Official NVIDIA handle for all things Robotics and Edge AI.
External link for NVIDIA Robotics
NVIDIA Robotics reposted this
ICYMI: Discover how developers are enhancing their #OpenUSD workflows with NVIDIA NIM Microservices. See how USD Code, USD Search, and USD Validate are transforming 3D environment creation using #generativeAI copilots & agents. Dave Tyner Paul Callender Alex Q. 📺 https://nvda.ws/3Z7Emiv
Join us on Sept. 19 at 10 AM PT for a deep dive into Milestone Systems AI-Bridge, and discover how #visionAI solution providers like IronYun and Lanner Electronics Inc. are leveraging this innovation to advance their AI applications. Uncover the latest in AI video analytics and VMS for AI ➡️ Explore NVIDIA's advancements in Video Language Models (VLMs) and Gen AI for IVA ➡️ Engage with experts and get your questions answered live Register Now > https://nvda.ws/3MtcNsE
Learn more about the NVIDIA Jetson AGX Orin and see why it's known as the leading platform for running transformer models like GPT-J, vision transformers, and stable diffusion at the edge.
Get a deeper look into the results submitted by NVIDIA in #MLPerf Inference v4.1, highlighted by the first ever submission using the NVIDIA Blackwell architecture, increased performance using the NVIDIA H200 Tensor Core GPU, and more. https://nvda.ws/4dGlpba
NVIDIA Robotics reposted this
Announcing the fourth annual Imagine conference! Join us on September 24th to learn how leading enterprises are solving their biggest challenges with edge AI at our flagship event. Hear from Arm, NVIDIA, Capgemini/Synapse Product Development, Pentatonic, Halma plc, and more, and get up to speed on the latest features, innovations, and best practices in edge AI. Don’t miss out on: • Insightful keynotes from industry leaders • Live demos showcasing edge AI in action • Networking opportunities with top professionals in the AI community Register now for this free event, either in person in Mountain View, CA, or online: edgeimpulse.com/imagine #imagine2024 #edgeAI #AI
NVIDIA Robotics reposted this
Take a deep dive with Walt Disney Imagineering on their robotic character platform that enables the rapid design of legged robotic characters that learn to imitate artist-specified animations with a combination of deep reinforcement learning and GPU-accelerated simulation.🤖 Replay this on-demand session from #OpenUSD Day at #SIGGRAPH2024 and learn about the first characters developed with Disney’s platform – the BDX series droids. Moritz Baecher https://nvda.ws/3yW0YrG
Introducing NVIDIA TAO 5.5: a powerful training framework with new foundation models, multi-modal sensor fusion, and advanced features like text prompt-based auto-labeling and knowledge distillation. #NVIDIAMetropolis Read the blog. 💡 https://nvda.ws/3z5JwRp
NVIDIA Jetson AGX Orin is revolutionizing #edgeAI and #robotics. 🤖 Capable of running diverse models such as LLMs, vision transformers, and Stable Diffusion locally, it turns sensor data into real-time, actionable insights. In the latest MLPerf benchmarks, Jetson AGX Orin achieved 6x the performance boost on the GPT-J LLM workload. 👏 Learn more.
In the latest round of #MLPerf industry benchmarks, the first ever submission of NVIDIA Blackwell delivered up to 4X more performance on #Llama 2 70B, while the NVIDIA H200 Tensor Core GPU continued to deliver outstanding results across every test in the data center category.
Explore how NVIDIA NIM accelerates the production of VLMs and #visualAI agents, from cloud to far-edge. Register for the webinar: https://nvda.ws/46ZsZLz
NVIDIA Robotics reposted this
Ready to enter the world of embodied agents and physical AI? Come along with us at Jetson AI Lab as we dive into Vision/Language Models and OpenVLA in sim, fine-tuning, and deployment https://lnkd.in/gR_CAQAz First we extended our lead streaming VLM stack with optimized quantization & inference for edge VLA's. Then integrated MimicGen from Yuke Zhu @Yuke Zhu and Jim Fan for an endless stream of training episodes for OpenVLA to learn new tasks from Karl Pertsch. We dialed in the workflow with test LoRA's onboard Jetson AGX Orin, and when the initial results were looking good hopped on some quick A100/H100 instances from Brev.dev The result: can stack blocks like a toddler 🤣🤖 But we cleared the workflow, learned a lot along the way training our own vision/language models, and the whole space is progressing rapidly - the latest last week was Crossformer. Kudos to RAIL Lab at Berkeley for all the amazing science they are sharing. Over time the workflow for robot learning will become streamlined to the point where it can adapt in-situ by end users, and perhaps robots will undergo nightly fine-tunings the same way that humans are thought to relive/simulate the day's experience during REM sleep. In the end we'll all become model tamers and robot mechanics ✌ Get the VLA training + deployment guide here! - https://lnkd.in/gR_CAQAz
NVIDIA Robotics reposted this
Our latest e-book has just been released! 📕 In it, you’ll find a complete guide to NVIDIA AI TAO computer vision models and how to use them on microcontrollers. We cover all the technical building blocks TAO has to offer, show you how to use TAO in Edge Impulse, then walk you through a step-by-step deployment onto the high-performance Renesas Electronics RA8D1 Evaluation Kit powered by the Arm Cortex-M85 MCU. Download and give it a read, here: https://lnkd.in/gtT7g9xh Deepu Talla Chintan Shah Adam Scraba Wei Xiao Debraj Sinha Brian Pickering Paul Williamson Mary Bennion Reinhard Keil Peter Fenn Monica Houston Jim Beneke Daryl Khoo Mohammed Umar Dogar Adam Benzion Amir Sherman