Neural network model construction method and device, and storage medium

Abstract

The invention discloses a construction method of a neural network model for realizing image classification. The construction method comprises the following steps: S1, constructing a unit structure search network, a system structure search network, an image training set and a random coding array; s2, generating a neural network model by using the unit structure search network, the system structuresearch network and the random coding array; s3, inputting the image training set into a neural network model to obtain an actual classification result; s4, judging whether an actual classification result meets a preset condition or not, and if not, performing a step S5; s5, updating the unit structure search network and the system structure search network according to the actual classification result and the theoretical classification of the image training set; s6, repeating S2; And S5, until the actual classification result obtained in the step S4 meets the preset condition. According to themethod disclosed by the invention, an original search space is converted into two spaces, namely a unit structure search space and an architecture search space, the optimal architecture of the architecture is searched in an automatic learning manner, and the flexibility of the generated model architecture is enhanced.

Description

technical field [0001] The present invention relates to the field of image classification, and more specifically, refers to a method, device and readable storage medium for constructing a neural network model. Background technique [0002] The neural network model is a model structure that can be stacked arbitrarily. The basic components include FC (full connection layer), Convolution (convolution layer), Polling (pooling layer), Activation (activation function), etc. The latter component is the previous one. The output of the component is used as the input, and different component connection methods and hyperparameter configuration methods have different effects in different application scenarios. Neural Architecture Search (NAS), the goal is to search for an optimal neural network model from a bunch of neural network components. Among them, common search methods include random search, Bayesian optimization, evolutionary algorithm, reinforcement learning, gradient-based al...

AI Technical Summary

Problems solved by technology

[0005] (1) Because there are too many combinations, the search space is huge and the cost of function calculation is huge;

[0006] (2) Manual design of model architecture, lack of flexibility

Images