Intrinsic’s Post

Jansen Huang's electrifying keynote at GTC event was mind-blowing🔥! Here are the game-changing highlights: ⭐️ Blackwell NVIDIA unveils the Blackwell GPU platform, setting a new standard for AI computing. Prepare for a revolution in performance and efficiency! ⭐️ Gen AI Microservices Introducing NVIDIA Inference Microservice (NIM), a cloud-native solution streamlining AI application development and deployment. Say hello to seamless integration! ⭐️ Super-AI Agents Huang unveils a groundbreaking approach to software development, leveraging a team of specialized AIs led by a SUPER-AI orchestrator. Goodbye to traditional coding, welcome to the future of #AI innovation! ⭐️ Humanoid Robots Enter Project GR00T, a groundbreaking Foundation Model empowering robots to learn and adapt like never before. With GR00T robots and Jetson Thor, #NVIDIA is reshaping the future of robotics. Get ready for a new era of intelligent machines! 🤖 #AIProduct

Nvidia's Project GROOT brings the human-robot future a significant step closer

Programming #robots for #industrial manufacturing pick-and-place tasks is challenging. This technical blog outlines a potential workflow to address this using NVIDIA Isaac Manipulator’s foundation models for grasp pose generation, Isaac Sim for evaluation in simulation, and Intrinsic Flowstate for real-world execution. #Automate2024 Learn more > https://nvda.ws/3JQ91Ig

Great guide on using Autodistill (https://lnkd.in/essrPMpx) to label data and train models for edge deployments

cuRobo is a CUDA accelerated library containing a suite of robotics algorithms that run significantly faster than existing implementations leveraging parallel compute. cuRobo currently provides the following algorithms: (1) forward and inverse kinematics, (2) collision checking between robot and world, with the world represented as Cuboids, Meshes, and Depth images, (3) numerical optimization with gradient descent, L-BFGS, and MPPI, (4) geometric planning, (5) trajectory optimization, (6) motion generation that combines inverse kinematics, geometric planning, and trajectory optimization to generate global motions within 30ms. cuRobo that provides robust and fast implementations of several core components of motion generation leveraging GPU compute. cuRobo leverages GPU compute to run optimization over many seeds in parallel, converging to good solutions for many robotics optimization problems including collision-free inverse kinematics and trajectory optimization. cuRobo can generate minimum-jerk collision-free trajectories for manipulators within 30ms on a NVIDIA RTX 4090 and within 100ms on a NVIDIA Jetson AGX Orin. cuRobo provides fast implementations of kinematics, collision-free inverse kinematics, trajectory optimization, graph planning, batched numerical optimization solvers (L-BFGS, MPPI), and global motion generation. cuRobo also provides fast collision functions to query signed distance between a robot and the world represented by cuboids, meshes (warp), and depth images (nvblox), leveraging several NVIDIA technologies all on the GPU! cuRobo is integrated with pyTorch to make it easy to add your own motion generation problems. https://lnkd.in/gnaXvq-Z

🎊 Utilizing #LlamaIndex within the Retrieval Augmented Generation(RAG) framework and MLC LLM as the inference server, a local RAG system has been deployed on NVIDIA Robotics Jetson Orin NX. Check out step-by-step wiki guidance to get started on your #reComputer Jetson device: https://lnkd.in/gHp7a3MV With ChromaDB as a local vector database, it stores information from the documents provided. When you ask a question, LlamaIndex converts the question into a vector through embedding layers and retrieves text from the vector database that is relevant to your question. This text serves as context, which is then passed to Llama2-7b, the large language model compressed using MLC LLM 4-bit quantization. Finally, the large language model generates an answer based on this text. This setup not only safeguards data privacy for users but also mitigates responding hallucinations commonly associated with large models. 👉GitHub repo for this project: https://lnkd.in/dvzXMkuK #nvidia #jetson #llamaspeak #RAG #LLM #MLC #GenerativeAI

💰 Want to save on labeling costs for your company? Here's a breakdown: Let's ballpark an average software engineer's salary at $100,000/year. A standard work year of 52 weeks at 40 hours/week is 2080 hours/year. Hourly Rate: $48.08/hour 🕒 If each engineer spends just 1 hour/week on labeling, that's 52 hours/year. Total Labeling Cost per Engineer: $2,499.16/year 💸 Multiply this by the number of engineers, and the costs quickly increase! But here's the game-changer: AutoDistill, a free open-sourced tool, can drastically reduce these costs. Imagine the time and money savings and how you can redirect those funds. #CostSavings #AutoDistill #Efficiency #Labeling"

GPU powered Virtual Desktops over Browsers, unlock the true potential of AI in Computer Vision Applications! 🔥 Real-time object detection for remote cameras ✅ 🔥 Image classification at lightning speed ✅ 🔥 Seamless AI and Robot integration ✅ Say goodbye to limitations and hello to innovation. 🌟 Let's explore the endless possibilities robolaunch brings to the table. 🚀🔗 See below, how robolaunch launches PyTorch and OpenCV based AI&CV application from browsers and get full power of NVIDIA Quadro A6000 GPU for video processing application in seconds. #robotics #AI #autonomousdriving #cloud #kubernetes

We are and we will part of this revolution with our perseverant work in VIDEO SYSTEMS Srl to support manufacturing sector to optimize their processes and produce high quality goods reducing the raw material and energy needs #zdm NVIDIA Robotics #teamwork #sustainability