NNEQ

FeatureFormer predictor improves neural network performance estimation for edge devices

Researchers have developed FeatureFormer, a novel neural performance predictor designed to accurately estimate latency and energy consumption for neural networks on resource-constrained edge devices. This predictor inco…