Training method and device for network structure optimizer of neural network and storage medium

• Main Authors: HUANG HANYU, ZHENG YIN, GUO YONG, LIN LEYU, HUANG YAN, WEI BINGZHENG, LU YANXIONG, WANG SU
• Format: Patent
• Language: Chinese; English
• Published: 11.02.2020
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention provides a training method and device for a network structure optimizer of a neural network, electronic equipment and a storage medium. The method comprises the steps of performing feature extraction on a network structure of a neural network through the network structure optimizer to obtain feature information corresponding to the network structure; predicting the feature information through the network structure optimizer, and determining a plurality of optimization modes for the network structure; updating the network structure of the neural network according to the optimization mode for the network structure to obtain an optimized network structure; and performing performance evaluation on the optimized network structure, and updating parameters of the network structure optimizer according to an evaluation result. According to the invention, the network structure of the neural network can be optimized according to the network structure optimizer, and the calculation cost is saved. 本发明提供了一种神经