Intel Software
@IntelSoftware
A place to engage with the worldwide community of software and hardware developers.
คุณอาจชื่นชอบ
#OpenVINO 2025.4 has arrived! New features include encrypted blob format for secure model deployment, simplified NPU deployment, improved LLM accuracy, and Windows ML gold support. ms.spr.ly/6019tbsJz
Video analytics at the edge requires detection and context. In this demo, Abhijit Sinha shows how Intel Edge AI on RHEL Image Mode fuses computer vision + LLM for keyframe extraction and summarization, all deployed locally on Intel CPU and Battlemage GPU. ms.spr.ly/6018tn2X0
youtube.com
YouTube
Intel Edge AI on RHEL Image Mode
Using AI to track open parking spaces? The #Metro AI Suite reference app shows how to build video analytics pipelines using #DLStreamer, #MQTT, and #NodeRED for dashboards in #Grafana. Watch the video: ms.spr.ly/6013tn2Xv
Write ahead logging (WAL) and database backup window operations can be bottlenecks for checksum performance. PostgreSQL 18 adds Intel® AVX-512 for CRC32C, replacing SSE4.2 for large buffers. Up to 8.1× faster algorithm-level performance and 13–17% throughput gains. Details:…
Multi-camera #visionAI is great… until video encoding drags inference down. One fix? Deep Learning Streamer, a microservice in the #OpenEdgePlatform that optimizes video encoding on the fly. Architecture: ms.spr.ly/6017tpyOT
Training AI models isn’t just about defaults. What really happens during deep learning? How do you pick the right model? When should you tweak hyperparameters? Explore the answers in Week 4 of the Geti™ Series with Guy Tamir: ms.spr.ly/6015tpyOR #Geti #DeepLearning…
Running LLMs efficiently isn’t trivial. Some queries need low latency and high throughput, others don't require as much power. In this demo, Alex Sin shows dynamic model routing with LiteLLM + vLLM on OpenShift AI to efficiently route requests to Intel Xeon CPUs and Gaudi AI…
youtube.com
YouTube
Supercharge LiteLLMs and vLLMs with Dynamic Routing Models
Multi-camera vision AI is great… until video encoding drags inference down. One fix? Deep Learning Streamer, a microservice in the #OpenEdgePlatform that optimizes video encoding on the fly. Architecture: ms.spr.ly/6012trEZO
Blender 5.0 is here! Congrats to the team @Blender for moving to a Vulkan backend optimized with Intel. This unlocks huge gains: ARC B580 up 80% on average (300% in certain scenes to 94 FPS), Core Ultra Series 2 up 45%, with thin-and-light notebooks hitting 27 FPS.
Ready to build proactive AI agents on the AI PC?Join our live DevCon session on creating next-gen agentic AI assistants using OpenVINO™.🗓 Nov 20 | ⏰ 10 AM PT 🔗 Register here: ms.spr.ly/6016trEZy #OpenVINO #AIPC #AgenticAI
What happens when you optimize PostgreSQL for Intel hardware on AWS? We tuned PostgreSQL 17 for EC2 m8i/c8i/r8i instances and saw up to 2.4x performance gains. From kernel tweaks to query engine optimizations—see how it works: ms.spr.ly/6018trEZK
How do you fuse multi-camera streams into a unified 3D scene for real-time insights? Guy Tamir walks us through the Smart Intersection reference app, an open-source solution powered by #DLStreamer, #OpenVINO, and #IntelSceneScape. Watch now: ms.spr.ly/6019trEZU
youtube.com
YouTube
Smart Intersection (Metro AI Suite | Open Edge Platform)
All hands on deck! Don't miss #IntelAI's post-#NeurIPS meetup at The Deck at Moonshine Flats. Join us for snacks, drinks, #OpenVINO and #AIPC demos, and a deck full of devs. The best connections happen after the conference. ms.spr.ly/6014tPuyY
Building computer vision models isn’t just about defaults. Why does patience matter in training? When should you ignore splits? How do you balance speed, size, and accuracy? Find out in Week 4 of the Geti™ Series from AI with Guy: ms.spr.ly/6013tOpPF #Geti #ComputerVision…
What matters more for computer vision: data quantity or data quality? This week, Guy Tamir digs into why high-res images and real-world match beat brute force, how annotation choices impact results, and why the 80/20 split still matters. Watch the latest episodes in the Geti™…
Making a renderer go fast takes more than running a tool. It means profiling correctly, interpreting counters, and understanding GPU architecture. We’ve been working with Microsoft to improve PIX, and on Nov 20 we’ll share how to read what PIX is telling you and apply it to…
Intel’s #Robotics AI Suite Fastmapping demo enables collaborative #VSLAM with @RealSenseai depth camera: intel.ly/3L3W01E
Starting a Vision AI project in Geti™? Guy Tamir walks through 3 ways to kick things off: • Start fresh • Import a ZIP • Create from a dataset (COCO, YOLO, Datumaro) Built for devs and domain experts—no PhD required. Watch the series: intel.ly/3LbZCPj or jump right…
What if your AI agents could work together, reasoning, sharing context, and taking action as a team? Join us at DevCon 2025 for a free workshop on how to build multi-agent GenAI applications using the Model–Context–Protocol (MCP) and open-source components like OpenVINO™ Model…
Do you know how we've partnered with @EA_DICE and @CriterionGames to optimize Battlefield 6 for PC hardware? Find out more on Nov 19th. #GPC #IntelArc #gamedev #BF6 intel.ly/3J1cwPz…
United States เทรนด์
- 1. #GivingTuesday 26.4K posts
- 2. #twitchrecap 7,729 posts
- 3. Larry 43.6K posts
- 4. Costco 52.4K posts
- 5. Cabinet 63.4K posts
- 6. So 79% 1,705 posts
- 7. NextNRG Inc. 3,460 posts
- 8. #AppleMusicReplay 9,751 posts
- 9. #Rashmer 29.7K posts
- 10. Jared Curtis 2,915 posts
- 11. Joe Schoen 6,639 posts
- 12. #DragRace 6,290 posts
- 13. Susan Dell 6,804 posts
- 14. Sabrina Carpenter 31.8K posts
- 15. Carton 29.6K posts
- 16. WFAN N/A
- 17. Sleepy Don 1,897 posts
- 18. King Von 1,985 posts
- 19. Lucario 30.2K posts
- 20. Pat Leonard N/A
คุณอาจชื่นชอบ
-
Windows
@Windows -
Sony
@Sony -
LIAC at EPFL
@SchwallerGroup -
Intel News
@intelnews -
Chrome
@googlechrome -
NVIDIAGameDev
@NVIDIAGameDev -
Microsoft Azure
@Azure -
Amazon Web Services
@awscloud -
NVIDIA Data Center
@NVIDIADC -
Intel Gaming
@IntelGaming -
NVIDIA HPC Developer
@NVIDIAHPCDev -
WD®
@WDCreators -
Intel
@intel -
NVIDIA Robotics
@NVIDIARobotics -
Google for Developers
@googledevs
Something went wrong.
Something went wrong.