Neural network model construction method based on heterogeneous chip energy consumption and related device

The invention discloses a neural network model construction method based on heterogeneous chip energy consumption and a related device, and the method comprises the steps: testing the operator energy consumption of model operators of different algorithm models on different heterogeneous chips through employing a preset test case, and obtaining a model operator energy consumption table, the model operator energy consumption table comprises operator names, chip models, execution time and energy consumption of model operators; constructing a reinforcement learning data set according to the model operators in the model operator energy consumption table; and based on the reinforcement learning strategy, searching an optimal energy consumption operator in the reinforcement learning data set according to a preset strategy network and a preset reward function, and constructing an optimal energy consumption neural network model. The neural network model constructed according to the optimal energy consumption operator