Microsoft shares what’s subsequent in machine studying at NVIDIA GTC | Azure Weblog and Updates


Placeholder

Discovering scalable options for right this moment’s international challenges requires forward-thinking, transformative instruments. As environmental, financial, and public well being considerations mount, Microsoft Azure is addressing these challenges head on with high-performance computing (HPC), AI, and machine studying. The behind-the-scenes energy for all the things from MRI scans to power administration and monetary companies, these applied sciences are equipping prospects and builders with modern options that break by the boundaries of what’s doable in knowledge and compute, paving the way in which for progress alternatives that span industries and purposes world wide.

Microsoft Azure is dedicated to unlocking these new alternatives for our prospects, offering the broadest vary of NVIDIA GPUs on the edge, on-premises, within the cloud, and for hybrid environments.

At NVIDIA GTC we’ll reveal this dedication by displaying how Azure’s superior HPC capabilities, and AI/machine studying within the cloud are driving transformation and making an impression along with NVIDIA’s newest know-how.

Microsoft Azure’s collaboration with NVIDIA was developed with our prospects in thoughts and centered on opening new doorways to innovation with graphics processing unit (GPU) acceleration within the cloud.

Study extra by registering right this moment for NVIDIA GTC, a free, on-line occasion working September 19 to 22, 2022.

Placeholder

Get an opportunity to win an NVIDIA Jetson Nano or swag field

In each of our classes you might have an opportunity to win a SWAG field full with a HPC t-shirt and mug or a Jetson Nano. Attend these classes and don’t neglect to search for the particular hyperlink to enter!

Microsoft Classes at NVIDIA GTC

The brand new SDK and CLI in Azure Machine Studying.

Bala Venkataraman, Principal Program Supervisor, Microsoft.

Video on demand

Azure Machine Studying is dedicated to simplifying the adoption of its platform for coaching and manufacturing. In 2022, we introduced the final availability of Azure Machine Studying CLI v2 and the preview of Azure Machine Studying Python SDK v2. Each launches reveal our continued deal with making workflows simpler and managing their total lifecycle ranging from coaching single jobs to pipelines and mannequin deployments. On this session, find out about the important thing enhancements in usability and productiveness, and the brand new options that include the command-line interpreter (CLI) and software program growth package (SDK) v2 of Azure Machine Studying.

Operationalize giant mannequin coaching on Azure Machine Studying utilizing multi-node NVIDIA A100 GPUs.

Sharmeelee Bijlani, Program Supervisor Azure Machine Studying, Microsoft; Razvan Tanase, Principal Engineering Supervisor Azure Machine Studying, Microsoft.

Wednesday, September 21, 10:00 to 10:50 AM PDT (1:00 to 1:50 PM EDT, 7:00 to 7:50 AM CEST)

In recent times, deep studying fashions have grown exponentially in measurement, demonstrating an acute want for purchasers to coach and fine-tune them utilizing large-scale knowledge infrastructure, superior GPUs, and an immense quantity of reminiscence. Luckily, builders can now use easy coaching pipelines on Azure Machine Studying to coach giant fashions working on the newest multi-node NVIDIA GPUs. This session will describe these software program improvements to prospects by Azure Machine Studying (together with a totally optimized PyTorch setting) that provides nice efficiency and an easy-to-use interface for large-scale coaching. We’ll additionally spotlight the ability of Azure Machine Studying by experiments utilizing 1,024 A100 Tensor Core GPUs to scale the coaching of a two-trillion parameter mannequin with a streamlined consumer expertise at 1,000 plus GPU scale.

Watch Get together #1: Operationalize large-model coaching on Azure Machine Studying utilizing multi-node NVIDIA A100 GPUs.

Mary Howell, NVIDIA.

Wednesday, Sep twenty first, 3:00 – 3:30 PM PDT

On this GTC Watch Get together, we shall be replaying our Operationalize Massive-Mannequin Coaching on Azure Machine Studying utilizing Multi-Node NVIDIA A100 GPUs session. Contributors shall be joined by consultants from throughout Microsoft and NVIDIA who deliver recent insights and experiences to the desk, taking the session to an entire new degree of understanding. Interplay is core to our GTC Watch Events, and we encourage you to hitch the dialogue with any feedback or questions. 

Watch Get together #2: Operationalize large-model coaching on Azure Machine Studying utilizing multi-node NVIDIA A100 GPUs.

Gabrielle Davelaar, AI Technical Specialist, Microsoft; Maxim Salnikov, Senior Azure GTM Supervisor, Microsoft; Henk Boelman, Senior Cloud Advocate–AI and Machine Studying, Microsoft; Alexander Younger, Technical Advertising and marketing Engineer, NVIDIA; Ulrich Knechtel, Microsoft Companion Supervisor (EMEA), NVIDIA.

Thursday, September 22, 2:00 to 3:30 PM CEST (5:00 to six:30 AM PDT, 8:00 to 9:30 AM EDT)

On this GTC Watch Get together, we shall be replaying our Operationalize Massive-Mannequin Coaching on Azure Machine Studying utilizing Multi-Node NVIDIA A100 GPUs session. Contributors shall be joined by consultants from throughout Microsoft and NVIDIA who deliver recent insights and experiences to the desk, taking the session to an entire new degree of understanding. Interplay is core to our GTC Watch Events, and we encourage you to hitch the dialogue with any feedback or questions.

Microsoft helps prospects throughout industries step up, reworking AI and machine studying on the Edge

Nuance’s Dragon Ambient eXperience helps docs doc care quicker with AI on Azure

Nuance developed an AI-based medical resolution that robotically turns doctor-patient conversations into correct medical notes. Constructed with Azure and PyTorch, this resolution saves docs transcription time, lowering administrative burdens and serving to them conduct extra centered, higher-quality interactions with their sufferers.

Vitality utility Elva builds a extremely safe DevOps platform with Azure infrastructure and community safety companies

Elva seemed to construct a safe, cloud-first DevOps platform that might meet Norway’s knowledge residency and compliance necessities, delivering automated companies that will assist develop community grid know-how. Utilizing Azure DDoS Safety, Azure Internet Utility Firewall, and Azure Kubernetes Service, Elva realized its objective, enhancing its in-house growth and knowledge integration capabilities. 

The Royal Financial institution of Canada creates customized presents whereas defending knowledge privateness with Azure confidential computing

The Royal Financial institution of Canada (RBC) partnered with Microsoft to create a privacy-preserving multi-party knowledge sharing platform constructed on Azure confidential computing. Referred to as VCR, this resolution allows RBC to personalize choices and shield privateness on the similar time, creating distinctive digital experiences that shoppers can belief.  

Placeholder

Recapping 2022 moments with Azure and NVIDIA applied sciences

Azure NC A100 v4-series

At Microsoft, our NC collection digital machines enable prospects entry to nearly limitless AI {hardware} infrastructure to allow them to be productive shortly. Final summer time, we leveled up, asserting the common availability of Azure NC A100 v4 collection digital machines. Powered by NVIDIA A100 80GB PCle Tensor Core GPUs and third Gen AMD EPYC™ processors, these digital machines assist our prospects achieve insights quicker, innovate with velocity, do extra with much less, and are the most performant and cost-competitive NC collection providing for a various set of workloads.

DeepSpeed on Azure

Azure Machine Studying makes use of giant fleets of the newest NVIDIA GPUs powered by NVIDIA Quantum InfiniBand interconnects to deal with large-scale AI coaching and tuning. Final July, we introduced a breakthrough in our software program stack, utilizing DeepSpeed and 1,024 NVIDIA A100 GPUs to scale the coaching of a two trillion parameter mannequin with a streamlined consumer expertise at 1,000 plus GPU scale. We’re bringing these software program improvements to you thru Azure Machine Studying (together with a totally optimized PyTorch setting) that provides nice efficiency and an easy-to-use interface for large-scale coaching.

NVads A10 v5 digital machines

Historically, graphics-heavy visualization workloads that run within the cloud require digital machines with full GPUs which can be each pricey and rigid. To fight this, we launched the primary GPU-partitioned (GPU-P) digital machine providing within the cloud, and simply final July, we introduced the final availability of NVads A10 v5 GPU accelerated digital machines. Azure is the primary public cloud to supply GPU partitioning on NVIDIA GPUs, and our new NVads A10 v5 digital machines are designed to supply the best alternative for any workload and supply optimum configurations for each single-user and multi-session environments. Dig into our newest digital machine innovation.

NVIDIA Jetson AGX Orin-powered edge AI units now obtainable

Microsoft is happy to announce that the NVIDIA Jetson AGX Orin SoM is now powering Azure Licensed edge units from industry-leading gadget builders together with AAEON, Advantech, and AVerMedia, together with the NVIDIA Jetson AGX Orin developer package.

Builders and resolution builders can now leverage highly effective NVIDIA Jetson AGX Orin devkits and manufacturing modules with Microsoft Azure to create, deploy, and function highly effective AI options on the edge, accelerating product growth and deployment at scale. The NVIDIA Orin Nano modules have set a brand new baseline for entry-level edge AI and robotics, constructing on the momentum behind the Jetson Orin platform worldwide. Keep tuned for brand spanking new Jetson Orin NX and Orin Nano associate merchandise launching to fulfill buyer wants in AI resolution growth.

NVIDIA DLI coaching powered by Azure

We’re proud to host NVIDIA deep studying institute (DLI) coaching at NVIDIA GTC once more this yr, with instructor-led workshops round accelerated computing, accelerated knowledge science, and deep studying. Hosted on Microsoft Azure, these classes allow and empower you to leverage NVIDIA GPUs on the Microsoft Azure platform to resolve the world’s most attention-grabbing and related issues. Register for a DLI workshop right this moment.

Be a part of us at NVIDIA GTC

In collaboration with NVIDIA, Microsoft delivers purpose-built AI, machine studying, and HPC options within the cloud to fulfill even essentially the most demanding real-world purposes at scale. Be a part of us at NVIDIA GTC September 19 to 22, to see how each enterprise can leverage the ability of GPUs on the edge, on-premises, within the cloud, and for hybrid options.

Study extra

Leave a Reply