Camvi Demonstrates Superior AI Inference Optimization in Dell Testing
Camvi Technologies announced today that it has demonstrated outstanding production AI inference optimization for resource-demanding large models such as ChatGPT.
BURLINGTON, Mass., March 14, 2023 /PRNewswire-PRWeb/ -- In response to accelerated production AI inference resource explosion due to the latest deployment of ChatGPT and the likes, US-based AI startup Camvi Technologies announced today that in addition to Camvi's NIST world #1 ranked Vision AI accuracy performance, it has also demonstrated outstanding AI cloud platform inference optimization.
Camvi's patent-pending deep-learning-based joint GPU-CPU scheduling, memory and power mode optimization reduces critical GPU resource requirements by over 90%. The continuous data driven optimization is critical to enable production AI inference at super low costs for resource-demanding large models such as ChatGPT. Camvi's unique asynchronous model-pipeline multiplexing in combination with adaptive quantization technology can further significantly reduce the number of GPUs required by Vision AI applications.
Dell Technologies Labs has validated that Camvi's NIST world #1 ranked Vision AI provides superior performance in accuracy, speed and scalability for large scale video and image recognition deployments. The Dell Technologies Labs tests aim to benchmark the real-world performance of Camvi Vision AI on a Dell server platform with Nvidia hardware acceleration to identify persons or objects from many channels against a large gallery and a long history, in real-time as well as recorded videos, at places such as airports, transit centers and casinos.
The tests were conducted using the Camvi's carrier-class InSight cloud native platform combined with a Dell R740xd server platform with two NVIDIA T4 GPU acceleration. The combination of software and hardware is shown to be effective in handling face recognition for 90 simultaneous 1080p video streams at 24 FPS with up to 10 faces in each frame and each face is identified from a gallery of 53 million faces in near real-time. In addition, Camvi InSight supports up to 3,000 simultaneous API calls on the Dell R740xd platform with sub-second response times. The sub-second matching performance with billions of images and superior cost effectiveness make the Camvi and Dell/Nvidia combination a powerful Vision AI platform.
With Camvi's AI cloud platform optimization, production AI inference for critical large AI models such as ChatGPT can be empowered via either a much smaller number of NVIDIA A100 GPUs or a set of cost and energy effective GPUs such as the Nvidia Tesla T4.
Camvi is growing rapidly and its superior Vision AI platform is being adopted in a wide range of industries. Camvi's customers benefit from the unparalleled accuracy, speed, scalability and joint CPU-GPU inference optimization of Camvi's state-of-the-art, patent-pending AI technology.
About Camvi Technologies Inc.
Camvi Technologies is a U.S.-based Artificial Intelligence company specializing in Vision AI technologies. Since 2015, Camvi Technologies has been at the forefront of Vision AI. Camvi makes the best-in-class Vision AI Platform-as-a-Service (PaaS) for Solution Architects and Developers who are looking for ready-to-deploy Vision AI solutions to embed into their own products and services. At the forefront of Vision AI, Camvi is constantly pushing the boundaries to radically improve performance, reduce cost and simplify deployment. Camvi is dedicated to the development of Vision AI solutions that allow our customers to ethically use advanced AI technology to address complex business issues. We provide white glove, white label delivery via a cost-effective SaaS model. Our Vision AI solution has been deployed in 11 countries across 5 continents.
Learn more about Camvi
Website: https://www.camvi.com
Follow on LinkedIn: http://www.linkedin.com/company/camvi-technologies/
Media Contact
John Chen, https://www.camvi.com/, 1 4083060658, [email protected]
SOURCE Camvi
Share this article