’s AI portfolio with an emphasis on solving deep learning problems
• Dataset preparation for model consumption – including preprocessing and data augmentation techniques
• Decision metrics for choosing a framework and network (topology)
• How to train and deploy deep learning models using Intel's AI portfolio

By the end of this course, you will have practical knowledge of:

• Preparing a dataset for model consumption
• Training a deep learning model using TensorFlow* with Intel optimizations
• Deploying on the CPU, integrated graphics and Intel® Neural Compute Stick 2 (Intel® NCS2) using the Intel® Distribution of OpenVINO™ toolkit

The estimated time to complete this course is five hours.

Tools and frameworks used in this course are part of the Intel® AI Analytics Toolkit^(Beta), powered by oneAPI.

• TensorFlow*
• Intel® Distribution of OpenVINO ™ Toolkit

Learn more about oneAPI.

Prerequisites

A basic understanding of AI principles, machine learning, and deep learning

Python* programming

Some exposure to different frameworks, such as TensorFlow* and Caffe*

Optional introduction courses:

Machine Learning

Deep Learning

Applied Deep Learning with TensorFlow