Lightning Talk: Every Millisecond Counts: The Fine-tuning Journey of an Ultra-Eff... Pavel Macenauer

This talk, presented by Pavel Macenauer from NXP Semiconductors, delves into the optimization journey of an ultra-efficient PyTorch model designed for edge AI applications. It examines the crucial role of local data processing, highlighting the significance of minimizing latency and power consumption in vision, audio, and language models. The speaker discusses various optimization techniques that enable these models to function effectively on resource-constrained devices, demonstrating the importance of real device profiling feedback to refine performance. This session emphasizes that every millisecond and milliwatt is vital in exploring new use cases and driving innovation in edge technology.