Neural network model processing method and device, equipment and storage medium

Wang Wenxiao,Zhao Shuai,Chen Minghao,Hu Jinming,Cai Deng,Liu Haifeng

2020-01-01

Abstract:The embodiment of the invention provides a neural network model processing method and device, equipment and a storage medium, and the method comprises the steps: obtaining a first neural network modelobtained through pre-training, wherein the first neural network model is used for image recognition; in response to the fact that a first sample image is input into the first neural network model, carrying out image recognition processing based on the output of each layer in the first neural network model, to acquire recognition accuracy corresponding to each layer; based on the identification accuracy corresponding to each layer in the first neural network model, determining the improvement amplitude of each layer for the identification accuracy of the first neural network model; deleting alayer of which the lifting amplitude is smaller than a preset threshold from the first neural network model to obtain a second neural network model; and training the second neural network model basedon the second sample image to obtain a target neural network model. According to the scheme provided by the embodiment of the invention, the image recognition speed of the neural network model can beimproved.