Structure search method, training method, device and storage medium of neural network

Abstract

The embodiment of the present invention provides a neural network structure search method, training method, device and storage medium. Wherein, the structure search method of the neural network includes: based on the preset network framework, using the structure search variables and the structure search ensemble to link each intermediate network layer in at least one intermediate network layer to obtain the output of the neural network to be determined; Wherein, the preset network architecture includes an input network layer, an output network layer, and at least one intermediate network layer; based on the output of the neural network to be determined and a training set, a first loss function is determined; based on the neural network to be determined The output of the network and the verification set determine the second loss function; update the model parameters of the neural network to be determined based on the first loss function; update the structural parameters of the neural network to be determined based on the second loss function; use the updated The model parameters and the updated structure parameters are used to determine the structure of the neural network to be determined.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to a neural network structure search method, training method, device and storage medium. Background technique [0002] Machine learning (ML, Machine Learning) is a branch of artificial intelligence. The purpose of machine learning is to allow machines to learn based on prior knowledge, so as to have the logical ability of classification and judgment. Machine learning models represented by neural networks continue to develop and are increasingly used in various industries. However, the design of the neural network in related technologies needs to combine expert experience and conduct a large number of parameter debugging experiments. The entire design process needs to be completed manually, which is time-consuming and labor-intensive. Contents of the invention [0003] In order to solve the existing technical problems, the embodiment of the present invention provides a ne...

AI Technical Summary

Problems solved by technology

However, the design of the neural network in related technologies requires the combination of expert experience and a large number of parameter debugging experiments. The entire design process needs to be completed manually, which is time-consuming and labor-intensive.

Images