Comprehensive Platform Enables Customers to Rapidly Create and Deploy Interoperable Solutions with Leading Edge AI Processors and Accelerators
Lanner and DeGirum, leaders in Edge AI, have partnered to introduce a software-based AI orchestration solution compatible with a growing array of Edge AI inferencing accelerators, such as GPUs and NPU ASICs. Unlike proprietary ecosystems, this platform emphasizes open interoperability, addressing challenges in software and data model portability.
Key Benefits
- Proven Edge AI Solution: Deploy computer vision, generative AI, and inferencing applications seamlessly, free from proprietary development restrictions.
- Scalability: Applications scale effortlessly with additional hardware, requiring no code changes.
- No Support Costs: Lanner customers receive one year of premium technical support from DeGirum at no cost.
- Hardware Performance Testing: Users can remotely benchmark models on various hardware configurations.
“We are excited to offer our customers a complete solution with our Edge AI systems, such as the MGX server family, which can be powered by DeGirum,” said Terence Chou, CEO of Lanner.
Also Read: BrowserStack Launches AI-Driven Test Platform
“The Lanner/DeGirum partnership expands our reach, integrating DeGirum’s PySDK and AI Hub with Lanner’s advanced server hardware,” said Winston Lee, CEO of DeGirum.
OrchestrAItor® Edge-AI Orchestration Platform
OrchestrAItor® accelerates AI prototyping without hardware setup. Developers can access hundreds of production-ready models or convert training checkpoints into deployable models for diverse hardware. Its model zoo enables edge devices to dynamically update models without bundling them in applications. OrchestrAItor® run-time modules operate directly on Lanner Edge AI servers.
Expanding Accessibility with JavaScript SDK
DeGirumJS, a JavaScript SDK, integrates AI capabilities into web and mobile applications, offering real-time object detection, image classification, and more directly in the browser.
Source: PRNewswire
