Koki Mitsunami and Ben Clark, Developer Advocates at Arm, demonstrate the company’s latest edge AI and vision technologies and products at the 2022 Embedded Vision Summit. Specifically, Mitsunami and Clark demonstrate machine learning on Arm technology.
Mitsunami shows how to improve PyTorch performance with Android NNAPI, by demonstrating models running on a mobile device using Streamline, Arm’s performance profiler. Developers can benefit from the NNAPI within the PyTorch ecosystem, and many mobile devices have models accelerated by the underlying Arm NN and Arm Compute Library.
Clark shows on-device face recognition on DTV, delivering security and privacy by keeping the function on-device while also delivering the CPU or GPU-processed performance required to do continuous multiple face recognitions at a distance. Use cases include personalization, parental control and teleconferencing.
