Library for modelling performance costs of different Neural Network workloads on NPU devices - View it on GitHub