How to Run Deep Learning Without Melting Your Phone -Iago Suarez

Deep learning and Artificial Intelligence (AI) are widespread technologies. They drive a wide range of applications, from search engines and autonomous vehicles to chatbots and extended reality headsets. However, these powerful neural networks pose a challenge: they are computationally demanding, especially when running on edge devices.
In this talk, Iago Suárez from Qualcomm XR Labs Europe will delve into techniques that enable the execution of these resource-intensive networks with minimal power consumption. A crucial step is converting the original model, typically trained in Python, into highly efficient C++ code that can run on a Neural Processing Engine, such as the Qualcomm® Hexagon™ Processor.