Join us for the OpenVINO™ Toolkit &
Media Analytics Hands-On Technical Training
This training was developed by the Intel® Network Builders University, offering technical professionals a limited-access opportunity to learn from Intel’s subject matter experts. The virtual hands-on training will help technical professionals in the network industry to improve their knowledge of key Intel® technologies, industry trends, and technical aspects of NFV deployments.
Technical Training Requirements
This training is intended for technical leaders and senior developers responsible for implementing or optimizing inference application on Intel’s latest generation platforms. You might be a Software Engineer, Network Architect, Dev Ops, etc.
Technical Expectations
- Attendees are expected to have technical familiarity with deep learning, video codecs and processing technology.
- Attendees should have experience in C++, python, Docker swarm or Kubernetes environments and SSL security tunneling. They should also understand typical networking operations on Intel two-socket servers.
- Attendees must have a computer or laptop with the ability to SSH to a lab environment hosted in a remote data center.
- Attendees should have basic knowledge of using git, command line shell (BASH) as well as working knowledge of linux system administration (how to use simple commands like cd, ls, mv, mkdir, etc) and least one Linux text editor (nano, pico, or vi/vim,) with which to edit simple scripts and text files.
- Prior knowledge of OpenVINO™ toolkit and its components. OpenVINO™ toolkit training completion is a pre-requisite to accessing the subsequent Media Analytics training and server credentials.
Syllabus
OpenVINO™ Syllabus
Intel® Smart Video/Computer vision Tools Overview
- An opportunity to learn about the features included in the Intel® Distribution of the OpenVINO™ toolkit, a deep learning inference which delivers fast, accurate real-world results using high-performance, AI and computer vision inference deployed into production across Intel® architecture from edge to cloud. Discover the key components of the OpenVINO™ Toolkit.
Model Optimizer
- Model Optimizer is a cross-platform command-line tool that facilitates the transition between the training and deployment environment, performs static model analysis, and adjusts deep learning models for optimal execution on end-point target devices. In this session, students will have the opportunity to learn what kind of optimization does Model Optimizer do to the model... and how much can you control it?
Inference Engine
- This session highlights the high-level inference API, Inference Engine which is implemented as dynamically loaded plugins for each hardware type and delivers advanced performance for each type without requiring users to implement and maintain multiple code pathways.
Multiple Models in One Application
- Students will examine how video analytics in Intel® Distribution of OpenVINO™ Toolkit support chaining multiple models together in one application.
Deep Learning (DL) Workbench & Deep Learning (DL) Streamer
- Take a look at the DL Workbench capabilities, data flow, interfaces with key components, as well as the new features, plus a DL Workbench walkthrough. DL Workbench is a web-based tool - UI extension of Intel® Distribution of OpenVINO™ toolkit functionality which visualizes performance data for topologies/layers to aid in model analysis.
- Intel® Distribution of OpenVINO™ toolkit Deep Learning (DL) Streamer is now part of the default installation package which enables developers to create and deploy optimized streaming media analytics pipelines across Intel® architecture from edge to cloud.
Intel® DevCloud for Edge Walkthrough
- Develop your computer vision applications using the Intel® DevCloud, which includes a preinstalled and preconfigured version of the Intel® Distribution of OpenVINO™ toolkit. Access reference implementations and pretrained models to help explore real-world workloads and hardware acceleration solutions.
Instructor
Shane Ye
Software Engineer, Intel
Follow on:
Shane Ye is a Developer Enabler at Intel, IOTG Developer Enabling Division. He is responsible for creating content, conducting technical training and supporting the developer community on Intel IoT technologies. He is skilled in developing product prototypes and demonstrations across a wide range of topics including computer vision, cloud-based services, board level communication protocols. He enjoys hiking and photographing with family in the nature.
Media Analytics Syllabus
Introduction to Media Analytics
- Media Analytics is an application of inference in media and video domain to gain intelligent insights for generating automated alerts and tagging. Modern cities protected with hundreds of cameras and thousands of hours of streaming services generate large amount of videos every day which is passed over to the edge and cloud. Learn how to apply deep learning to these videos on architecture and design flexible solutions for emerging applications like smart cities.
Media Analytics Hardware Platforms
- This session will dive into the Intel hardware platforms and products that are best suited for media analytics solution and how to leverage them for your workload optimizations. This section will specifically discuss the 2nd gen Intel® Xeon® Scalable processors and the Visual Cloud Acceleration Card for Analytics (VCAC-A).
Software Tools for Media Analytics
- In this section, we discuss software components designed to optimize inference by leverage neural network libraries and OpenVINO toolkit. We will go deeper on how to develop media analytics pipeline through open source multi-media framework, Ffmpeg and GStreamer part of open visual cloud software stack with a hands-on exercise.
Open Visual Cloud Pipelines
- This session provides design details for Smart City application, one of the open Visual Cloud reference application using media analytics. We go over multiple building blocks of the pipeline like camera management, edge analytics, transcoding for storage through hands-on exercise and code review.
Next Steps: How to get started
- If the training has piqued your interest in media analytics this session will provide a list of available resources and next steps to further your education and deepen your engagements with Intel.
Instructor
Surbhi Madan
Solution Architect at Intel Corporation Follow on:
Passionate about technology that enhances multimedia consumption and innovating solutions in media processing, Surbhi Madan is a solution architect in the data center group at Intel. Her role involves working with customers and partners to apply deep learning on media as a data, driving optimized solution on Intel architecture. She holds a master's degree in electrical engineering from the University of Southern California specializing in image and video processing and has held multiple roles in the last 6 years at Intel. She is currently based in Phoenix, Arizona.
Srikanth Ramakrishna
Platform Applications Engineer at Intel Corporation Follow on:
Srikanth Ramakrishna is a Platform Applications Engineer for Media Analytics at the Data Center Group at Intel. His role primarily involves in supporting customers with resolving hardware and software issues, maintaining collateral documents and customer training and interaction. He is very passionate and enthusiastic about deep learning and computer vision technologies. He holds a master’s degree in Electrical and Computer Engineering from University of Illinois at Chicago with focus in Deep Learning and Signal Processing. He is currently based out of Folsom, California.
Office Hours & Certificate
After enrollment is confirmed, students will be provided opportunities to schedule a brief call with your instructors to answer questions and support next steps. (This is not a requirement for completion.)
When the training modules, labs, and quizzes are complete, a Certificate of Completion will be provided.