Workshop method

Getting Started with Intel® Optimization for PyTorch*

Intel® and Analytics India Magazine hosted a oneAPI AI Analytics Toolkit workshop – a masterclass in Intel® optimization techniques for accelerating deep learning workloads – on March 25, 2022from 3:00 p.m. to 5:00 p.m..

The Intel® Extension for PyTorch is part of the Intel® Optimization for PyTorch with optimizations for additional performance on Intel® hardware. The company said most optimizations will be available in stock versions of PyTorch. This includes AVX-512 Vector Neural Network Instructions (AVX512 VNNI) and Intel® Advanced Matrix Extensions (Intel® AMX).

Intel® has released Intel® Extension for PyTorch as an open source project on GitHub. Find the installation guide here. Additionally, users can use the pip command on Linux to install the Intel® Extension for PyTorch.

Register for the workshop here.

Intel® has worked closely with Facebook (now Meta) to bring optimizations to the PyTorch community and is committed to continuously optimizing PyTorch with future advancements in Intel® hardware.

PyTorch brings a modular design with a registration API that allows third parties to extend its functionality, including core optimizations, graphics optimization passes, custom operations, etc., with an extension library which can be dynamically loaded as a Python module or from C++ as a shared module. library.

IPEX is one such PyTorch extension library, an open source project maintained by Intel® and released as part of Intel® AI Analytics Toolkit powered by oneAPI.

What are you going to learn?

The workshop will introduce attendees to the Intel® Calibrated Optimizations for PyTorch, Installation Guide, and Performance Improvement Number. Additionally, attendees will learn easy-to-use Python API, vectorization, parallelism, quantization, operator merging, constant folding, and more.

Intel® will have a demo on:

  • Intel® Optimization for PyTorch
  • Intel® Extension for PyTorch
  • The Intel® Extension for PyTorch targets optimizations on the AVX-512 instruction set.
  • Intel® Optimization for PyTorch released in oneAPI AI Analytics Toolkit.

Who should be present?

  • AI and ML developers
  • Data Scientists
  • AI enthusiasts
  • AI researchers
  • GPU and HPC programmers

______________________________________________

Exclusive contests — Participate and win!

Prize Draw Contest

  • Analytics India Magazine organizes a Draw where at the end of the workshop 10 lucky participants will have the chance to WIN an Amazon voucher worth INR 2,000/- each.

To note: Winners will be selected based on their engagement on Discord throughout the workshop. https://discord.gg/ycwqTP6>

  • In addition, all oneAPI workshop attendees will receive a flat 20% discount on The rise 2022 conference pass.

______________________________________________

Register on Intel® DevCloud account, a free development sandbox with access to the latest Intel® hardware and oneAPI software, and use event code oneAPI25March to register and follow the emailed instructions. Join the Discord channel for live Q&A during the workshop.

Agenda:

Session Contents Duration Owner
introduction Developer Ecosystem Program 3:00 p.m. – 3:10 p.m. Kavita Arour
Overview Introducing the oneAPI AI Analytics toolkit 3:10 p.m. – 3:25 p.m. Aditya Sirvaiya
DevCloud setup 3:25 p.m. – 3:40 p.m. Aditya Sirvaiya
Intel® Extension for PyTorch Introducing Intel® Optimization for PyTorch 3:40 p.m. – 4:00 p.m. Jing Xu
Intel® Extension for PyTorch Intel® Extension for Pytorch + Case Study 4:00 p.m. – 4:30 p.m. Jing Xu
QA + practice Hands-on with RetinaNet medical image detection workload 4:30 p.m. – 5:00 p.m. Jing Xu and Aditya Sirvaiya

Prerequisites for the workshop:

  • Please register for a Intel® DevCloud account, a free development sandbox with access to the latest Intel® hardware and oneAPI software. You must use the event code oneAPIMarch 25 to register, then follow the instructions sent by e-mail.
  • Please save this email with the username, node name and URL embedded in the email and bring it to the workshop. (Check your spam folder if this email does not reach your inbox).
  • Here is a Handy DevCloud setup guide
  • For a live Q&A during the workshop, please join us on the Discord channel https://discord.gg/ycwqTP6.

Session speakers:

Kavita Arour

Kavita Arour is Developer Marketing Lead- APJ at Intel® and focuses on developing software solutions for the Asia-Pacific market. She leads developer outreach, engagements, and programs in GPU, HPC, and AI. With 17 years of brand and field marketing experience, Kavita has garnered the engagement of many customers and users across Asia-Pacific.

Jing Xu

Jing Xu is a Senior Technical Consulting Engineer working as an AI Specialist in a developer product consulting and support organization within the Intel® Software Group. It enabled global developers, enterprise users, engineers, and researchers to use Intel® software tools and high-performance libraries. His research areas include machine learning, deep learning, performance optimization and data analytics R&D.

Aditya Sirvaiya

Aditya Sirvaiya is an AI Technical Consulting Engineer with the Intel® Software Group. Aditya holds a BSc in Engineering Physics from IIT, Delhi and a MSc in Computer Science with specialization in AI from IIT, Mumbai.

So what are you waiting for? Click on here to register for the oneAPI workshop on “Getting Started with Intel® Optimization for PyTorch”.