Brandon Lloyd

Are you tired of hand-tuning each model you develop? What if you could describe the architecture once and let a system apply graph transformations and optimized kernels? NVIDIA TensorRT LLM AutoDeploy marks a shift toward approaching inference optimization as a compiler and runtime responsibility rather than a burden on the model author. This approach enables faster experimentation, broader model coverage, and a cleaner separation between model design and deployment. Learn more from our documentation, examples scripts, and the blog, ‘Automating Inference Optimizations with NVIDIA TensorRT LLM AutoDeploy’ Read the full post: https://lnkd.in/dYC2eAkA #PyTorch #OpenSourceAI #AI #Inference #Innovation

138

The ExecuTorch 1.0 GA release is here, and it’s redefining what’s possible for AI at the edge. Built on a unified PyTorch workflow and optimized for Arm-based platforms, ExecuTorch 1.0 makes developing and deploying edge AI faster, with higher performance and less fragmentation across billions of edge devices. By using the same PyTorch models end-to-end across mobile, embedded, and edge platforms, workflows are streamlined and time-to-market is accelerated, alongside automatic performance optimizations through Arm KleidiAI, TOSA, and CMSIS-NN integrations. And the best part? It’s available right now, so you can bring edge AI to life everywhere, for everyone: https://okt.to/mS5u8R

75

2 Comments

🏫 UC San Diego's Hao AI Lab now has full access to an #NVIDIADGX B200 system at the San Diego Supercomputer Center, opening up new research opportunities and elevating critical work in LLM #inference. ➡️ https://nvda.ws/4pObI0s The DGX B200 is already accelerating two Hao AI Lab projects: FastVideo and the Lmgame benchmark (tests LLMs using Tetris and Super Mario Bros). In addition to these projects, research in healthcare and biology are underway. Professor Hao Zhang says DGX B200 “enables us to prototype and experiment much faster than using previous-generation hardware,” with performance among the best in the world. 👏 We're excited to see the results of their research and projects.

182

3 Comments

In this T‑RO paper, the authors propose a task‑driven co‑design framework that optimizes sensors, planners & compute under safety, cost, energy & weight constraints—demonstrated on AVs to show how task complexity affects autonomy stack choices. https://lnkd.in/gEHEPKrU Paper title: CODEI: Resource-Efficient Task-Driven Co-Design of Perception and Decision Making for Mobile Robots Applied to Autonomous Vehicles Paper authors: Dejan Milojevic, Gioele Zardini, Miriam Elser, Andrea Censi, and Emilio Frazzoli #RobotSensingSystems #AutonomousVehicles #LIDARSensor #MobileRobot

77

2 Comments