Nvidia issued its anticipated raft of stories at CES this week, right here’s an summary of bulletins for the HPC-AI sector:
‘Mega’ Omniverse Blueprint for Industrial Robotic Fleet Digital Twins
The corporate mentioned Mega is an omniverse framework for next-gen industrial AI and robotic simulation by means of software-defined testing and optimization of factories and warehouses.
Citing details and figures – there are 10 million factories, almost 200,000 warehouses and 40 million miles of highways – the corporate mentioned this huge industrial community of manufacturing services and distribution facilities remains to be laboriously and manually designed, operated and optimized.
Mega is a blueprint designed for creating, testing and optimizing bodily AI and robotic fleets at scale in a digital twin earlier than deployment into real-world services. That is for superior warehouses and factories that use fleets of autonomous cell robots, robotic arm manipulators and humanoids working alongside folks.
Nvidia mentioned Mega provides a reference structure of accelerated computing, AI, Nvidia Isaac and Nvidia Omniverse applied sciences for creating and testing digital twins for AI-powered robotic brains that drive robots, video analytics AI brokers, tools and extra for dealing with huge complexity and scale. The brand new framework brings software-defined capabilities to bodily services, enabling steady improvement, testing, optimization and deployment.
Nvidia Omniverse with Generative Bodily AI
Nvidia announced generative AI fashions and blueprints that increase its Omniverse integration into bodily AI functions, corresponding to robotics, autonomous autos and imaginative and prescient AI.
The corporate mentioned Accenture, Altair, Ansys, Cadence, Foretellix, Microsoft and Neural Idea are among the many first to combine Omniverse into their software program services and products. Industrial automation firm Siemens additionally introduced the supply of Teamcenter Digital Actuality Viewer — the primary Siemens Xcelerator software powered by NVIDIA Omniverse libraries.
“Bodily AI will revolutionize the $50 trillion manufacturing and logistics industries. The whole lot that strikes — from vehicles and vehicles to factories and warehouses — can be robotic and embodied by AI,” mentioned Jensen Huang, founder and CEO at NVIDIA. “NVIDIA’s Omniverse digital twin working system and Cosmos bodily AI function the foundational libraries for digitalizing the world’s bodily industries.”
Nvidia mentioned the USD Code and USD Search NVIDIA NIM microservices are actually usually obtainable, they’re designed to let builders use textual content prompts to generate or seek for OpenUSD belongings. A brand new NVIDIA Edify SimReady generative AI mannequin unveiled in the present day can mechanically label present 3D belongings with attributes like physics or supplies, enabling builders to course of 1,000 3D objects in minutes as a substitute of over 40 hours manually, based on the corporate.
Undertaking DIGITS With Grace Blackwell 10 Superchip Debuts as AI Supercomputer
Nvidia introduced Project DIGITS, which the corporate referred to as a private AI supercomputer designed to supply AI researchers, information scientists and college students entry to Grace Blackwell platform, introduced final March on the firm’s GTC convention.
Undertaking DIGITS provides a petaflop of GB10 Superchip computing efficiency for prototyping, fine-tuning and working massive AI fashions. The corporate mentioned customers can develop and run inference on fashions utilizing their desktop system, then deploy the fashions on accelerated cloud or information middle infrastructure.
The corporate mentioned Undertaking DIGITS delivers GB10 efficiency utilizing solely a regular electrical outlet. Every Undertaking DIGITS options 128GB of unified, coherent reminiscence and as much as 4TB of NVMe storage. With the supercomputer, builders can run as much as 200-billion-parameter massive language fashions. As well as, utilizing Nvidia ConnectX networking, two Undertaking DIGITS AI supercomputers might be linked to run as much as 405-billion-parameter fashions.
Cosmos World Basis Mannequin Platform for Bodily AI Growth
Nvidia introduced Cosmos, a platform comprised of generative basis fashions, tokenizers, guardrails and an accelerated video processing pipeline constructed for improvement of bodily AI methods corresponding to autonomous autos (AVs) and robots.
Cosmos fashions can be obtainable beneath an open mannequin license to assist speed up the work of the robotics and AV group. Builders can preview the primary fashions on the NVIDIA API catalog, or obtain the household of fashions and fine-tuning framework from the NVIDIA NGC catalog or Hugging Face.
Robotics and automotive firms, together with 1X, Agile Robots, Agility, Determine AI, Foretellix, Fourier, Galbot, Hillbot, IntBot, Neura Robotics, Skild AI, Digital Incision, Waabi and XPENG, together with ridesharing big Uber, are among the many first to undertake Cosmos.
“The ChatGPT second for robotics is coming. Like massive language fashions, world basis fashions are basic to advancing robotic and AV improvement, but not all builders have the experience and sources to coach their very own,” mentioned Huang. “We created Cosmos to democratize bodily AI and put normal robotics in attain of each developer.”
All Nvidia CES-related bulletins and blogs can be found here.
