Raymond Lo, OpenVINO Edge AI Software Evangelist at Intel, presents the “Fearless AI Challenge: Can You Quickly Deploy AI Inference to Billions of Devices?” tutorial at the May 2021 Embedded Vision Summit.
Have you ever been inspired by a new AI research paper and tried to replicate it using open source code, only to end up with a “bricked” development system? This is a common experience. The AI ecosystem is fragmented; many solutions require specialized hardware, such as GPUs, or specific frameworks, libraries, APIs or tools that may conflict with your current development environment. Intel identified this gap and created the OpenVINO toolkit and OpenVINO Notebooks to address these challenges.
In this talk, Lo shows how to get AI inference running in 10 minutes or less—avoiding common pitfalls by simplifying download, setup and runtime. And, using OpenVINO, you can deploy your solution to billions of existing computing devices—including embedded hardware like VPUs—without recompiling your code! At the end of this session, you can replicate Lo’s work on your own machine.